Ride-hailing order automatic billing method and device based on three-end space-time evidence chain

Abstract

The present application relates to a three-terminal space-time evidence chain-based online car-hailing order automatic billing method and device, applied to the online car-hailing billing technical field, comprising: continuously receiving vehicle terminal, driver terminal and passenger terminal three-terminal positioning data and performing positioning quality evaluation, combining the starting point and the target scene type to generate adaptive starting fence and ending fence, respectively, on this basis, performing multiple evidence independent determination such as vehicle arrival / destination arrival, vehicle stay / parking, passenger approach, same trip movement, path start / finish and passenger separation, and adopting a continuous observation window confirmation mechanism to trigger billing start and end, effectively avoiding billing false triggering under GPS drift, complex road levels and single distance threshold or single positioning; thereby comprehensively improving the accuracy, reliability and fairness of online car-hailing billing in complex operating environments.

Description

Technical Field

[0001] This invention relates to the field of ride-hailing billing technology, specifically to a method and apparatus for automatic billing of ride-hailing orders based on a three-terminal spatiotemporal evidence chain. Background Technology

[0002] Currently, the start and end times of billing for ride-hailing orders mainly rely on drivers manually clicking "Start Billing" and "End Billing" on their app. In actual operation, due to factors such as drivers forgetting to perform the operation, network latency, location drift, inaccurate pick-up points, complex entrances to destination buildings, and complex scenarios such as parking lots, airports, and train stations, the start or end times of billing may not match the actual service time.

[0003] Existing solutions employ two main approaches. One type uses a fixed distance threshold to determine if a vehicle is close to the pick-up point or destination, but it cannot handle complex situations such as large shopping malls, hospitals, airports, elevated highway levels, underground parking lots, and GPS drift. Another type relies on passenger confirmation or driver manual termination, but passengers may not activate the function in time, and drivers may maliciously advance or delay billing. Post-incident manual determination of responsibility is costly and provides a poor user experience.

[0004] Therefore, there is a need for an automatic billing start and end determination method that can integrate location data from the vehicle terminal, driver terminal, and passenger terminal, navigation path, road topology, and vehicle motion status, so that the billing behavior is real-time, accurate, and traceable. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide an automatic billing method and device for ride-hailing orders based on a three-terminal spatiotemporal evidence chain, which aims to solve the problems in the prior art that rely on passenger confirmation or driver manual termination, but passengers may not operate in time, and drivers may maliciously advance or delay billing, resulting in high costs and poor user experience due to manual judgment afterward.

[0006] According to a first aspect of the present invention, an automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain is provided, the method comprising:

[0007] Obtain ride-hailing order information and determine the pick-up scenario type and target scenario type based on the pick-up point and destination in the ride-hailing order information; During the pick-up process, three-terminal positioning data are continuously acquired, including vehicle-end positioning, driver-end positioning, and passenger-end positioning, and the quality of the three-terminal positioning is determined. The system determines the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality. It then determines whether the vehicle has arrived at the boarding point based on the three-terminal positioning data, the boarding scenario type, and the dynamic fence. Based on the three-terminal positioning data and the boarding scenario type, it determines whether the vehicle is stopping or waiting at low speed within the dynamic fence of the boarding point. Based on the three-terminal positioning data, it determines whether the distance between the vehicle and the passenger terminal continuously decreases to a preset range within the observation window, thus determining whether the passenger is approaching. Based on the three-terminal positioning data, it determines whether the vehicle and the passenger terminal meet preset consistency conditions after leaving the boarding point, thus determining whether they are moving together. Finally, based on the three-terminal positioning data and the navigation route, it determines whether a path has been initiated. When the arrival of the vehicle at the pick-up point and the vehicle's stay or slow waiting within the dynamic fence of the pick-up point are both determined to be true, and the three items of passenger approach, co-movement, and path initiation meet the preset number corresponding to the pick-up scenario type, and all of the above determinations are true within a continuous observation window, automatic billing is triggered. After entering the billing state, the target dynamic fence is determined based on the target scene type and the three-terminal positioning quality; the destination is reached based on the three-terminal positioning data and the target dynamic fence; the vehicle is stopped or waiting at low speed within the target dynamic fence based on the three-terminal positioning data and the target scene type; the completion ratio of the vehicle's traveled path relative to the planned path is determined based on the navigation route to determine whether the path is completed; and the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops is determined based on the three-terminal positioning data to determine whether the passenger has been separated. When the arrival at the destination and the vehicle's stay or slow-speed waiting within the target dynamic fence are both determined to be true, and the completion of the route and the separation of passengers meet the preset number corresponding to the destination scenario type, and all of the above determinations are true within a continuous observation window, automatic termination of billing is triggered.

[0008] Preferably, The step of determining the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality includes: Based on the boarding point location, different basic radii are preset according to different boarding scenario types; The base radius is corrected based on the positioning accuracy of the vehicle terminal and the passenger terminal. The lower the positioning accuracy, the larger the corrected radius. Invalid positioning is not included in the correction and the corresponding invalid positioning is marked as unavailable. The corrected fence radius does not exceed the first upper limit value. If the boarding point is a multi-entrance building or park, then the entrance set of the boarding point, the historical boarding points on the same road side, and the parking area will generate multiple candidate entrances respectively. The radius of each candidate entrance is a preset value. When the number of samples that have been manually reviewed and approved by any candidate entrance in the historical period reaches a preset number and the review pass rate reaches a preset ratio, it will be used as a candidate boarding point dynamic fence. Acquire the near-field connection signal between the vehicle and the passenger terminal as valid near-field evidence; If valid near-field evidence is obtained during the pick-up process, the radius of the dynamic fence at the current pick-up point will be reduced by a preset ratio, and the reduced radius will not be lower than the preset lower limit.

[0009] Preferably, Before determining whether the vehicle has arrived at the pick-up point or the destination, a road accessibility check is also included, which includes: Calculate the road network reachable distance from the vehicle location to the pick-up point or destination. If the road network reachable distance is greater than a preset multiple of the vehicle's straight-line distance, then even if the vehicle's straight-line distance is within the radius of the dynamic fence at the pick-up point or the target dynamic fence, the vehicle is not determined to have arrived. If the road where the vehicle is currently located is on a different road level than the road where the pick-up point or destination is located, and the navigation route indicates that the vehicle cannot reach the side of the road where the pick-up point or destination is located within the preset distance, then the vehicle will not be considered to have arrived.

[0010] Preferably, It also includes: risk control, irregular arbitration, and fee freezing. The risk control includes: In the automatic start billing process, after receiving the start billing operation instruction from the driver, the system determines the arrival of the pick-up point, the vehicle's stop or slow-speed wait within the dynamic fence of the pick-up point, the passenger's approach, the number of valid near-field evidence and the number of co-movements, and determines the risk level of starting early based on the number of missing items. When the risk level is high, the system blocks the start of billing and requests the passenger's confirmation. In the automatic billing termination process, the system determines the number of five pieces of evidence that are valid: the vehicle enters the target dynamic fence area, the vehicle stops or waits at low speed within the target dynamic fence, the route is completed, the passenger is separated, and the driver has not ended the billing for more than a preset time. The risk level of delayed termination is determined based on the number of valid evidence items. When the risk level is high, a countdown is started, and the billing is forcibly terminated after the countdown ends. Among them, if the vehicle detours at low speed within a preset range around the destination for more than a preset time, or if the vehicle moves away from the target dynamic fence for more than a preset distance while the passenger terminal is still within the target fence, it is directly determined to be high risk. When at least two of the preset abnormal conditions are simultaneously detected, automatic billing is suspended and the fee for the corresponding time period is marked as pending arbitration. The preset abnormal conditions include: invalid three-terminal positioning data, inconsistent road level between the vehicle location and the pick-up point or destination, and unreachable road. The fees marked as pending arbitration, or the dispute fees corresponding to the risks of early commencement, or the dispute fees corresponding to the risks of delayed termination, shall be frozen from the time of the first risk triggering until the time of risk elimination or manual review.

[0011] Preferably, The determination of whether a vehicle is stopped or waiting at low speed within the pick-up point or target dynamic fence includes: Obtain the duration for which the vehicle speed is below a preset speed threshold; Obtain the dwell time threshold and location fluctuation limit corresponding to the current boarding scenario type or the target scenario type; Obtain the positioning quality of the vehicle and correct the dwell time threshold based on the positioning quality; The vehicle's position fluctuation radius is obtained within the duration. When the duration reaches the corrected dwell time threshold and the position fluctuation radius does not exceed the upper limit of the position fluctuation, it is determined that the vehicle is staying or waiting at low speed.

[0012] Preferably, The step of determining whether the distance between the vehicle terminal and the passenger terminal continuously decreases to a preset range within the observation window based on the three-terminal positioning data, thereby determining whether a passenger is approaching, includes: Different proximity distance judgment radii are determined based on the positioning quality of the passenger terminal; Taking the current moment as the endpoint, take the first observation window forward. If the cumulative reduction or reduction ratio of the distance between the passenger terminal and the vehicle end within the first observation window reaches a preset value, and the current distance between the two is not greater than the proximity distance determination radius, then the passenger proximity is determined to be established. If passenger terminal location is marked as unavailable, valid near-field evidence or passenger confirmation actions will be used instead of passenger proximity determination.

[0013] Preferably, The step of determining whether the vehicle and the passenger terminal meet preset consistency conditions after the vehicle leaves the boarding point based on the three-terminal positioning data, thereby determining whether they are moving together, includes: The second observation window is taken as the starting point when the vehicle leaves the dynamic fence of the boarding point. If the number of valid positioning sampling points in the second observation window is less than the preset number, then no same-movement judgment is made. The differences in motion parameters between the vehicle terminal and the passenger terminal within the second observation window are calculated, including speed difference, direction difference, trajectory similarity and acceleration correlation. When each parameter meets its corresponding preset threshold, it is counted as evidence of consistency. The required number of consistent evidences is determined based on the positioning quality of the passenger terminal. The lower the positioning quality, the higher the required number of consistent evidences. When the required number of consistent evidences is reached, it is determined that the co-movement is established.

[0014] Preferably, The step of determining whether passenger separation has occurred based on the three-terminal positioning data, specifically whether the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops, meets preset separation conditions, includes: The third observation window is taken as the starting point of the vehicle parking time; the distance between the vehicle end and the passenger terminal is continuously monitored within the third observation window. When the distance continuously increases beyond the preset distance and the passenger terminal position is within the target dynamic fence range, the passenger separation is determined to be successful. The vehicle and passenger terminal positions are continuously monitored within the third observation window. When the vehicle moves away from the target dynamic fence by more than a preset distance while the passenger terminal position remains within the target dynamic fence for more than a preset time, the passenger separation is determined to be successful.

[0015] Preferably, When automatic billing is triggered, start time backtracking is also performed: the time when the vehicle leaves the boarding point dynamic fence and the same movement is established, the time when the passenger confirms the operation, and the time when valid near-field evidence is first established, and the latest of these times is used as the billing start time; When automatic billing termination is triggered, the termination time is also backtracked: the time when the vehicle first enters the target dynamic fence, the time when the vehicle first stops, the time when the route completion ratio reaches the threshold, the time when passenger separation is established, and the time when the driver manually terminates the operation are recorded respectively. At least three times that are established at the same time are selected from the above times, and the latest one is taken as the billing termination time. If the driver manually ends the operation later than the billing end time by more than a preset deviation, the fees between the two will be marked as pending freezing.

[0016] According to a second aspect of the present invention, an automatic billing device for ride-hailing orders based on a three-terminal spatiotemporal evidence chain is provided, the device comprising: Scene classification module: used to obtain ride-hailing order information and determine the pick-up scene type and target scene type based on the pick-up point and destination in the ride-hailing order information; Location acquisition module: used to continuously acquire three-terminal location data during the pick-up process, including vehicle-end location, driver terminal location, and passenger terminal location, and to determine the quality of the three-terminal location; The billing evidence acquisition module is used to: determine the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality; determine whether the vehicle has arrived at the boarding point based on the three-terminal positioning data, the boarding scenario type, and the dynamic fence of the boarding point; determine whether the vehicle is staying or waiting at low speed within the dynamic fence of the boarding point based on the three-terminal positioning data and the boarding scenario type; determine whether the distance between the vehicle and the passenger terminal continuously decreases to a preset range within the observation window based on the three-terminal positioning data, thereby determining whether the passenger is approaching; determine whether the vehicle and the passenger terminal meet preset consistency conditions after leaving the boarding point based on the three-terminal positioning data, thereby determining whether they are moving together; and determine whether the route has been initiated based on the three-terminal positioning data and the navigation route. The billing start trigger module is used to trigger automatic billing when the arrival of the vehicle at the pick-up point and the vehicle stops or waits at low speed within the dynamic fence of the pick-up point are both determined to be true, and the three items of passenger approach, co-movement and path start meet the preset number corresponding to the pick-up scenario type, and the above determinations are all true within the continuous observation window. End of Billing Evidence Acquisition Module: After entering the start of billing state, this module determines the target dynamic fence based on the target scene type and the three-terminal positioning quality; determines whether the destination has been reached based on the three-terminal positioning data and the target dynamic fence; determines whether the vehicle is staying or waiting at low speed within the target dynamic fence based on the three-terminal positioning data and the target scene type; determines whether the completion ratio of the vehicle's traveled path relative to the planned path reaches a threshold based on the navigation route, thereby determining whether the path is completed; and determines whether the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops meets a preset separation condition based on the three-terminal positioning data, thereby determining whether the passenger has been separated. End-of-billing trigger module: When the arrival at the destination and the vehicle's stay or slow-speed waiting within the target dynamic fence are both determined to be true, and the completion of the route and separation of the passenger meet the preset number corresponding to the destination scenario type, and all of the above determinations are true within a continuous observation window, the automatic end-of-billing is triggered.

[0017] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This application continuously receives positioning data from the vehicle, driver, and passenger terminals and performs positioning quality assessment. It generates adaptive start and end fences based on the scenario type of the pick-up point and the target. On this basis, it performs independent multi-evidence judgments for vehicle arrival / destination arrival, vehicle stop / parking, passenger approach, peer movement, route start / completion, and passenger separation. It also adopts a continuous observation window confirmation mechanism to trigger the start and end of billing, effectively avoiding billing errors caused by single distance thresholds or single positioning under GPS drift and complex road layers. Thus, it comprehensively improves the accuracy, reliability, and fairness of ride-hailing billing in complex operating environments.

[0018] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0020] Figure 1 This is a flowchart illustrating an automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain, according to an exemplary embodiment. Figure 2 This is a schematic diagram illustrating the process of automatically starting billing according to another exemplary embodiment; Figure 3 This is a schematic diagram illustrating the process of automatically ending billing according to another exemplary embodiment; Figure 4 This is a schematic diagram of an automatic billing device for ride-hailing orders based on a three-terminal spatiotemporal evidence chain, according to another exemplary embodiment. In the attached diagram: 1-Scene classification module, 2-Location acquisition module, 3-Start billing evidence acquisition module, 4-Start billing trigger module, 5-End billing evidence acquisition module, 6-End billing trigger module. Detailed Implementation

[0021] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0022] Example 1: Figure 1This is a flowchart illustrating an automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain, according to an exemplary embodiment. Figure 1 As shown, the method includes: S1, obtain ride-hailing order information, and determine the pick-up scenario type and target scenario type based on the pick-up point and destination in the ride-hailing order information; S2, during the pick-up process, continuously acquire three-terminal positioning data, including vehicle-end positioning, driver-end positioning and passenger-end positioning, and determine the quality of the three-terminal positioning; S3. Determine the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality. Determine whether the vehicle has arrived at the boarding point based on the three-terminal positioning data, the boarding scenario type, and the dynamic fence of the boarding point. Determine whether the vehicle is staying or waiting at low speed within the dynamic fence of the boarding point based on the three-terminal positioning data and the boarding scenario type. Determine whether the distance between the vehicle and the passenger terminal continuously shrinks to a preset range within the observation window based on the three-terminal positioning data, thereby determining whether the passenger is approaching. Determine whether the vehicle and the passenger terminal meet preset consistency conditions after leaving the boarding point based on the three-terminal positioning data, thereby determining whether they are moving together. Determine whether the route is initiated based on the three-terminal positioning data and the navigation route. S4. When the arrival of the vehicle at the pick-up point and the vehicle stopping or waiting at low speed within the dynamic fence of the pick-up point are both determined to be true, and the three items of passenger approach, traveling together and path initiation meet the preset number corresponding to the pick-up scenario type, and the above determinations are all true within the continuous observation window, automatic billing is triggered. S5, after entering the billing start state, determine the target dynamic fence based on the target scene type and the three-terminal positioning quality; determine whether the destination has been reached based on the three-terminal positioning data and the target dynamic fence; determine whether the vehicle is staying or waiting at low speed within the target dynamic fence based on the three-terminal positioning data and the target scene type; determine whether the completion ratio of the vehicle's traveled path relative to the planned path has reached a threshold based on the navigation route, thereby determining whether the path has been completed; determine whether the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops based on the three-terminal positioning data meets the preset separation condition, thereby determining whether the passenger has been separated. S6, when the arrival of the destination and the vehicle stop or wait at low speed within the target dynamic fence are both determined to be true, and the completion of the route and the separation of the passenger meet the preset number corresponding to the destination scenario type, and the above determinations are all true within the continuous observation window, the automatic end of billing is triggered. It is understood that the following definitions are given first in this embodiment to facilitate a better understanding of this application, specifically including: Positioning quality grading rules: If the positioning accuracy radius is no greater than 15 meters, the sampling delay is no greater than 3 seconds, and the speed change between two consecutive points does not exceed 6 meters per second, it is marked as a high-quality positioning. If the positioning accuracy radius is greater than 15 meters but no greater than 50 meters, or the sampling delay is greater than 3 seconds but no greater than 8 seconds, it is marked as a medium-quality positioning. If the positioning accuracy radius is greater than 50 meters but no greater than 100 meters, or the sampling delay is greater than 8 seconds but no greater than 20 seconds, it is marked as a low-quality positioning. If the positioning accuracy radius is greater than 100 meters, the sampling delay is greater than 20 seconds, the position jump exceeds 80 meters within 3 seconds and the vehicle speed does not support the jump, or the map matching road is opposite to the driving direction, the positioning point is marked as an invalid positioning.

[0023] Scene complexity classification rules: Ordinary roads, entrances to residential areas, and clearly marked roadside pick-up points are marked as ordinary scenarios; commercial districts with multiple entrances, hospital parks, school parks, office building parks, airports, train stations, underground parking lots, elevated roads with multiple levels, parallel main and auxiliary roads, or areas with weak GPS are marked as complex scenarios; if the same location simultaneously meets any two of the following conditions: multiple entrances, unclear road levels, and historical error greater than 40 meters, it is marked as a highly complex scenario.

[0024] Dynamic fence generation: Both the start and end of billing are based on dynamic geofences as evidence, but the geofence results only constitute part of the confidence level. The initial fence determination rules are as follows: If the road network reachable distance from the vehicle to the pick-up point or any candidate pick-up entrance is no greater than the current starting fence radius, and the vehicle's road level, roadside, and direction of travel are consistent with the pick-up point, then the starting fence evidence is valid. If the vehicle's straight-line distance is within the fence but the road network reachable distance is greater than 1.5 times the fence radius, or if the vehicle is located on an oncoming road, on an elevated upper level, in a tunnel, or on a closed road, then the starting fence evidence is invalid. If the vehicle distance is no greater than 50% of the fence radius, it is considered strong fence evidence; greater than 50% but no greater than 100%, it is considered ordinary fence evidence; exceeding the fence radius, it is considered outside fence evidence. If the fence is expanded due to medium or low quality positioning, the passenger approach rule or the valid near-field evidence rule must be satisfied simultaneously; otherwise, the vehicle cannot enter the subsequent billing state solely based on the fence.

[0025] The rules for determining when to terminate a fence are as follows: If the road network reachable distance from the vehicle to the destination or any candidate drop-off area is no greater than the current end-of-fence radius, and the road level where the vehicle is located is consistent with the destination, and the vehicle can stop or drop off passengers on that road, then the end-of-fence evidence is valid. If the vehicle is on an elevated road, tunnel, closed road, on the opposite side of the road, or on a no-stopping main road, and the navigation cannot reach the destination drop-off area within 300 meters, then the end-of-fence evidence is invalid. If the vehicle distance is no greater than 50% of the end-of-fence radius, it is considered strong arrival evidence; if it is greater than 50% but no greater than 100%, it is considered ordinary arrival evidence; if it exceeds the fence radius, it is considered not arrived. If the vehicle-side positioning is invalid, the end-of-fence cannot be valid. If the passenger-side positioning is invalid but the vehicle-side positioning is valid, then vehicle-side, route completion, and parking evidence can be used to continue the judgment.

[0026] Rules for Evidence of Stay: Low speed refers to a vehicle speed of no more than 5 kilometers per hour; stopping refers to a vehicle speed of no more than 1 kilometer per hour. For ordinary boarding scenarios, the cumulative low speed or stopping time must be no less than 10 seconds with a positional fluctuation radius of no more than 15 meters; for complex boarding scenarios, the cumulative low speed or stopping time must be no less than 20 seconds with a positional fluctuation radius of no more than 25 meters; for highly complex boarding scenarios, the cumulative low speed or stopping time must be no less than 30 seconds with a positional fluctuation radius of no more than 35 meters. For ordinary alighting scenarios, the cumulative stopping or low speed time must be no less than 15 seconds; for complex alighting scenarios, no less than 25 seconds; for highly complex alighting scenarios, no less than 35 seconds.

[0027] Passenger approach rules: Within a 30-second observation window, if the distance between the passenger and vehicle ends decreases by at least 15 meters, or by at least 30%, and the current distance is not greater than the vehicle-passenger proximity radius, then passenger proximity is considered established. The proximity radius is 25 meters for high-quality passenger positioning, 40 meters for medium-quality positioning, and 60 meters for low-quality positioning; passenger proximity is not used as a necessary condition for invalid passenger positioning.

[0028] Near-field evidence rules: The following conditions must be met for an evidence to be considered valid near-field evidence: The Bluetooth connection lasts for at least 5 seconds and the signal strength is not lower than -75dBm; The passenger terminal must be connected to the in-vehicle Wi-Fi or a hotspot of the same device for at least 10 seconds; the passenger must click "I have boarded" and the time between the click and the vehicle's arrival at the boarding point must not exceed 120 seconds. The distance between passengers in the vehicle shall not exceed 30 meters within 30 seconds after the car door is opened; The driver's call or text message confirmation record with the passenger occurred within 180 seconds before or after the vehicle arrived at the pick-up point.

[0029] Rules of Parallel Movement: Within a 60-second window after the vehicle leaves the boarding point, if the average speed difference is no greater than 3 meters per second, the average direction difference is no greater than 35 degrees, the trajectory shape similarity is no less than 0.75, and the acceleration change correlation is no less than 0.60, these are respectively recorded as four pieces of evidence for traveling together. Under high-quality or medium-quality passenger positioning, traveling together is considered valid when at least three of the four conditions are met; under low-quality passenger positioning, traveling together is considered valid when at least two of the four conditions are met and there is valid near-field evidence; when passenger positioning is invalid, traveling together is not a necessary condition, and instead, consistency between the driver's terminal and the vehicle's terminal, near-field evidence, and path initiation evidence are used together as substitutes.

[0030] Path and road reachability rules: If the vehicle deviates from the navigation path by no more than 30 meters and the directional difference is no more than 45 degrees, the path is considered consistent. In complex road scenarios, the deviation distance is allowed to be relaxed to 50 meters, but the road level must be consistent. The path initiation condition is that the vehicle moves at least 50 meters from outside the boarding point fence within 90 seconds, and the angle between the moving direction and the destination direction is no more than 60 degrees. The path completion condition is that the planned path completion rate is no less than 90%, and the remaining navigation distance is no more than 300 meters. In complex destination scenarios, a completion rate of no less than 85% and the vehicle entering the candidate drop-off area are used as alternative conditions.

[0031] Passenger separation rules: If, within 120 seconds after the vehicle stops, the distance between the passenger and vehicle continuously increases by at least 30 meters, or if the vehicle leaves the candidate drop-off area while the passenger remains within the candidate drop-off area for at least 60 seconds, then passenger separation is deemed established. If the passenger's location is invalid, but the vehicle has stopped, the route has been completed, and the driver has not finished for more than 60 seconds, and there is passenger confirmation or a historical drop-off point match, then this will be treated as alternative evidence of separation.

[0032] Billing begins following these rules: Among vehicle arrival, vehicle stop, passenger approach or replacement confirmation, route initiation, and evidence of companion movement or replacement, vehicle arrival and vehicle stop are mandatory, and at least two of the other three must be met; at the same time, the risk of early initiation has not triggered the high-risk condition; when the above results are met for two consecutive observation windows, the system will automatically start billing.

[0033] End of billing rules: Among the evidence of arrival at the destination, vehicle slow speed or stopping, route completion, and passenger separation or alternative separation, arrival at the destination and vehicle slow speed or stopping are required, and at least one of the other two must be true; if the driver has not ended for more than 60 seconds or the vehicle has moved more than 80 meters away from the candidate drop-off area of ​​the destination, the countdown will be automatically terminated; when the above results are true for two consecutive observation windows, the system will automatically end or backtrack the end time.

[0034] Risk classification rules: The high-risk conditions for early start are: after the driver clicks start, any three of the following five conditions must be met: the vehicle has not entered the start fence, has not stopped, the passenger has not approached, no valid near-field evidence has appeared, and no co-movement has occurred. The high-risk conditions for delayed end are: any four of the following five conditions must be met: the vehicle has reached the destination candidate drop-off area, has stopped or is moving at low speed, the route has been completed, the passenger has separated or a replacement separation has been established, and the driver has not ended the process for more than 60 seconds. High-risk situations directly enter the pending confirmation, countdown, or arbitration phases; medium-risk situations occur when less than one of the above conditions is met, and the process enters the observation phase; low-risk situations occur when only one of the above abnormalities is met, and the process continues in the normal state machine.

[0035] Freezing Fee Rules: If the system enters abnormal arbitration, the freeze period starts from the time of the first risk trigger and ends at the time of manual review or automatic evidence recovery. If the risk of delayed termination is established, the freeze period starts from the reasonable termination time to the driver's actual termination time; if the risk of early start is established, the freeze period starts from the time the driver manually starts to the time the system confirms the reasonable start time.

[0036] Billing status transition rules: Read the order status from the previous moment; read the boarding evidence, alighting evidence, start billing confidence level, end billing confidence level, early start risk, delayed end risk, driver operation, and passenger confirmation within the current time window.

[0037] If the order is in the pick-up phase, and the distance from the vehicle to the pick-up point to the road network decreases by no less than 50 meters and the path deviation is no more than 30 meters within two consecutive 30-second windows, then the order will be maintained or will enter the driver approach phase.

[0038] If a vehicle enters the dynamic fence of the pick-up point, the road level is consistent, the directional difference is no more than 45 degrees, and the low speed or parking meets the criteria, then it enters the driver arrival candidate state.

[0039] If any two of the following four conditions are met: passenger proximity is established, valid near-field evidence is established, driver clicks start and vehicle arrival and stop are established, and vehicle is at least 50 meters away from the pick-up point, then the vehicle enters the suspected boarding state.

[0040] If the necessary conditions for vehicle arrival and vehicle stop are met, and at least one of the following is met—passenger approach or substitution confirmation, route initiation, or evidence of companion movement or substitution—but the rules for starting billing have not yet been met, the vehicle enters a waiting state for billing to begin, and a 30-second observation window is set.

[0041] If the billing start rule is met for two consecutive observation windows and the early start high risk is not triggered, then the billing process will proceed.

[0042] If a vehicle enters the candidate drop-off area of ​​the destination, the road level is consistent, the route completion rate reaches 90% or the complex scenario reaches 85%, and the low speed or parking determination is established, it enters the suspected arrival state.

[0043] If any two of the following conditions are met: passenger separation is confirmed, alternative separation is confirmed, driver has not ended the process for more than 60 seconds, or vehicle has left the destination drop-off area for more than 80 meters, the system will enter a waiting-to-end-of-charge state and prompt the driver to end the process.

[0044] If the rules for ending billing are met for two consecutive observation windows, and the driver has not ended the billing process within the 30-second countdown, the billing will automatically end and the process will be back to the time when the destination arrival, parking, and route completion were first met simultaneously. If the passenger separation occurs after the parking time, the process will be back to the time when the passenger separation was met.

[0045] If any two of the following four types of anomalies are met simultaneously: invalid location, inconsistent road level, conflict between driver operation and evidence, and conflict between passenger and vehicle evidence, then the system will enter an abnormal arbitration state, freeze the fees for the corresponding disputed area, and preserve the chain of evidence.

[0046] Having clarified the above rules, the specific procedures for this application include: As attached Figure 2 The automatic billing start process is shown below: Initialize the order and pick-up point model: After a driver accepts an order, the system reads basic order information, including the pick-up point, pick-up navigation, scene tags, historical pick-up point errors, roadside information, available parking areas, and pricing rules. The system then generates a basic model for determining the start of billing for the current order. The system retains the pick-up point source, map matching results, and dynamic fence parameters for subsequent recalculation.

[0047] Collect evidence from three points: The system continuously collects vehicle location, driver location, and passenger location data, and simultaneously reads vehicle speed, direction, acceleration, driver operation, passenger confirmation, network status, and positioning accuracy. It maps data from different sources to the same time window and marks positioning quality, network latency, and abnormal jumps.

[0048] Generate dynamic fences for boarding points: The system reads the boarding point type, GPS accuracy, road topology, historical errors, and near-field evidence. If the boarding point is in a normal scenario with no historical error configuration, the initial fence radius is 30 meters; for complex scenarios, it's 50 meters; for highly complex scenarios, it's 80 meters. If the vehicle or passenger has medium-quality positioning, the fence radius increases by 20 meters; if either has low-quality positioning, it increases by 40 meters. Invalid positioning does not participate in fence scaling, and the fence's maximum height is 120 meters. If the boarding point is located at an airport, train station, hospital, commercial area, or industrial park, the system generates candidate entrances based on the entrance set and manually verified samples from the past 30 days. Entrances are only activated when there are at least 20 candidate entrance samples and a manual verification pass rate of at least 85%. If valid near-field evidence exists, the fence radius is tightened to 70% of the current radius, but not less than 25 meters. If the median error between the pick-up location and the map point of the sample manually verified in the past 30 days is greater than 40 meters, then historical error compensation will be added to the fence; the compensation amount will be 50% of the median error, and a single compensation will not exceed 30 meters; if valid near-field evidence is found, the current starting fence radius will be tightened to 70%, but not less than 25 meters; at the same time, the vehicle stay must still meet the minimum duration of the corresponding scenario, and billing is not allowed to start solely based on near-field evidence.

[0049] If the fence is expanded due to improved positioning quality, stricter evidence rules will be applied simultaneously: For medium-quality positioning, the vehicle dwell time must be increased by 10 seconds, and at least two of the following four criteria must be met: passenger approach, near-field evidence, path initiation, and co-location movement; for low-quality positioning, the vehicle dwell time must be increased by 20 seconds, and at least three of the above four criteria must be met. Invalid positioning will not be included in the positive judgment for initiating billing and must be replaced by near-field evidence and path initiation. Determine the vehicle's actual arrival at the pick-up point: Determine if the vehicle's reachable distance from the road network is no greater than the starting fence radius; determine if the vehicle's road level is consistent with the boarding point, if the directional difference is no greater than 45 degrees, and if the path deviation is no greater than 30 meters; for complex road scenarios, the path deviation can be relaxed to 50 meters, but the road level must be consistent; determine if the vehicle meets the minimum low-speed or parking time requirements; no less than 10 seconds for normal scenarios, no less than 20 seconds for complex scenarios, and no less than 30 seconds for highly complex scenarios; if the vehicle is on an elevated upper level, in a tunnel, on an opposite road, on a closed road, or if the reachable distance from the road network exceeds 1.5 times the fence radius, arrival is not determined.

[0050] Determining whether a passenger is approaching a vehicle or a vehicle is approaching a passenger. The system calculates the distance change between the passenger and vehicle ends within a 30-second window. If the distance decreases by at least 15 meters, or by at least 30%, and the current distance is not greater than the proximity radius, then the passenger is considered to be close. The proximity radius is determined based on the passenger's positioning quality: 25 meters for high quality, 40 meters for medium quality, and 60 meters for low quality. If the passenger stays continuously within the boarding point fence for at least 30 seconds, the position fluctuation radius is not greater than 20 meters, and the distance between the vehicle and the passenger decreases by at least 20 meters, then the passenger is also considered close. If the passenger's positioning is invalid, the passenger's closeness is not used as positive evidence; instead, near-field evidence, passenger confirmation, or confirmation from the actual passenger must be used.

[0051] Determine whether a vehicle is stationary or waiting at low speed: The system calculates the cumulative time spent at speeds not exceeding 5 km / h and the parking time at speeds not exceeding 1 km / h. For ordinary boarding scenarios, the cumulative time spent at low speed or while parked must be at least 10 seconds with a position fluctuation radius not exceeding 15 meters. For complex scenarios, the requirement is at least 20 seconds with a fluctuation radius not exceeding 25 meters. For highly complex scenarios, the requirement is at least 30 seconds with a fluctuation radius not exceeding 35 meters. If the positioning quality is medium, the required dwell time is increased by 10 seconds; if it is low, it is increased by 20 seconds. If the positioning is invalid, the evidence of dwelling is not valid. If the speed meets the low-speed requirement but the position fluctuation exceeds the corresponding upper limit, the system determines it as low-speed passage and does not count it as dwelling evidence.

[0052] Obtain near-field evidence or passenger confirmation: Near-field evidence is established when the Bluetooth connection lasts for at least 5 seconds and the signal strength is not lower than -75dBm. Near-field evidence is established when the passenger terminal is connected to the in-vehicle Wi-Fi or a hotspot of the same device for at least 10 seconds. If a passenger clicks "I'm on the bus" and the time between the click and the vehicle arriving at the boarding point is no more than 120 seconds, the near-field evidence is established. If the distance between the vehicle and passengers is no more than 30 meters within 30 seconds after the vehicle door is opened, or if the driver and passengers make a phone call or text message confirmation within 180 seconds before or after the vehicle arrives, it will also be considered as valid near-field or confirmation evidence. Near-field evidence cannot trigger billing on its own; it must be used in conjunction with vehicle arrival and vehicle stop.

[0053] Determine the same movement after leaving the boarding point: The system determines whether the vehicle has moved at least 50 meters outside the boarding point fence within 90 seconds; whether the angle between the vehicle's direction of movement and the destination direction is no greater than 60 degrees, or whether the overlap ratio between the replanned route and the navigation route in the first 300 meters is no less than 70%; and calculates the speed difference, direction difference, trajectory similarity, and acceleration correlation within a 60-second window. For high-quality or medium-quality passenger positioning, if at least three of the four pieces of evidence for traveling together are valid, the system determines that traveling together is valid. For low-quality passenger positioning, if at least two pieces of evidence are valid and valid near-field evidence is also valid, the system determines that traveling together is valid. In the case of invalid passenger positioning, the system uses driver terminal and vehicle terminal consistency, vehicle IMU, and near-field evidence as substitutes; if at least two of the substitute pieces of evidence are valid, the system determines that traveling together is valid.

[0054] Begin risk identification and prevention in advance: After the driver clicks "start," five pieces of evidence are checked: vehicle entering the start area, vehicle stopping, passenger approach, valid near-field evidence, and movement of other passengers. If any three of these five items are missing, the system determines that starting early is high-risk, blocks the start of billing, and enters a waiting period for billing or abnormal arbitration. If any two of these five items are missing, the system determines that starting early is medium-risk, sends a confirmation request to the passenger, and continues to observe for 30 seconds. If only one item is missing and both vehicle arrival and vehicle stopping are confirmed, the system determines that starting early is low-risk, and allows the start of billing to proceed according to the rules. Fees incurred during high-risk or medium-risk periods between the driver's manual start and a reasonable start date are frozen.

[0055] Automatically start billing and determine a reasonable start time: The system requires two essential conditions to be met: vehicle arrival and vehicle stay. At least two of the following three conditions must be met: passenger approach or replacement confirmation, route initiation, and evidence of companion movement or replacement. In complex or low-quality positioning scenarios, all three conditions must be met. The above results must be met for two consecutive 30-second observation windows, and the high-risk event must not be triggered if started early. The reasonable start time is prioritized based on the time when the vehicle leaves the pick-up point and companion movement is confirmed. If passenger confirmation exists and its time is later than the vehicle departure time, the passenger confirmation time is used. If the passenger terminal is unavailable and valid near-field evidence is established, the time of the first establishment of near-field evidence is used. The order status is switched to "billing in progress," and the evidence chain for starting billing, candidate times, and the final selection reason are saved.

[0056] As attached Figure 3 The automatic billing termination process shown includes the following: Collect evidence near the destination: During order billing, the system continuously collects the locations of the vehicle, driver, and passenger, and reads the navigation route completion rate, remaining navigation distance, vehicle speed, driver's operation completion, passenger confirmation, GPS accuracy, and network status. When the remaining navigation distance is no more than 500 meters, the route completion rate is no less than 80%, or the vehicle enters a range of 1.5 times the radius of the end fence, the system increases the frequency of end billing judgment to once every 5 seconds. According to the above-mentioned positioning quality classification rules, the system marks the positioning as high quality, medium quality, low quality, or invalid, and records the sampling delay and map matching results.

[0057] Generate target dynamic fence: The system reads the destination type, available parking area, roadside location, historical drop-off points, GPS accuracy, and scene tags. If the destination is a normal scene and no radius is configured, the end fence is set to 50 meters; for complex scenes, it's 80 meters; for highly complex scenes, it's 120 meters. If the vehicle or passenger has medium-quality positioning, the fence radius increases by 20 meters; if either has low-quality positioning, it increases by 40 meters; if the vehicle has invalid positioning, automatic termination is not triggered. For large parks, airports, train stations, and commercial areas, only candidate drop-off areas that meet the following criteria are included in the end fence: at least 20 orders have passed manual review in the past 30 days, are consistent with the roadside location, and have a complaint rate below 5%. The maximum end fence radius is 180 meters; if this limit is exceeded, 180 meters is used as the maximum judgment range.

[0058] Determine road accessibility and actual arrival at the destination: Determine if the vehicle's reachable distance from the road network is no greater than the end fence radius or if it has entered any candidate drop-off area; determine if the vehicle's road level matches the destination and if the vehicle is located in a parking or drop-off area; if the vehicle is on an elevated road, tunnel, closed road, on the opposite side of the road, or on a no-parking main road and the navigation cannot reach the drop-off area within 300 meters, then even if the straight-line distance is less than the fence radius, arrival is not determined; if the vehicle distance is no greater than 50% of the end fence radius, it is recorded as strong arrival evidence; if it is greater than 50% but no greater than 100%, it is recorded as ordinary arrival evidence; if it exceeds the fence radius, it is recorded as not arriving.

[0059] Determining if a vehicle is moving at low speed or stopped: The system determines whether the vehicle speed is no more than 5 km / h and the stopping speed is no more than 1 km / h. For ordinary alighting scenarios, the cumulative time at low speed or stopping should be no less than 15 seconds; for complex alighting scenarios, no less than 25 seconds; and for highly complex alighting scenarios, no less than 35 seconds. If the vehicle speed is higher than 5 km / h, or the position fluctuation radius during low speed exceeds 50 meters, the system determines that the destination has been passed and the process will not end automatically. The stopping time is recorded as one of the candidate ending times. If the subsequent passenger separation time is later than the stopping time, the passenger separation time will be used first.

[0060] Determine path completion rate: The system calculates the completion rate of the actual driving trajectory relative to the planned or recalculated route. A route is considered complete when the completion rate is at least 90% and the remaining navigation distance is no more than 300 meters. In complex destination scenarios, a route is also considered complete when the completion rate is at least 85%, the vehicle enters a candidate drop-off area, and the road level is consistent. If the completion rate is below 70%, the system determines it as passing through the area midway or an abnormal location, and automatic termination is not allowed. When the completion rate is between 70% and 90%, the system enters observation mode and must simultaneously meet the conditions of stopping for at least 35 seconds and passenger separation before continuing the judgment process.

[0061] Determining whether a passenger disembarks or the passenger vehicle separates: If, within 120 seconds after the vehicle stops, the distance between the passenger and vehicle continuously increases by at least 30 meters, and the passenger is located within the candidate drop-off area of ​​the destination, passenger separation is determined. If the vehicle moves more than 80 meters away from the candidate drop-off area of ​​the destination, and the passenger remains within the candidate drop-off area of ​​the destination for at least 60 seconds, passenger separation is determined. If the passenger's position fluctuation radius within the candidate drop-off area of ​​the destination is no greater than 25 meters, and the vehicle continues to move more than 100 meters, passenger separation is also determined. When the passenger's location is invalid, alternative separation evidence is used only if the vehicle stops, the route is completed, the driver has not finished for more than 60 seconds, and there is passenger confirmation or historical drop-off point matching.

[0062] Identify the risk of driver delaying the end of the journey: The system checks five pieces of evidence: vehicle entering the destination's candidate drop-off area, slow speed or stopping, route completion, passenger separation or alternative separation, and driver not ending the process for more than 60 seconds. If any four of these five evidences are present, the system is deemed to have a high risk of delayed termination, and a 30-second countdown is initiated. If any three of these evidences are present, the system is deemed to have a medium risk of delayed termination, and the system prompts the driver to end the process and continues to observe for another 60 seconds. If the vehicle slows down within 300 meters of the destination for more than 180 seconds, or leaves the destination's candidate drop-off area by more than 80 meters while the passenger is still within the destination's candidate drop-off area, the system is directly deemed to have a high risk of delayed termination.

[0063] Calculate the reasonable end time for billing: A set of candidate end times is generated, including the time of first entry into the candidate drop-off area of ​​the destination, the time of first parking, the time of route completion, the time of passenger separation, the time of driver manual end, and the time of passenger confirmation. The candidate time must satisfy at least three pieces of end evidence simultaneously. A time point that satisfies only a single piece of evidence cannot be used as a reasonable end time. If the parking time, route completion time and passenger separation time differ by no more than 60 seconds, the system selects the latest of the three as the reasonable end time to avoid ending before the passenger has disembarked. If the driver's manual end time is more than 60 seconds later than the time of passenger separation, the system selects the time of passenger separation as the reasonable end time and freezes subsequent charges. If the passenger terminal is unavailable and the alternative separation evidence is valid, the latest of the vehicle parking time, route completion time and the time when the driver has not ended for 60 seconds is selected.

[0064] Automatically end or initiate confirmation: The system requires two essential conditions to be met: arrival at the destination and vehicle slow speed or stopping. It also requires at least one of the following: route completion, passenger separation, or alternative separation evidence. For complex or low-quality positioning scenarios, both conditions must be met. If the driver has not ended the process for more than 60 seconds or the vehicle has moved more than 80 meters away from the candidate drop-off area, the system starts a 30-second countdown. If the driver has not ended the process after the countdown expires and the termination rule still applies, the system automatically terminates. If any two of the following conditions are met: invalid positioning data from all three terminals, unreachable roads, inconsistent road levels, or positioning jumps, the system does not automatically terminate and enters an abnormal arbitration process. If only one conflict exists, the system initiates a confirmation process with the driver and passenger, with a confirmation waiting time of 60 seconds. If no confirmation is received within the time limit and the termination rule still applies, the system either automatically terminates or enters arbitration.

[0065] Settle and freeze disputed fees and write back: The billing interval is determined based on a reasonable start time and a reasonable end time; the base fee is calculated by reading the actual trajectory and pricing rules; if there is a risk of starting early, the fee between the driver's manual start time and the reasonable start time is frozen; if there is a risk of ending late, the fee between the reasonable end time and the driver's actual end time is frozen; if an abnormal arbitration is initiated, the frozen interval starts from the time of the first risk trigger and ends at the time of manual review or automatic evidence recovery; the automatic termination evidence chain, driver operation, passenger confirmation, complaint results, and manual review conclusions are written back to the historical sample database; the candidate drop-off area and threshold are only allowed to be updated when there are no fewer than 20 manually reviewed samples for the same POI in the past 30 days.

[0066] This embodiment also provides anomaly handling risk management strategies, specifically including: The driver starts billing in advance: If, after the driver clicks "Start Billing," any three of the following five criteria are missing: vehicle entering the start zone, vehicle stopping, passenger approach, valid near-field evidence, and accompanying movement, the system determines this as a high-risk premature start. The system immediately blocks the start of billing, sends a 60-second confirmation request to the passenger, and freezes the fare between the driver's manual start time and the system's confirmed reasonable start time. If the passenger confirms within 60 seconds and both vehicle arrival and stopping are confirmed, the system enters a waiting period for billing to begin. If the passenger does not confirm and the evidence is still missing, the system initiates an abnormal arbitration process and records the driver's risk sample.

[0067] Driver delays ending billing: If any four of the following conditions are met: the vehicle enters the destination's candidate drop-off area, slows down or stops, the route is completed, passengers are separated or replaced, or the driver has not finished for more than 60 seconds, the vehicle is deemed to be at high risk of delayed termination, and a 30-second countdown to termination is issued to the driver. If the countdown has not ended by then, the vehicle will automatically terminate and a reasonable termination time will be selected according to the above rules. If the driver's actual termination time is more than 60 seconds later than the reasonable termination time, the fee for that time period will be frozen and the driver's risk sample will be recorded.

[0068] GPS drift and weak network scenarios: If any location source meets any of the following conditions, it is determined to be a location drift or invalid: the location accuracy radius is greater than 100 meters; the sampling delay is greater than 20 seconds; the location jump exceeds 80 meters within 3 seconds and the vehicle speed does not support the jump; the map matching road is opposite to the vehicle's travel direction; the distance between the vehicle and the driver exceeds 200 meters within the same time window and the driver does not show a departure status. Invalid locations will be circuit-broken and will not be included in positive evidence. Low-quality locations are only allowed as auxiliary evidence. If a weak network causes a delay of 8 to 20 seconds, the two most recent reliable points will be used for linear interpolation over time. If the delay exceeds 20 seconds, no interpolation will be performed, and the system will wait for a new reliable point before recalculating.

[0069] Elevated roads, main and auxiliary roads, and underground parking lots: When the straight-line distance between the vehicle and the pick-up point or destination is no greater than the radius of the fence, but the road level of the vehicle is inconsistent with the road level of the target location, or the vehicle is located on an elevated road, tunnel, closed road, opposite road, or no-parking main road and cannot be navigated to the target location within 300 meters, the start or end of billing shall not be triggered. In the case of underground parking lot, if GPS positioning is invalid, the system must simultaneously meet two of the following three conditions: parking for no less than 30 seconds, establishment of near-field evidence, matching of historical parking areas, or passenger confirmation, before allowing the billing judgment to continue.

[0070] Scenarios where ride-hailing services and passenger apps are unavailable: For scenarios involving proxy orders, passengers not carrying their mobile phones, passengers with location services disabled, or passengers with no valid location data for 20 consecutive seconds, the system will mark the passenger's location evidence as unavailable. To start billing, at least three of the following four conditions must be met: vehicle arrival, vehicle stoppage, driver confirmation or actual passenger confirmation, and route initiation. To end billing, at least three of the following four conditions must be met: destination arrival, vehicle stoppage, route completion, driver not ending the trip for more than 60 seconds, or actual passenger confirmation. Actual passenger confirmation via SMS or phone must occur within 180 seconds before or after vehicle arrival to be valid.

[0071] Passengers maliciously terminate the trip early or file malicious complaints: If the passenger's location shows that they have reached the destination, but the route completion rate is less than 70%, the vehicle and driver are still traveling on the same road segment, the vehicle speed is greater than 15 kilometers per hour, and the remaining navigation distance to the destination is greater than 1 kilometer, then the passenger's location is considered to be drifting or at risk of premature termination. Billing will not be automatically terminated, and this type of scenario will enter a 60-second observation period. If the vehicle and driver continue to maintain the same route during the observation period and the passenger does not restore a reliable location, then an abnormal arbitration will be initiated and all evidence chains will be preserved.

[0072] After each order is completed, the following is also included: Historical Samples and Dispute Replay Database: The historical sample and dispute write-back database is used to record samples from automatic start, automatic end, passenger complaints, driver appeals, manual liability determination, GPS drift, route deviation, destination detour, and special scenarios. The steps for generating a single sample are as follows: Record the scene type, pick-up point type, and destination type corresponding to the sample; Preserve the boarding evidence, disembarking evidence, risk evidence, and state machine results generated at that time; Save the system's automatic judgment results, subsequent actions by the driver or passenger, complaint results, appeal results, and manual review results; Adjustments to storage costs, confidence level of evidence, and sample update time.

[0073] Samples are used to update the dynamic fence radius, evidence weight, risk threshold, and scene parameters. Updates are only allowed when there are no fewer than 20 manually reviewed samples for the same POI in the past 30 days. The fence radius adjustment cannot exceed 20 meters per time, and the weight adjustment cannot exceed 10% of the original weight per time.

[0074] Example 2: Figure 4 This is a schematic diagram of a ride-hailing order automatic billing device based on a three-terminal spatiotemporal evidence chain, according to another exemplary embodiment. The device includes: Scene classification module 1: Used to obtain ride-hailing order information and determine the pick-up scene type and target scene type based on the pick-up point and destination in the ride-hailing order information; Location acquisition module 2: used to continuously acquire three-terminal location data during the pick-up process, the three-terminal location data including: vehicle terminal location, driver terminal location and passenger terminal location, and determine the quality of the three-terminal location; The billing evidence acquisition module 3 is used to: determine the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality; determine whether the vehicle has arrived at the boarding point based on the three-terminal positioning data, the boarding scenario type, and the dynamic fence of the boarding point; determine whether the vehicle is staying or waiting at low speed within the dynamic fence of the boarding point based on the three-terminal positioning data and the boarding scenario type; determine whether the distance between the vehicle and the passenger terminal continuously shrinks to a preset range within the observation window based on the three-terminal positioning data, thereby determining whether the passenger is approaching; determine whether the vehicle and the passenger terminal meet preset consistency conditions after leaving the boarding point based on the three-terminal positioning data, thereby determining whether they are moving together; and determine whether the route has been initiated based on the three-terminal positioning data and the navigation route. The billing start trigger module 4 is used to trigger automatic billing when the arrival of the vehicle at the pick-up point and the vehicle stops or waits at low speed within the dynamic fence of the pick-up point are both determined to be true, and the three items of passenger approach, traveling together and path start meet the preset number corresponding to the pick-up scenario type, and the above determinations are all true within the continuous observation window. End of Billing Evidence Acquisition Module 5: After entering the start of billing state, it is used to determine the target dynamic fence based on the target scene type and the three-terminal positioning quality; determine whether the destination has been reached based on the three-terminal positioning data and the target dynamic fence; determine whether the vehicle is staying or waiting at low speed within the target dynamic fence based on the three-terminal positioning data and the target scene type; determine whether the completion ratio of the vehicle's traveled path relative to the planned path reaches a threshold based on the navigation route, thereby determining whether the path is completed; and determine whether the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops meets the preset separation condition based on the three-terminal positioning data, thereby determining whether the passenger has been separated. End-of-billing trigger module 6: When the arrival at the destination and the vehicle's stay or slow-speed waiting within the target dynamic fence are both determined to be true, and the completion of the route and separation of the passenger meet the preset number corresponding to the destination scenario type, and all of the above determinations are true within a continuous observation window, the automatic end-of-billing is triggered.

[0075] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0076] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.

[0077] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0078] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0079] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0080] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0081] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0082] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0083] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for automatic billing of ride-hailing orders based on a three-terminal spatiotemporal evidence chain, characterized in that, The method includes: Obtain ride-hailing order information and determine the pick-up scenario type and target scenario type based on the pick-up point and destination in the ride-hailing order information; During the pick-up process, three-terminal positioning data are continuously acquired, including vehicle-end positioning, driver-end positioning, and passenger-end positioning, and the quality of the three-terminal positioning is determined. The system determines the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality. It then determines whether the vehicle has arrived at the boarding point based on the three-terminal positioning data, the boarding scenario type, and the dynamic fence. Based on the three-terminal positioning data and the boarding scenario type, it determines whether the vehicle is stopping or waiting at low speed within the dynamic fence of the boarding point. Based on the three-terminal positioning data, it determines whether the distance between the vehicle and the passenger terminal continuously decreases to a preset range within the observation window, thus determining whether the passenger is approaching. Based on the three-terminal positioning data, it determines whether the vehicle and the passenger terminal meet preset consistency conditions after leaving the boarding point, thus determining whether they are moving together. Finally, based on the three-terminal positioning data and the navigation route, it determines whether a path has been initiated. When the arrival of the vehicle at the pick-up point and the vehicle's stay or slow waiting within the dynamic fence of the pick-up point are both determined to be true, and the three items of passenger approach, co-movement, and path initiation meet the preset number corresponding to the pick-up scenario type, and all of the above determinations are true within a continuous observation window, automatic billing is triggered. After entering the billing state, the target dynamic fence is determined based on the target scene type and the three-terminal positioning quality; the destination is reached based on the three-terminal positioning data and the target dynamic fence; the vehicle is stopped or waiting at low speed within the target dynamic fence based on the three-terminal positioning data and the target scene type; the completion ratio of the vehicle's traveled path relative to the planned path is determined based on the navigation route to determine whether the path is completed; and the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops is determined based on the three-terminal positioning data to determine whether the passenger has been separated. When the arrival at the destination and the vehicle's stay or slow-speed waiting within the target dynamic fence are both determined to be true, and the completion of the route and the separation of passengers meet the preset number corresponding to the destination scenario type, and all of the above determinations are true within a continuous observation window, automatic termination of billing is triggered.

2. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 1, characterized in that, The step of determining the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality includes: Based on the boarding point location, different basic radii are preset according to different boarding scenario types; The base radius is corrected based on the positioning accuracy of the vehicle terminal and the passenger terminal. The lower the positioning accuracy, the larger the corrected radius. Invalid positioning is not included in the correction and the corresponding invalid positioning is marked as unavailable. The corrected fence radius does not exceed the first upper limit value. If the boarding point is a multi-entrance building or park, then the entrance set of the boarding point, the historical boarding points on the same road side, and the parking area will generate multiple candidate entrances respectively. The radius of each candidate entrance is a preset value. When the number of samples that have been manually reviewed and approved by any candidate entrance in the historical period reaches a preset number and the review pass rate reaches a preset ratio, it will be used as a candidate boarding point dynamic fence. Acquire the near-field connection signal between the vehicle and the passenger terminal as valid near-field evidence; If valid near-field evidence is obtained during the pick-up process, the radius of the dynamic fence at the current pick-up point will be reduced by a preset ratio, and the reduced radius will not be lower than the preset lower limit.

3. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 1, characterized in that, Before determining whether the vehicle has arrived at the pick-up point or the destination, a road accessibility check is also included, which includes: Calculate the road network reachable distance from the vehicle location to the pick-up point or destination. If the road network reachable distance is greater than a preset multiple of the vehicle's straight-line distance, then even if the vehicle's straight-line distance is within the radius of the dynamic fence at the pick-up point or the target dynamic fence, the vehicle is not determined to have arrived. If the road where the vehicle is currently located is on a different road level than the road where the pick-up point or destination is located, and the navigation route indicates that the vehicle cannot reach the side of the road where the pick-up point or destination is located within the preset distance, then the vehicle will not be considered to have arrived.

4. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 3, characterized in that, It also includes: risk control, irregular arbitration, and fee freezing. The risk control includes: In the automatic start billing process, after receiving the start billing operation instruction from the driver, the system determines the arrival of the pick-up point, the vehicle's stop or slow-speed wait within the dynamic fence of the pick-up point, the passenger's approach, the number of valid near-field evidence and the number of co-movements, and determines the risk level of starting early based on the number of missing items. When the risk level is high, the system blocks the start of billing and requests the passenger's confirmation. In the automatic billing termination process, the system determines the number of five pieces of evidence that are valid: the vehicle enters the target dynamic fence area, the vehicle stops or waits at low speed within the target dynamic fence, the route is completed, the passenger is separated, and the driver has not ended the billing for more than a preset time. The risk level of delayed termination is determined based on the number of valid evidence items. When the risk level is high, a countdown is started, and the billing is forcibly terminated after the countdown ends. Among them, if the vehicle detours at low speed within a preset range around the destination for more than a preset time, or if the vehicle moves away from the target dynamic fence for more than a preset distance while the passenger terminal is still within the target fence, it is directly determined to be high risk. When at least two of the preset abnormal conditions are simultaneously detected, automatic billing is suspended and the fee for the corresponding time period is marked as pending arbitration. The preset abnormal conditions include: invalid three-terminal positioning data, inconsistent road level between the vehicle location and the pick-up point or destination, and unreachable road. The fees marked as pending arbitration, or the dispute fees corresponding to the risks of early commencement, or the dispute fees corresponding to the risks of delayed termination, shall be frozen from the time of the first risk triggering until the time of risk elimination or manual review.

5. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 4, characterized in that, The determination of whether a vehicle is stopped or waiting at low speed within the pick-up point or target dynamic fence includes: Obtain the duration for which the vehicle speed is below a preset speed threshold; Obtain the dwell time threshold and location fluctuation limit corresponding to the current boarding scenario type or the target scenario type; Obtain the positioning quality of the vehicle and correct the dwell time threshold based on the positioning quality; The vehicle's position fluctuation radius is obtained within the duration. When the duration reaches the corrected dwell time threshold and the position fluctuation radius does not exceed the upper limit of the position fluctuation, it is determined that the vehicle is staying or waiting at low speed.

6. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 5, characterized in that, The step of determining whether the distance between the vehicle terminal and the passenger terminal continuously decreases to a preset range within the observation window based on the three-terminal positioning data, thereby determining whether a passenger is approaching, includes: Different proximity distance judgment radii are determined based on the positioning quality of the passenger terminal; Taking the current moment as the endpoint, take the first observation window forward. If the cumulative reduction or reduction ratio of the distance between the passenger terminal and the vehicle end within the first observation window reaches a preset value, and the current distance between the two is not greater than the proximity distance determination radius, then the passenger proximity is determined to be established. If passenger terminal location is marked as unavailable, valid near-field evidence or passenger confirmation actions will be used instead of passenger proximity determination.

7. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 6, characterized in that, The step of determining whether the vehicle and the passenger terminal meet preset consistency conditions after the vehicle leaves the boarding point based on the three-terminal positioning data, thereby determining whether they are moving together, includes: The second observation window is taken as the starting point when the vehicle leaves the dynamic fence of the boarding point. If the number of valid positioning sampling points in the second observation window is less than the preset number, then no same-movement judgment is made. The differences in motion parameters between the vehicle terminal and the passenger terminal within the second observation window are calculated, including speed difference, direction difference, trajectory similarity and acceleration correlation. When each parameter meets its corresponding preset threshold, it is counted as evidence of consistency. The required number of consistent evidences is determined based on the positioning quality of the passenger terminal. The lower the positioning quality, the higher the required number of consistent evidences. When the required number of consistent evidences is reached, it is determined that the co-movement is established.

8. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 7, characterized in that, The step of determining whether passenger separation has occurred based on the three-terminal positioning data, specifically whether the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops, meets preset separation conditions, includes: The third observation window is taken as the starting point of the vehicle parking time; the distance between the vehicle end and the passenger terminal is continuously monitored within the third observation window. When the distance continuously increases beyond the preset distance and the passenger terminal position is within the target dynamic fence range, the passenger separation is determined to be successful. The vehicle and passenger terminal positions are continuously monitored within the third observation window. When the vehicle moves away from the target dynamic fence by more than a preset distance while the passenger terminal position remains within the target dynamic fence for more than a preset time, the passenger separation is determined to be successful.

9. The automatic billing method for ride-hailing orders based on a three-terminal spatiotemporal evidence chain according to claim 8, characterized in that, When automatic billing is triggered, start time backtracking is also performed: the time when the vehicle leaves the boarding point dynamic fence and the same movement is established, the time when the passenger confirms the operation, and the time when valid near-field evidence is first established, and the latest of these times is used as the billing start time; When automatic billing termination is triggered, the termination time is also backtracked: the time when the vehicle first enters the target dynamic fence, the time when the vehicle first stops, the time when the route completion ratio reaches the threshold, the time when passenger separation is established, and the time when the driver manually terminates the operation are recorded respectively. At least three times that are established at the same time are selected from the above times, and the latest one is taken as the billing termination time. If the driver manually ends the operation later than the billing end time by more than a preset deviation, the fees between the two will be marked as pending freezing.

10. An automatic fare calculation device for ride-hailing orders based on a three-terminal spatiotemporal evidence chain, characterized in that, The device includes: Scene classification module: used to obtain ride-hailing order information and determine the pick-up scene type and target scene type based on the pick-up point and destination in the ride-hailing order information; Location acquisition module: used to continuously acquire three-terminal location data during the pick-up process, including vehicle-end location, driver terminal location, and passenger terminal location, and to determine the quality of the three-terminal location; The billing evidence acquisition module is used to: determine the dynamic fence of the boarding point based on the boarding scenario type and the three-terminal positioning quality; determine whether the vehicle has arrived at the boarding point based on the three-terminal positioning data, the boarding scenario type, and the dynamic fence of the boarding point; determine whether the vehicle is staying or waiting at low speed within the dynamic fence of the boarding point based on the three-terminal positioning data and the boarding scenario type; determine whether the distance between the vehicle and the passenger terminal continuously decreases to a preset range within the observation window based on the three-terminal positioning data, thereby determining whether the passenger is approaching; determine whether the vehicle and the passenger terminal meet preset consistency conditions after leaving the boarding point based on the three-terminal positioning data, thereby determining whether they are moving together; and determine whether the route has been initiated based on the three-terminal positioning data and the navigation route. The billing start trigger module is used to trigger automatic billing when the arrival of the vehicle at the pick-up point and the vehicle stops or waits at low speed within the dynamic fence of the pick-up point are both determined to be true, and the three items of passenger approach, co-movement and path start meet the preset number corresponding to the pick-up scenario type, and the above determinations are all true within the continuous observation window. End of Billing Evidence Acquisition Module: After entering the start of billing state, this module determines the target dynamic fence based on the target scene type and the three-terminal positioning quality; determines whether the destination has been reached based on the three-terminal positioning data and the target dynamic fence; determines whether the vehicle is staying or waiting at low speed within the target dynamic fence based on the three-terminal positioning data and the target scene type; determines whether the completion ratio of the vehicle's traveled path relative to the planned path reaches a threshold based on the navigation route, thereby determining whether the path is completed; and determines whether the distance change between the vehicle and the passenger terminal within a preset time after the vehicle stops meets a preset separation condition based on the three-terminal positioning data, thereby determining whether the passenger has been separated. End-of-billing trigger module: When the arrival at the destination and the vehicle's stay or slow-speed waiting within the target dynamic fence are both determined to be true, and the completion of the route and separation of the passenger meet the preset number corresponding to the destination scenario type, and all of the above determinations are true within a continuous observation window, the automatic end-of-billing is triggered.

Images