Method, apparatus, and system for providing two-wheeled vehicle sharing platform service utilizing idle two-wheeled vehicle and space information

Abstract

One embodiment of the present invention relates to a method, apparatus, and system for providing a motorcycle sharing platform service utilizing idle motorcycle and space information, wherein driver authentication, selection of sharing candidate motorcycles and sharing candidate spaces, borrower reliability evaluation, input / output space suitability evaluation, driving range simulation, idle time matching rate analysis, and calculation of combined sharing costs are performed, and based on the results, rental condition information including whether use is approved, usage fees, warranty conditions, insurance coverage conditions, and input / output processing conditions is automatically determined.

Description

Technology Field

[0001] The following embodiments relate to a method, apparatus, and system for providing a motorcycle sharing platform service utilizing idle motorcycle and space information, wherein driver authentication, selection of sharing candidate motorcycles and sharing candidate spaces, borrower reliability evaluation, input / output space suitability evaluation, driving range simulation, idle time matching rate analysis, and calculation of combined sharing costs are performed, and based on the results, rental condition information including whether use is approved, usage fees, warranty conditions, insurance coverage conditions, and input / output processing conditions is automatically determined. Background Technology

[0002] With the recent rise of the sharing economy, the demand for personal mobility and two-wheeled vehicles (motorcycles, scooters, electric two-wheeled vehicles, etc.) is surging. However, privately owned two-wheeled vehicles are left idle in specific locations for most of the day, and due to the lack of dedicated parking spaces for two-wheeled vehicles in urban areas, the problem of illegal parking on side streets is becoming serious.

[0003] While some peer-to-peer (P2P) rental and sharing services are being attempted to utilize these idle resources, the unique physical and operational characteristics of two-wheeled vehicles, which differ from four-wheeled vehicles, are not reflected in the system at all. For example, although it may be impossible to move heavy, high-displacement two-wheeled vehicles by manually pulling them after dismounting depending on the slope of the parking area floor or the condition of the road surface, conventional systems uniformly match vehicles and spaces based solely on location coordinates.

[0004] Furthermore, it is difficult to encourage voluntary participation from vehicle owners due to the lack of a reasonable compensation system regarding the risk of damage when lending motorcycles—which are valuable personal assets—to others, as well as the wear and tear on consumables (tires, brake pads, etc.) caused by harsh driving. Moreover, in decentralized models where the motorcycle owner and the owner of the idle parking space may be different, there is currently a lack of platform technology capable of transparently and fairly settling costs among multiple parties by integrating factors such as electric charging costs, contributions to night security (CCTV), and depreciation. Prior art literature

[0005] (Patent Document 0001) KR 10-2835508 B (Patent Document 0002) KR 10-2455030 B (Patent Document 0003) KR 10-2308288 B (Patent Document 0004) KR 10-2886534 B The problem to be solved

[0006] The problem that one embodiment of the present invention aims to solve is to provide a method, device, and system for providing a motorcycle sharing platform service utilizing idle motorcycle and space information, which performs optimal rental matching and multi-party customized profit settlement by comprehensively analyzing the borrower's reliability, the physical suitability of the motorcycle specifications and the entry / exit space, the risk of power depletion according to the expected driving environment, and the depreciation of parts, in order to overcome the limitations of the conventional simple location-based P2P sharing matching and unidirectional billing system as described above. means of solving the problem

[0007] According to one embodiment, a method for providing a two-wheeled vehicle sharing platform service utilizing idle two-wheeled vehicle and space information, wherein a server includes a processor, memory, a communication module, and a non-transient storage medium, and the processor executes a program stored in the non-transient storage medium, the method comprises: receiving service request information from a driver terminal including driver identification information, two-wheeled vehicle usage request information, and usage condition information; performing authentication for the driver based on the driver identification information and obtaining driver information including usage qualification information, usage history information, and payment method information associated with the driver identification information; and selecting a target two-wheeled vehicle and a target inbound / outbound space corresponding to the service request information based on operational resource information including two-wheeled vehicle information and storage space information. The method provides a motorcycle sharing platform service utilizing idle motorcycle and space information, comprising: a step of calculating rental condition information including at least one of approval status for use, usage fee, warranty condition, insurance coverage condition, and entry / exit processing condition for the target motorcycle, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target entry / exit space information, and pre-set operation policy information; and a step of selecting the target motorcycle and target entry / exit space, comprising: a step of selecting sharing candidate motorcycles and sharing candidate spaces by reflecting available time intervals, available area intervals, charging or fuel status, vehicle type suitability, and space access convenience corresponding to the motorcycle usage request information, based on a plurality of idle motorcycle information and idle space information included in the operation resource information; and a step of determining the target motorcycle and target entry / exit space to be assigned to the driver based on the driver information, sharing candidate motorcycle, sharing candidate space, and pre-set operation policy information.

[0008] In addition, the step of determining the target motorcycle and target entry / exit space comprises: deriving a borrower reliability index by quantifying the history of past motorcycle damage, traffic law violations, return delay frequency, reservation cancellation frequency, and payment failure history included in the driver information, and determining the primary matching eligibility for each sharing candidate motorcycle by comparing the borrower reliability index with a minimum reliability threshold for rental allowance previously set by the owner of the sharing candidate motorcycle; A step of calculating the suitability of a motorcycle-specific entry / exit space reflecting entry / exit stability, parking suitability, and storage stability by comparing the floor slope, road surface flatness, inclusion of security CCTV coverage, availability of charging infrastructure for electric motorcycles, availability of nighttime illumination, and availability of a rain-blocking structure of the sharing selection space with the drive method, body weight, vehicle type, wheelbase, and steering angle specifications of the sharing selection motorcycle; A step of simulating the actual driving range by reflecting the current battery state of charge or remaining fuel amount, recent energy consumption or fuel consumption per unit distance, and correction coefficients by drive method of the shared selected motorcycle, based on the estimated driving distance, scheduled time of use, elevation change profile of the scheduled driving route, estimated load, and ambient temperature information included in the above motorcycle usage request information, and calculating the risk of power depletion during driving based on the difference between the actual driving range and the estimated driving distance; a step of calculating the estimated idle duration, which is the time interval expected to be available until the next reservation, retrieval, maintenance, and owner usage schedule of the shared selected motorcycle begins, and deriving an idle time match rate by calculating the temporal intersection ratio between the estimated idle duration and the driver's scheduled rental time;A step of deriving a combined sharing cost by summing the motorcycle sharing revenue unit price set by the owner of the selected sharing motorcycle, the space sharing revenue unit price set by the owner of the selected sharing space, the estimated charging cost or fuel preservation cost, and the estimated recovery cost of the selected sharing motorcycle; a step of calculating a final sharing matching score for each combination of the selected sharing motorcycle and the selected sharing space by weighted summing the borrower reliability index, the suitability of the motorcycle-specialized entry / exit space, the reverse penalty corresponding to the risk of power depletion, the idle time matching rate, the combined sharing cost, the driver's past preferred vehicle type information, and the driver's past preferred return area information; and, for a plurality of combinations with the highest final sharing matching scores, verify whether there are duplicate reservations in the same time slot, whether there is a possibility of exceeding the occupancy of the same space, and whether there is a conflict with the next reservation, and then determine the combination with the highest final matching score reflecting the verification results as the target motorcycle and target entry / exit space; may be included.;

[0009] And, the step of calculating the above rental condition information comprises: a step of calculating a basic usage fee by adding a pre-set platform brokerage fee, payment processing fee, and operational management fee to the sharing combination cost derived during the process of determining the above target motorcycle and target entry / exit space, and deriving a final usage fee to be charged to the borrower by applying a dynamic discount rate proportional to the idle deadlock resolution contribution of the above target motorcycle and target entry / exit space to the above basic usage fee if the above idle time matching rate is above a pre-set optimization threshold; and a step of calculating the space sharing revenue to be settled to the owner of the above target entry / exit space based on the space sharing revenue unit price included in the above sharing combination cost, wherein at least one of 1) an electricity charge amount based on the estimated charging amount if the above target motorcycle is an electric motorcycle, 2) a security weighting cost proportional to the occupancy time of a security CCTV if the scheduled usage time of the target entry / exit space corresponds to nighttime hours, and 3) a management surcharge corresponding to the entry restriction time of the target entry / exit space is additionally reflected to finally determine the above space sharing revenue; A step of calculating the motorcycle sharing revenue to be settled to the owner of the target motorcycle based on the motorcycle sharing revenue unit price included in the above-mentioned sharing combination cost, and calculating a depreciation adjustment cost by reflecting at least one of the predicted tire wear, brake pad consumption, battery charge / discharge degradation, and drivetrain load index based on the estimated driving distance, planned usage time, estimated load, and elevation change profile of the planned driving route included in the motorcycle usage request information, and finally determining the motorcycle sharing revenue by reducing the platform brokerage fee as the depreciation adjustment cost increases and adding an amount corresponding to the reduction to the motorcycle sharing revenue unit price;The method may include the step of calculating a reliability deviation between the borrower reliability index and a minimum reliability threshold for rental allowance set by the owner of the target motorcycle, and if the reliability deviation is less than a set safety margin range, generating insurance and guarantee conditions including at least one of the following to hedge sharing risk: increasing the self-damage deductible applied to the borrower in the event of an accident, a condition for subscribing to a platform-exclusive short-term insurance special clause, a condition for depositing an additional security deposit, and a condition restricting use during high-risk hours; and the step of finally determining rental condition information by combining the final usage fee, space sharing revenue, motorcycle sharing revenue, platform brokerage fee, and insurance and guarantee conditions, wherein the cost burden relationship, profit distribution relationship, scope of usage liability, and scope of compensation liability in the event of an accident are defined among the borrower, motorcycle owner, space owner, and platform.

[0010] In addition, the step of calculating the rental condition information comprises: a step of calculating driver risk information based on the driver information and the motorcycle usage request information; and a step of calculating rental fees and insurance coverage conditions based on the driver risk information, grade information of the target motorcycle, and operation policy information; and the step of calculating the driver risk information comprises: a step of extracting driver static profile data from the driver information, including the driver's age, license type, elapsed time since obtaining the license, past driving mileage by motorcycle displacement range, insurance accident history, traffic law violation history, number of rental uses during a recent predetermined period, past return delay history, payment failure history, and whether a commercial delivery platform account is linked. A step of extracting regional information from the above-mentioned motorcycle usage request information, comprising a planned departure area, planned return area, and planned driving area, including road type, road gradient, intersection density, inclusion of speed limit zones, inclusion of areas prone to frequent rain, inclusion of areas with poor nighttime lighting, inclusion of steep slope areas, inclusion of areas prone to frequent poor road surfaces, and inclusion of areas prone to motorcycle accidents; a step of generating dynamic context data including the regional information, purpose of use, planned rental start time, planned rental end time, planned usage time, and estimated driving distance by extracting the purpose of use, planned rental start time, planned rental end time, planned usage time, and estimated driving distance from the above-mentioned motorcycle usage request information; and a step of calculating a model proficiency index representing the degree of experience the driver has actually experienced driving a vehicle of the same or similar performance group as the target motorcycle by correlating and comparing the driver's past driving mileage by motorcycle displacement range with the target motorcycle's displacement, body weight, vehicle type, drive method, and maximum output information.A step of generating a data set for risk analysis by mapping accident correlations, severe accident contributions, exposure risks by time of day, and risk coefficients by region for each element of the driver static profile data and each element of the dynamic context data by referring to a pre-established motorcycle accident factor analysis database, and configuring the mapping results into a numerical vector having an array form of risk scores and risk coefficients for each element; a step of assigning commercial operation weights corresponding to the frequency of repeated stops and starts, the possibility of multiple short-distance trips, and the proportion of driving during peak hours to the data set for risk analysis if the driver’s commercial delivery platform account linkage is true, and assigning non-commercial operation weights corresponding to personal movement patterns if the purpose of use is non-commercial; and a step of assigning environmental risk correction coefficients to region-related elements included in the dynamic context data if the planned driving area includes at least one of a rain-prone area, a night illumination-vulnerable area, a steep slope area, a road surface-poor area, and a motorcycle accident-prone area. The method may include the step of inputting the above-mentioned risk analysis dataset, model proficiency index, commercial operation weight or non-commercial operation weight, and environmental risk correction coefficient into a pre-trained machine learning-based risk prediction model to calculate risk component values ​​including a basic driving risk score, an hourly accident exposure frequency weight, an environmental risk surcharge coefficient, and a model non-conformity penalty coefficient; and the step of combining the above-mentioned risk component values ​​to generate driver risk information including a final risk score corresponding to the driver, a risk grade, and a contribution by risk factor.

[0011] In addition, the step of calculating the rental fee and insurance application conditions comprises: a step of deriving a basic rental fee and a basic insurance premium by reflecting the basic depreciation rate of the target motorcycle, the standard accident exposure level by vehicle group, and the standard demand coefficient by time of day, by referring to pre-established standard billing rate information based on the grade information of the target motorcycle, the purpose of use included in the motorcycle usage request information, the scheduled start time of rental, the scheduled end time of rental, the scheduled time of use, and the estimated mileage; a step of calculating a dynamic rate weight by applying a sigmoid function to the final risk score included in the driver risk information, and calculating a rental fee correction coefficient and an insurance premium correction coefficient applied to the basic rental fee and the basic insurance premium, respectively, based on the dynamic rate weight; A step of generating an insurance guarantee condition correction coefficient that determines at least one of the following when at least one of the above basic driving risk score, hourly accident exposure frequency weighting, environmental risk surcharge coefficient, and model non-conformity penalty coefficient exceeds a preset threshold: the minimum applicable value of the self-vehicle damage deductible applied in the event of an accident, the deposit addition ratio, whether to reduce the personal injury and property damage coverage limits, and whether to activate a conditional exemption clause for high-risk driving; a step of generating conditional special clause data including a remote activation condition for a maximum speed limit function through the drive control unit of the above target motorcycle, a rapid acceleration limit condition, a night driving allowance time limit condition, and a condition for shortening the driving monitoring cycle when entering a motorcycle accident-prone area or a steep slope area when the risk grade of the above driver risk information corresponds to a preset management requirement grade; and a step of calculating an environment-linked correction rate defined in the above operation policy information when the scheduled rental time of the above target motorcycle corresponds to at least one of the commuting peak time, late-night time, rain forecast time, and time of rapid temperature drop.The method may include: a step of calculating a corrected rental fee and a corrected insurance premium by applying a rental fee correction coefficient and an insurance premium correction coefficient, respectively, to the basic rental fee and the basic insurance premium, and then applying an environment-linked correction rate; and a step of determining rental condition information by combining the deposit and deductible amount information reflecting the insurance guarantee condition correction coefficient, the conditional special clause data, and the corrected rental fee and the corrected insurance premium.

[0012] A device (server) according to one embodiment may be combined with hardware and controlled by a computer program stored on a medium to execute the method of any one of the methods described above. Effects of the invention

[0013] According to one embodiment, beyond simply assigning rental vehicles in response to a driver's request for use, by integrally mapping the idle time of individual vehicles and information on idle parking spaces owned by others, it has the effect of simultaneously maximizing the utilization of abandoned motorcycle resources and empty parking spaces within the city and creating a new multilateral sharing economy ecosystem.

[0014] Furthermore, by cross-validating a reliability index—quantifying the driver's past traffic violations and history of causing damage—with an allowable threshold set by the vehicle owner, the system systematically resolves the psychological and financial anxiety of owners renting their vehicles to others. In particular, by calculating a "suitability for motorcycle-specific entry / exit spaces" by comparing the floor slope and road surface flatness of the sharing space with the motorcycle's body weight, wheelbase, and steering angle specifications, it can prevent accidents such as rollovers or storage instability that may occur when manually parking heavy motorcycles. In addition, by precisely simulating the actual driving range by reflecting elevation changes along the planned route, load capacity, and ambient temperature, it provides an advanced matching system that drastically reduces the critical risk of battery or fuel depletion during rental.

[0015] Furthermore, a smart settlement system is completed that derives a combined cost by aggregating the desired sharing unit price, estimated charging costs, and recovery costs of both the motorcycle owner and the space owner, and transparently distributes profits among multiple parties based on this. If a borrower's scheduled rental time contributes to resolving the motorcycle's idle dead-time, a dynamic discount rate is applied to increase the matching success rate. Additionally, by preemptively predicting tire and brake pad wear caused by harsh conditions, such as altitude changes in mountainous terrain along the driving route, platform fees are reduced and paid to the vehicle owner as a depreciation adjustment cost, thereby fundamentally preventing complaints regarding wear and tear on consumables. Finally, if a borrower's reliability falls below a standard, the system dynamically defends against risks arising from peer-to-peer transactions—such as increasing the own vehicle damage deductible or mandating the deposit of a security deposit—thereby significantly enhancing the sustainability and stability of the sharing service.

[0016] According to one embodiment, by assigning a target motorcycle optimized for the driver based on the driver's identification information and usage request information and calculating customized rental condition information, the service satisfaction is increased by providing reasonable rates to safe drivers, and the efficiency of asset operation is maximized for platform operators.

[0017] Furthermore, the accuracy of the assessment is dramatically improved by analyzing risk through the fusion of dynamic context—such as road conditions in the planned driving area, time of day, and whether commercial delivery is involved—with static profiles like the driver's age and past accident history. In particular, by deriving a 'model proficiency index' through cross-validation of the driver's past mileage by engine displacement with the specifications of the rental motorcycle and applying it to a machine learning model, it is possible to precisely predict in advance the probability of accidents caused by inexperienced drivers unfamiliar with high-displacement motorcycles.

[0018] Furthermore, not only are basic rental fees and insurance premiums adjusted in proportion to the calculated dynamic risk, but the platform's financial risks are strongly defended by increasing the deductible for own damage or adding a deposit if the vehicle is identified as a high-risk group. In addition, by remotely linking conditional special clauses—such as maximum speed limits, rapid acceleration restrictions, and night driving time restrictions based on the risk grade—with the motorcycle's control system, it demonstrates an excellent safety management effect that goes beyond simple fee differentiation to actively deter actual accidents. Brief explanation of the drawing

[0019] FIG. 1 is a schematic diagram showing a two-wheeled vehicle sharing platform service provision system utilizing idle two-wheeled vehicles and space information according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating a method for providing a two-wheeled vehicle sharing platform service utilizing idle two-wheeled vehicles and space information according to an embodiment of the present invention. FIG. 3 is a flowchart illustrating the step of selecting a target motorcycle and a target inbound / outbound space of a method for providing a motorcycle sharing platform service utilizing idle motorcycle and space information according to an embodiment of the present invention. FIG. 4 is a flowchart illustrating the step of determining the target motorcycle and target inbound / outbound space to be assigned to the driver in a method for providing a motorcycle sharing platform service utilizing idle motorcycle and space information according to an embodiment of the present invention. Specific details for implementing the invention

[0020] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0021] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to the specific disclosed forms, and the scope of this specification includes modifications, equivalents, or substitutions that fall within the technical concept.

[0022] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0023] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or coupled with that other component, or that there may be other components in between.

[0024] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0025] In particular, where a 'step' in this specification is described as 'comprising' one or more detailed steps or sub-steps, said 'step' may be interpreted as including its own basic processing step while simultaneously performing the described detailed steps as well.

[0026] For example, if it is stated that 'a step of doing B to A' includes 'a step of doing D to C; a step of doing F to E; and a step of doing H to G,' the 'step of doing B to A' may be interpreted not merely as the basic operation of doing B to A, but as a configuration that performs detailed procedures together, such as a step of doing D to C, a step of doing F to E, and a step of doing H to G.

[0027] Accordingly, the above configuration does not exclude various sub-procedures included within the scope of execution of the corresponding step, and may be included within the scope of the present invention even if other procedures or means performing substantially the same or equivalent functions are substituted.

[0028] Expressions such as 'end part', 'both ends', 'one end', 'other end', and 'side end' of a component can be interpreted as referring to at least / any one of the end parts of that component.

[0029] In the description of the present invention, 'a method in which a device comprises a processor, a memory, a communication module, and a non-transient storage medium, and a program stored in the non-transient storage medium is executed by the processor,' the term 'method' may be interpreted as referring to the program stored in the non-transient storage medium itself or a part of the program.

[0030] The term 'Return' as used in the description of the present invention may refer to a result value being output, returned, or returned from a method, procedure, function, etc. used in a given program language / structure.

[0031] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application. For example, the term 'artificial intelligence model' may be selected from one or more of known general artificial intelligence models.

[0032] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. In describing the embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiments, such detailed description is omitted.

[0033] According to one embodiment, a server includes a processor, memory, a communication module, and a non-transient storage medium, and a method for providing a short-to-medium term rental platform service for two-wheeled vehicles through the calculation of rental fees and insurance premiums based on driver information, which is performed by the processor by executing a program stored in the non-transient storage medium, comprises: receiving service request information including driver identification information, two-wheeled vehicle usage request information, and usage condition information from a driver terminal; performing authentication for the driver based on the driver identification information and obtaining driver information including usage qualification information, usage history information, and payment method information linked to the driver identification information; and selecting a target two-wheeled vehicle and a target inbound / outbound space corresponding to the service request information based on operational resource information including two-wheeled vehicle information and storage space information. The method provides a short-to-medium term motorcycle rental platform service through the calculation of rental fees and insurance premiums based on driver information, comprising: a step of calculating rental condition information including at least one of whether the use of the target motorcycle is approved, a usage fee, a warranty condition, an insurance coverage condition, and an entry / exit processing condition, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target entry / exit space information, and pre-set operation policy information; wherein the step of calculating rental condition information includes: a step of calculating driver risk information based on the driver information and the motorcycle usage request information; and a step of calculating rental fees and insurance coverage conditions based on the driver risk information, grade information of the target motorcycle, and operation policy information.

[0034] The step of receiving service request information is a step in which the server receives service request information, which is basic data for the start of rental, from a driver terminal connected by a driver who wishes to rent a two-wheeled vehicle for the short to medium term, and forms standard data for subsequent authentication procedures, operational resource allocation procedures, and rental condition calculation procedures.

[0035] A driver terminal may include smart devices such as smartphones, tablets, PCs, vehicle terminals, or wearable devices that run a dedicated application or webpage for a two-wheeled vehicle rental platform and can transmit and receive data with a server via a wired or wireless communication network. Driver identification information refers to identification data capable of uniquely identifying a driver connected to the platform, and may include, for example, at least one of a login account, name, contact information, mobile phone number, device unique identification number, identification token corresponding to biometric authentication, driver's license identification value, or member unique number. Additionally, two-wheeled vehicle usage request information refers to information indicating the specific usage details of the rental desired by the driver, and may include the rental start time, rental end time, scheduled usage time, estimated driving distance, purpose of use, desired vehicle type or engine displacement range, scheduled departure location, scheduled return location, etc. Usage condition information may include the insurance level selected in advance by the driver or assigned as a default by the platform, deposit limit, preferred pick-up / drop-off method, permission for night driving, use for delivery, rental of accessories such as helmets or smartphone holders, and information on the selection of additional special clauses. For example, if a driver requests a medium-sized electric motorcycle for travel of approximately 40 km between 6 PM and 10 PM and selects a deposit-reducing insurance condition and a helmet rental option, the server can receive the entire request as a single service request information and use it as an input value for subsequent steps.

[0036] The step of performing driver authentication and acquiring driver information involves verifying whether the driver is a legitimate user with actual service usage rights based on the received driver identification information, and calling up historical data associated with the verified driver to form foundational information for driver-centric risk analysis and condition calculation.

[0037] The authentication procedure may be performed using at least one of the following methods: password input, one-time authentication code input, biometric recognition, verification of registered terminal, or identity authentication through a telecommunications carrier. Driver information refers to comprehensive profile information that is pre-registered or accumulated and stored in a platform database in conjunction with the driver identification information, and may include eligibility information, usage history information, and payment method information. Here, eligibility information refers to information indicating whether a driver can legally and safely operate a two-wheeled vehicle, and may include the type of driver's license held, date of license acquisition, validity of the license, whether the license is suspended or revoked, whether age criteria are met, status of agreement to terms and conditions, existence of unpaid fees, and status of restrictions on platform usage. Usage history information refers to records of the driver having previously rented a two-wheeled vehicle through the platform, and may include the number of past rentals, accumulated mileage, number of return delays, history of accident or damage reports, driving patterns by time of day, information on previously selected vehicle models, history of penalty imposition, and history of payment failures. Payment method information may include credit card information, linked account information, simple payment identifiers, or automatic payment approval status registered in advance by the driver to charge service usage fees, deposits, penalties, and additional charges. At this stage, rather than simply verifying membership registration status, multi-layered verification can be performed to determine whether the driver is in a state to actually use the service level requested by the driver. For example, if the first user holds a Class 2 small vehicle license, has no history of payment failure in the past year, and has no outstanding fees, whereas the second user has outstanding fees or their license has expired, the server can complete normal authentication only for the first user and proceed to the subsequent step.

[0038] The step of selecting target motorcycles and target inbound / outbound spaces is a step of searching for motorcycles and storage spaces available for allocation to drivers at the current time and at the time of request, based on the real-time resource status included in the operational resource information, and extracting the combination most suitable for the above service request information.

[0039] Operational resource information refers to an integrated database of multiple motorcycles and multiple inbound / outbound locations registered on the platform, and may include motorcycle information and storage space information. Motorcycle information may include the vehicle type, engine displacement, drive system, body weight, current battery charge status or remaining fuel level, current location, availability, accumulated mileage, maintenance requirements, scheduled collision history, and insurance risk group classification information for each motorcycle. Storage space information may include the location coordinates of each inbound / outbound location, the number of currently available parking spaces, availability of charging facilities, nighttime illumination conditions, availability of security CCTVs, space occupancy status, and return availability. Target motorcycles may refer to motorcycles selected as actual allocation candidates in correspondence with service request information and driver information among the above operational resource information, and target inbound / outbound spaces may be any one of a storage space, parking space, charging space, or designated location where the target motorcycles are released and returned. For example, if a driver requests a rental with the conditions of a 125 cc scooter, pickup at 2:00 PM, and near a specific subway station, the server can search for a 125 cc scooter within a 500 m radius of the subway station that currently has a battery charged to 80% or more, is in good condition, and has no conflicts with reservations at the same time, and select it as the final target motorcycle and target pickup space. Additionally, even if there are three motorcycles available at the same time, if one has a low battery charge, another has a tight scheduled return time, and the remaining one best matches the requested distance, the server can select the remaining one as the target motorcycle and also select a space closest to the scheduled return area that has no conflicts with occupancy at the same time as the target pickup space.

[0040] The step of calculating rental condition information is a step of automatically determining approval conditions and cost conditions to be applied to the actual service contract by combining driver information, motorcycle usage request information, grade information of the target motorcycle, target entry / exit space information, and pre-set operation policy information.

[0041] Rental condition information is not limited to simple usage fee information but may be a comprehensive set of conditions including usage approval status, usage fees, warranty conditions, insurance coverage conditions, vehicle entry / exit processing conditions, speed limit conditions, permission for night driving, whether special clauses are added, and a liability sharing structure in the event of an accident. Class information of the target motorcycle may include vehicle group, displacement range, whether it is electric or internal combustion engine, body weight, new vehicle or depreciation stage, and insurance risk group classification. Operation policy information may include basic rental unit prices by time of day, approval criteria by vehicle type, minimum required driving experience by age group, insurance policies by risk grade, surcharge rates by time of day during adverse weather conditions, restricted driving conditions by region, deposit policies, and promotion rules. Through this step, the server determines whether the driver is suitable to operate the target motorcycle and grants final usage approval, calculates the basic usage fee based on the estimated rental time, and may generate rental condition information including the self-damage deductible to be borne by the driver in the event of an accident, the deposit amount, permission for night driving, and vehicle entry / exit processing conditions. For example, even for the same medium-sized motorcycle, conditions with a low deposit and wide insurance coverage may be provided to users with a low risk level, while rental condition information including an additional deposit, increased deductible, and speed limit conditions may be provided to users with a high risk level. In this way, at this stage, rather than simply referencing each of the above input values, the final conditions can be determined in the form of a package by considering the driver risk information described later together with the rental fee and insurance coverage conditions.

[0042] The step of calculating driver risk information is a detailed process for calculating the rental condition information, which involves analyzing static characteristics corresponding to a certified driver and dynamic situations corresponding to the current time of the request for use together to quantify the level of relative accident risk expected when the driver uses the currently requested target motorcycle.

[0043] Here, driver risk information may be a set of values ​​directly used for subsequent billing and insurance calculations, such as the final risk score, risk grade, contribution by risk factor, base driving risk score, weighting of accident exposure frequency per hour, environmental risk surcharge factor, and model non-conformity penalty factor. Being based on the above driver information and the above motorcycle usage request information may mean that it does not merely look at the driver's past records, but also reflects the currently requested time of day, purpose of use, planned driving area, estimated distance, and suitability with the performance group of the target motorcycle. For example, a higher risk score may be calculated when the same user requests a long-distance drive including an accident-prone area for delivery during late-night hours, compared to when they request a short-distance trip for non-commercial use in a flat area during daytime hours. The driver risk information generated at this stage may be used as input values ​​for generating rental fee adjustment factors, insurance premium adjustment factors, deductible levels, and conditional special clauses in subsequent stages.

[0044] The step of calculating rental fees and insurance coverage conditions involves calculating the basic rental fee and basic insurance premium by combining the driver risk information, the grade information of the target motorcycle, and the operational policy information, and then determining the final cost and insurance conditions by applying risk-based correction values ​​and operational environment correction values. For example, the server can first derive the basic rental fee and basic insurance premium based on the basic depreciation rate of the target motorcycle, the standard accident exposure by vehicle group, and the standard demand coefficient by time of day. Subsequently, a dynamic rate weight is calculated by applying a sigmoid function-based transformation to the final risk score, and this can be branched and applied as the rental fee correction coefficient and the insurance premium correction coefficient, respectively. Here, the sigmoid function may be a non-linear normalization function designed to induce gradual cost changes in the low-risk score range and to cause costs to increase sharply in the range above a certain risk score. In addition, if at least one of the basic driving risk score, the weighting of accident exposure frequency per hour, the environmental risk surcharge factor, and the model non-conformity penalty factor exceeds a threshold, the minimum applicable value of the own vehicle damage deductible, the setting of the deposit addition ratio, the determination of whether to reduce the personal injury and property damage coverage limits, and the decision of whether to activate the conditional exemption clause for high-risk driving may be performed simultaneously. Along with this, if the risk grade of the above driver risk information corresponds to a management requirement grade, conditional special clause data may be generated, including conditions for remote activation of the maximum speed limit function, conditions for limiting rapid acceleration, conditions for limiting the permitted time for night driving, or conditions for shortening the driving monitoring cycle when entering accident-prone zones or steep slope zones. For example, if a user's final risk score is calculated as 0.82 and a request for late-night operation during a rain forecast period is included, the server can calculate the adjusted rental fee and adjusted insurance premium by applying a risk-based correction rate and an environment-linked correction rate based on a basic rental fee of 20,000 won and a basic insurance premium of 6,000 won, and determine rental condition information including an additional deposit and speed limit conditions.

[0045] The step of calculating rental condition information may be a step of processing so that the above-mentioned corrected rental fee and corrected insurance premium, as well as the deposit and deductible amount information reflecting the insurance guarantee condition correction coefficient, conditional special clause data, inbound / outbound processing conditions and approval maintenance conditions are combined into one final condition set.

[0046] In other words, this step is not merely about calculating two amounts, but rather about digitizing and finalizing the entire set of service operation rules that will actually be applied between the driver and the platform. Accordingly, the server can notify the driver of the possibility of approval and simultaneously provide a comprehensive package including the target motorcycle, target pick-up / drop-off space, available time range, fees, insurance, deposit, driving restrictions, and post-return liability structure. Based on this structure, the short-to-medium-term motorcycle rental service can be implemented not as a simple resource allocation service, but as a dynamic contractual platform service based on driver risk analysis.

[0047] Additionally, the step of calculating the driver risk information comprises: a step of extracting driver static profile data from the driver information, including the driver's age, license type, elapsed time since obtaining the license, past driving mileage by motorcycle displacement range, insurance accident history, traffic law violation history, number of rental uses during a recent predetermined period, past return delay history, payment failure history, and whether a commercial delivery platform account is linked; a step of extracting region information from the motorcycle usage request information, including the road type, road gradient, intersection density, inclusion of speed limit zones, inclusion of areas prone to rain, inclusion of areas with poor nighttime illumination, inclusion of steep slope areas, inclusion of areas prone to poor road surface conditions, and inclusion of areas prone to motorcycle accidents, wherein the planned usage area consists of a planned departure area, a planned return area, and a planned driving area; and a step of generating dynamic context data including the region information, the purpose of use, the planned rental start time, the planned rental end time, the planned usage time period, and the estimated driving distance, by extracting the purpose of use, the planned rental start time, the planned rental end time, the planned usage time period, and the estimated driving distance from the motorcycle usage request information. A step of calculating a model proficiency index indicating the degree of experience the driver has actually had in driving a vehicle of the same or similar performance group as the target motorcycle by correlating and comparing the driver’s past driving mileage by motorcycle displacement range with the engine displacement, body weight, vehicle type, drive system, and maximum output information of the target motorcycle; and a step of generating a data set for risk analysis by mapping accident correlations, severe accident contributions, exposure risks by time period, and risk coefficients by region for each element of the driver static profile data and each element of the dynamic context data by referring to a pre-established motorcycle accident factor analysis database, and configuring the mapping results into a numerical vector having an array form of risk scores and risk coefficients by element.Regarding the above-mentioned risk analysis dataset, if the driver’s commercial delivery platform account linkage is true, a commercial operation weight corresponding to the frequency of repeated stops and starts, the possibility of multiple short-distance drives, and the proportion of driving during peak hours is assigned, and if the purpose of use is non-commercial, a non-commercial operation weight corresponding to personal movement patterns is assigned; if the above-mentioned planned driving area includes at least one of a rain-prone area, a night illumination-vulnerable area, a steep slope area, a road surface-poor area, and a motorcycle accident-prone area, a step of assigning an environmental risk correction coefficient to the area-related elements included in the above-mentioned dynamic context data; and inputting the above-mentioned risk analysis dataset, the model proficiency index, the commercial operation weight or the non-commercial operation weight, and the environmental risk correction coefficient into a pre-trained machine learning-based risk prediction model to calculate a risk component value including a basic driving risk score, an hourly accident exposure frequency weight, an environmental risk surcharge coefficient, and a model non-conformity penalty coefficient. and may include the step of combining the above risk component values ​​to generate driver risk information including a final risk score corresponding to the driver, a risk grade, and a contribution by risk factor.

[0048] The step of extracting driver static profile data is to collect fixed or semi-fixed risk factors that do not change drastically over time from a database among the unique history and characteristics of a driver registered on the platform, and to structure them into basic input data representing the driver's basic risk propensity.

[0049] Here, driver static profile data may refer to a set of characteristic values ​​that are attributed to the driver themselves over a long period or accumulated and managed regardless of current rental requests. For example, the driver's age can be used as a basic variable to distinguish between novice and skilled driver groups, and the type of license can be used as a qualification variable indicating the range of motorcycles that can be operated, such as a moped license or a Class 2 small motorcycle license. Additionally, the time elapsed since obtaining the license can be a variable indicating the potential for accumulating actual road driving experience, and past driving mileage data by motorcycle displacement range can be a key input value for determining whether the driver has primarily operated small motorcycles of 125 cc or less, or if their experience includes driving large manual motorcycles of 500 cc or more. Insurance accident history, traffic law violation history, number of rental uses during a recent specified period, past return delay history, and payment failure history can be used as quantitative indicators representing the driver's driving tendencies, punctuality tendencies, and transaction stability, and whether a commercial delivery platform account is linked can be used as an attribute value indirectly suggesting the possibility that the driver has a commercial driving pattern involving frequent stop-and-go, a possibility of a rapid increase in driving distance within a short period, or a tendency toward reckless driving. At this stage, the server may not simply store each of the above items as raw values, but rather query them in correspondence with the driver identification key in the database and perform processes such as missing value correction, categorization, normalization, or coding for risk analysis to convert them into driver static profile data in a form suitable for subsequent analysis.For example, if a driver is 24 years old, has been licensed for 14 months, has a driving history of 1,800 km in the 50 cc to 125 cc range over the past 2 years, has 1 accident, 2 traffic violations, has used rental services 11 times in the past 3 months, has delayed return 1 time, has 0 payment failures, and is linked to a delivery platform account, the server can extract each of the above items and configure them into a single static profile record.

[0050] The step of extracting regional information is a step of retrieving data from a map database, road attribute database, weather risk map, accident statistics map, or spatial information system to convert it into structured regional information, based on the planned departure area, planned return area, and planned driving area included in the motorcycle usage request information, in order to finely identify the environmental risk of the physical space where the driver is highly likely to actually operate.

[0051] Here, the term "area of ​​planned use" may not refer to a single point, but may be a concept encompassing the starting point, destination, the planned driving section between them, the optimal expected route, or a driving range pre-specified by the driver. The area information may be data that quantifies or classifies motorcycle-specific risk factors existing within the aforementioned area of ​​planned use in the form of flags, and may include road type, road gradient, intersection density, inclusion of speed limit zones, inclusion of areas prone to frequent rain, inclusion of areas vulnerable to nighttime illumination, inclusion of steep slope zones, inclusion of areas prone to frequent poor road surfaces, and inclusion of areas prone to motorcycle accidents. Road type may be a distinguishing value indicating the basic characteristics of the driving environment, such as urban arterial roads, residential side roads, mountain roads, industrial complex roads, or underpass entrances. Intersection density may be a value calculated based on the number of intersections within a certain radius or driving distance, and inclusion of areas vulnerable to nighttime illumination may be a binary variable indicating whether sections with low public lighting density or illumination measurements below a standard value are included. Due to the physical characteristic of having two wheels, two-wheeled vehicles are much more vulnerable to road surface conditions or gradients than four-wheeled vehicles; therefore, whether a region includes areas with frequent poor road surfaces or steep slopes can serve as a very important variable in estimating the probability of an accident. For example, if the planned departure area is an urban commercial district and the planned return area is a hilly residential area, and the planned driving route includes steep roads and accident-prone intersections, the server can identify the planned usage area from the above request and extract multiple corresponding regional risk attribute values ​​to form a regional information record.

[0052] The step of generating dynamic context data is to extract the purpose of use, the scheduled start time of the rental, the scheduled end time of the rental, the scheduled time of use, and the estimated driving distance from the motorcycle usage request information, and to combine this with the regional information to form a dynamic input data set representing the driving situation at the specific point in time and on the specific route.

[0053] Here, dynamic context data refers to a set of information that represents the temporal, spatial, and purposeful driving context, which varies according to the current request, rather than characteristics fixed to the driver itself. The purpose of use can be categorized into delivery, commuting, personal errands, sightseeing travel, or test drives, and the scheduled time of use can be expressed as categorical information that directly influences changes in risk, such as daytime, peak commuting hours, late-night hours, or rain forecast hours. The estimated driving distance may be a value representing the driving intensity likely to be consumed in the aforementioned request, and when combined with the aforementioned regional information, it can be interpreted as context data reflecting actual driving difficulty rather than simple distance information. For example, even for the same distance of 35 km, the meaning of risk may differ between driving on a flat urban arterial road during the day and driving on a mountain slope section late at night; therefore, the server can generate unique situational data specific to this particular drive by combining the time of day, regional characteristics, and purpose of use. In other words, even for the same driver, the risk of accidents can be completely different depending on whether they are driving for leisure in the suburbs within 50 km during the daytime on weekends or driving for short-distance delivery purposes while passing through rain-prone areas during the late-night hours on weekdays; therefore, dynamic context data can be generated by fusing each of the above elements. The dynamic context data generated at this stage can be combined with a model proficiency index in a subsequent stage or used as an input value for a machine learning-based risk prediction model.

[0054] The step of calculating the model proficiency index is a step of quantifying the degree of experience the driver has actually experienced driving a vehicle of the same or similar performance group as the target motorcycle to be currently assigned by correlating and comparing the driver's past driving mileage by motorcycle displacement range with the displacement, body weight, vehicle type, drive system, and maximum output information of the target motorcycle.

[0055] Here, "corresponding and comparing" means not merely comparing a single displacement value, but calculating a multi-dimensional performance group similarity by considering displacement ranges, body weight, whether it is a scooter or a manual type, whether it is electric motor-driven or internal combustion engine-driven, and the maximum output level together. The model proficiency index may be an indicator converted into a percentage or a specific score range that measures how closely the physical and mechanical characteristics of the target motorcycle to be rented match the specifications of vehicles the driver has previously driven. For example, if a driver who has operated only 110 cc scooters for a long period requests a 300 cc internal combustion engine manual motorcycle, a low model proficiency index may be calculated; conversely, if a driver with experience repeatedly driving electric motorcycles with output and body weight ranges similar to a 125 cc electric scooter requests an equivalent electric motorcycle, a high model proficiency index may be calculated. As another example, if a driver's past driving mileage is 100% concentrated on low-displacement automatic scooters, whereas the target motorcycle to be picked up is a high-output manual motorcycle with a displacement of 1000 cc, a body weight of 250 kg, and a manual transmission drive system, the mismatch between the past experience and the target vehicle's performance group is very large, so the model proficiency index may be calculated as a very low value, such as 0.1. The server can calculate the above similarity in the form of a weighted sum of displacement suitability, weight suitability, vehicle type suitability, drive system suitability, and output suitability, and finally generate the model proficiency index by normalizing it to a continuous value between 0 and 1 or between 0 and 100. The above model proficiency index is a value that reflects the actual level of skill required for the driver to operate the requested target motorcycle, and can be used as an input value or correction criterion for the model incompatibility penalty coefficient in a subsequent step.

[0056] The step of generating a data set for risk analysis involves mapping accident correlations, severe accident contributions, exposure risks by time period, and risk coefficients by region for each element of the driver static profile data and each element of the dynamic context data by referring to a pre-established motorcycle accident factor analysis database, and configuring the mapping results into numerical vectors having the form of arrays of risk scores and risk coefficients for each element, thereby processing the collected qualitative and quantitative raw data into a standardized mathematical form that a machine learning model can immediately compute.

[0057] The existing database for analyzing motorcycle accident factors may be a repository that statistically analyzes the impact of specific factors on accident frequency or severity based on hundreds of thousands of past actual motorcycle accident cases, rental operation cases, legal violation cases, and insurance claim cases. Additionally, mapping may refer to the process of connecting each input element to corresponding risk parameters stored in the database. For example, being in one's early 20s may correspond to a specific accident correlation value, having held a license for less than one year may correspond to a relatively high value in terms of contribution to severe accidents, and late-night hours may correspond to a weighted coefficient in terms of exposure risk by time of day. Furthermore, if the inclusion of an accident-prone area is true, the regional risk coefficient may be mapped upward, and in the case of a delivery request, a high value for exposure risk due to repeated stopping and starting may be assigned. A numeric vector refers to textual and categorical elements, such as early 20s, sports replica car models, and late-night hours, being replaced with numbers in the form of a one-dimensional array, such as a risk score of 7, an accident correlation of 0.8, and an exposure risk of 1.2, respectively. The server can construct the numeric vector by placing the mapping results for each of the above elements into individual dimensions of the vector, such as the first dimension being an age-based risk score, the second dimension being a license history-based risk score, the third dimension being an accident history-based risk coefficient, the fourth dimension being a late-night driving coefficient, and the fifth dimension being an accident-prone area coefficient. The risk analysis dataset generated in this way can serve as a normalized feature vector suitable for subsequent weighting and input to a machine learning model.

[0058] The step of assigning commercial operation weights or non-commercial operation weights is a step that allows differences in actual usage patterns to be reflected even for the same basic risk vector by adding additional interpretation values ​​for each operation purpose to the data set for risk analysis.

[0059] If the driver's commercial delivery platform account is linked, the server may assign a commercial operation weight corresponding to the frequency of repeated stops and starts, the possibility of multiple short-distance trips, and the proportion of driving during peak hours. Here, the frequency of repeated stops and starts may be a value quantifying the probability of repeated stops and restarts per unit of time, and the possibility of multiple short-distance trips may be a value indicating the probability of repeated short-distance travel. The commercial operation weight may be a penalty coefficient in the nature of an additive point that quantifies driving environments unique to delivery drivers, such as frequent rapid acceleration in alleyways, the risk of passing traffic signals recklessly, and the frequency of repeated stops and starts in the middle of the road. Conversely, if the purpose of use is non-commercial, a non-commercial operation weight corresponding to personal movement patterns may be assigned; this may be a value reflecting usage patterns with relatively fewer stops and starts or lower time pressure, such as repeated long-distance commuting, simple weekend travel, or sightseeing trips. For example, even for the same 20 km driving request, if a commercial delivery account is linked, a commercial driving weighting value greater than 1 may be assigned, which can increase the subsequent risk score, whereas for non-commercial daytime driving for commuting purposes, a non-commercial driving weighting value may be applied at a relatively lower level. This allows for protection against the phenomenon where general rental rates are abused for delivery purposes, causing a surge in the platform's overall accident rate.

[0060] The step of assigning an environmental risk correction factor is a step of additionally correcting the risk of region-related elements included in the dynamic context data to reflect external road environment and weather risks beyond the driver's control in the final risk calculation process, when the planned driving area includes at least one of a rain-prone area, a night illumination-vulnerable area, a steep slope area, a road surface-poor area, and a motorcycle accident-prone area.

[0061] Here, the environmental risk correction factor is a coefficient that reflects the amplifying effect of region-based risks on the actual probability of an accident. It can also serve as a multiplicative factor that adds to the risk level when physical environmental obstacles—such as those that drastically reduce motorcycle tire grip or obstruct the rider's vision—are present along the driving path. For example, a planned driving area containing both steep slopes and frequent poor road surface conditions during late-night hours may be assigned a higher environmental risk correction factor compared to flat urban roads during the day. The server can assign individual coefficients based on the inclusion of each risk zone and then integrate them using summation or multiplication. In cases where multiple risk zones are included simultaneously, cumulative addition correction or upper limit restriction correction may be applied. For instance, if both the inclusion of a frequent rain zone and a nighttime illumination-poor zone are true, a composite correction factor reflecting mutual synergy is assigned instead of simply summing the individual coefficients, thereby more realistically reflecting the actual increase in risk. In addition, if a steep section with a gradient of 15% or more overlaps with an illuminance-vulnerable area due to a lack of streetlights at night in the driver's planned driving route, the environmental accident risk can be numerically amplified by multiplying the elements related to that area by an environmental risk correction factor of 1.5 or more.

[0062] The step of calculating risk component values ​​is to input the above-mentioned risk analysis dataset, model proficiency index, commercial operation weight or non-commercial operation weight, and environmental risk correction coefficient into a pre-trained machine learning-based risk prediction model to calculate risk component values ​​including a basic driving risk score, an hourly accident exposure frequency weight, an environmental risk surcharge coefficient, and a model non-conformity penalty coefficient.

[0063] Here, the machine learning-based risk prediction model may be a classification model or a regression model, such as a multi-layer perceptron or an ensemble learning model, trained on historical rental driving data, accident data, traffic violation data, regional road risk data, and insurance claim data. The basic driving risk score may be a value representing the driver's inherent tendency to cause accidents, and the hourly accident exposure frequency weight may be a value indicating how high the risk exposure is per unit of time during the scheduled usage period. The environmental risk surcharge coefficient may represent the increase in external environmental risk based on the aforementioned planned driving area and time of day, and the vehicle incompatibility penalty coefficient may be a penalty value that increases as the vehicle proficiency index decreases. At this stage, the artificial intelligence algorithm can precisely calculate the non-linear interactions between numerous data variables collected and corrected from various angles. For example, in a situation where a novice driver rents a high-displacement vehicle and the rain-prone area and late-night hours overlap simultaneously, the model can identify synergistic risks that increase exponentially rather than simply adding up the risk levels. In addition, among the calculated risk component values, the basic driving risk score refers to the intrinsic and statistical risk of the driver itself, and the model incompatibility penalty coefficient may refer to a value that independently quantifies the risk of inexperienced rollover caused by the discrepancy between the driver's past experience and the performance of the rental vehicle. For example, if the base risk value of the risk analysis dataset is high, the commercial operation weight is 1.35, the environmental risk correction coefficient is 1.2, and the model proficiency index is 0.42, the server can input the above input values ​​into a model to calculate risk component values ​​such as a basic driving risk score of 0.61, an hourly accident exposure frequency weight of 1.18, an environmental risk surcharge coefficient of 1.24, and a model incompatibility penalty coefficient of 0.27.

[0064] The step of generating driver risk information is to calculate comprehensive risk information including a final risk score corresponding to the driver, a risk grade, and a contribution by risk factor by combining the above risk component values.

[0065] The final risk score may be a representative value combining the above-mentioned basic driving risk score, the weighting of accident exposure frequency per hour, the environmental risk surcharge factor, and the model non-conformity penalty factor; for example, it may be calculated as a continuous value between 0 and 1 by normalizing after weighted summation or by compressing it using a non-linear function. The risk grade may be a categorical result obtained by correlating the above-mentioned final risk score with a number of pre-set standard intervals to classify it into low-risk, medium-risk, high-risk, or management requirement grades. The contribution by risk factor refers to an explainability indicator representing how much each element influenced the formation of the above-mentioned final risk score, and may be a value generated in the form of explainable artificial intelligence data. For example, it may be calculated in a manner such as 18% for the late-night time factor, 22% for the accident-prone area factor, 15% for the delivery purpose factor, and 20% for the lack of model proficiency factor; this can be utilized as supporting data to objectively and transparently inform drivers of the reasons for rental fee and insurance premium surcharges in the future. In addition, the final risk score can be calculated as a score out of 100, and the risk grade can be categorized into alphabetical grades such as S, A, B, and C. The driver risk information generated at this stage can be used as a direct input value for calculating rental fee adjustment factors, insurance premium adjustment factors, security deposit surcharge ratios, deductible levels, and conditional special clause data in subsequent stages. In other words, this stage is not merely a procedure for evaluating risk, but can be a stage for generating key judgment information to dynamically determine the overall contract conditions thereafter.

[0066] And, the step of calculating the rental fee and insurance application conditions comprises: a step of deriving a basic rental fee and a basic insurance premium by reflecting the basic depreciation rate of the target motorcycle, the standard accident exposure level by vehicle group, and the standard demand coefficient by time of day, by referring to pre-established standard billing rate information based on the grade information of the target motorcycle, the purpose of use included in the motorcycle usage request information, the scheduled start time of rental, the scheduled end time of rental, the scheduled time of use, and the estimated mileage; a step of calculating a dynamic rate weight by applying a sigmoid function to the final risk score included in the driver risk information, and calculating a rental fee correction coefficient and an insurance premium correction coefficient applied to the basic rental fee and the basic insurance premium, respectively, based on the dynamic rate weight; A step of generating an insurance guarantee condition correction coefficient that determines at least one of the following when at least one of the above basic driving risk score, hourly accident exposure frequency weighting, environmental risk surcharge coefficient, and model non-conformity penalty coefficient exceeds a preset threshold: the minimum applicable value of the self-vehicle damage deductible applied in the event of an accident, the deposit addition ratio, whether to reduce the personal injury and property damage coverage limits, and whether to activate a conditional exemption clause for high-risk driving; a step of generating conditional special clause data including a remote activation condition for a maximum speed limit function through the drive control unit of the above target motorcycle, a rapid acceleration limit condition, a night driving allowance time limit condition, and a condition for shortening the driving monitoring cycle when entering a motorcycle accident-prone area or a steep slope area when the risk grade of the above driver risk information corresponds to a preset management requirement grade; and a step of calculating an environment-linked correction rate defined in the above operation policy information when the scheduled rental time of the above target motorcycle corresponds to at least one of the commuting peak time, late-night time, rain forecast time, and time of rapid temperature drop.The method may include: a step of calculating a corrected rental fee and a corrected insurance premium by applying a rental fee correction coefficient and an insurance premium correction coefficient, respectively, to the basic rental fee and the basic insurance premium, and then applying an environment-linked correction rate; and a step of determining rental condition information by combining the deposit and deductible amount information reflecting the insurance guarantee condition correction coefficient, the conditional special clause data, and the corrected rental fee and the corrected insurance premium.

[0067] The step of deriving the basic rental fee and basic insurance premium is a step of calculating the basic rental fee and basic insurance premium, which are standard prices before the driver's individual risk level is reflected, in the system by referring to established standard charging rate information based on the grade information of the target motorcycle, the purpose of use included in the motorcycle usage request information, the scheduled start time of rental, the scheduled end time of rental, the scheduled time of use, and the estimated driving distance.

[0068] Here, the standard billing rate information refers to a standard rate table pre-mapped in a database by the platform operator based on the type of vehicle and rental environment. The basic depreciation rate for the target motorcycle may be a figure representing the percentage of the statistical decrease in vehicle value that occurs as the motorcycle is driven, and it may be set higher for motorcycles with larger displacement or higher-priced brands. The standard accident exposure rate by vehicle group may refer to the average probability of accidents calculated based on statistical data or insurance statistics for each vehicle type and displacement group, such as scooters, manual, touring, or electric motorcycles. The standard demand coefficient by time of day may be a constant used to flexibly apply rates by reflecting whether the rental start time is a peak period with high demand or a relatively quiet idle period. The server can derive a billing amount based on pure material criteria, excluding individual driver factors, by synthesizing these objective indicators. For example, even for the same 125 cc class scooter, the calculation results for the basic rental fee and basic insurance premium may differ depending on whether it is rented for 6 hours during peak commuting hours or for 2 hours during idle hours on weekday mornings. In other words, this step may be a step for calculating an initial charge based on the pure decrease in asset value, the average accident exposure by vehicle class, and the demand environment by time of day, while excluding the individual driver's risk propensity.

[0069] The step of calculating dynamic rate weights, rental fee correction coefficients, and insurance premium correction coefficients is a step of generating numerical coefficients to reflect the individual driver's risk level in the form of a multiplication operation to the previously calculated base rate.

[0070] Applying a sigmoid function to the final risk score included in the above driver risk information may mean controlling the extent of risk reflection using an S-shaped non-linear mathematical function that transforms an input arbitrary real value into a certain range between 0 and 1 or a corresponding limit range. Accordingly, even if the final risk score increases linearly, it is possible to prevent the rate weight from exponentially skyrocketing or turning negative, and to control the correction rate to converge smoothly within the limits of the maximum surcharge rate and maximum discount rate specified by the platform. The server can generate the above dynamic rate weight separately as correction coefficients corresponding to the basic rental fee and basic insurance premium, respectively, thereby allowing different reflection ratios to be applied to the rental fee and insurance premium even for the same risk score. For example, a rental fee correction coefficient of 0.8 and an insurance premium correction coefficient of 0.85 may be assigned to a prime driver with a very low final risk score to apply a discount effect, while a rental fee correction coefficient of 1.3 and an insurance premium correction coefficient of 1.5 may be assigned to a driver with a high risk score to apply a surcharge. Through this structure, the server can precisely reflect the individual driver's risk level in the fee system while distinguishing and controlling insurance risk and asset usage risk.

[0071] The step of generating an insurance guarantee condition adjustment coefficient is a step of generating a control value to reduce the scope of insurance coverage or increase the collateral burden for high-risk situations where it is difficult to sufficiently defend against the platform's financial risk by raising fees alone.

[0072] Here, the case where at least one of the base driving risk score, the weighting of accident exposure frequency per hour, the environmental risk surcharge factor, and the model non-conformity penalty factor exceeds a preset threshold may refer to a state where there is concern that the probability of an accident or the scale of loss in the event of an accident may exceed the platform's permissible limits, rather than a simple average level of risk. The minimum applicable value for the own vehicle damage deductible may refer to the lower limit of the minimum amount that the driver must mandatorily bear out of pocket when the motorcycle is damaged in an accident, and this lower limit amount may be significantly increased as the risk level increases. The deposit surcharge ratio may be a percentage added to the base deposit, and the reduction of bodily injury and property damage coverage limits may be a value indicating whether to limit the maximum compensation amount borne by the insurer. Additionally, the activation of the high-risk driving conditional waiver clause may be data indicating whether to incorporate a special clause into the rental contract that restricts all or part of insurance coverage in the event of an accident under specific risk conditions, such as when the model non-conformity penalty factor is very high or when driving in a high-risk area late at night. For example, if the model non-conformity penalty coefficient and the environmental risk surcharge coefficient simultaneously exceed a threshold, the server may generate insurance guarantee condition adjustment coefficients by raising the minimum applicable deductible value, increasing the deposit surcharge ratio, and activating high-risk conditional exemption clauses. As such, this step may be an adjustment step intended to conservatively redesign the loss-sharing structure itself in the event of an accident, rather than a simple fee surcharge.

[0073] The step of generating conditional special clause data is a step of reflecting special clauses related to vehicle control and driving restrictions in the contract conditions in response to the driver's risk level when the risk grade of the driver risk information corresponds to a pre-set management requirement grade.

[0074] Here, the "management requirement grade" may refer to a grade classified for driver groups with a high probability of immediate accidents due to an overlap of high-risk environmental factors and driver risk factors, or cases involving a history of multiple traffic violations or attempts to rent unskilled models. If the above risk grade corresponds to the management requirement grade, the server may generate conditional special agreement data including conditions for remote activation of the maximum speed limit function via the target motorcycle's drive control unit, conditions for limiting rapid acceleration, conditions for limiting permitted night driving hours, and conditions for shortening the driving monitoring cycle when entering a motorcycle accident-prone area or a steep slope area. The conditions for remote activation of the maximum speed limit function and the rapid acceleration limit may be special agreements for consent to forcibly software-lock the throttle opening amount or maximum output limit by communicating with the motorcycle's ECU upon signing the rental contract. The condition for shortening the driving monitoring cycle may refer to a special clause that mandates precise control by changing the transmission frequency of GPS location and speed data—which is normally sent to the server every 10 seconds—to every 1 second upon entering an accident-prone area, thereby increasing the server's data communication volume. For example, if a driver requests a high-displacement vehicle on a route including steep slopes during late-night hours and the risk level is determined to be at a management requirement level, the server may generate conditional special clause data for said driver that includes a maximum speed limit of 80 km / h, restrictions on rapid acceleration, a reduction in permitted night driving hours, and a shortened monitoring cycle upon entering a high-risk section. In other words, this step may be a stage that contractually grants hardware control authority when the driver's risk level is at its worst.

[0075] The step of calculating the environment-linked adjustment rate is intended to additionally reflect macroscopic risks resulting from external environmental variability and factors of demand surges in the fare, rather than the driver's personal risk.

[0076] Here, if the scheduled rental time of the target motorcycle corresponds to at least one of the following: peak commuting hours, late-night hours, rain forecast hours, or hours of rapid temperature drop, the server may calculate an environment-linked correction rate defined in the operational policy information. Peak commuting hours may refer to a time period where road congestion and demand concentration occur simultaneously, and late-night hours may refer to a time period where external risk factors, such as reduced lighting and an increase in drunk driving vehicles, increase. Rain forecast hours refer to a time period with a high probability of precipitation, which may reflect an environment where road grip is reduced and braking distance increases. Hours of rapid temperature drop may refer to specific time periods linked with public data from the Korea Meteorological Administration, where temperatures drop sharply below 0°C or there is a high probability of black ice formation; during these times, road friction is extremely reduced, which can lead to a sharp increase in the probability of accidents. The environment-linked correction rate may refer to an additional surcharge percentage applied to protect against asset loss of the platform during such temporary and rapid external environmental changes, while simultaneously responding to a surge in demand. For example, if the late-night hours and the rain forecast hours overlap at the same time, the server can calculate an environment-linked correction rate that reflects both environmental factors in combination and apply it to the subsequent billing stage.

[0077] The step of calculating the adjusted rental fee and adjusted insurance premium is to calculate the actual amount to be charged by multiplying the previously derived basic amount by the individual driver's risk coefficient and finally adding an environment-linked adjustment rate.

[0078] The server can apply a rental fee correction factor to the base rental fee and an insurance premium correction factor to the base insurance premium, and then apply an environment-linked correction rate to each of the calculated results. While the order of operations may vary depending on the operational policy, a sequential multiplication structure may be adopted where external environment correction is added after reflecting the individual driver's risk level. For example, if the base rental fee is 10,000 won, the rental fee correction factor is 1.2, and the environment-linked correction rate is 1.1, the calculation process can be performed in such a way that the final corrected rental fee is determined as 13,200 won by sequentially multiplying 10,000 won by 1.2 and 1.1. Similarly, if the base insurance premium is 6,000 won, the insurance premium correction factor is 1.4, and the same environment-linked correction rate is 1.1, the corrected insurance premium can be calculated as 9,240 won. According to this method, even for the same vehicle and usage time, differences in driver risk and environmental risk can be finely reflected in the actual charges.

[0079] The step of finalizing rental condition information is a step of completing the final contract data set by combining the deposit and deductible amount information reflecting the insurance guarantee condition correction coefficient, the conditional special clause data, and the correction rental fee and correction insurance premium, and bundling all billing data, rights and obligations data, and hardware control constraints derived by separate calculations into one package.

[0080] Here, rental condition information may refer to comprehensive contract information including the cost burden relationship, profit distribution relationship, scope of usage liability, and scope of compensation liability in the event of an accident between the borrower, the motorcycle owner, the space owner, and the platform. The finalized rental condition information may be transmitted from the server to the driver's terminal, and the process may proceed to the physical rental approval, unlocking, or delivery procedure of the target motorcycle by the driver approving the conditions through electronic signature and payment on the application. For example, the server may generate rental condition information for a driver that includes adjusted rental fees, adjusted insurance premiums, a deposit, a deductible, a maximum speed limit special clause, and nighttime driving restriction conditions, display this on the driver's terminal, and perform the final usage approval procedure only when the driver's electronic consent is obtained. As such, this step is not merely a procedure of listing a few amounts, but may be a step that completes the final contract structure that takes actual effect by integrating the risk analysis results, physical asset conditions, external environmental conditions, and insurance constraints.

[0081] According to one embodiment, the method comprises: a step of collecting road surface environment data obtained from a front recognition sensor of a target motorcycle and body attitude data obtained from a 6-axis or 9-axis inertial measurement unit (IMU) in real time; a step of analyzing the road surface environment data and body attitude data to calculate motorcycle-specific accident occurrence probability information, including the risk of a stoppie rollover due to straight-line braking of the target motorcycle and the risk of a slip (low-side or high-side) due to the banking angle during cornering; and a step of performing driving assistance control by intervening in the electronic control unit (ECU) of the target motorcycle to vary the braking force distribution ratio between the front and rear wheels or dynamically limit the throttle output of the driving power source when the motorcycle-specific accident occurrence probability exceeds a preset active control threshold. The present invention provides a driving assistance method specialized for two-wheeled vehicles for accident prevention, comprising the step of recording an ADAS event log including the intervention history of the driving assistance control, banking angle data at the time of intervention, and impact amount data.

[0082] The step of collecting road surface environment data and vehicle body posture data in real time is a step of continuously acquiring basic sensor data to identify the surrounding environment and the physical behavioral state of the vehicle while the target motorcycle is driving, and providing it as an input value for subsequent risk prediction calculations.

[0083] The target two-wheeled vehicle may include a scooter, manual motorcycle, electric two-wheeled vehicle, or high-displacement two-wheeled vehicle equipped with or linked to the driving assistance function of the present invention, and may refer to a two-wheeled vehicle currently being driven by the driver. The forward recognition sensor may refer to a group of sensors positioned in front of or in a forward downward direction of the target two-wheeled vehicle to scan road conditions and road surface conditions ahead of the driving path, and may be composed of a camera, an infrared sensor, a lidar, a radar, or a combination thereof. Road surface environment data refers to condition information of the road ahead obtained through the forward recognition sensor, and may include whether the road surface is wet, whether it is frozen, whether potholes exist, whether fine sand or gravel is distributed, whether protruding obstacles exist, lane boundary shape, or road surface reflectivity. Additionally, the 6-axis or 9-axis inertial measurement sensor (IMU) may refer to a hardware module that precisely measures the acceleration in the x-axis, y-axis, and z-axis directions of the target two-wheeled vehicle, as well as roll, pitch, and yaw data, which are rotational angular velocities around each axis. Vehicle body attitude data may refer to dynamic numerical information including pitch angular velocity, roll angular velocity, yaw rate, longitudinal acceleration, lateral acceleration, vertical impact amount, banking angle, vehicle body tilt, degree of tilt, and vibration amount, which are measured or calculated in real time through the IMU. For example, when a target motorcycle enters a curved section at 65 km / h, a server or in-vehicle controller may simultaneously collect minute changes in reflected light from the road surface ahead and changes in the roll angular velocity of the vehicle body and use them as basic data for calculating the slip risk.

[0084] The step of calculating accident probability information specific to two-wheeled vehicles is a step of analyzing the aforementioned road surface environment data and vehicle body posture data to quantify the possibility of a fatal accident that may occur due to the physical instability unique to the target two-wheeled vehicle.

[0085] Here, the jackknife rollover risk refers to the probability that the vehicle body will flip forward as the rear wheels lift off the ground during sudden straight-line braking, due to an excessive increase in front wheel contact load and a rapid decrease in rear wheel contact load. The banking angle refers to the angle at which a motorcycle tilts its body inward to counteract centrifugal force while turning a corner; the low-side slip risk may be a value indicating the possibility of the vehicle body sliding inward and overturning when the tire's grip or frictional force exceeds physical limits while in the banking angle state. The high-side slip risk may be a value indicating the risk of rollover where the tire, which was sliding outward, suddenly regains grip, causing the vehicle body to be violently propelled in the opposite direction of the tilt due to the recoil. Motorcycle-specific accident probability information refers to a multidimensional result calculated by comprehensively determining the actual probability of occurrence of the aforementioned jackknife rollover risk and low-side or high-side slip risks, reflecting the dynamic characteristics of motorcycles, which are extremely sensitive to body tilt and front-to-rear weight transfer unlike four-wheeled vehicles. At this stage, instead of merely determining whether the sensor value exceeds a threshold, real-time accident probability information can be calculated by comprehensively reflecting the degree of road surface friction reduction, the rate of increase in banking angle, lateral behavioral instability, and braking and driving conditions. For example, if a rear wheel load reduction pattern is detected while the pitch rise rate increases rapidly during sudden braking, the risk of a jackknife rollover may increase, and if lateral slippage is detected with an increased banking angle on a wet road surface, the risk of a low-side slip may increase. The accident probability information specific to motorcycles calculated in this way can be used as a direct input value to determine whether to intervene in subsequent driving assistance control.

[0086] The step of performing driving assistance control is a step in which, when the probability of an accident specific to the two-wheeled vehicle exceeds a preset active control threshold, the system intervenes in the electronic control unit (ECU) of the target two-wheeled vehicle to physically avoid an accident or mitigate damage by correcting the driver's operation when the risk of an accident reaches a substantial limit.

[0087] The active control threshold may be a pre-set probabilistic risk value to initiate intervention in driving assistance control, and it may be set variably depending on the vehicle group, vehicle type, road surface conditions, driving speed, or driver mode settings. The electronic control unit may be a device acting as the brain that electronically integrates and controls the engine, brake system, suspension, etc., of a two-wheeled vehicle. Varying the braking force distribution ratio between the front and rear wheels may mean, for example, that even if the driver applies the front brake hard, if a jackknife risk is detected, the system lowers the hydraulic pressure of the front brake caliper and raises the hydraulic pressure of the rear brake to forcibly suppress the pitch rise of the vehicle body and maintain balance. Dynamically limiting the throttle output of the driving power source may refer to an operation that induces the recovery of tire grip by forcibly reducing the driving torque delivered to the rear wheels—such as by reducing the amount of fuel injected into the engine or air intake, or delaying the ignition timing—even if the driver pulls the accelerator lever when a slip risk is detected during cornering. For example, if the banking angle increases sharply and signs of rear-wheel slip are detected while cornering on a wet road surface late at night, the ECU can immediately intervene to restore tire grip by limiting rear-wheel drive torque and correcting cornering braking force. According to this method, unlike a approach that merely provides a warning notification, the vehicle's behavior can be actively stabilized before an accident actually unfolds.

[0088] The step of recording ADAS event logs involves storing key vehicle behavior data at the moment of intervention of the driving assistance control and before and after it in a structured form, similar to a black box, so that it can be utilized for subsequent vehicle status analysis, server-linked return inspection, maintenance determination, or accident cause tracing.

[0089] Here, the intervention history of driving assistance control may be a detailed record indicating when, due to what risk factor, what mechanical control was performed and with what intensity. Impact data may be a three-dimensional numerical value regarding the intensity and direction of the impact calculated by an inertial measurement sensor when an external physical strike is applied to the motorcycle body, such as hitting a bump or falling over in place, despite control intervention. The ADAS event log may be integrated record data that groups these control intervention history, banking angle data at the time of intervention, impact data, front and rear braking force adjustment amounts, throttle output limit amounts, intervention duration, and risk types in chronological order. A server or in-vehicle storage unit stores the above ADAS event log in chronological order and, if necessary, can manage it in conjunction with driver identification information, target motorcycle information, and GPS-based location information. For example, if a specific driver experiences two jackknife rollover risk controls and one cornering slip risk control due to sudden braking during a rental, ADAS event logs including banking angle, impact force, intervention intensity, and location information at the time of each control intervention can be accumulated and stored. This log data can be utilized as core foundational data for analyzing the cause of accidents, as well as for tracking invisible damaged parts of the vehicle or objectively evaluating the driver's driving reliability by linking with a motorcycle rental platform server in the future. Furthermore, the ADAS event logs can be used as supporting data for controlling the PTZ camera to prioritize filming suspected accident areas during unmanned returns, as reference data for additional billing settlements, or as data for updating driver risk history.

[0090] Additionally, the aforementioned front recognition sensor includes: a multi-channel vision sensor and a LiDAR sensor, and the step of calculating the two-wheeled vehicle-specific accident occurrence probability information comprises: identifying hazardous objects including the wetness, freezing, presence of potholes, distribution of fine sand, and presence of protruding obstacles on the road surface ahead by analyzing the road surface texture pattern, light reflectance, point cloud density, point cloud height deviation, and changes in road surface slope with respect to the road surface environment data, and generating a dynamic road surface friction coefficient profile for each section of the driving path based on the location, area, and distribution of the hazardous objects; A step of analyzing longitudinal acceleration information, pitch angular velocity information, and vertical impact information obtained from the above 6-axis or 9-axis inertial measurement sensor, real-time compression displacement data of the suspension stroke sensor equipped on the front fork of the target motorcycle, rotational speed data of the front wheel speed sensor, and rotational speed data of the rear wheel speed sensor to calculate the dynamic load transfer concentrated on the front wheel side during sudden braking, the front wheel contact load increase rate, and the rear wheel contact load decrease rate; a step of simulating the rear wheel lift-off threshold point at which the vertical load on the road surface of the rear tire reaches zero (0) based on the dynamic load transfer, the front wheel contact load increase rate, and the rear wheel contact load decrease rate, and deriving the jackknife rollover risk of the target motorcycle based on the remaining time margin until the threshold point and the body pitch rise rate; A step of analyzing the roll angular velocity, yaw rate, and lateral acceleration information of the above 6-axis or 9-axis inertial measurement sensor and the rotational speed phase difference between the front wheel speed sensor and the rear wheel speed sensor to calculate the real-time banking angle, lateral slip angle, tire friction loss rate, and cornering stability degradation rate;A step of establishing a traction circle limit model acceptable to the tire on the current road surface based on the dynamic road surface friction coefficient profile, and generating a state vector defined by the real-time banking angle, lateral slip angle, and friction force loss rate; a step of calculating a low-side slip risk based on the speed at which the coordinates of the state vector approach the boundary region of the traction circle limit model; a step of calculating a high-side rollover weight in which the vehicle body is thrown off when the tire friction force is instantaneously recovered while the lateral slip angle has developed above a preset threshold, and at least one of a reverse snap peak sign of the yaw rate, vehicle body rebound acceleration, and a pattern of rapid increase in rear wheel rotation speed is detected, and correcting the low-side slip risk based on the high-side rollover weight; and may include the step of combining the jackknife rollover risk, low-side slip risk, and high-side rollover weight with the current vehicle speed, occupant load distribution, load information, and center of gravity elevation displacement coefficient of the target motorcycle to finally calculate the motorcycle-specific accident probability information, in which the real-time accident probability, expected impact location, and risk type are defined as a multidimensional matrix.

[0091] A forward recognition sensor including a multi-channel vision sensor and a LiDAR sensor may be a composite sensor group for mutually complementarily detecting road surface conditions and obstacle conditions present in the forward driving path of a target two-wheeled vehicle.

[0092] Here, a multi-channel vision sensor refers to a device that collects light of various wavelengths, including not only the visible light region but also the infrared region, and acquires visual images through a camera lens; it can be used to identify changes in road surface color, light reflectivity, texture patterns, shadow patterns, gloss characteristics of wet surfaces, and surface contamination status. A LiDAR sensor may be a sensor that emits a laser into the space in front of a target motorcycle and receives reflected waves to acquire information on distance, shape, and elevation difference; it can be used to generate point cloud data to precisely measure irregularities, depressions, protrusions, speed bumps, and changes in surface inclination of the road surface. Since the multi-channel vision sensor is advantageous for precisely identifying color and optical texture, and the LiDAR sensor is advantageous for precisely identifying distance and shape information, a server or in-vehicle controller can recognize the road surface ahead more reliably by using the two types of sensors simultaneously. For example, even in rainy nighttime conditions where it is difficult to clearly distinguish between water film and shadows using only a vision sensor, combining this with LiDAR point cloud height deviation information can more accurately identify actual wet zones and pothole zones.

[0093] The step of identifying hazardous objects by analyzing the texture pattern, light reflectance, point cloud density, point cloud height deviation, and changes in the slope of the road surface and generating a dynamic road surface friction coefficient profile for each section is a step of quantitatively identifying critical risk factors for two-wheeled vehicle driving present on the current driving path and converting them into a numerical value called friction force by analyzing visual features and three-dimensional shape features together with the road surface environment data obtained from the forward recognition sensor.

[0094] Texture patterns can appear as different image patterns on dry asphalt, wet asphalt, gravel-mixed areas, or cracked road surfaces, respectively, and light reflectance can serve as a key indicator for distinguishing between wet or icy conditions of the road surface. Point cloud density may be a value indicating how densely LiDAR reflection points are distributed in space, and point cloud height deviation may be a value indicating the height difference from the reference plane in the presence of potholes or protruding obstacles. Changes in road surface slope may be an indicator of the presence of steep approach ramps, uneven sections, or subsidence sections. In particular, fine sand can be identified as a very significant hazard object because it causes almost a loss of friction during motorcycle cornering. A server or in-vehicle controller can synthesize each of the above features to identify hazard objects, including the wet or icy condition of the road surface ahead, the presence of potholes, the distribution of fine sand, and the presence of protruding obstacles. Subsequently, based on the location, area, and distribution of the aforementioned hazardous objects, the driving path can be divided into multiple detailed sections, and a dynamic road surface friction coefficient profile representing the expected frictional force level for each section can be generated. Here, the dynamic road surface friction coefficient profile refers to the generation of limit values ​​of frictional force that a tire can generate with the ground in the form of a continuous graph or array data, based on the identified hazardous objects, in 1-meter units or specific sections along the planned driving path. For example, a dynamic road surface friction coefficient profile for each section can be formed such that, within the forward 30-meter section, the friction coefficient for the 0-10-meter section is 0.82, for the 10-18-meter section is 0.54 due to accumulated rain, for the 18-22-meter section is 0.41 due to fine sand distribution, and for the 22-30-meter section is 0.79 due to a normal road surface; in the wet road section, a profile may be generated in which the basic friction coefficient of 0.8 is lowered to a level of 0.5.

[0095] The step of calculating the dynamic load transfer amount, the front wheel contact load increase rate and the rear wheel contact load decrease rate by analyzing longitudinal acceleration information, pitch angular velocity information, vertical impulse information, real-time compression displacement data of the suspension stroke sensor, rotational speed data of the front wheel speed sensor and rotational speed data of the rear wheel speed sensor is a step of estimating in real time how the load is transferred between the front and rear axles of the vehicle during sudden braking, and accurately calculating the phenomenon of the center of gravity shifting forward by fusing the vehicle's deceleration state and the degree of physical contraction of the suspension.

[0096] Longitudinal acceleration information can indicate the absolute magnitude of deceleration, pitch angular velocity information can indicate the rotational speed at which the vehicle body tilts forward, and vertical impact information can indicate the instantaneous impact response to uneven surfaces or load changes during braking. A suspension stroke sensor equipped on the front fork refers to hardware that measures, in millimeters, the amount of compression in the suspension—the shock absorber supporting the front wheel of a two-wheeled vehicle—caused by braking force. The real-time compression displacement data from the suspension stroke sensor indicates how much the front fork is compressed during braking, which can indirectly reflect the magnitude of the load concentrated on the front wheel. Additionally, rotational speed data from the front and rear wheel speed sensors can serve as an indicator showing whether there is an imbalance in rotational deceleration between the front and rear wheels during braking. Dynamic load transfer can be a numerical representation of the physical phenomenon where the weight of the vehicle and occupants shifts from the rear wheels to the front wheels due to the law of inertia during sudden braking. When the brakes are applied to a two-wheeled vehicle, pitch occurs and the front suspension is strongly compressed; at this time, the front wheel contact load increases rapidly while the rear wheel contact load decreases, and this step can track this rate of change through real-time calculation. A server or controller can combine the sensor values ​​to calculate a dynamic load transfer amount indicating how quickly the center of gravity of the target two-wheeled vehicle shifts toward the front wheel side due to sudden braking, and simultaneously derive the rate of increase in front wheel contact load and the rate of decrease in rear wheel contact load. For example, if the amount of compression of the front wheel fork increases rapidly at 0.3 seconds after the start of braking and the rotational speed of the rear wheel does not decrease as rapidly as that of the front wheel, the controller can determine that the front wheel contact load is rapidly increasing and the rear wheel contact load is rapidly decreasing.

[0097] The step of simulating the rear wheel lift-off critical point and deriving the jackknife rollover risk based on the remaining time margin and the vehicle body pitch rise rate is a step of predicting a future point in time when the vertical load on the road surface of the rear tires reaches zero due to sudden braking, and quantifying the rollover risk using the time remaining until that point and the forward rotation trend of the vehicle body.

[0098] Here, the rear wheel lift-off critical point may be a mathematically simulated value representing the exact future point in time when the vertical load pressing the rear tires against the ground becomes zero due to the sustained reduction rate of the rear wheel contact load—that is, when the rear wheels effectively begin to lift into the air. The remaining time margin may refer to the time remaining from the current point in time until the aforementioned rear wheel lift-off critical point; a smaller value may indicate an imminent risk of rollover. The body pitch rise rate may be a value indicating how quickly the forward tilt of the vehicle body increases during braking. A server or controller can simulate the rear wheel lift-off critical point by performing a vehicle dynamics model based on the amount of dynamic load transfer, the front wheel contact load increase rate, and the rear wheel contact load reduction rate, and can derive a high risk of jackknife rollover if both the remaining time margin to the critical point and the body pitch rise rate simultaneously satisfy the risk level. For example, if the rate of increase in body pitch, where the rear part of the vehicle body is lifted upward, increases rapidly when the remaining time margin is 0.3 seconds, it may be determined that the risk of jackknife rollover is very high, and if the remaining time margin is 0.18 seconds or less and the rate of increase in body pitch rises rapidly above a preset threshold value, it may be determined that the risk of rollover is extremely high.

[0099] The step of calculating the real-time banking angle, lateral slip angle, tire friction loss rate, and cornering stability degradation rate by analyzing roll angular velocity, yaw rate, lateral acceleration information, and the rotational speed phase difference between the front wheel speed sensor and the rear wheel speed sensor is a step of dynamically analyzing how much the target motorcycle is tilted, how much slipping is progressing, and how much current traction is lost during cornering.

[0100] Roll angular velocity refers to the speed at which the vehicle body leans to the left or right, and by integrating this, the real-time banking angle—the angle at which the two-wheeled vehicle is tilted relative to the road surface—can be calculated. Yaw rate refers to the speed at which the vehicle body attempts to rotate around a vertical axis. Lateral slip angle refers to the angular difference between the target driving direction indicated by the steering angle and the direction in which the vehicle body is actually being pushed forward due to inertia. The rotational speed phase difference between the front and rear wheel speed sensors refers to a phenomenon where the rolling speeds of the front and rear wheels do not match; this data intuitively indicates that excessive driving force has been transmitted to the rear wheels while cornering, causing the tires to slip and spin idly. The tire friction loss rate is a ratio indicating how much the actual grip between the road surface and the tires has decreased compared to a normal state, while the cornering stability degradation rate is an indicator representing the degree to which the vehicle body loses cornering stability and enters a corner unstably. The phase difference in front and rear wheel rotation speeds can indicate how unevenly the rotational behavior of the front and rear wheels proceeds under the same trajectory conditions, and if the rear wheel shows a larger phase change, it can be interpreted as an increased possibility of slip. For example, if the real-time banking angle increases to 38 degrees, the lateral slip angle continuously expands, and the phase difference in rear wheel rotation speeds increases rapidly, the controller can determine that the rate of decline in cornering stability and the rate of loss of friction force are increasing rapidly.

[0101] The step of establishing a friction source limit model acceptable to the tire on the current road surface and generating a state vector based on a dynamic road surface friction coefficient profile is a step for mathematically defining the friction margin range of the current driving section and expressing the real-time behavior of the target two-wheeled vehicle as coordinate values ​​within the model.

[0102] Here, the friction circle limit model refers to a physical model that represents the sum of lateral and longitudinal friction forces—which a tire can withstand without losing grip—as a virtual circle or ellipse. According to the previously calculated dynamic road surface friction coefficient profile, the radius or boundary size of this friction circle is significantly reduced when the road surface is slippery. A state vector can be data that quantifies the real-time banking angle, lateral slip angle, and tire friction loss rate as points on a single multidimensional coordinate system. For example, a 3D state vector can be generated with the banking angle component, slip angle component, and friction loss component each as a single axis. This generated state vector can be used as a reference value to determine whether the target motorcycle is currently located in a stable zone within the friction circle limit model or is approaching an unstable zone near the boundary.

[0103] The step of calculating the low-side slip risk based on the speed at which the coordinates of the state vector approach the boundary region of the friction circle limit model is a step of quantifying the low-side slip risk by considering not only how close the current vehicle state is to the ground contact limit, but also how quickly it is moving toward that boundary.

[0104] When the coordinates of the state vector are at the center of the friction circle, it signifies a stable driving state; however, as the approach speed at which these coordinates move toward the limit line—the outer boundary of the circle—due to a deep banking angle or excessive acceleration, the risk of low-side slip—a phenomenon in which the tires lose traction and the vehicle body tilts inward and slides—can be calculated to increase steeply. Since the mere fact that the state vector is close to the boundary is insufficient to adequately explain the urgency of actual slip occurrence, the rate of change of time or the velocity vector as the state vector's coordinates approach the friction circle limit boundary can be calculated to determine whether there is a sudden entry into instability. For example, even if the current state vector is still inside the friction circle, if it is moving fast enough to reach the boundary within 0.1 seconds, the risk of low-side slip can be calculated as high. Conversely, if the state vector is near the boundary but the approach speed is very gradual and the rate of increase of the banking angle is low, the risk of low-side slip can be calculated as relatively low. According to this method, the controller can preemptively recognize the risk at the stage immediately preceding actual slippage.

[0105] The step of calculating the high-side rollover weight and correcting the low-side slip risk is a step for separately reflecting the possibility of a high-side rollover that throws the rider into the air beyond simple slipping when the tire friction force is instantaneously recovered while the lateral slip angle has developed above a threshold, and at least one of the reverse snap peak sign of the yaw rate, body rebound acceleration, and rear wheel rotation speed rapid increase pattern is detected.

[0106] An expansion of the lateral slip angle beyond a critical threshold may indicate that the rear wheels are sliding severely outward. In this state, if the driver suddenly closes the throttle or encounters a hard road surface and tire friction is instantaneously restored, the inertial energy of the sliding may suddenly stop, causing the suspension to compress strongly and then spring back up. Signs of a reverse snap peak in the yaw rate may refer to a pattern where a sudden change in rotation attempts to occur in the opposite direction to the existing rotation. Body rebound acceleration may be an acceleration response where the vehicle body springs back up in a rebounding manner due to the sudden recovery of rear wheel traction, while a pattern of rapid increase in rear wheel rotational speed may indicate a state where the rear wheels exhibit a sudden rotational response as driving torque and traction recovery intertwine. When these complex signs are detected, the server or controller can calculate a high-side rollover weighting for the vehicle body to be thrown back and apply this to the existing low-side slip risk to generate a corrected rollover risk value. For example, even if the risk of a simple low-side slip is at a moderate level, if a rapid recovery of rear wheel grip and a reverse snap peak are detected simultaneously, a high-side rollover weight is assigned significantly, and the final risk can be much higher.

[0107] The step of finally calculating the probability of accidents specific to two-wheeled vehicles by combining the jackknife rollover risk, low-side slip risk, and high-side rollover weight with the current vehicle speed, occupant load distribution, load information, and center of gravity elevation displacement coefficient is a step of integrating the preceding individual risk values ​​together with the vehicle condition and loading conditions to generate the final accident probability result.

[0108] Current vehicle speed can be a variable that significantly alters the severity of risk even in the same slip or rollover situation, and occupant load distribution can be a variable reflecting the driver's posture, the presence of a passenger, or the degree of front-to-rear center of gravity bias. Load information may include data on the weight and location of delivery boxes, luggage, or additional cargo, which can act as a factor raising the center of gravity or shifting it to the rear. The center of gravity elevation displacement coefficient may be a coefficient indicating how much higher the center of gravity of the motorcycle has shifted from the ground due to these occupants and cargo. The higher the center of gravity, the greater the probability of a jackknife rollover even with the same brake pressure, and the greater the risk of slipping even at the same cornering speed. The server or controller can combine the jackknife rollover risk, low-side slip risk, and high-side rollover weight with the vehicle state variables to finally calculate motorcycle-specific accident probability information, where real-time accident probability, expected impact location, and risk type are defined as a multidimensional matrix. For example, the real-time accident probability can be defined as 0.86, the expected impact area as the front wheel side and right cowl, and the risk type as a high-side dominant composite risk; other examples may generate detailed results such as an accident probability of 95%, a risk type of jackknife, and expected impact areas as the front cowl and front fork. The final calculated motorcycle-specific accident probability information can be used as key judgment data to determine which control logic to apply first in the subsequent driving assistance control stage.

[0109] And, the step of performing the driving assistance control comprises: a step of determining whether the currently imminent main risk type is one of a jackknife rollover risk, a low-side slip risk, a high-side rollover risk, and a combined risk, based on the real-time accident occurrence probability, expected impact location, and risk type included in the two-wheeled vehicle-specific accident occurrence probability information; and a first control step in which, if the main risk type is determined to be the jackknife rollover risk, the braking hydraulic pressure applied to the front brake caliper is independently depressurized based on a depressurization coefficient inversely proportional to the remaining time margin until the rear wheel lift-off threshold point, while simultaneously increasing the pressure of the rear brake hydraulic pressure to suppress the body pitch rise rate, and the compression damping force of the front electronically controlled suspension is increased, and the throttle output of the driving power source is gradually limited to suppress excessive nose-dive and rollover progression. A second control step in which, when the main risk type is determined to be the low-side slip risk, a cornering ABS (C-ABS) map is activated to downward correct the upper limit of the maximum allowable braking force of the front and rear wheels based on a braking force correction coefficient inversely proportional to the real-time banking angle, and tire grip is restored within the friction circle limit model by performing at least one of reducing the opening amount of the electronic throttle valve (Ride-by-Wire), delaying the ignition timing, and reducing the rear-wheel drive torque based on the friction force loss rate and the lateral slip angle; a third control step in which, when the main risk type is determined to be the high-side rollover risk, the rebound damping force of the rear electronic control suspension is momentarily opened to the lowest level for a preset limit time (Soft-mode) to absorb the body reaction energy generated during the rapid recovery of traction of the rear tires, the increase in rear-wheel drive torque is further limited, and asymmetric braking intervention is performed to suppress the development of bouncing of the vehicle body;If the above-mentioned main risk type is determined to be a complex risk, a fourth control step of performing two or more of the above-mentioned first to third control steps sequentially or in parallel according to a priority table; a step of outputting asymmetric vibration feedback to a left or right haptic vibration module of a steering handlebar mapped to the direction of travel of the above-mentioned lateral slip angle in synchronization with the braking intervention and driving torque intervention of the electronic control unit, and inducing intuitive counter-steering and weight shifting of the driver by varying the vibration intensity, vibration frequency, and vibration duration of the asymmetric vibration feedback according to the above-mentioned risk type and risk level; and a step of gradually fading out the intervention intensity of the driving assistance control and terminating when the real-time banking angle, vehicle body posture data, yaw rate, friction loss rate, and rear wheel lift-off risk level are restored to a normal range or satisfy a stable state continuously for longer than a preset stabilization maintenance time; may be included.

[0110] The step of determining the main risk type is to comprehensively interpret the real-time accident probability, expected impact area, and risk type included in the above-mentioned motorcycle-specific accident probability information to determine whether the most imminent accident pattern currently corresponds to one of the jackknife rollover risk, low-side slip risk, high-side rollover risk, and complex risk.

[0111] Here, the real-time accident probability may be a probability value calculated independently for each risk type, and the expected impact area may refer to a region highly likely to collide or make contact first in the event of an accident, such as the front cowl, front fork, left or right side of the body, handlebars, or the area around the rear wheel. A risk type may be a classification value indicating which accident mechanism the current vehicle behavior is most likely to unfold into, going beyond simple slipping or braking instability. A complex risk may refer to an extreme situation where two or more fatal dynamic phenomena overlap, such as the rear wheel lifting into the air while the vehicle slides sideways, occurring simultaneously with sudden braking and deep turning. Based on data in the form of a multidimensional matrix derived from the preceding computation process, the server or in-vehicle controller can determine which defensive control logic to prioritize; by comparing the probability values ​​for each risk type, it can identify the highest risk as the current primary risk type, and if two or more risks are detected simultaneously at a high level, it can be identified as a complex risk. For example, if the risk of rear wheel lift-off increases rapidly during sudden braking on a straight section, the risk of jackknife rollover may be determined as the primary risk type, and if the rate of friction loss increases and the lateral slip angle expands during cornering, the risk of low-side slip may be determined as the primary risk type. On the other hand, if a rapid recovery of rear wheel traction and signs of reverse snap are detected simultaneously while slipping is in progress, the risk of high-side rollover may be determined as the primary risk type.

[0112] The first control step is a step in which, when the main risk type is determined to be a jackknife rollover risk, the braking hydraulic pressure applied to the front brake caliper is independently reduced based on a reduction coefficient inversely proportional to the remaining time margin until the rear wheel lift-off threshold point, and at the same time, the rear brake hydraulic pressure is increased to suppress the body pitch rise rate, the compression damping force of the front electronically controlled suspension is increased, and the throttle output of the driving power source is gradually limited to suppress excessive nose dive and rollover progression.

[0113] Here, the pressure reduction coefficient may refer to a control factor that is applied more significantly as the remaining time margin decreases; it can signify a mathematical ratio that releases the pressure of the front brakes more strongly and quickly as the time remaining until the rear wheels are completely lifted off the ground—that is, as the risk of rollover becomes more imminent. Independently controlling the braking hydraulic pressure applied to the front brake calipers implies that the ECU forcibly reduces the front brake pressure via the hydraulic control module, even if the driver is gripping the front brake lever tightly. Simultaneously increasing the pressure of the rear brake hydraulic pressure may refer to a control mechanism that utilizes rear traction to disperse deceleration force and assist in pulling the rear of the vehicle body toward the ground. Nose dive refers to the phenomenon where the front of the vehicle body sinks excessively toward the ground due to braking inertia, and increasing the compression damping force of the front electronically controlled suspension may be a control mechanism intended to suppress this nose dive by slowing down the speed at which the front wheels are excessively compressed during sudden braking. In addition, by gradually limiting the throttle output of the driving power source, the engine brake or motor drive response can be controlled to mitigate the front-to-rear load transfer of the entire vehicle body. For example, if the risk of a jackknife rollover increases rapidly during sudden braking from a speed of 90 km / h, the controller can suppress the progression of the vehicle body rollover by reducing the hydraulic pressure of the front wheels, reinforcing the hydraulic pressure of the rear wheels, and increasing the suspension damping force in 0.1-second intervals. In other words, this step may be a control stage that mechanically prevents the phenomenon where the center of gravity of the vehicle body shifts forward and the rear wheels flip over, thereby safely performing the most optimized maximum deceleration within a physical limit where rollover does not occur.

[0114] The second control step is a step of, when the main risk type is determined to be a low side slip risk, activating a cornering ABS map that lowers the upper limit of the maximum allowable braking force of the front and rear wheels based on a braking force correction coefficient inversely proportional to the real-time banking angle, and restoring the tire grip to within the friction circle limit model by performing at least one of reducing the opening amount of the electronic throttle valve, delaying the ignition timing, and reducing the rear wheel drive torque based on the friction force loss rate and the lateral slip angle.

[0115] Here, as the real-time banking angle increases, the motorcycle is already utilizing a significant amount of lateral friction; therefore, it can easily exceed its grip limit even with the application of only a small additional braking force. Consequently, the braking force correction factor can be set to a smaller value as the banking angle increases, thereby automatically lowering the maximum braking force allowed for the front and rear brakes. A cornering ABS map may be a table of braking force allowance characteristics set more conservatively than standard straight-line braking to ensure braking stability during cornering, or it may be a brake lock-prevention control rule exclusive to motorcycles that reflects the vehicle's tilt in real time. Furthermore, an electronic throttle valve may be a device that controls the acceleration response of an engine or motor solely through electrical signals without physical steel cables; reducing its opening amount may imply a control mechanism that reduces the actual air intake or output command even if the driver maintains acceleration. Ignition timing delay can be a control that reduces the instantaneous torque delivered to the rear wheels by delaying the engine ignition timing, and rear-wheel drive torque reduction can be achieved by methods such as limiting inverter output in the case of electric motorcycles, or adjusting fuel injection amount or ignition delay in the case of internal combustion engine motorcycles. For example, if the banking angle is 35 degrees or more on a wet road surface and the lateral slip angle continuously increases, the controller can immediately activate the cornering ABS map and gradually reduce the rear-wheel drive torque to restore the vehicle body state back to a stable region within the friction circle limit model. This step can be a stage that actively controls a situation where the tires slip outward during cornering and the vehicle body tips inward, inducing the tires, which were spinning and being pushed outward, to regain grip on the road surface and return to a stable trajectory.

[0116] The third control step is a step in which, when the main risk type is determined to be a high-side rollover risk, the rebound damping force of the rear electronically controlled suspension is momentarily opened to the lowest level for a preset limit time to absorb the body reaction energy generated during the rapid recovery of traction of the rear tires, additionally limiting the increase in rear-wheel drive torque, and performing asymmetric braking intervention to suppress the bouncing development of the vehicle body.

[0117] Here, since the risk of a high-side rollover can lead to much more abrupt and fatal behavior than a simple slip, the controller can perform faster and more aggressive posture stabilization control than in a low-side situation. A high-side can occur when the rear wheels slide significantly laterally and then suddenly regain road grip; at this moment, the rear suspension is compressed hard and then explosively springs upward, potentially launching the driver into the air. Momentarily opening the rebound damping force of the rear electronically controlled suspension to the lowest level implies a control mechanism that temporarily imparts very soft recovery characteristics to suppress or absorb the repulsive force attempting to spring back up after rapid compression. For example, switching the rebound damping force to Soft-mode allows the compressed suspension to extend softly and loosely instead of stiffly, enabling the suspension itself to absorb a significant portion of the upward rebound energy applied to the vehicle body. Further limiting the increase in rear-wheel drive torque may be a control measure intended to prevent the amplification of vehicle body rebound, as a sudden re-transmission of drive torque the moment rear grip is restored could occur. Asymmetrical braking intervention may refer to control that mitigates the rotation and bouncing of the vehicle body by adjusting the braking force on one side to be greater depending on the contact situation of the front and rear wheels or the left and right wheels. For example, in a situation where the rear wheels slide significantly outward while turning to the right and then suddenly regain contact, the controller can suppress the driver from being thrown upward by opening the rear wheel rebound damping force to Soft-mode, further reducing the driving torque, and, if necessary, performing asymmetrical braking intervention that reflects the left-right balance.

[0118] The fourth control step is a step of performing two or more of the first to third control steps sequentially or in parallel according to a priority table when the main risk type is determined to be a composite risk.

[0119] Here, "complex risk" may refer to a state where high levels of jackknife rollover risk and slip risk exist simultaneously at the same time, or where signs of low-side slip and high-side rollover are detected consecutively at short time intervals. A priority table may be a policy table that defines which control to perform first by considering the lethality of each risk type, remaining time margin, expected impact location, current vehicle speed, and center of gravity status. For example, it may include advanced judgment criteria, such as when jackknife risk and low-side slip risk occur simultaneously due to excessive braking caused by overspeeding upon corner entry; in this case, if the jackknife phenomenon is judged to be more lethal as it causes a complete loss of steering power, pitch suppression control is activated 0.1 seconds earlier, followed immediately by banking angle control intervention. "Sequential execution" may mean continuing control in the order of first mitigating the rollover risk by depressurizing the front wheels and increasing pressure on the rear wheels, followed by reducing driving torque and activating the cornering ABS map. Parallel execution refers to a control method that suppresses multiple risks within a short period of time by simultaneously initiating braking force adjustment, driving torque reduction, and suspension damping force control. For example, in the event that sudden braking and rear wheel slip occur simultaneously on a downhill curve, the controller can refer to a priority table to perform jackknife response control and low-side response control simultaneously or with a very short time difference. According to this structure, it is possible to flexibly respond to complex motorcycle behaviors that are difficult to handle with control logic that assumes only a single risk.

[0120] The step of outputting asymmetric vibration feedback is to output asymmetric vibration feedback to a left or right haptic vibration module of a steering handlebar mapped to the direction of travel of the lateral slip angle, synchronized with the braking intervention and driving torque intervention of the electronic control unit, and to induce intuitive counter-steering and weight shifting of the driver by varying the vibration intensity, vibration frequency, and vibration duration of the asymmetric vibration feedback according to the risk type and risk level.

[0121] Here, asymmetric vibration feedback refers to a feedback method that applies stronger vibrations to one side, either the left or the right handlebar, or assigns different vibration timings and frequencies between the left and right sides. Mapping the direction of travel to the lateral slip angle means, for example, that if the vehicle body slides to the right, a handlebar vibration in a specific direction is output to induce steering to the opposite side, thereby providing the driver with an intuitive response direction. The vibration intensity can be set stronger as the risk level increases, the vibration frequency can vary depending on the urgency of the slip or the type of risk, and the vibration duration can be linked to the duration of the intervention. For instance, if the risk of a low-side slip is progressing to the right, a relatively stronger vibration can be applied to the left handlebar to induce the driver to reflexively attempt counter-steering and weight shifting. This stage goes beyond simply automatically controlling the vehicle; it may be a human-machine cooperative control phase where the system directly induces the driver's physical and instinctive crisis response behaviors alongside hardware-based electronic control, allowing the driver and the vehicle control system to cooperate simultaneously to avoid an accident.

[0122] The step of gradually attenuating the intervention intensity of the driving assistance control and terminating it is a step of naturally terminating the system intervention while preventing secondary instability caused by sudden release of control when the real-time banking angle, vehicle body posture data, yaw rate, friction loss rate, and rear wheel lift-off risk are restored to a normal range or when a stable state is continuously satisfied for longer than a preset stabilization maintenance time.

[0123] Here, the "normal range" may refer to a pre-set safety tolerance zone for each state variable; for example, it may signify a state where the rate of change in banking angle stabilizes below a reference level, and the yaw rate and friction loss rate return to the normal driving range. The "stabilization maintenance time" may be a time interval designed to verify whether a stable state has been maintained continuously for a certain period, rather than simply entering the normal range instantaneously. Gradually reducing the intervention intensity may mean progressively decreasing the amount of front or rear braking force correction, the throttle limit level, the amount of suspension damping force correction, and the intensity of haptic feedback. This is because if mechanical limitations or forced operations applied to the brake caliper or throttle valve are released instantaneously, the vehicle's body balance may be shaken again due to the resulting reaction. For instance, even if it is determined that a dangerous condition has been resolved, instead of immediately releasing all control, the correction amount may be gradually reduced over 0.5 to 2 seconds to return full control to the driver. According to this phased termination method, it is possible to prevent slipping from recurring or body balance from changing abruptly immediately after control release, and to maintain overall driving stability more smoothly.

[0124] In addition, prior to the step of collecting the above-mentioned vehicle body posture data in real time: a step of transmitting driver identification information, status information of the target motorcycle, and current location information to a server through the communication module of the target motorcycle; and a step of receiving confirmation from the server regarding whether the target motorcycle usage is approved, target entry / exit space information, and rental condition information corresponding to the driver, and determining whether the driving power source through the electronic control unit (ECU) of the target motorcycle is allowed to start based on the starting permission condition and entry / exit processing condition included in the rental condition information; wherein the server further comprises: a step of receiving service request information including driver identification information, motorcycle usage request information, and usage condition information from a driver terminal; a step of performing authentication for the driver based on the driver identification information and obtaining driver information including user qualification information, usage history information, and payment method information linked to the driver identification information; and a step of selecting a target motorcycle and a target entry / exit space corresponding to the service request information based on operational resource information including motorcycle information and storage space information. A control method is controlled according to the following steps: a step of calculating rental condition information including at least one of whether usage is approved, usage fees, warranty conditions, insurance coverage conditions, and entry / exit processing conditions for the target motorcycle, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target entry / exit space information, and pre-set operation policy information; and a step of automatically controlling the unmanned entry / exit of the target motorcycle through network linkage with the target entry / exit space and the target motorcycle; wherein the step of automatically controlling the unmanned entry / exit of the target motorcycle comprises: a step of generating unmanned exit control information including whether unmanned exit is approved for the driver, electronic unlock conditions, ignition permission conditions, helmet storage box opening conditions, and exit permission conditions, based on the driver identification information, driver information, and rental condition information;The method comprises the step of generating unmanned return control information including whether the target motorcycle has been received, whether it is in the correct position within a designated area, whether there is exterior damage, and whether accessories have been returned, based on image information or sensor information received from at least one of a camera, proximity sensor, weight sensor, and position sensor disposed in the target entry / exit space; wherein the step of generating unmanned return control information comprises: a step of verifying whether the target motorcycle has been received in the correct position by determining whether the wheelbase trajectory, center of gravity position, and entry angle of the front and rear wheels of the target motorcycle satisfy an allowable error range within a pre-set designated area dedicated to motorcycles, by mutually linking load distribution data obtained from a multi-channel weight sensor arranged in a grid pattern on the floor surface and front and rear wheel detection data obtained from a proximity sensor embedded in a wheel stopper when the target motorcycle enters the target entry / exit space; When the verification of whether the vehicle has been received at the correct location is completed, the method comprises the steps of: approving the authority to initiate the return by comparing first biometric information, which includes at least one of fingerprint information, facial information, and vein information included in the driver information, with second biometric information of the returner obtained through an authentication terminal installed on the target motorcycle or the target entry / exit space; and, when the authority to initiate the return is approved, switching at least one of the immobilizer, electronic lock, and mechanical steering lock of the target motorcycle to an engaged state; and analyzing the 6-axis or 9-axis inertial measurement unit (IMU) information, impact sensor information, and driving log information of the target motorcycle obtained during the use of the target motorcycle, and, when abnormal banking angle data, impact amount data exceeding a threshold, estimation of power delivery, and estimation of left / right slip direction data are extracted, deriving a suspected accident area based on the extracted data.When the suspected accident area is identified, a control signal is transmitted to a pan-tilt-zoom (PTZ) camera placed in the target entry / exit space to vary the shooting angle, shooting order, focal length, and magnification ratio for the suspected accident area, thereby acquiring multi-angle entry exterior image data with an increased proportion of high-resolution focused imaging of the suspected accident area; input the multi-angle entry exterior image data and reference exterior image data captured at the time of rental into a pre-trained visual intelligence-based damage judgment model to perform damage segmentation analysis by area, thereby quantifying at least one of the scratch area, damage depth, crack length, and physical damage rate of the vehicle body exterior; and calculate the weight difference between the current weight of the target motorcycle measured through the multi-channel weight sensor and the reference weight at the time of release, and determine whether accessories including a helmet, side mirror, charging cable, and detachable battery pack have been properly returned by synthesizing the weight difference value, the infrared sensor scan results inside the helmet storage compartment, the storage compartment opening / closing log, and the accessory tag recognition results. and may include the step of finally generating unmanned return control information for billing settlement, comprising at least one of normal return approval information, damage inspection request information, missing accessory warning information, additional billing settlement information, and administrator confirmation request information, by combining the above-mentioned receipt status at the correct location, return initiation authority approval status, quantified damage information, and accessory normal return status.;

[0125] The step of transmitting driver identification information, status information of the target motorcycle, and current location information to the server is a step in which, immediately before the driver boards the vehicle and begins actual driving, the hardware terminal of the target motorcycle initiates a communication session with the platform server via a network and transmits basic verification data so that the server can determine whether the driver and the vehicle match the current rental contract.

[0126] Here, the communication module of the target motorcycle may refer to a vehicle-mounted communication device capable of transmitting and receiving data with an external server via LTE, 5G, Wi-Fi, Bluetooth, or a dedicated vehicle communication network. Driver identification information may refer to unique data indicating who the user is that has accessed the motorcycle, such as a platform account identifier, mobile authentication token, biometric authentication linked identifier, or a unique member number. Status information of the target motorcycle may refer to hardware diagnostic data including current battery charge or remaining fuel level, ECU active status, current lock status, immobilizer status, accumulated mileage, recent fault codes, sensor abnormalities, and history of external impacts to the vehicle. Current location information may include GNSS-based coordinates, relative distance to a previously registered delivery point, location within a vehicle zone, or parking area identifier. According to this step, the server can determine whether the correct renter has boarded the correct vehicle and can perform a preliminary review to determine whether the vehicle is located at the actual reserved location, has not been moved without authorization by another user, or requires maintenance. For example, if the reserved vehicle is within a 5 m radius of the designated delivery point, the battery charge is 78%, and the lock status is normal, the server can proceed with the subsequent approval process. On the other hand, if the vehicle is out of position or an ECU error code is detected, the approval process for use may be suspended.

[0127] The step of receiving confirmation from the server regarding whether the target motorcycle usage is approved in accordance with the driver, the target entry / exit space information, and the rental condition information, and determining whether to allow the starting of the driving power source through the electronic control unit of the target motorcycle based on the starting permission condition and entry / exit processing condition included in the rental condition information, is a step in which the motorcycle terminal determines whether to allow or block the operation of the engine or motor on its own under the control of the server.

[0128] Here, the target entry / exit space information may include available exit points, scheduled return points, vehicle zone identification values, designated parking space numbers, and entry / exit time restriction information. The rental condition information may be a comprehensive set of conditions including usage approval status, usage time range, usage fees, warranty conditions, insurance coverage conditions, entry / exit processing conditions, unmanned exit approval conditions, and engine start permission conditions. Engine start permission conditions may refer to electronic authorization defined so that the ECU opens the start relay or motor drive permission signal only when the driver's risk assessment has been successfully passed, the payment method is validly registered, an attempt to exit from an approved vehicle zone is made within the approved time range, biometric authentication or app-based authentication has been validly completed, and the vehicle status is normal. Entry / exit processing conditions may include the time limit for leaving the designated point, the availability of opening the helmet storage locker, exit gate linkage conditions, and conditions for verifying the location at the time of return. If the conditions are not met, the ECU can prevent unauthorized driving by forcibly cutting off power to the throttle or starter motor, even if the driver possesses a physical smart key. For example, if a driver approaches the vehicle 20 minutes before the reservation start time, the server may determine, based on rental condition information, that the conditions for allowing the start have not yet been met and block the ECU's signal to allow the start. Conversely, if all conditions are met—reaching the reservation start time, completing app authentication, matching the designated vehicle zone, and the vehicle status being normal—the ECU can allow the start.

[0129] The step of receiving service request information including driver identification information, motorcycle usage request information, and usage condition information from a driver terminal is a step of receiving initial rental reception data from a user operating a motorcycle rental platform application on the server side and securing raw input data necessary for subsequent resource allocation and unmanned dispatch approval.

[0130] Here, the motorcycle usage request information may include the rental start time, rental end time, estimated usage time, planned departure location, planned return location, desired vehicle type, and estimated mileage. Usage condition information may refer to detailed contract conditions encompassing the scope of insurance coverage requested by the driver, rental time, or additional services, such as whether to rent an additional helmet. This stage may not be a simple reservation acceptance stage, but rather a stage that forms an initial contract data set that serves as a reference value for calculating subsequent rental conditions and unmanned entry / exit control.

[0131] The step of performing authentication for the driver based on the driver identification information and obtaining driver information including user qualification information, usage history information, and payment method information associated with the driver identification information is a step of querying the platform database to verify whether the requester has the legal qualification and past trustworthiness to rent a two-wheeled vehicle.

[0132] User eligibility information may include license type, license validity, satisfaction of age criteria, and service usage restriction status. Usage history information may include past rental counts, return delay history, payment failure history, accident or damage history, and driving pattern history. Payment method information may include pre-registered card information, account information, automatic payment approval status, and deposit withdrawal availability status. According to this stage, the server can filter out unqualified users or users with outstanding payments at the initial stage and can also reflect the aforementioned driver information in determining whether to approve unmanned vehicle release thereafter.

[0133] The step of selecting a target motorcycle and a target entry / exit space corresponding to the service request information based on operational resource information including motorcycle information and storage space information is a step of determining the optimal rental target by matching the driver's rental conditions, namely conditions such as location, time, and vehicle type, with vehicles and parking zones currently waiting in an idle state for rental within the platform.

[0134] Here, operational resource information may include the current location, availability, charge or fuel status, vehicle type, and maintenance status of each of multiple motorcycles, as well as the occupancy status, access availability, CCTV availability, lighting conditions, and vehicle zone allocation status of each of multiple entry / exit spaces. The server can select a target motorcycle and a target entry / exit space by searching for a combination of vehicle and space that best matches the service request information. For example, it can select an electric scooter that is closest to the scheduled return location, has sufficient battery charge, and has no reservation conflicts during the designated time slot, along with a designated vehicle zone where the vehicle can be released. This selection process can provide a spatial reference point for subsequent unmanned release approval and unmanned return inspection.

[0135] The step of calculating rental condition information including at least one of the following: approval of use of the target motorcycle, usage fee, warranty condition, insurance coverage condition, and inbound / outbound processing condition, based on the above driver information, motorcycle usage request information, grade information of the target motorcycle, target inbound / outbound space information, and pre-set operation policy information, is a step of calculating the final charge amount, the insurance exemption limit, and detailed constraint conditions to be applied to unmanned inbound / outbound as specific numerical or logical conditions by comprehensively calculating the confirmed vehicle and driver data.

[0136] Here, operational policy information may include approval criteria by vehicle group, rates by time of day, deposit policies, special insurance policy terms, and rules allowing unmanned pickup and return. For example, high-risk drivers may be subject to additional deposits and restricted ignition conditions, and additional inbound / outbound processing conditions may be added, such as restricting pickup to locations other than designated bases. The rental condition information generated at this stage can subsequently be used as input values ​​for determining whether to allow vehicle ignition, generating unmanned pickup control information, and unmanned return inspection policies.

[0137] The step of automatically controlling the unmanned entry and exit of the target motorcycle through network linkage with the target entry and exit space and the target motorcycle is a step in which the server remotely controls various hardware equipment of the rental zone and the communication module of the vehicle to complete the exit and return procedures non-face-to-face, even if the manager is not present at the site.

[0138] This step may be a series of remote integrated control procedures that unlock the vehicle by verifying driver authentication and condition fulfillment upon release, and automatically inspect the receiving location, vehicle condition, and accessory condition upon return. According to this step, a series of procedures, including vehicle release approval, ignition permission, return verification, and billing settlement, can be automatically performed even without platform operation personnel being physically present on-site.

[0139] According to one embodiment, a server includes a processor, memory, a communication module, and a non-transient storage medium, and a method for providing an unmanned entry / exit automation service for two-wheeled vehicle rental based on driver identification information, wherein the method is performed by the processor by executing a program stored in the non-transient storage medium, the method comprises: receiving service request information from a driver terminal including driver identification information, two-wheeled vehicle usage request information, and usage condition information; performing authentication for the driver based on the driver identification information and obtaining driver information including usage qualification information, usage history information, and payment method information linked to the driver identification information; selecting a target two-wheeled vehicle and a target entry / exit space corresponding to the service request information based on operational resource information including two-wheeled vehicle information and storage space information; and calculating rental condition information including at least one of whether the use of the target two-wheeled vehicle is approved, usage fees, warranty conditions, insurance coverage conditions, and entry / exit processing conditions based on the driver information, two-wheeled vehicle usage request information, grade information of the target two-wheeled vehicle, target entry / exit space information, and pre-set operational policy information. The method comprises the step of automatically controlling the unmanned entry and exit of the target motorcycle through network linkage with the target entry / exit space and the target motorcycle; wherein the step of automatically controlling the unmanned entry and exit of the target motorcycle includes: the step of generating unmanned exit control information including whether unmanned exit is approved for the driver, electronic unlocking conditions, ignition permission conditions, helmet storage box opening conditions, and vehicle exit permission conditions, based on the driver identification information, driver information, and rental condition information; and the step of generating unmanned return control information including whether the target motorcycle is entered, whether it is positioned correctly in a designated area, whether there is external damage, and whether accessories are returned, based on image information or sensor information received from at least one of a camera, proximity sensor, weight sensor, and position sensor disposed in the target entry / exit space.The present invention provides a method for providing an unmanned entry and exit automation service for two-wheeled vehicle rental based on driver identification information, including

[0140] The step of receiving service request information is a step in which the server receives service request information including driver identification information, two-wheeled vehicle usage request information, and usage condition information from a driver terminal, and forms reference data for the authentication procedure, vehicle and space allocation procedure, rental condition calculation procedure, and unmanned entry / exit control procedure that are subsequently performed.

[0141] Here, the term "driver terminal" may refer to a smart device, such as a smartphone, tablet, PC, or vehicle-linked terminal, that accesses a two-wheeled vehicle rental platform application or web interface to request a rental. The term "driver identification information" may refer to data intended to identify the user, such as a platform account identifier, a unique member number, a personal authentication token, a biometric authentication-linked identifier, or a mobile phone-based authentication value. The term "two-wheeled vehicle usage request information" may refer to information indicating the intention to use the vehicle, such as the scheduled rental start time, scheduled rental end time, estimated usage time, scheduled departure location, scheduled return location, desired vehicle type, estimated driving distance, or whether the vehicle is an electric two-wheeled vehicle. The term "usage condition information" may include detailed contract conditions, such as the scope of insurance coverage, deposit limits, permission for night driving, helmet rental, provision of charging cables, or permission for unmanned returns. For example, if a driver wishes to use a 125 cc scooter from an urban commercial district to a residential area from 2:00 PM to 6:00 PM and selects basic insurance and helmet rental options, the server may receive the entire request as a single service request information and use it as an input for subsequent procedures. This step may not be a simple reservation acceptance step, but rather a step that forms an initial data set serving as the starting point for the entire rental contract process.

[0142] The step of performing authentication for a driver and obtaining driver information is to verify whether the actual rental requester is a legitimate user based on the driver identification information, and to retrieve driver information from a database that includes user qualification information, usage history information, and payment method information associated with the driver identification information.

[0143] Here, eligibility information may refer to information used to determine whether a motorcycle can be legally rented, such as license type, license validity, fulfillment of age criteria, status of service usage restrictions, and completion of identity verification. Usage history information may include accumulated data used to assess user reliability, such as the number of past rentals, history of delayed returns, history of payment failures, history of accidents or damage, history of returns outside designated areas, or driving pattern history. Payment method information may include pre-registered card information, the status of simple payment integration, the availability of deposit withdrawal, and the status of automatic payment approval. By verifying the above driver information, the server can filter out unqualified users, users with outstanding fees, or users with usage restrictions at an initial stage, and can subsequently reflect this driver information in determining whether to approve unmanned vehicle release and the conditions for allowing the engine to start. For example, if a user holds a Class 2 small motorcycle license, has a properly registered payment method, and has no history of delayed returns or accidents, the server may determine that user as a properly authenticated driver and proceed with subsequent procedures. On the other hand, if the license has expired or there are outstanding fees, the approval of use for the aforementioned driver may be restricted.

[0144] The step of selecting target motorcycles and target inbound / outbound spaces is a step of determining the actual assets to be allocated corresponding to the service request information based on operational resource information including motorcycle information and storage space information.

[0145] Here, operational resource information may include the current location, availability, charge status or remaining fuel level, vehicle type, maintenance status, lock status, and recent history of exterior defects for each of the multiple two-wheeled vehicles, as well as the occupancy status, entry / exit availability, vehicle zone identification value, lighting conditions, presence of CCTV, presence of charging facilities, and return permission status for each of the multiple entry / exit spaces. The server can select a target two-wheeled vehicle and a target entry / exit space by searching for the combination of vehicle and space that best matches the service request information. For example, it can select an electric scooter that is closest to the planned return location, has sufficient battery charge, and has no reservation conflicts during the designated time slot, along with a designated vehicle zone where that vehicle can be released. Furthermore, even if multiple vehicles exist at the same time, the most suitable vehicle and space can be finally determined by comprehensively considering vehicle condition defects, charge status, reservation conflicts, and the efficiency of the planned return route. This selection process can provide a spatial reference point for subsequent unmanned release approval and unmanned return inspection.

[0146] The step of calculating rental condition information is a step of generating rental condition information including at least one of the following: approval of use of the target motorcycle, usage fee, warranty condition, insurance coverage condition, and inbound / outbound processing condition, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target inbound / outbound space information, and pre-set operation policy information.

[0147] Here, the grade information of the target motorcycle may include displacement ranges, whether it is electric or internal combustion engine, vehicle price ranges, and classification values ​​for vehicle groups or insurance risk groups. Operational policy information may include rates by time of day, approval criteria by vehicle type, deposit policies, special insurance coverage policies, rules allowing unmanned pickup and return, and rules allowing returns outside designated locations. Rental condition information may include not only final usage approval but also specific usage fees, deductible levels, deposit ranges, unmanned pickup approval conditions, engine start permission conditions, entry / exit processing conditions, and inspection criteria upon return. For example, high-risk drivers may be subject to additional deposits and restricted engine start permission conditions, and entry / exit processing conditions restricting pickup outside designated locations may be added. The rental condition information generated at this stage can subsequently be utilized as input values ​​for determining whether the vehicle's engine is allowed, generating unmanned pickup control information, and unmanned return inspection policies. In other words, the above rental condition information may not be simple billing data, but rather control standard data that operates the entire unmanned rental platform.

[0148] The step of automatically controlling the unmanned entry and exit of the target motorcycle is a step of opening the vehicle by verifying driver authentication and condition fulfillment during exit through network linkage with the target entry and exit space and the target motorcycle, and automatically inspecting the entry location, vehicle condition, and accessory condition during return.

[0149] Here, unmanned automatic entry and exit control refers to a system where a server remotely controls various hardware equipment in the rental zone and the vehicle's communication module to complete the exit and return procedures non-face-to-face, even without a manager being present on-site. For example, during exit, the process may include unlocking the vehicle, opening the helmet storage compartment, controlling the exit barrier, and allowing the ECU to start; during return, the process may include verifying the vehicle's position, verifying biometric authentication, locking the vehicle, inspecting for exterior damage, detecting missing accessories, and settling additional charges. According to this stage, a series of procedures—from vehicle exit approval and start-up permission to return verification and charge settlement—can be automatically performed even without platform operators physically present on-site. Therefore, this stage may not be merely simple vehicle opening and closing control, but rather an integrated control stage designed to realize server-centric unmanned rental operations.

[0150] The step of generating unmanned outbound control information is to generate unmanned outbound control information including whether unmanned outbound is approved for the driver, electronic unlocking conditions, ignition permission conditions, helmet storage box opening conditions, and vehicle discharge permission conditions, based on the driver identification information, driver information, and rental condition information.

[0151] Here, the approval of unmanned vehicle release may be the result of a comprehensive assessment of factors such as whether the driver is the reserved party, whether eligibility is valid, whether payment or a deposit has been secured, and whether the release time has arrived. The electronic unlock condition may be a condition that unlocks the vehicle upon completion of app authentication or successful near-field communication authentication. The engine start permission condition may be a condition set so that the ECU allows engine or motor operation only when the electronic usage authority granted by the server is activated. The helmet storage compartment opening condition may refer to the authority to release the electronic latch only after driver identity verification is complete, and the vehicle exit permission condition may include conditions allowing spatial departure, such as unlocking the vehicle zone barrier or unmanned stopper. For example, if the server determines the approval of unmanned vehicle release to be true, the vehicle may unlock the electronic lock, open the helmet storage compartment, and activate the ECU's engine start permission condition. The unmanned vehicle release control information generated in this way can be utilized as a reference value to link the vehicle-side control logic with the space-side exit control logic at the actual time of release.

[0152] The step of generating unmanned return control information is to generate unmanned return control information including whether the target motorcycle has been received, whether it has been positioned in a designated area, whether its exterior has been damaged, and whether accessories have been returned, based on image information or sensor information received from at least one of a camera, a proximity sensor, a weight sensor, and a position sensor placed in the target entry / exit space.

[0153] Here, unmanned return control information may refer to integrated processing result data including return approval status, damage inspection request information, missing parts warning information, additional charge settlement information, and administrator verification request information. This step may not merely verify whether the vehicle has been returned, but may also review the appropriateness of the return location, the legitimacy of the actual returner, the presence of exterior damage, and whether the parts have been returned properly. In other words, the unmanned return control information may function as an approval signal for the completion of the return procedure, as well as a signal to initiate additional compensation for damages or maintenance measures. For example, if the vehicle is parked correctly in the designated area, the returner's biometric information matches the registration information, and no damage or missing parts are detected, only normal return approval information may be generated. On the other hand, if the vehicle is parked in an abnormal location or if damage to the right cowl and a missing charging cable are detected together, unmanned return control information including damage inspection request information, missing parts warning information, and additional charge settlement information may be generated. The unmanned return control information generated in this way is transmitted to a server and can be used for contract termination, settlement processing, administrator notifications, and subsequent procedures to restore the rental availability status.

[0154] In addition, the step of generating the unmanned return control information comprises: a step of verifying whether the target motorcycle has entered the correct position by determining whether the wheelbase trajectory, center of gravity position, and entry angle of the front and rear wheels of the target motorcycle satisfy the allowable error range within a pre-set designated area dedicated to motorcycles, by mutually linking load distribution data obtained from a multi-channel weight sensor arranged in a grid pattern on the floor surface and front and rear wheel detection data obtained from a proximity sensor embedded in a wheel stopper when the target motorcycle enters the target entry / exit space; and a step of approving the authority to initiate the return by comparing first biometric information including at least one of fingerprint information, facial information, and vein information included in the driver information with second biometric information of the returner obtained through an authentication terminal installed on the target motorcycle or the target entry / exit space when the authority to initiate the return is approved, and switching at least one of the immobilizer, electronic lock, and mechanical steering lock of the target motorcycle to an engaged state. A step of analyzing 6-axis or 9-axis inertial measurement unit (IMU) information, impact sensor information, and driving log information of the target motorcycle acquired during the use of the target motorcycle, and if abnormal banking angle data, impact amount data exceeding a threshold, spurt degree estimation data, and left / right slip direction estimation data are extracted during driving, deriving a suspected accident area based on the extracted data; and a step of, if the suspected accident area is derived, transmitting a control signal to a pan-tilt-zoom (PTZ) camera placed in the target entry / exit space to vary the shooting angle, shooting order, focal length, and magnification ratio for the suspected accident area, thereby acquiring multi-angle entry exterior image data with an increased proportion of high-resolution focused imaging of the suspected accident area;The method may include: a step of inputting the multi-angle receiving exterior image data and the reference exterior image data captured at the time of rental into a pre-trained visual intelligence-based damage judgment model to perform damage segmentation analysis by part, thereby quantifying at least one of the scratch area, damage depth, crack length, and physical damage rate of the vehicle body exterior; a step of calculating the weight difference value between the current weight of the target motorcycle measured through the multi-channel weight sensor and the reference weight at the time of delivery, and synthesizing the weight difference value, the infrared sensor scan result inside the helmet storage compartment, the storage compartment opening / closing log, and the accessory tag recognition result to determine whether accessories including a helmet, side mirror, charging cable, and detachable battery pack have been returned normally; and a step of combining the receipt status at the correct location, the return initiation authority approval status, the quantified damage information, and the normal return status of accessories to finally generate unmanned return control information for billing settlement, which includes at least one of normal return approval information, damage inspection request information, missing accessory warning information, additional billing settlement information, and administrator confirmation request information.

[0155] The step of verifying whether the vehicle has entered the correct position is a step of determining whether the vehicle has stopped in the correct position within the designated return area by mutually linking load distribution data obtained from multi-channel weight sensors arranged in a grid pattern on the floor surface and front and rear wheel detection data obtained from proximity sensors embedded in the wheel stoppers when the vehicle enters the target entry / exit space.

[0156] Here, a multi-channel weight sensor may refer to multiple load detection cells densely embedded in a two-dimensional grid format on the floor surface of the return zone, and each sensor cell can form a load distribution map by measuring the magnitude and distribution location of the load pressing on the corresponding location in real time. Load distribution data may be positional weight data indicating how much weight is loaded at which location on the front wheel, rear wheel, side stand, or main stand. A proximity sensor embedded in the wheel stopper may refer to a sensor installed in front of or behind the wheel stopper to detect whether the front wheel and rear wheel have reached a reference position, and may be implemented using infrared, ultrasonic, magnetic field, or optical sensor methods. Front and rear wheel detection data may be distance values ​​or contact data indicating how close each wheel is to the wheel stopper reference point. A server or field controller may mutually link the load distribution data and the front and rear wheel detection data to determine whether the wheelbase trajectory, body center position, and entry angle of the target motorcycle's front and rear wheels satisfy the allowable error range within a pre-set designated area dedicated to motorcycles. Here, the wheelbase trajectory may refer to the linear arrangement connecting the center points of the front and rear wheels, the body center position may be a value indicating how far the center of the total load deviates from the center of the parking space, and the entry angle may refer to the angle formed by the major axis of the two-wheeled vehicle with the reference axis of the parking space. The system calculates the coordinates where the load is detected to determine whether the vehicle is parked diagonally or outside the designated space, and can approve the entry into the correct position and activate the subsequent return procedure only if it is parked straight within the allowable tolerance range. For example, entry into the correct position can be determined only if the center points of the front and rear wheels are each located within 3 cm of the center line of the designated space, the body center position is within 5 cm of the center of the space, and the entry angle shows a deviation of 2 degrees or less relative to the reference axis.Conversely, if the vehicle is parked diagonally or the rear wheels overlap the outer line of the zone, it can be determined that the vehicle has been parked in the correct position. According to this step, hardware verification can be performed using multiple sensors to verify whether the driver has parked the vehicle neatly and accurately in the return zone for the next user, and the reference location for subsequent exterior inspection photography and additional charge determination can be unified.

[0157] The step of approving the authority to initiate the return and switching the locking device to a locked state is, when the verification of receipt at the correct location is completed, comparing the first biometric information included in the driver information with the second biometric information obtained on-site at the time of return to confirm whether the actual returner is a legitimate rental user, and, once the verification is complete, immediately switching the target motorcycle to a non-operable state.

[0158] Here, the first biometric information may be reference biometric data including at least one of fingerprint information, facial information, and vein information registered at the time of the driver's membership registration or rental approval. The second biometric information may be biometric data acquired in real time at the moment of return through an authentication terminal installed on the target motorcycle or the target entry / exit space. The authentication terminal may be composed of a fingerprint recognition panel, a facial recognition camera, a back-of-hand vein recognition device, or a combination thereof. A server or field controller may determine whether to approve the authority to initiate the return by comparing the first biometric information and the second biometric information, and may approve the authority to initiate the return only if the matching rate is greater than or equal to a preset threshold value. Here, the authority to initiate the return may refer to an electronic authority that can actually and effectively start the return procedure. By comparing this, malicious acts such as having a third party return the vehicle on behalf of the driver after an accident or stealing the rental account can be fundamentally blocked. If the authority to initiate the return is approved, the controller may switch at least one of the immobilizer, electronic lock, and mechanical steering lock of the target motorcycle to an engaged state. An immobilizer may refer to an electronic security device that blocks engine or motor operation in the absence of a registered authorization signal; an electronic lock may be a device that engages wheel locks or seat locks via a remote signal; and a mechanical steering lock may be a physical lock that fixes the steering wheel rotation itself. Once authorization to initiate the return is granted, the system may block engine ignition via the immobilizer and engage the mechanical steering lock of the steering system—namely, the steering lock—to switch to a physical security state that prevents the vehicle from being moved until the next authorized person appears. For example, if the driver's fingerprint information matches the fingerprint information recognized at the time of return by 98% or more, the controller may immediately activate the immobilizer and engage the steering lock to prevent the vehicle from being moved without authorization after the return.Conversely, if the biometric information does not match, the authority to initiate the return may not be approved, and the status may be switched to a request for administrator verification. This step prevents fraudulent returns or unauthorized abandonment of the vehicle by third parties and allows the vehicle to be secured in a safe storage state immediately after return.

[0159] The step of deriving suspected accident areas is to analyze IMU information, impact sensor information, and driving log information obtained during the use of the above-mentioned target motorcycle, and when abnormal banking angle data, impact amount data exceeding a threshold, estimation of power delivery, and estimation of left and right slip direction data are extracted, to identify body parts that are highly likely to have external damage based on the extracted data.

[0160] Here, IMU information may refer to vehicle body attitude and vibration data based on 3-axis acceleration and 3-axis angular velocity, and impact sensor information may refer to data indicating the intensity and direction of a physical impact received by the vehicle body at a specific point in time. Driving log information may refer to driving records including hourly speed, location, whether sudden braking or acceleration occurred, whether cornering occurred, whether road surface conditions changed, and past ADAS intervention history. Abnormal banking angle data may refer to records of angles at which the vehicle body tilted beyond normal cornering levels to the point of nearly touching the ground. The estimation of sudden tilt and the estimation of left-right slip direction may be data reconstructed three-dimensionally by integrating the logs of the acceleration and angular velocity sensors to determine whether the vehicle body tipped over in place or slid to the left or right while driving. A server or in-vehicle controller can combine the above data to infer the suspected accident site. For example, if an abnormal banking angle of more than 70 degrees to the right and a rightward impact force are recorded simultaneously, the right cowl, right turn signal, right side mirror, and right handle end can be identified as suspected accident areas. Conversely, if front wheel impact and signs of a jackknife rollover are detected, the front fork, front cowl, and front lamp cover can be set as priority inspection areas. According to this step, vehicle dynamic data recorded during driving can be directly linked to a post-accident unmanned inspection system to logically trace back damaged areas, and instead of photographing and analyzing the entire exterior with the same intensity, areas with a high probability of damage can be set as priority inspection targets, thereby simultaneously increasing the accuracy and efficiency of return inspection.

[0161] The step of acquiring multi-angle incoming exterior image data is a step of acquiring multi-angle incoming exterior image data with an increased proportion of high-resolution focused imaging of the suspected accident area by varying the shooting angle, shooting order, focal length, and magnification ratio of the suspected accident area by transmitting a control signal to a PTZ camera placed in the target incoming / outgoing space when the suspected accident area is identified.

[0162] Here, a PTZ camera refers to a camera capable of performing Pan, Tilt, and Zoom functions, and can precisely photograph specific areas through horizontal rotation, vertical tilt adjustment, and optical or digital zoom. The shooting angle refers to the angle in which the camera faces the target motorcycle, and the shooting order refers to a control sequence indicating which area to photograph first if multiple suspected accident areas exist. The focal length refers to the distance at which the camera lens is focused, and the magnification is a value that adjusts the level of image magnification to identify even fine scratches or cracks. During a typical unmanned return, only approximate photos of the vehicle's entire front, back, left, and right sides can be taken; however, since a suspected accident area has been identified in this stage, the shooting algorithm can be varied to clearly capture even fine scratches on the paint surface or cracks in the plastic by rotating the lens to focus on that area, for example, the lower right cowl, and increasing the magnification. For example, if the right cowl and right turn signal are identified as suspected accident sites, the PTZ camera can be controlled to first rotate to the right to photograph the entire cowl at 1x magnification, and then to magnify and photograph the turn signal housing at 3x or 5x magnification. Multi-angle incoming exterior video data may refer to a set of video data capturing the vehicle's exterior at the time of return in various directions, at various angles, and at various magnification levels. According to this step, detailed damage areas that are difficult to identify with standardized front and rear photography alone can be precisely secured, and the reliability of subsequent video analysis can be improved while deterring attempts by the driver to conceal damage by applying tape or covering it with foreign substances.

[0163] The step of quantifying at least one of the scratch area, damage depth, crack length, and physical damage rate of the turn signal on the vehicle body exterior is a step of converting exterior damage occurring at the time of return into a quantitative value by inputting the multi-angle incoming exterior image data and the reference exterior image data taken at the time of rental into a pre-trained visual intelligence-based damage judgment model and performing damage segmentation analysis by part.

[0164] Here, reference exterior image data may refer to reference images captured and stored showing the vehicle's exterior in a normal state at the time of rental delivery. The visual intelligence-based damage judgment model may be a deep learning-based image analysis model capable of separating candidate damage areas by part at the pixel level and estimating the type and severity of the damage. Damage segmentation analysis by part may refer to an analysis procedure that separates each part—such as the vehicle body exterior, turn signals, mirrors, and lamp covers—into individual regions, and then segments candidates for scratches, cracks, breakage, deformation, or detachment within those regions. The scratch area can be calculated by converting the number of pixels in the extracted damage area into an actual area, and the severity of damage can be estimated by determining whether the paint is simply peeled off, if the panel is deeply gouged, or if internal structural components are exposed, based on the image's brightness gradient, shadow patterns, degree of contour deformation, and depth differences between multiple viewpoints. The crack length may be a value calculated in mm or cm through skeletal analysis after extracting the crack path within the damage area. The physical damage rate of the turn signal may be a value converted into a percentage of the damaged area, deformation ratio, reduction ratio of the light-emitting part area, or displacement ratio of the turn signal lens or housing. For example, if a scratch of 18 cm² is detected on the right cowl and the deformation ratio of the turn signal housing is calculated to be 35%, the controller may record this as return damage information. According to this step, subsequent billing settlement and administrator review can be performed based on quantified damage figures rather than simple visual inspection.

[0165] The step of determining whether accessories have been properly returned is to calculate the weight difference between the current weight of the target motorcycle measured through the multi-channel weight sensor and the weight based on the time of shipment, and to determine whether accessories including a helmet, side mirror, charging cable, and detachable battery pack have been properly returned by combining the weight difference value, the infrared sensor scan result inside the helmet storage compartment, the storage compartment opening / closing log, and the accessory tag recognition result.

[0166] Here, the standard weight at the time of shipment may be the standard weight measured prior to shipment, including the subject motorcycle in a normal condition and all standard accessories. The current weight may be the actual measured weight at the time of return, and the weight difference value, representing the difference between the two weights, may serve as a primary judgment value indicating whether accessories are missing. The infrared sensor scan result inside the helmet compartment may be the result of detecting whether a helmet or a specific accessory is actually present inside the compartment. The compartment opening / closing log may indicate the history of the compartment being opened and closed immediately before or during the return process, and the accessory tag recognition result may be data indicating whether the corresponding accessory is present in the vehicle or compartment when an identifier such as an RFID, NFC, or Bluetooth tag is attached to each accessory. At this stage, the theft or loss of internal accessories that are difficult to verify visually can be determined through cross-verification of multiple sensor values. For example, if the current weight at the time of return has decreased by 1.8 kg or 3 kg or more compared to the weight at the time of shipment, the charging cable tag is not recognized, and only the helmet is scanned by the storage compartment's infrared sensor, the controller may interpret this as the charging cable not being returned. Additionally, since it may be difficult to identify side mirrors based solely on weight changes, the tag recognition results and the external image analysis results can be referenced together to determine whether they have not been returned or are missing. According to this step, the status of missing accessories that are difficult to confirm visually can also be cross-verified through multiple sensors.

[0167] The step of finally generating unmanned return control information for billing settlement is to combine the correct position receiving status, return initiation authority approval status, quantified damage information, and accessory normal return status to finally generate unmanned return control information for billing settlement that includes at least one of normal return approval information, damage inspection request information, accessory missing warning information, additional billing settlement information, and administrator confirmation request information.

[0168] Here, "Normal Return Approval Information" may refer to data indicating that the rental contract can be terminated without further action and the vehicle can be transitioned to a next rental availability state. "Damage Inspection Request Information" may refer to data indicating that quantified damage information exceeds a pre-set threshold, requiring further review by an administrator or an advanced analysis module. "Missing Parts Warning Information" may refer to warning data indicating that specific parts were not returned properly. "Additional Charge Settlement Information" may include information on additional costs automatically calculated based on the degree of damage, the type of missing parts, failure to return to the correct location, or violation of return conditions. "Administrator Verification Request Information" may refer to verification request data transmitted to the administrator terminal when the damage or missing status is critical or data inconsistencies occur, making termination impossible through automatic approval alone. All positive and negative inspection results derived from sensors and AI models can be aggregated to finalize system control commands and financial settlement data; if all inspection results are normal, "Normal Return Approval Information" can be generated to notify the driver of the rental termination. Conversely, if the damage settlement index for the right cowl exceeds the threshold and the charging cable is missing, damage inspection request information, missing accessory warning information, and additional charge settlement information may be generated together. Additionally, if biometric authentication matches but failure to return to the correct location and damage information are detected simultaneously, administrator verification request information may be added. The unmanned return control information generated in this manner for charge settlement is transmitted to a server and can be utilized for contract termination, automatic payment, administrator notifications, vehicle maintenance reservations, and subsequent procedures for restoring the vehicle to a rental-ready state.

[0169] And, the step of quantifying at least one of the scratch area, damage depth, crack length, and physical damage rate of the turn signal of the vehicle body exterior comprises: a step of performing image alignment preprocessing including lens distortion correction, illumination normalization, shadow removal, reflection suppression, and alignment of the vehicle body reference coordinate system on the reference exterior image data and the multi-angle incoming exterior image data; a step of dividing a plurality of inspection target areas, including a vehicle body exterior area, a turn signal area, a mirror area, and a lamp cover area, from each of the reference exterior image data and the multi-angle incoming exterior image data on which the image alignment preprocessing has been performed, and generating a pixel-unit difference map, an edge change map, and a texture anomaly map for each inspection target area; and a step of inputting the difference map, the edge change map, and the texture anomaly map into a visual intelligence-based damage judgment model to calculate a damage candidate mask, a damage type classification value, and a damage severity estimate value for each inspection target area. A step of calculating the scratch area of ​​the vehicle body exterior using the number of pixels of the damaged area included in the damage candidate mask and a pre-set image-to-measurement conversion coefficient, estimating the damage depth using the brightness gradient, contour deformation, and discontinuity information between multiple viewpoints of the damaged area, and calculating the crack length based on the skeletonization results of the crack path; a step of quantifying the physical damage rate of the turn signal by calculating the reduction rate of the light-emitting part area, the housing damage ratio, and the fixed position deviation of the exterior image data at the time of return relative to the reference exterior image data for the turn signal area; and a step of calculating a settlement reflection score for each damage by referring to a part-specific importance table, a damage-type-specific depreciation rate table, and a safety-related component weighting table for the scratch area, damage depth, crack length, and physical damage rate of the turn signal.and may include the step of deriving a final damage settlement index by weighting the damage-specific settlement reflection scores, and generating damage judgment result information that is reflected in at least one of normal return approval information, damage inspection request information, additional charge settlement information, and administrator confirmation request information based on the final damage settlement index.;

[0170] The step of performing image registration preprocessing is a step of correcting the reference exterior image data and the multi-angle incoming exterior image data so that they can be compared on the same standard even if they were acquired under different shooting conditions.

[0171] Here, the reference exterior image data may be a reference image of a normal state captured at the time of rental, and the multi-angle incoming exterior image data may be a set of exterior image data acquired through a PTZ camera or multiple cameras at the time of return. Since the shooting angle, illumination, shadow formation method, degree of lens distortion, and reflected light conditions may differ between the time of rental start and the time of return, performing a damage comparison while maintaining these differences may result in undamaged areas being mistaken for damaged, or conversely, actual damage being obscured by background changes and failing to be detected. Lens distortion correction may be a process of correcting barrel distortion or pincushion distortion occurring at the edges of the image due to the use of wide-angle or zoom lenses. Illumination normalization may be a process of mitigating the phenomenon where the same area appears excessively bright or dark by correcting the differences in brightness distribution and histograms between the time of rental and the time of return. Shadow removal may be a process of isolating or removing shadow areas caused by people, structures, or changes in the direction of sunlight around the vehicle, and shading changes unrelated to damage can be mitigated through pixel interpolation or light source model-based correction. Reflection suppression may be a process of correcting bright areas that could be mistaken for damage due to strong reflections on the vehicle body paint surface or turn signal lenses. Vehicle body reference coordinate system alignment may be a process of aligning the reference exterior image and the exterior image at the time of return to the same vehicle body reference point; for example, the two images can be placed in the same coordinate system using reference points such as the center of the front axle, the end point of the handlebars, the corner of the turn signal, the position of the license plate, or the outer edge of the lamp cover. According to this step, comparison errors caused by differences in shooting angles, lighting environments, and camera magnification before and after rental can be reduced, thereby improving the precision of subsequent damage assessment.

[0172] The step of dividing multiple inspection target areas and generating pixel-unit difference maps, edge change maps, and texture anomaly maps for each inspection target area is a step of subdividing the reference appearance image with completed alignment preprocessing and the appearance image at the time of return into the same parts, and generating feature maps that quantitatively express the appearance change patterns that occurred in each part.

[0173] Here, the inspection target area may refer to multiple parts requiring damage inspection, such as the vehicle body exterior area, turn signal area, mirror area, and lamp cover area. The vehicle body exterior area may include cowls, fenders, side panels, or front cowls, and the turn signal area may include the lenses and housings of the front, rear, left, and right turn signals. The mirror area may include the outer edges and connecting parts of the left and right side mirrors, and the lamp cover area may include the transparent cover area protecting the headlights or taillights. Dividing multiple inspection target areas may mean separating parts with different materials, shapes, and functions into individual objects through semantic segmentation techniques or object segmentation algorithms. The pixel-level difference map may be a map generated by calculating the difference in pixel values ​​at the same coordinates between the reference exterior image and the exterior image at the time of return, and can intuitively display noticeable scratches, paint peeling, or color changes. The edge change map can be a map that extracts differences in edge information where shape changes are prominent, such as the outlines of a part or crack lines, and can sensitively detect cases where new fracture lines that did not originally exist are created due to the cowl being dented or plastic breaking. The texture anomaly map can be a feature map that emphasizes abnormal roughness, damage patterns, or changes in reflective properties compared to normal surface texture, and can precisely represent changes such as traces of surface abrasion caused by slipping or fine wear marks. For example, in the vehicle body exterior area, the pixel difference map can represent paint color changes and scratch marks, the edge change map can represent new linear boundaries caused by cracks, and the texture anomaly map can represent changes in surface texture of gouged or scraped areas. According to this step, damage candidates for each part can be represented as a combination of multiple feature maps rather than a single image value, thereby increasing the precision of damage type identification.

[0174] The step of calculating damage candidate masks, damage type classification values, and damage severity estimates is to input the difference map, edge change map, and texture anomaly map into a visual intelligence-based damage determination model to separate areas with actual potential damage in each inspection target area and quantitatively determine the type and degree thereof.

[0175] Here, a damage candidate mask refers to mask data that displays only pixel regions suspected of damage within an image in the form of binary values ​​or multi-classes. In other words, it may be the result of artificial intelligence extracting only pixel clusters with a high probability of actual damage by synthesizing the features of various change maps. The damage type classification value may be a classification result indicating which type of damage it corresponds to, such as scratches, cracks, breakage, dents, detachment, or deformation. The damage severity estimate may be a stepwise or continuous value reflecting the level of surface paint damage, the level of exterior plastic breakage, the possibility of internal structure exposure, or the possibility of functional component damage; it may also be an estimate of whether the damage is merely a cosmetic defect or significant hardware damage that could impair actual driving safety. The visual intelligence-based damage judgment model may have a structure combining a deep learning-based segmentation network and a classification network, and can generate individual damage candidates for each of multiple inspection target areas. For example, in the vehicle body exterior area, a linear damage candidate with a length of 7 cm and a scratch mask with an area of ​​12 cm² can be detected simultaneously, and in the turn signal area, a housing damage type and an estimate of high-severity damage can be calculated. According to this step, it is possible to more accurately distinguish between actual damage, which is difficult to distinguish based on simple image differences alone, and changes caused by shooting errors or contaminants.

[0176] The step of calculating the scratch area, damage depth, and crack length of the vehicle body exterior is to calculate the scratch area of ​​the vehicle body exterior using the number of pixels of the damage area included in the damage candidate mask and a preset image-to-measurement conversion coefficient, to estimate the damage depth using the brightness gradient of the damage area, contour deformation, and disparity information between multiple viewpoints, and to calculate the crack length based on the skeletalization result of the crack path.

[0177] Here, the image-to-measurement conversion factor may be a factor that converts pixel units in an image into actual physical length or area by reflecting the camera's shooting distance, lens focal length, magnification, and correction parameters. The scratch area may be a value converted into cm² units by applying the above conversion factor to the number of pixels classified as scratches within the damage candidate mask. The damage depth may be a depth indicator that estimates whether the damage is merely a surface scratch, whether the exterior material is deeply gouged, or whether there is a possibility of internal structure exposure, by utilizing changes in the brightness gradient of the damaged area, the degree of indentation or protrusion of the outer contour, and parallax information between multiple viewpoints. For example, by analyzing the minute parallax between images captured by a stereo camera or from multiple angles, it is possible to calculate whether the scratch is gouged to a depth of several millimeters or if only the paint surface has been peeled off. The crack length may be a value calculated by determining the centerline length after extracting the connected crack paths from the crack candidate area and performing skeletal analysis. Skeletonization is a technique that thins cracks represented by thick pixel areas into centerlines 1 pixel thick, allowing for a more accurate calculation of the actual length of curved cracks. For example, a scratch area of ​​18 cm², a damage depth of 2.4 mm, and a crack length of 6.8 cm can be calculated simultaneously on the right cowl. According to this step, damage can be converted into concrete numerical values ​​that can be used for actual assessment and maintenance decisions, rather than simply being judged as existing.

[0178] The step of quantifying the physical damage rate of the turn signal involves calculating the reduction rate of the light-emitting part area, the housing damage ratio, and the fixed position deviation of the exterior image data at the time of return relative to the reference exterior image data for the said turn signal area, thereby quantifying the physical damage rate of the turn signal. Since turn signals are often positioned protruding from the outer surface of the vehicle body, they may be one of the safety-related components that are most likely to be damaged first in the event of a rollover in place or a side slip.

[0179] Here, the light-emitting area reduction rate may be a value indicating how much the effective area of ​​the lens or cover through which actual light passes has decreased compared to the normal state. The housing damage ratio may refer to the ratio of the area within the outer housing of the turn signal where cracks, breakage, or detachment have occurred. The fixed position deviation may be a value indicating how much the turn signal has been displaced from the reference position due to rotation, tilting, or detachment. For example, even if the turn signal is not completely broken, if the mount area is torn and sags downward or is bent sideways, the corresponding position deviation may be calculated significantly. The server or image analysis module can calculate the physical damage rate of the turn signal as a percentage by weightedly combining the three values ​​mentioned above. For example, if the light-emitting area reduction rate is 20%, the housing damage ratio is 35%, and the fixed position deviation exceeds the allowable standard by twice, the final turn signal damage rate may be calculated as 42%. According to this step, turn signal damage can be expressed as a numerical value that reflects not only whether it is simply broken, but also the degree of functional and structural loss.

[0180] The step of calculating the settlement reflection score by damage is a step of calculating the settlement reflection score by damage by referring to the importance table by part, the depreciation rate table by damage type, and the safety-related component weighting table for the scratch area, damage depth, crack length, and physical damage rate of the direction indicator.

[0181] Here, the importance table by part refers to a table in which the settlement impact is set differently depending on the location where the damage occurs, even if the damage is the same. For example, minor scratches on the lower part of the vehicle body exterior may have a relatively low importance, while damage to areas around turn signals, headlights, or the steering system, or structural parts directly related to passenger safety, may have a high importance. The depreciation rate table by damage type refers to a table that determines the reduction in vehicle value or the reflection rate of repair costs for each type of damage, such as scratches, cracks, dents, or detachment. For example, a higher depreciation rate may be assigned to penetrating cracks or broken detachment than to shallow surface scratches. The safety-related part weighting table refers to a table that assigns additional weights to parts that directly affect safe driving, such as turn signals, lamp covers, mirrors, or brake levers. The server can calculate a settlement reflection score for each type of damage by querying the table values ​​corresponding to each of the above damage figures and combining them. For example, a 35% housing damage to a turn signal may be assigned a higher settlement score than an 18 cm² scratch on the right cowl. According to this step, even for damage of the same area, differential settlement is possible depending on the location, type, and impact on safety.

[0182] The step of deriving a final damage settlement index and generating damage determination result information is to derive a final damage settlement index by weighting the settlement reflection scores for each damage, and to generate damage determination result information that is reflected in at least one of normal return approval information, damage inspection request information, additional charge settlement information, and administrator confirmation request information based on the final damage settlement index.

[0183] Here, the final damage settlement index may be a value obtained by integrating the individual settlement reflection scores for multiple types of damage into a single comprehensive score. A weighted average may be calculated by reflecting the importance, safety impact, and whether each type of damage involves a functional part. For example, damage to turn signals and cracks may be reflected with higher weights than simple exterior scratches. Damage assessment result information may be generated based on which of the pre-set standard ranges the final damage settlement index falls into. For example, if the final damage settlement index is very low, it may be reflected in the normal return approval information; if it is at a medium level or higher, it may be reflected in the damage inspection request information and additional charge settlement information; and if it is extremely high, or if the damage types are complex and concentrated on high-risk parts, administrator verification request information may also be generated. For example, if the final damage settlement index is 10 points or less, it can be processed as a normal return approval; if it exceeds 10 points but is 40 points or less, additional charge settlement information can be generated; and if it exceeds 40 points or if the turn signal and lamp cover are simultaneously determined to be severely damaged, administrator verification request information and damage inspection request information can be generated simultaneously. According to this step, the final approval of the unmanned return procedure, additional cost charges, and whether an administrator intervenes can be consistently determined based on a single indicator that aggregates individual damage items.

[0184] According to one embodiment, a method for providing a two-wheeled vehicle sharing platform service utilizing idle two-wheeled vehicle and space information, wherein a server includes a processor, memory, a communication module, and a non-transient storage medium, and the processor executes a program stored in the non-transient storage medium, the method comprises: receiving service request information from a driver terminal including driver identification information, two-wheeled vehicle usage request information, and usage condition information; performing authentication for the driver based on the driver identification information and obtaining driver information including usage qualification information, usage history information, and payment method information associated with the driver identification information; and selecting a target two-wheeled vehicle and a target inbound / outbound space corresponding to the service request information based on operational resource information including two-wheeled vehicle information and storage space information. The method provides a motorcycle sharing platform service utilizing idle motorcycle and space information, comprising: a step of calculating rental condition information including at least one of approval status for use, usage fee, warranty condition, insurance coverage condition, and entry / exit processing condition for the target motorcycle, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target entry / exit space information, and pre-set operation policy information; and a step of selecting the target motorcycle and target entry / exit space, comprising: a step of selecting sharing candidate motorcycles and sharing candidate spaces by reflecting available time intervals, available area intervals, charging or fuel status, vehicle type suitability, and space access convenience corresponding to the motorcycle usage request information, based on a plurality of idle motorcycle information and idle space information included in the operation resource information; and a step of determining the target motorcycle and target entry / exit space to be assigned to the driver based on the driver information, sharing candidate motorcycle, sharing candidate space, and pre-set operation policy information.

[0185] The step of receiving service request information is a step in which the server receives service request information including driver identification information, motorcycle usage request information, and usage condition information from a driver terminal, and forms reference data for the authentication procedure, sharing candidate resource selection procedure, target motorcycle and target entry / exit space determination procedure, and rental condition information calculation procedure performed thereafter.

[0186] Here, the term "driver terminal" may refer to a smart device, such as a smartphone, tablet, PC, or vehicle-linked terminal, that requests a rental by accessing a dedicated application or web interface of a two-wheeled vehicle sharing platform. The term "driver identification information" may refer to data intended to identify the user, such as a platform account identifier, a unique member number, a personal authentication token, a mobile phone-based authentication value, a terminal's unique identification number, or an identifier linked to biometric authentication. The term "two-wheeled vehicle usage request information" may include the scheduled rental start time, scheduled rental end time, estimated usage time, scheduled departure location, scheduled return location, desired vehicle type, estimated driving distance, preference for electric two-wheeled vehicles, or estimated load capacity information. The term "usage condition information" may include the scope of insurance coverage requested by the borrower, deposit limits, permission for night use, provision of charging cables, return method, restrictions on return locations, preferred parking conditions, or rental of additional equipment such as helmets. For example, if a driver requests a 125 cc scooter for urban business travel from 9 a.m. to 1 p.m., allows electric two-wheeled vehicles, designates a return location near a station, and selects basic insurance and a medium deposit condition, the server can receive the entire request as a single service request information and use it as an input value for subsequent operations. This step may not be a simple reservation acceptance step, but rather a step that forms an initial data set serving as a judgment criterion for the entire sharing service.

[0187] The step of performing authentication for a driver and obtaining driver information is to verify whether the actual sharing requester is a legitimate user based on the driver identification information, and to retrieve driver information from a database that includes user eligibility information, usage history information, and payment method information associated with the driver identification information.

[0188] Here, eligibility information may refer to information used to determine whether a two-wheeled vehicle can be used legally, such as license type, license validity, fulfillment of age criteria, status of service usage restrictions, account status, and completion of identity verification. Usage history information may include accumulated behavioral history such as the number of past two-wheeled vehicle uses, return delay history, reservation cancellation frequency, accident or damage history, payment failure history, return history after leaving the designated location, traffic law violation records, or preferred vehicle types and preferred return locations. Payment method information may include pre-registered card information, simple payment integration status, deposit withdrawal availability, automatic payment approval status, and the cumulative number of payment failures. By verifying the above driver information, the server can filter out unqualified users, repeat non-payers, or users with usage restrictions at an initial stage, and can also reflect the above driver information in the subsequent calculation of the borrower trust index and the confirmation of rental condition information. For example, if a user holds a Class 2 small vehicle license, has a payment method properly registered, has no accident history during the past 12 uses, and has one or fewer delayed returns, the server may determine that the user is a properly authenticated driver and continue to perform subsequent procedures. On the other hand, if the license has expired or there is a history of repeated payment failures, the server may restrict the approval of use for the said driver or require additional guarantee conditions.

[0189] The step of selecting target motorcycles and target inbound / outbound spaces is a step of determining the actual assets to be allocated corresponding to the service request information based on operational resource information including motorcycle information and storage space information.

[0190] Here, operational resource information may refer to a database that combines information on motorcycles, which includes the current location, availability, charging status or remaining fuel amount, vehicle type, maintenance status, whether they are scheduled for recent retrieval, and information on the owner's planned use for each of multiple idle motorcycles registered by ordinary individuals to lend out during times when they are not using them, rather than assets directly owned by the platform company; and information on storage spaces, which includes the occupancy status, accessibility, type of space, availability of charging facilities, availability of CCTV, nighttime illumination conditions, and accessibility convenience for each of multiple idle spaces registered to provide as rental locations, such as parking lots, yards, or vacant spaces owned by individuals. Based on the information on multiple idle motorcycles and idle spaces included in the operational resource information, the step of selecting candidate motorcycles and candidate spaces for sharing by reflecting the available time interval, available area interval, charging or fuel status, vehicle type suitability, and space accessibility convenience corresponding to the motorcycle usage request information may be a step of forming a primary candidate group that meets the conditions from the entire resource group. Here, the available time interval may refer to the time period during which the requested rental time overlaps with the actual non-occupancy time of the vehicle or space, and the available area interval may refer to the distance or route suitability range between the planned departure and return locations. Vehicle suitability may be a value determining whether the vehicle has an engine displacement the borrower can handle, is suitable for the purpose of use and expected load, or matches the driver's past preferred vehicle types. Space accessibility convenience may be a value reflecting whether the space is located in a position easily accessible to the borrower on foot or by public transportation, whether the difficulty of entering the parking area is not excessive, and whether the surrounding traffic congestion or security environment is satisfactory.The system can generate a list of multiple candidate two-wheeled vehicles and multiple candidate spaces for sharing by comprehensively filtering factors such as whether the borrower's desired rental time overlaps with the sharing allowance time set by the vehicle / space owner, whether the battery charge or fuel status is sufficient, and the suitability of the distance from the requested area. For example, two scooters and one electric two-wheeled vehicle located within 800 m of the planned return location, with a battery charge of 72% or more, a normal recent maintenance history, and no other reservations from the planned departure time to the planned return time, can be selected as candidate two-wheeled vehicles for sharing, and three spaces with high accessibility, installed CCTVs, and meeting nighttime illumination standards can be selected as candidate spaces for sharing.

[0191] Based on driver information, sharing candidate motorcycles, sharing candidate spaces, and pre-set operation policy information, the step of determining the target motorcycle and target entry / exit space to be assigned to the driver is the step of finally selecting the optimal combination that most appropriately satisfies the borrower's reliability history, vehicle characteristics, and physical conditions of the space from among the primary selected candidate groups.

[0192] Here, operational policy information may include vehicle allocation priority, space allocation priority, priority allocation rules for electric motorcycles, owner protection criteria, space utilization optimization criteria, safety criteria, and settlement policies pre-set by the platform. In the aforementioned confirmation stage, the final allocation combination can be derived by comprehensively considering not only distance or remaining capacity, but also the driver's past usage history, preferred vehicle type, return tendencies, the owner's schedule of candidate vehicles, the occupancy potential of candidate spaces, and operational efficiency. Instead of matching vehicles solely based on proximity, the system can verify from various angles whether the borrower's accident history and reliability level included in the driver information meet the safety standards required by the vehicle owner, whether there is a risk of rollover during pickup or return if a heavy, large motorcycle candidate is matched to a candidate space with a steep incline, or whether there is a possibility of collision with the next reservation at the time of return. For example, even if two candidate vehicles are available at the same time, if one vehicle matches the borrower's past preferred vehicle type and the other vehicle is scheduled for the owner's next use soon, the server may confirm the former as the priority allocation target. Similarly, even if both candidate spaces are available for return, if one space is brighter, equipped with CCTV, and does not violate entry restriction hours, the server can confirm that space as the target entry / exit space. The combination that passes verification and possesses the highest matching suitability can be confirmed as the final target motorcycle and target entry / exit space and assigned to the borrower. According to this step, it becomes possible to determine the optimal sharing resource that considers safety, operational efficiency, owner benefits, and borrower convenience together, rather than simply assigning resources of proximity.

[0193] The step of calculating rental condition information is a step of generating rental condition information including at least one of the following: approval of use of the target motorcycle, usage fee, warranty condition, insurance coverage condition, and inbound / outbound processing condition, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target inbound / outbound space information, and pre-set operation policy information.

[0194] Here, the grade information of the target motorcycle may include displacement ranges, whether it is electric or internal combustion engine, vehicle price range, drive system, and classification values ​​for settlement criteria groups or insurance risk groups. The information on the target entry / exit space may include the type of space, owner information, accessibility, security level, availability of charging infrastructure, and usage restrictions by time of day. The pre-configured operational policy information may be system rules predefined, such as the commission rate received by the platform for mediating transactions, regulations on rate surcharges based on the borrower's risk level, the profit sharing ratio between the vehicle owner and the space owner, deposit policies, insurance special clause policies, and entry / exit restriction rules. Whether usage is approved may be determined by comprehensively considering the qualifications and reliability of the driver and the conditions of the target motorcycle and the target entry / exit space, and the usage fee may be calculated by reflecting the combined cost of sharing, platform brokerage fees, idle time matching rate, and operational policies. Guarantee conditions may include whether an additional deposit is required, the level of deductibles, and restrictions based on risk grades, while insurance coverage conditions may include whether a platform-exclusive short-term insurance rider is subscribed to, the scope of coverage, and restrictions on use during high-risk hours. Deposit and withdrawal processing conditions may include available withdrawal times, permitted return spaces, space access rules, restrictions on space occupancy time, and conditions regarding costs to be settled with the space owner. For example, a basic deposit and standard insurance conditions may apply to drivers with a high borrower reliability index; however, if the reliability deviation is small or usage during high-risk hours is involved, conditions requiring an additional deposit and the subscription to an insurance rider may be applied together. Furthermore, if the target deposit and withdrawal space is available only during nighttime hours, return conditions within that timeframe and security weighting costs may be reflected in the rental condition information.This stage is not merely a procedure for calculating fees, but may be a stage for establishing the rights and obligations, cost sharing, and risk sharing structures among borrowers, motorcycle owners, space owners, and the platform in the form of data.

[0195] In addition, the step of determining the target motorcycle and target entry / exit space comprises: deriving a borrower reliability index by quantifying the history of past motorcycle damage, traffic law violations, return delay frequency, reservation cancellation frequency, and payment failure history included in the driver information, and determining the primary matching eligibility for each sharing candidate motorcycle by comparing the borrower reliability index with a minimum reliability threshold for rental allowance previously set by the owner of the sharing candidate motorcycle; A step of calculating the suitability of a motorcycle-specific entry / exit space reflecting entry / exit stability, parking suitability, and storage stability by comparing the floor slope, road surface flatness, inclusion of security CCTV coverage, availability of charging infrastructure for electric motorcycles, availability of nighttime illumination, and availability of a rain-blocking structure of the sharing selection space with the drive method, body weight, vehicle type, wheelbase, and steering angle specifications of the sharing selection motorcycle; A step of simulating the actual driving range by reflecting the current battery state of charge or remaining fuel amount, recent energy consumption or fuel consumption per unit distance, and correction coefficients by drive method of the shared selected motorcycle, based on the estimated driving distance, scheduled time of use, elevation change profile of the scheduled driving route, estimated load, and ambient temperature information included in the above motorcycle usage request information, and calculating the risk of power depletion during driving based on the difference between the actual driving range and the estimated driving distance; a step of calculating the estimated idle duration, which is the time interval expected to be available until the next reservation, retrieval, maintenance, and owner usage schedule of the shared selected motorcycle begins, and deriving an idle time match rate by calculating the temporal intersection ratio between the estimated idle duration and the driver's scheduled rental time;A step of deriving a combined sharing cost by summing the motorcycle sharing revenue unit price set by the owner of the selected sharing motorcycle, the space sharing revenue unit price set by the owner of the selected sharing space, the estimated charging cost or fuel preservation cost, and the estimated recovery cost of the selected sharing motorcycle; a step of calculating a final sharing matching score for each combination of the selected sharing motorcycle and the selected sharing space by weighted summing the borrower reliability index, the suitability of the motorcycle-specialized entry / exit space, the reverse penalty corresponding to the risk of power depletion, the idle time matching rate, the combined sharing cost, the driver's past preferred vehicle type information, and the driver's past preferred return area information; and, for a plurality of combinations with the highest final sharing matching scores, verify whether there are duplicate reservations in the same time slot, whether there is a possibility of exceeding the occupancy of the same space, and whether there is a conflict with the next reservation, and then determine the combination with the highest final matching score reflecting the verification results as the target motorcycle and target entry / exit space; may be included.;

[0196] The step of deriving a borrower reliability index and selecting a sharing screening motorcycle involves quantifying the borrower's transaction stability and operational reliability by quantifying the history of past motorcycle damage, traffic law violations, return delays, reservation cancellations, and payment failures included in the driver information, and comparing this with the minimum reliability threshold for rental allowance set by the owner of each sharing candidate motorcycle to primarily select only the vehicles eligible for matching.

[0197] Here, the borrower reliability index may be a comprehensive score indicating the likelihood that a borrower will safely and faithfully use and return a motorcycle, which is the asset of another party, and may be quantitative data converted to a maximum of 100 points or a pre-established standard range. The history of motorcycle damage may be a value reflecting the frequency or severity of past events that caused asset damage, such as accidents, rollovers, cosmetic damage, or loss of parts, and the history of traffic violations may be a value reflecting the cumulative degree of road traffic violations, such as speeding, signal violations, or failure to wear safety equipment. The frequency of return delays and reservation cancellations may be indicators representing adherence to time commitments and the faithfulness of contract fulfillment, and the history of payment failures may be an indicator representing financial reliability. The server may derive the borrower reliability index after normalizing each of the above items by assigning pre-established weights; for example, higher penalty weights may be assigned to accident history and payment failure history, while relatively lower penalty weights may be assigned to simple reservation cancellations. The minimum trust threshold for rental permission may be a minimum trust baseline pre-set by each vehicle owner when registering their motorcycle on a sharing platform. For example, owners of expensive, large motorcycles may set a threshold to allow rentals only to drivers with a trust index of 85 points or higher, while owners of small, everyday scooters may set it at 65 points or higher. The server compares the borrower's trust index with the aforementioned threshold to determine the initial matching eligibility of each candidate motorcycle, and can select only those vehicles that meet the threshold as motorcycles for sharing. For instance, if a borrower's trust index is calculated at 70 points and a specific vehicle owner has set the threshold at 80 points, that vehicle is eliminated during the initial matching eligibility determination, and only vehicles that pass this criterion can advance to the matching pool in the subsequent stage. This process prevents indiscriminate matching between drivers and vehicles and allows for the reflection of vehicle owners' asset protection needs at an early stage.

[0198] The step of calculating the suitability of a specialized loading and unloading space for two-wheeled vehicles is a step of quantitatively evaluating whether the space is suitable for actual loading, unloading, and storage by cross-comparing the physical conditions of the space and the specifications of the vehicle with respect to the sharing-selected two-wheeled vehicles and the corresponding sharing-selecting space.

[0199] Here, the term "sharing selection space" may refer to a space among the aforementioned sharing candidate spaces that is subject to review in combination with a specific sharing selection motorcycle. The floor slope of the aforementioned sharing selection space may be a value that affects the likelihood of the motorcycle tipping over when stopped or released, and the flatness of the road surface may be a value that affects stand support stability and wheel alignment stability. The inclusion of security CCTV coverage may be related to the effect of deterring vehicle theft or damage, and the availability of charging infrastructure for electric motorcycles may directly affect the possibility of re-operating the electric motorcycle after return. The availability of nighttime illumination may be a value that affects safety during late-night returns or releases, and the availability of a rain-blocking structure may be a value that affects the suitability of storage for vehicles or electric motorcycles that are vulnerable to external exposure. The server may compare the aforementioned space attributes by correlating them with the drive system, body weight, vehicle type, wheelbase, and steering angle specifications of the aforementioned sharing selection motorcycle. For example, large manual motorcycles with heavy body weight may face a high risk of losing balance when exiting in steeply sloped spaces, and vehicles with long wheelbases may struggle to secure a turning radius in confined areas. Furthermore, if a large manual motorcycle weighing over 200 kg is assigned to an idle space in an alleyway with a slope of 15 degrees or more, the system may deem it unsuitable because there is a high risk that the rider will overturn in place due to the weight while attempting to manually push the vehicle after dismounting. Conversely, while lightweight electric scooters may be suitable for small spaces, their operational efficiency may be reduced in areas lacking charging infrastructure. The suitability rating for motorcycle-specific entry and exit spaces is a value calculated as a comprehensive percentage score or normalized figure by cross-analyzing the motorcycle's inherent physical specifications with the physical environment of the private parking space to determine whether the vehicle can safely enter and exit that space.According to this step, the risk of tipping, difficulty in delivery, or damage during storage can be reduced by selecting a space that matches the vehicle's physical properties, rather than allocating it simply because it is empty.

[0200] The step of calculating the risk of power depletion during driving is to simulate the actual driving distance based on the current battery charge status or remaining fuel amount of the sharing selected motorcycle, the recent energy consumption or fuel consumption per unit distance, and correction coefficients by driving method, based on the estimated driving distance, scheduled time of use, elevation change profile of the planned driving route, estimated load, and ambient temperature information included in the motorcycle usage request information, and to calculate the risk of power depletion during driving based on the difference between the actual driving distance and the estimated driving distance.

[0201] Here, the current battery charge status may be a value representing the remaining battery charge as a percentage for electric motorcycles, and the remaining fuel amount may be a value representing the amount of fuel remaining in the fuel tank in liters for internal combustion engine motorcycles. The recent energy consumption or fuel consumption per unit distance may refer to the actual energy or fuel consumption per kilometer consumed by the motorcycle under recent operating conditions. The elevation change profile of the planned route may be data representing the ratio of uphill, downhill, and flat sections, and the estimated load may refer to additional loads caused by delivery boxes, luggage, or the presence of a passenger. The ambient temperature may be an environmental variable that can affect battery efficiency or fuel consumption efficiency. The correction factor by drive mode may be a coefficient intended to reflect the difference in power consumption characteristics between electric motor-driven vehicles and internal combustion engine-driven vehicles. The elevation change of the planned route, the estimated load, and the ambient temperature may be key variables that drastically increase the energy consumption of the motorcycle. For example, under ambient temperature conditions of minus 10°C, power may be consumed much faster than the manufacturer's certified efficiency when driving on a route that includes steep mountain roads with an increased number of passengers or cargo loads. The server can simulate the actual driving range by combining the above variables, and if the actual driving range is sufficiently greater than the estimated driving range, the risk of power depletion can be assessed as low. Conversely, if the difference between the two distances is small or negative, the likelihood of discharge or fuel shortage during travel can be assessed as high. For example, if the estimated driving range is 42 km and the simulated actual driving range is 38 km, the server can assign a high risk of power depletion to this combination. Additionally, if the actual driving range is shorter than the estimated driving range or the margin is less than 10%, the risk of power depletion can be calculated to increase sharply.According to this step, it is possible to prevent situations where the vehicle stops during the rental period and contribute to selecting only combinations that match the user's driving purpose and the vehicle's actual energy status.

[0202] The step of calculating the expected duration of idle time and deriving the idle time match rate is to calculate how long the sharing selected motorcycle is expected to be available for additional use before the next reservation, retrieval, maintenance, and owner usage schedule begins, and to calculate the degree of match between that time interval and the borrower's expected rental time.

[0203] Here, the estimated idle duration refers to the time interval during which the aforementioned shared selected motorcycle is expected to be available for use from the current point in time until it can no longer be provided for sharing due to a next reservation or retrieval schedule. The next reservation may be the start time of another borrower's reservation, the retrieval schedule may be the time when the vehicle owner retrieves the vehicle for personal use, and the maintenance schedule may be the time scheduled by the platform or the owner to retrieve the vehicle for inspection. Additionally, the estimated idle duration may refer to the pure available time during which the vehicle can be fully rented to others, excluding essential buffer time required for vehicle retrieval or battery charging, starting from the deadline time set by the vehicle owner for their own use. The server may calculate the estimated idle duration based on the earliest point in time among the aforementioned schedules. Subsequently, the idle duration match rate can be derived by calculating the ratio of the temporal intersection between the estimated idle duration and the driver's scheduled rental time. For example, if a vehicle is scheduled to remain idle for the next 6 hours and the borrower's scheduled rental time is 4 hours, the idle time match rate can be calculated as relatively high. Additionally, if the vehicle's scheduled idle time is a total of 3 hours from 1:00 PM to 4:00 PM, but the borrower requests to use the vehicle for 2 hours and 30 minutes from 1:00 PM to 3:30 PM, this can result in a very high idle time match rate score because it consumes most of the total idle time. Conversely, if the vehicle's scheduled idle time is 3 hours and the borrower's scheduled rental time is 5 hours, the idle time match rate may be low or close to zero; the match rate may also be low if only 30 minutes of the 3 hours of idle time are requested. According to this step, it is possible to quantify how well the requested usage time matches the vehicle's actual available time, and it is possible to prevent vehicles with insufficient remaining idle time from being misassigned or idle time from being inefficiently fragmented.

[0204] The step of deriving the cost of the sharing combination is to calculate the total cost that the platform must consider for using the combination of the vehicle and space by summing the unit price of the sharing revenue set by the owner of the sharing selected two-wheeled vehicle, the unit price of the space sharing revenue set by the owner of the sharing selected space, the estimated charging cost or fuel preservation cost, and the estimated recovery cost of the sharing selected two-wheeled vehicle.

[0205] Here, the two-wheeled vehicle sharing revenue unit price may be the basic rental revenue unit price to be settled with the vehicle owner, and the space sharing revenue unit price may be the space usage revenue unit price to be settled with the space owner. The estimated charging cost may be the electricity cost required for charging after returning the electric two-wheeled vehicle, and the fuel maintenance cost may be the cost required for refueling after returning the internal combustion engine vehicle. The estimated recovery cost may be the transportation cost that the platform or owner may bear if the vehicle needs to be moved to another location or recovered after being left in the designated return space. The above-mentioned combined sharing cost may refer to standard cost data obtained by summing up the compensation requested by the vehicle owner and the space owner, respectively, as well as the incidental costs required for the recovery of the vehicle after operation. The server may calculate the combined sharing cost by summing the above cost elements, and this cost may be directly reflected in the calculation of subsequent usage fees and the final matching score. For example, if a vehicle sharing revenue unit price of 9,000 won, a space sharing revenue unit price of 2,500 won, an estimated charging cost of 800 won, and an estimated recovery cost of 1,500 won are required, the combined cost of the sharing combination can be derived as 13,800 won. According to this step, a reasonable matching becomes possible that considers not only convenience from the user's perspective but also actual operating costs and settlement burdens.

[0206] The step of calculating the final sharing matching score is to calculate the final sharing matching score for each combination of the sharing selected motorcycle and the sharing selected space by weighting and summing the borrower reliability index, the suitability of the motorcycle-specific entry / exit space, the reverse penalty corresponding to the risk of power depletion, the idle time matching rate, the sharing combined cost, the driver's past preferred vehicle type information, and the driver's past preferred return area information.

[0207] Here, the reverse penalty corresponding to the risk of power depletion refers to a penalty value in the form of a deduction from the final score as the risk of power depletion increases. Information on a driver's past preferred vehicle types may be data indicating a tendency to repeatedly select specific vehicle groups, and information on a driver's past preferred return locations may be data indicating frequently returned locations or locations with high return success rates. Weighted summation can be performed by reflecting the importance each element has on the quality of the final match; for example, in a policy that prioritizes vehicle owner protection, the weight of the borrower trust index can be set high, while in a policy that prioritizes operational efficiency, the weights of the sharing combined cost and idle time match rate can be set high. Additionally, in a safety-first policy, greater weight can be assigned to the suitability of the two-wheeled vehicle-specific entry / exit space and the reverse penalty corresponding to the risk of power depletion. For example, under a policy where accident prevention is the top priority, high weights may be multiplied to spatial suitability and reverse penalty, while under a policy where securing platform revenue or margins is the priority, high weights may be assigned to combinations with lower sharing combination costs. The server can calculate a final sharing matching score by synthesizing the above factors for each combination. For instance, since one combination may have high reliability and preferred vehicle suitability but high costs and a low idle time match rate, and another combination may have low costs and a high time match rate but low spatial suitability, the score can be calculated as a balanced value that reflects all these factors. According to this step, an optimized matching score can be generated that integrates reliability, safety, energy suitability, costs, and user preferences, rather than fragmentary conditions.

[0208] The step of determining the combination with the highest final matching score involves verifying, for multiple combinations with the highest sharing final matching scores, whether there are duplicate reservations in the same time slot, whether there is a possibility of exceeding the occupancy limit of the same space, and whether there is a conflict with the next reservation; and then determining the combination with the highest final matching score reflecting the verification results as the target motorcycle and target entry / exit space. Here, whether there are duplicate reservations in the same time slot may refer to whether a specific vehicle or a specific space is already assigned to another reservation within the same time interval. Whether there is a possibility of exceeding the occupancy limit of the same space may refer to whether there is a possibility that multiple vehicles can be placed in a single space simultaneously, and whether there is a conflict with the next reservation may refer to whether the completion of the current assignment will overlap temporally with the next reservation or the owner's usage schedule. This verification may be a procedure to exclude from the final assignment even high-scoring combinations if there is a conflict between the actual timetable and the reality of the physical space. For example, checking for duplicate reservations in the same time slot may be a procedure to verify whether the vehicle that was ranked first until just before has already been claimed by another user due to network transmission delays or payments made by other users that occurred almost simultaneously. The verification of whether occupancy exceeds the capacity of the same space may be a procedure to check whether return reservations are scheduled for more than the actual capacity of a specific private parking space. For example, if two vehicles scheduled for return at the same time already exist in a space capable of accommodating two vehicles, assigning one additional vehicle would result in physical space over-occupancy, so that combination may be excluded. The server primarily performs the aforementioned collision verification on multiple combinations that have received the highest final sharing matching scores, and may apply penalty points or exclude combinations where a collision is confirmed. Subsequently, reflecting the verification results, the combination maintaining the highest final score may be determined as the target motorcycle and target entry / exit space.For example, if a combination with the highest initial score is excluded due to confirmed duplicate bookings for the same time slot, and the next-ranked combination satisfies all conditions without conflict, the server can confirm the latter as the final combination. According to this step, unrealistic combinations with only high scores are prevented from being selected, and only combinations capable of actual service operation can be confirmed as the final allocation targets.

[0209] And, the step of calculating the above rental condition information comprises: a step of calculating a basic usage fee by adding a pre-set platform brokerage fee, payment processing fee, and operational management fee to the sharing combination cost derived during the process of determining the above target motorcycle and target entry / exit space, and deriving a final usage fee to be charged to the borrower by applying a dynamic discount rate proportional to the idle deadlock resolution contribution of the above target motorcycle and target entry / exit space to the above basic usage fee if the above idle time matching rate is above a pre-set optimization threshold; and a step of calculating the space sharing revenue to be settled to the owner of the above target entry / exit space based on the space sharing revenue unit price included in the above sharing combination cost, wherein at least one of 1) an electricity charge amount based on the estimated charging amount if the above target motorcycle is an electric motorcycle, 2) a security weighting cost proportional to the occupancy time of a security CCTV if the scheduled usage time of the target entry / exit space corresponds to nighttime hours, and 3) a management surcharge corresponding to the entry restriction time of the target entry / exit space is additionally reflected to finally determine the above space sharing revenue; A step of calculating the motorcycle sharing revenue to be settled to the owner of the target motorcycle based on the motorcycle sharing revenue unit price included in the above-mentioned sharing combination cost, and calculating a depreciation adjustment cost by reflecting at least one of the predicted tire wear, brake pad consumption, battery charge / discharge degradation, and drivetrain load index based on the estimated driving distance, planned usage time, estimated load, and elevation change profile of the planned driving route included in the motorcycle usage request information, and finally determining the motorcycle sharing revenue by reducing the platform brokerage fee as the depreciation adjustment cost increases and adding an amount corresponding to the reduction to the motorcycle sharing revenue unit price;The method may include the step of calculating a reliability deviation between the borrower reliability index and a minimum reliability threshold for rental allowance set by the owner of the target motorcycle, and if the reliability deviation is less than a set safety margin range, generating insurance and guarantee conditions including at least one of the following to hedge sharing risk: increasing the self-damage deductible applied to the borrower in the event of an accident, a condition for subscribing to a platform-exclusive short-term insurance special clause, a condition for depositing an additional security deposit, and a condition restricting use during high-risk hours; and the step of finally determining rental condition information by combining the final usage fee, space sharing revenue, motorcycle sharing revenue, platform brokerage fee, and insurance and guarantee conditions, wherein the cost burden relationship, profit distribution relationship, scope of usage liability, and scope of compensation liability in the event of an accident are defined among the borrower, motorcycle owner, space owner, and platform.

[0210] The step of calculating the basic usage fee and deriving the final usage fee is to calculate the basic usage fee to be initially presented to the borrower by adding the pre-set platform brokerage fee, payment processing fee, and operation management fee to the sharing combination cost derived during the process of determining the target motorcycle and target loading / unloading space, and to derive the final usage fee to be actually charged to the borrower by applying a dynamic discount rate proportional to the idle deadlock resolution contribution of the target motorcycle and target loading / unloading space when the idle time matching rate is above a pre-set optimization threshold.

[0211] Here, the combined sharing cost may be a standard cost calculated by summing the unit price of two-wheeled vehicle sharing revenue paid to vehicle owners, the unit price of space sharing revenue paid to space owners, estimated charging or fuel conservation costs, and estimated recovery costs. Platform brokerage fees, payment processing fees, and operational management costs may refer to the essential margins and expenses required for the platform operator to provide the matching system, maintain the payment network, and operate servers and customer centers. The server may calculate the basic usage fee to be charged to the borrower by adding the above elements to the combined sharing cost allocated to the vehicle and space providers. An idle time matching rate exceeding a pre-set optimization threshold may refer to a case where a borrower's rental request efficiently fills the idle time of vehicles and spaces, thereby contributing to reducing resource fragmentation from the platform's perspective. The idle deadlock resolution contribution may be a numerical value indicating how effectively the borrower occupied and utilized the spare time segments where vehicles or spaces might otherwise remain unrented. For example, if the optimization threshold is 70% and the borrower's idle time matching rate is calculated to be high at 90%, this can be evaluated as a contribution that significantly increased the platform's resource turnover. Accordingly, the final usage fee can be calculated by subtracting a certain percentage of a dynamic discount rate from the base usage fee; for example, if the base usage fee is 15,000 won and the idle deadlock resolution contribution is calculated to be high, a dynamic discount rate of 8% or 15% can be applied to derive a final usage fee of 13,800 won or an equivalent level. According to this step, the platform can stably recover operating costs while providing price incentives to borrowers who increase resource utilization.

[0212] The step of finally determining the space sharing revenue is to calculate the space sharing revenue to be settled to the owner of the target entry / exit space based on the space sharing revenue unit price included in the sharing combination cost, and to finally determine the space sharing revenue by additionally reflecting at least one of the following: a power charge amount based on the expected charging amount when the target motorcycle is an electric motorcycle, a security weighting cost proportional to the occupancy time of a security CCTV when the scheduled time of use of the target entry / exit space corresponds to nighttime hours, and a management surcharge corresponding to the time of restricted entry of the target entry / exit space.

[0213] Here, the unit price of space sharing revenue may be a base amount set by the space owner in exchange for providing their parking lot, yard, idle space in front of a commercial building, or vacant space on an hourly or per-transaction basis. If the subject motorcycle is an electric motorcycle, the electricity charge based on the estimated charging requirement may be an additional compensation amount reflecting the amount of charging power and electricity costs necessary for re-operation after return, or a value covering the actual costs associated with using the personal outlet or charger provided by the space owner. The security weighting cost may be a cost reflecting the additional value or risk burden associated with operating security CCTVs during nighttime hours or providing lighting and security infrastructure, and the management surcharge may be a management fee charged when vehicle access is restricted during specific times or when greater intervention by the space operator is required. For example, if the space is scheduled to be returned after 11:00 PM and the space also provides electric motorcycle charging facilities, the final space sharing revenue can be determined at 3,700 won by adding a power charge of 700 won and a security weighting cost of 500 won to the space sharing revenue unit price of 2,500 won. According to this step, the space owner can receive a settlement that reflects not only simple space rent but also actual operational burdens and value provided, and as a result, can increase the voluntary participation rate of idle space providers and secure a stable storage base.

[0214] The step of finally determining the two-wheeled vehicle sharing revenue is to calculate the two-wheeled vehicle sharing revenue to be settled to the owner of the target two-wheeled vehicle based on the two-wheeled vehicle sharing revenue unit price included in the sharing combination cost, and to calculate a depreciation correction cost by reflecting at least one of the predicted tire wear, brake pad consumption, battery charge / discharge degradation, and drivetrain load index based on the estimated driving distance, planned usage time, estimated load, and elevation change profile of the planned driving route included in the two-wheeled vehicle usage request information, and to reduce the platform brokerage fee as the depreciation correction cost increases and add an amount corresponding to the reduction to the two-wheeled vehicle sharing revenue unit price to finally determine the two-wheeled vehicle sharing revenue.

[0215] Here, the unit price of revenue for two-wheeled vehicle sharing may be the base settlement amount set by the vehicle owner when registering their two-wheeled vehicle on the platform. The depreciation adjustment cost may be a cost that quantifies the wear burden imposed by a specific rental on the vehicle's tires, brakes, battery, or drivetrain. Tire wear, brake pad wear, battery charge / discharge degradation, and drivetrain load index, predicted based on estimated mileage or altitude change profiles, may refer to physical wear figures calculated by statistical or predictive models regarding how much faster and more severely the two-wheeled vehicle's drivetrain consumables wear out compared to normal conditions when a borrower drives on steep mountain roads or travels long distances carrying heavy loads or passengers. The server may calculate the depreciation adjustment cost by reflecting the above factors, and as the depreciation adjustment cost is calculated to be higher, it may reduce a portion of the brokerage fee taken by the platform and transfer an amount corresponding to that reduction to the vehicle owner, thereby adjusting the revenue from two-wheeled vehicle sharing upward. For example, if the basic revenue unit price for two-wheeled vehicle sharing is 9,000 won and the depreciation adjustment cost is calculated to be high, resulting in a reduction of 400 won from the platform brokerage fee of 1,200 won, the aforementioned 400 won is added to the revenue unit price, thereby finalizing the revenue for two-wheeled vehicle sharing at 9,400 won. According to this process, vehicle owners can receive higher compensation in transactions where their vehicles are expected to wear out significantly, and the platform can increase the incentive for owners to participate. In other words, by voluntarily reducing a portion of the brokerage margin that the platform company would otherwise take and transferring it to the vehicle owner's settlement amount, a mutually beneficial settlement structure can be implemented that promotes the sharing of high-value vehicles or vehicles with high operating loads.

[0216] The step of creating insurance and guarantee conditions is to calculate the reliability deviation between the borrower reliability index and the minimum reliability threshold for rental allowance set by the owner of the target motorcycle, and if the reliability deviation is less than a set safety margin range, to increase the self-damage deductible applied to the borrower in the event of an accident to hedge sharing risk, or to create insurance and guarantee conditions including at least one of the conditions for subscribing to a platform-exclusive short-term insurance special clause, conditions for depositing an additional security deposit, and conditions for restricting use during high-risk hours.

[0217] Here, reliability deviation may refer to the difference between the borrower's reliability index and the minimum reliability threshold set by the vehicle owner. For example, if the threshold is 80 points and the borrower's reliability index is 83 points, the reliability deviation may be calculated as 3 points. The safety margin zone may be a reference range where it is determined that additional risk mitigation is still necessary for borrowers who barely exceed the threshold. For example, if the safety margin zone is set to 15 points, a borrower with a reliability deviation of 5 points or less may still be classified as a borderline high-risk user even if the initial matching has been approved. The own damage waiver may be the minimum amount of damage that the borrower must directly bear in the event of an accident, and the condition for subscribing to a platform-exclusive short-term insurance rider may be an insurance requirement necessary to reinforce coverage or diversify risk for specific transactions. The condition requiring an additional deposit may refer to supplementary collateral to cover situations such as vehicle damage or failure to return, while the condition restricting use during high-risk hours may limit usage during high-risk time periods, such as late-night hours or times when severe weather is expected. For example, if the reliability deviation is very small (2 points) and nighttime usage is included, the server may increase the deductible and create conditions requiring both an additional deposit and a platform-exclusive short-term insurance rider. Furthermore, in cases of very high risk, conditions may be created to restrict the vehicle's starting permission entirely, preventing the engine from being started during late-night high-risk hours. According to this approach, the platform can defend against losses on transactions that barely pass the threshold, and vehicle owners can also share their assets more safely.

[0218] The step of finalizing rental condition information is a step of finalizing rental condition information that combines the above-mentioned final usage fee, space sharing revenue, two-wheeler sharing revenue, platform brokerage fee, and insurance and warranty conditions to define the cost burden relationship, revenue distribution relationship, scope of usage liability, and scope of compensation liability in the event of an accident among the borrower, two-wheeler owner, space owner, and platform.

[0219] Here, the final usage fee may be the amount actually borne by the borrower, and the space sharing revenue and motorcycle sharing revenue may be amounts to be settled with the space owner and the vehicle owner, respectively. The platform brokerage fee may be compensation for the platform providing transaction brokerage, control, and settlement services. Insurance and warranty conditions may be contract data defining the liability sharing structure and pre-collateral structure in the event of an accident. The server may combine the above elements into a single rental condition information package to present to the borrower and simultaneously record it as settlement information for the vehicle owner and the space owner. The cost burden relationship may refer to the total payment amount to be paid by the borrower and the deposit structure, and the profit distribution relationship may be accounting data indicating the proportion in which the said amount is settled to the motorcycle owner, the space owner, and the platform, respectively. The scope of usage liability and the scope of compensation liability in the event of an accident may be contract rules defining how and at whose expense compensation or reimbursement will be provided in the event of tire wear, illegal parking fines, vehicle damage, or failure to return the vehicle during the rental period. For example, if for a transaction, the borrower's final usage fee is 13,800 won, the space sharing revenue is 3,700 won, the motorcycle sharing revenue is 9,400 won, the platform brokerage fee is 700 won, and the additional deposit is 30,000 won, the server can generate final rental condition information including the above values ​​and related insurance conditions. According to this step, beyond a simple rental approval process, a final contractual state can be formed in which costs, profits, and responsibilities between each participant are clearly defined, and complex motorcycle sharing transactions involving distributed individual owners can be concluded in the most transparent and automated manner.

[0220] An apparatus according to one embodiment includes a processor and memory. The processor may include at least one apparatus described above through the drawings or perform at least one method described above through the drawings. The memory may store information related to the method described above or store a program in which the method described above is implemented. The memory may be volatile memory or non-volatile memory.

[0221] The processor can execute a program and control the device. The code of the program executed by the processor can be stored in memory. The device can be connected to an external device (e.g., a personal computer or a network) through an input / output device (not shown in the drawing) and exchange data.

[0222] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0223] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0224] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or instruct the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium, or device so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0225] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0226] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below. Explanation of the symbols

[0227] S100: Step of receiving service request information S200: Step for acquiring driver information S300: Step for selecting target motorcycles and target loading / unloading spaces S400: Step for calculating rental condition information S410: Step for calculating driver risk information S411: Step to extract driver static profile data S412: Step to extract regional information S413: Step to generate dynamic context data S414: Step for calculating the model proficiency index S415: Step to generate a data set for risk analysis S416: Step of assigning commercial operation weights or non-commercial operation weights S417: Step of applying environmental risk correction factors S418: Step for calculating risk component values S419: Step for generating driver risk information S420: Step for calculating rental fees and insurance coverage conditions S100: A step of transmitting driver identification information, status information of the target motorcycle, and current location information to a server. S200: A step for determining whether to allow the starting of the driving power source S300: The stage of collecting vehicle body posture data in real time S400: Step for calculating accident probability information specific to two-wheeled vehicles S410: Step of generating a dynamic road surface friction coefficient profile for each section S420: Step for calculating dynamic load transfer amount, front wheel ground contact load increase rate, and rear wheel ground contact load decrease rate S430: Step to determine jackknife overturning risk S440: A step for calculating real-time banking angle, lateral slip angle, tire friction loss rate, and cornering stability degradation rate. S450: Step to generate state vector S460: Step for calculating low-side slip risk S470: Step of correcting low-side slip risk based on high-side rollover weighting S480: The stage for finally calculating accident probability information specific to two-wheeled vehicles S500: Step to perform driving assistance control S600: Step for recording ADAS event logs S100: Step of receiving service request information S200: Step for acquiring driver information S300: Step for selecting target motorcycles and target loading / unloading spaces S400: Step for calculating rental condition information S500: A step for automatically controlling the unmanned entry and exit of target motorcycles. S510: Step of generating unmanned dispatch control information S520: Step of generating unmanned return control information S521: Step to verify whether the target motorcycle has been entered into the correct location. S522: Step of switching at least one of the immobilizer, electronic lock, and mechanical steering lock of the target two-wheeled vehicle to a locked state S523: Step for identifying suspected accident areas S524: Step of acquiring multi-angle incoming exterior video data S525: A step of quantifying at least one of the scratch area, damage depth, crack length, and physical damage rate of the turn signal on the vehicle body exterior. S526: Step for determining whether accessories have been returned normally S527: Step for finally generating unmanned return control information for billing settlement S100: Step of receiving service request information S200: Step for acquiring driver information S300: Step for selecting target motorcycles and target loading / unloading spaces S310: Step for selecting sharing candidate motorcycles and sharing candidate spaces S320: Step for determining the target motorcycle and target loading / unloading space to be assigned to the driver S321: Step for selecting sharing-selected motorcycles S322: Step for calculating suitability of loading and unloading spaces specialized for two-wheeled vehicles S323: Step for calculating the risk of power depletion while driving S324: Step to derive idle time matching rate S325: Step to derive sharing combination costs S326: Step for calculating the final sharing matching score by combination S327: The step of determining the combination with the highest final matching score as the target motorcycle and target loading / unloading space. S400: Step for calculating rental condition information

Claims

Claim 1 A method for providing a two-wheeled vehicle sharing platform service utilizing idle two-wheeled vehicle and space information, wherein a server includes a processor, memory, a communication module, and a non-transient storage medium, and the program stored in the non-transient storage medium is executed by the processor, comprising: receiving service request information from a driver terminal including driver identification information, two-wheeled vehicle usage request information, and usage condition information; performing authentication for the driver based on the driver identification information and obtaining driver information including usage qualification information, usage history information, and payment method information associated with the driver identification information; and selecting a target two-wheeled vehicle and a target inbound / outbound space corresponding to the service request information based on operational resource information including two-wheeled vehicle information and storage space information. The method comprises the step of calculating rental condition information including at least one of the following: approval of use, usage fee, warranty condition, insurance coverage condition, and entry / exit processing condition for the target motorcycle, based on the driver information, motorcycle usage request information, grade information of the target motorcycle, target entry / exit space information, and pre-set operational policy information; and the step of selecting the target motorcycle and target entry / exit space includes: the step of selecting sharing candidate motorcycles and sharing candidate spaces by reflecting available time intervals, available area intervals, charging or fuel status, vehicle type suitability, and space access convenience corresponding to the motorcycle usage request information, based on a plurality of idle motorcycle information and idle space information included in the operational resource information; and the step of determining the target motorcycle and target entry / exit space to be assigned to the driver based on the driver information, sharing candidate motorcycle, sharing candidate space, and pre-set operational policy information.The step of determining the target motorcycle and target entry / exit space includes: quantifying the history of past motorcycle damage, history of traffic law violations, frequency of return delays, frequency of reservation cancellations, and history of payment failures included in the driver information to derive a borrower reliability index, and determining the primary matching eligibility for each sharing candidate motorcycle by comparing the borrower reliability index with a minimum reliability threshold for rental allowance previously set by the owner of the sharing candidate motorcycle, thereby selecting a sharing candidate motorcycle; A step of calculating the suitability of a motorcycle-specific entry / exit space reflecting entry / exit stability, parking suitability, and storage stability by comparing the floor slope, road surface flatness, inclusion of security CCTV coverage, availability of charging infrastructure for electric motorcycles, availability of nighttime illumination, and availability of a rain-blocking structure of the sharing selection space with the drive method, body weight, vehicle type, wheelbase, and steering angle specifications of the sharing selection motorcycle; A step of simulating the actual driving range by reflecting the current battery state of charge or remaining fuel amount, recent energy consumption or fuel consumption per unit distance, and correction coefficients by drive method of the shared selected motorcycle, based on the estimated driving distance, scheduled time of use, elevation change profile of the scheduled driving route, estimated load, and ambient temperature information included in the above motorcycle usage request information, and calculating the risk of power depletion during driving based on the difference between the actual driving range and the estimated driving distance; a step of calculating the estimated idle duration, which is the time interval expected to be available until the next reservation, retrieval, maintenance, and owner usage schedule of the shared selected motorcycle begins, and deriving an idle time match rate by calculating the temporal intersection ratio between the estimated idle duration and the driver's scheduled rental time;A step of deriving a combined sharing cost by summing the motorcycle sharing revenue unit price set by the owner of the selected sharing motorcycle, the space sharing revenue unit price set by the owner of the selected sharing space, the estimated charging cost or fuel preservation cost, and the estimated recovery cost of the selected sharing motorcycle; a step of calculating a final sharing matching score for each combination of the selected sharing motorcycle and the selected sharing space by weighted summing the borrower reliability index, the suitability of the motorcycle-specialized entry / exit space, the reverse penalty corresponding to the risk of power depletion, the idle time matching rate, the combined sharing cost, the driver's past preferred vehicle type information, and the driver's past preferred return area information; A method for providing a motorcycle sharing platform service utilizing idle motorcycle and space information, comprising: a step of verifying, for a plurality of combinations with the highest final matching scores in the sharing, whether there are duplicate reservations in the same time slot, whether there is a possibility of exceeding the occupancy of the same space, and whether there is a conflict with the next reservation, and determining the combination with the highest final matching score reflecting the verification results as the target motorcycle and target entry / exit space; Claim 2 In claim 1, the step of calculating the rental condition information comprises: a step of calculating a basic usage fee by adding a pre-set platform brokerage fee, a payment processing fee, and an operation management fee to the sharing combined cost derived during the process of determining the target motorcycle and the target loading / unloading space, and, if the idle time matching rate is greater than or equal to a pre-set optimization threshold, a dynamic discount rate proportional to the idle deadlock resolution contribution of the target motorcycle and the target loading / unloading space is applied to the basic usage fee to derive a final usage fee to be charged to the borrower; A step of calculating the space sharing revenue to be settled to the owner of the target entry / exit space based on the space sharing revenue unit price included in the above-mentioned sharing combination cost, wherein at least one of 1) an electricity charge amount based on the estimated charging requirement if the target motorcycle is an electric motorcycle, 2) a security weighting cost proportional to the occupancy time of a security CCTV if the scheduled usage time of the target entry / exit space corresponds to nighttime hours, and 3) a management surcharge corresponding to the entry restriction time of the target entry / exit space is additionally reflected to finally determine the space sharing revenue; A step of calculating the motorcycle sharing revenue to be settled to the owner of the target motorcycle based on the motorcycle sharing revenue unit price included in the above-mentioned sharing combination cost, and calculating a depreciation adjustment cost by reflecting at least one of the predicted tire wear, brake pad consumption, battery charge / discharge degradation, and drivetrain load index based on the estimated driving distance, planned usage time, estimated load, and elevation change profile of the planned driving route included in the motorcycle usage request information, and finally determining the motorcycle sharing revenue by reducing the platform brokerage fee as the depreciation adjustment cost increases and adding an amount corresponding to the reduction to the motorcycle sharing revenue unit price;A method for providing a motorcycle sharing platform service utilizing idle motorcycle and space information, comprising: a step of calculating a reliability deviation between the borrower reliability index and a minimum reliability threshold for rental allowance pre-set by the owner of the target motorcycle, and, if the reliability deviation is less than a pre-set safety margin range, generating insurance and guarantee conditions including at least one of an upward adjustment of the self-vehicle damage deductible applied to the borrower in the event of an accident for hedging sharing risk, a condition for subscribing to a platform-exclusive short-term insurance special clause, a condition for depositing an additional security deposit, and a condition restricting use during high-risk hours; and a step of finally determining rental condition information in which the cost burden relationship, profit distribution relationship, scope of usage liability, and scope of compensation liability in the event of an accident are defined among the borrower, motorcycle owner, space owner, and platform by combining the final usage fee, space sharing revenue, motorcycle sharing revenue, platform brokerage fee, and insurance and guarantee conditions; Claim 3 delete

Images