A car surrounding commodity recommendation method based on an internet of things

Abstract

The present application relates to the field of internet of things technology and business data processing, and discloses a vehicle surrounding commodity recommendation method based on internet of things, comprising: obtaining a future trajectory line of a vehicle containing a time-arrival probability distribution and supply information of a merchant, calculating an expected probability of a space-time encounter between the two to generate an expected encounter event, and comprising a dynamic robustness maintenance mechanism for discarding the calculation result and activating an implicit trajectory deduction mode when the vehicle deviates from the trajectory, the present application changes an instant query structure initiated by a single party only after demand is generated by establishing a two-way information guidance mode based on future event prediction to cope with the supply-demand mismatch problem in a mobile environment, and the dynamic robustness maintenance mechanism contained maintains the continuity and reliability of the service when the vehicle driving path is uncertain.

Description

Technical Field

[0001] This invention relates to a method for recommending vehicle-related products based on the Internet of Things (IoT), belonging to the field of IoT technology and business data processing. Background Technology

[0002] In current IoT applications, providing location-based service recommendations for moving vehicles is a common technological approach. By acquiring the vehicle's real-time geographic location and pushing nearby commercial information, it meets some of the driver's immediate needs. However, when this approach, which relies on a single-vehicle perspective and immediate response, is placed in a large-scale, highly dynamic real-world traffic environment, especially in scenarios with significant passenger flow tidal effects, such as highway service areas or popular tourist destinations, its design limitations lead to future information barriers between the supply and demand sides. On the one hand, when drivers approach a service point, they are unaware of the service point's real-time capacity and potential queuing situation, resulting in uncertainty in their consumption decisions. On the other hand, service providers cannot predict the scale of passenger flow and their demand preferences in the near future, and can only rely on experience-based inventory preparation and passive service. This directly leads to the misallocation of service resources and the loss of business opportunities.

[0003] To improve this situation, the industry has tried to predict demand by increasing the frequency of data interaction or building large-scale centralized platforms. However, the former only speeds up the updating of real-time information and does not address the prediction of future states, while the latter introduces excessive system complexity, computational costs, and the risk of leakage of user travel privacy. Therefore, these conventional improvement ideas have failed to provide universally applicable solutions.

[0004] Specifically, existing technologies suffer from the following shortcomings: 1. The lack of a predictive mechanism for the future spatiotemporal state of vehicles results in delayed and reactive recommendation behavior, failing to provide proactive supply and demand guidance; 2. Information exchange between the supply and demand sides is unidirectional and static, unable to reflect real-time changes in service capabilities or form a closed-loop coordinated adjustment; 3. The failure to transform the individual future trajectories of massive numbers of vehicles into collective predictive data usable for macro-level situational awareness makes it impossible to anticipate and avoid service point congestion caused by demand convergence. Therefore, establishing a lightweight predictive mechanism that, while protecting user privacy, effectively links the future trajectories of individual vehicles with the dynamic service capabilities of businesses, enabling two-way proactive guidance of information between supply and demand, and ensuring the mechanism's reliable operation under various uncertain conditions of real-world driving, has become a technical challenge that needs to be addressed by those skilled in the art. Summary of the Invention

[0005] This invention provides a vehicle-related product recommendation method based on the Internet of Things. Its main purpose is to solve the problem that existing technologies cannot predict the future spatiotemporal state of vehicles, resulting in one-way and static supply and demand information, making it difficult to achieve proactive collaborative guidance and lacking robustness under real driving conditions.

[0006] To achieve the above objectives, the present invention provides a method for recommending vehicle-related products based on the Internet of Things, the method comprising the following steps:

[0007] Step a, obtain the future trajectory line generated by the vehicle based on its navigation destination. The future trajectory line is a data structure that contains a series of future spatiotemporal points and the time arrival probability distribution of the corresponding future spatiotemporal points.

[0008] Step b: Obtain supply information published by the merchant, including service content and effective time and space window;

[0009] Step c: Calculate the expected probability of the vehicle's future trajectory line encountering the merchant's supply information in time and space, so as to generate at least one expected encounter event.

[0010] Step d involves pushing guidance information to vehicles or merchants based on the probability of expected encounters. The method also includes a dynamic robustness maintenance mechanism, which performs the following operations: continuously comparing the vehicle's real-time location with its future trajectory; determining a trajectory deviation event when the lateral offset distance between the vehicle's real-time location and the future trajectory exceeds a baseline threshold; discarding all calculation results for expected encounters based on the future trajectory in response to the trajectory deviation event; and activating an implicit trajectory inference mode. This implicit trajectory inference mode generates a probability vector field representing the probability of the vehicle's short-term driving direction, based on historical gravity points determined by cluster analysis of the vehicle's local historical trajectory data, and selecting the historical gravity point with the highest weight in conjunction with the current time and date context. This probability vector field is used to maintain the product recommendation service.

[0011] Preferably, the dynamic robustness maintenance mechanism also performs the following operations: after activating the implicit trajectory extrapolation mode, it continuously monitors whether there is a re-anchoring event where the vehicle trajectory is re-anchored to a new or existing navigation route, and in response to the re-anchoring event, it automatically resumes the execution of steps a to d.

[0012] Preferably, after step d, the method further includes: aggregating the future trajectory lines of multiple vehicles pointing to the same merchant and whose time windows overlap, to calculate an expected encounter traffic density that characterizes future passenger flow density; comparing the expected encounter traffic density with the merchant's service capacity to determine an encounter quality level; and adjusting the content of the guidance information pushed to the vehicles in step d based on the encounter quality level.

[0013] Preferably, the time arrival probability distribution in step a is dynamically calculated and updated based on the vehicle's current speed and real-time road condition information using a normal distribution model.

[0014] Preferably, the calculation of the expected probability in step c involves calculating the probability that each discrete spatiotemporal point on the future trajectory falls within the merchant's effective spatiotemporal window, and then summing the probabilities of each point by weight.

[0015] Preferably, the guidance information pushed to the vehicle in step d is an opportunity reminder that includes the expected probability of encountering the vehicle and suggestions for booking or navigation detours.

[0016] Preferably, the guidance information pushed to the merchant in step d is a future passenger flow density heat map generated based on the expected encounter probability of multiple vehicles.

[0017] Preferably, the method further includes a dynamic modulation mechanism for supply-side service capacity, which performs the following operations: when an expected encounter event is confirmed as a real transaction, an encounter confirmation signal associated with the real transaction is received; based on the encounter confirmation signal, the service capacity count value in the supply information corresponding to the merchant is reduced in real time; and based on the reduced service capacity count value, the subsequent expected probability is calculated.

[0018] Preferably, the guidance information pushed to the vehicle in step d is a binary timing opportunity containing at least two alternative encounter events that differ in their core value propositions; the method also includes: receiving behavioral feedback signals generated by the user's selection behavior of the binary timing opportunity; constructing a personalized profile for the user based on the accumulated behavioral feedback signals, including time sensitivity and price sensitivity weight factors; and using the weight factors in the personalized profile to sort or filter the subsequently generated expected encounter events.

[0019] Preferably, the calculation of the expected encounter flow density follows these rules: ,in, To anticipate the traffic density, In the time window The future trajectory lines all point to the total number of vehicles belonging to the merchant. For the first The expected probability of a vehicle encountering a business.

[0020] Compared with the prior art, the beneficial effects of the present invention are:

[0021] 1. By acquiring the future trajectory line generated by the vehicle based on the navigation destination and combining it with the effective spatiotemporal window supply information released by the merchant, the expected probability of the two parties encountering each other in the future is calculated before the vehicle and the merchant make physical contact. Based on this expected probability, the method pushes guidance information to both the vehicle and the merchant, establishing a two-way information guidance method based on the prediction of future events, which changes the information interaction structure of instant query initiated by only one party after the demand is generated.

[0022] 2. The dynamic robustness maintenance mechanism included in this method first discards the calculation results based on the old trajectory line when a vehicle trajectory deviation event is determined, and then activates the implicit trajectory inference mode based on the vehicle's local historical trajectory data. This design combines adaptability to the aimless state and maintains the continuity of the product recommendation service in two different scenarios where the core input of the main scheme is missing or fails. By monitoring the trajectory resetting event, this method switches between different inference modes, forming an operating logic that covers multiple driving states.

[0023] 3. This invention couples multiple technical mechanisms to form a multi-dimensional information verification and adjustment loop. On the one hand, by aggregating the future trajectory lines of multiple vehicles pointing to the same merchant, the expected traffic density is calculated, thereby predicting possible micro-congestion in the future and adjusting the content of the guidance information accordingly. On the other hand, by receiving the encounter confirmation signal after a real transaction, the service capacity count value in the merchant supply information is updated in real time. The final recommendation decision is simultaneously constrained by both the congestion assessment based on group prediction and the real-time service capacity based on transaction feedback, thus verifying the practical feasibility of the recommendation results. Attached Figure Description

[0024] Figure 1 This is a sequence diagram of the interaction flow of a vehicle-related product recommendation method based on the Internet of Things according to the present invention;

[0025] Figure 2 This is a core logic functional block diagram of a vehicle-related product recommendation method based on the Internet of Things according to the present invention;

[0026] Figure 3 This is a functional interaction use case diagram of a vehicle-related product recommendation system based on the Internet of Things according to the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are only some embodiments of this invention, not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0028] This invention provides a method for recommending vehicle-related products based on the Internet of Things (IoT). This method is built on a system comprising vehicle edge computing nodes, merchant information publishing terminals, and cloud or regional aggregation nodes. Its data flow begins with a probabilistic prediction of the vehicle's future state, and through spatiotemporal calculations of supply and demand information, it provides pre-emptive information guidance to both supply and demand sides. Specifically, the method first obtains a future trajectory line generated based on the navigation destination from the vehicle side, and obtains supply information including dynamic service capabilities from the merchant side. Then, the cloud or regional aggregation node performs spatiotemporal calculations on both to generate an expected encounter event representing the probability of future meetings, and recommends products based on the probability of this event. The method pushes guidance information to vehicles or merchants and also integrates a dynamic robustness maintenance mechanism to deal with situations where vehicles deviate from the predetermined navigation route during driving, thereby maintaining service continuity in a real driving environment. In a specific application scenario, such as a car traveling along a highway, the driver has set the final destination through the in-vehicle navigation system. Due to the lack of prediction of future passenger flow, there is often a mismatch between the supply capacity of commercial services such as catering and refueling in the highway service areas and the demand of vehicles entering the service areas. To address this supply and demand imbalance caused by the lack of future information transparency, the method provided by this invention is configured to perform the following steps.

[0029] Step a) Obtain the future trajectory line generated by the vehicle based on its navigation destination. The future trajectory line is a data structure containing a series of future spatiotemporal points and the corresponding time arrival probability distributions of these points. Specifically, the vehicle edge computing node uses the final destination and planned path provided by the navigation system as basic input to generate a series of discrete spatiotemporal point sequences (t+1, loc+1), (t+2, loc+2) that the vehicle will pass through within a future time period, such as 30 minutes. For each spatiotemporal point, to quantify the uncertainty of the vehicle's actual arrival time, this method uses a normal distribution model to characterize the probability of its arrival time, that is, to calculate a dynamically adjusted time arrival probability distribution for each point. The expectation of this distribution The calculation is based on the current vehicle speed, the remaining distance of the navigation route, and real-time traffic information, while the variance... This reflects the degree of volatility in current road conditions. This method uses this approach to probabilistically predict the future spatiotemporal state of individual vehicles. Step b involves obtaining supply information published by merchants, which includes service content and effective spatiotemporal windows. The merchants' supply information is designed as a data structure containing core service content, real-time service capabilities, and effective spatiotemporal windows. Merchants publish their service information through an application deployed on smartphones or dedicated terminals. For example, a coffee shop in a service area publishes freshly ground lattes, valid for the next 2 hours, with a service capacity of 5 cups per minute. This information is packaged into a supply vector and broadcast to the regional server via wireless communication networks such as 4G or 5G. This supply vector contains digital service resource information that can be dynamically calculated and scheduled by the system.

[0030] The system calculates the expected probability of a vehicle's future trajectory line encountering the merchant's supply information in space and time, generating at least one expected encounter event. To achieve this calculation, the system simplifies the integral process of solving the spatiotemporal encounter probability of the two into a discrete probability multiplication and accumulation. Specifically, it calculates the probability of each discrete spatiotemporal point on the future trajectory line falling within the merchant's effective spatiotemporal window. If the time arrival probability distribution of a future spatiotemporal point is... Furthermore, the effective time and space window for merchant supply information is and geographical scope The probability of a valid encounter occurring at that point is... And its geographical location falls into The system will then map all future points on the trajectory line. We perform a weighted summation to obtain a total expected encounter probability. When the probability value exceeds a preset trigger threshold, such as 75%, the calculation ultimately generates an expected encounter event containing vehicle, merchant, and probability information for subsequent guidance decisions. Step d involves pushing guidance information to vehicles or merchants based on the probability of the expected encounter event. For vehicles, the pushed guidance information is an opportunity reminder, which may include the estimated encounter probability and suggested actions, such as an 85% probability that brand A coffee will not require queuing at a service area 15 kilometers ahead, prompting the user to make a reservation. This interaction is achieved through the vehicle's central control screen or mobile device. For merchants, the system aggregates the expected encounter probabilities of multiple vehicles and generates a future passenger flow density heat map in the form of anonymous statistics, which is then pushed to the merchant's terminal. For example, it is predicted that approximately 120 vehicles with a preference for coffee consumption will pass by within the next hour, suggesting adjustments to inventory. This two-way information interaction method is used to achieve information collaboration between supply and demand based on future predictions.

[0031] To address situations where drivers deviate from the planned navigation route during real-world driving, rendering predictions based on the original trajectory inapplicable, this method also includes a dynamic robustness maintenance mechanism. This mechanism performs the following operations: First, the vehicle's real-time position is continuously compared with the future trajectory. The comparison procedure is defined as follows: the onboard edge nodes compare the vehicle's real-time GPS position with the geometric path of the future trajectory at a frequency of 1Hz. When the vehicle's lateral deviation distance continuously exceeds a baseline threshold (e.g., 50 meters) set to filter out signal drift or normal lane changes for a specific duration (e.g., 3 seconds), the system determines that a trajectory deviation event has occurred. In response to the trajectory deviation event, to avoid guidance based on the invalid trajectory, the system performs a circuit breaker operation, i.e., immediately discarding all calculation results of expected encounters based on the future trajectory. Subsequently, to maintain service continuity, the system activates an implicit trajectory extrapolation mode. This mode is used to probabilistically extrapolate the vehicle's short-term driving intentions when the vehicle is in a state without a clear navigation destination. The specific procedure is as follows: When the vehicle is not in motion, the edge computing node performs density clustering analysis on the locally stored historical trajectory data to automatically identify the user's historical gravity points. This process is completed locally to protect user privacy. During driving, the mode combines the current time and date context, such as 8:00 AM on a weekday, to select the historical gravity point with the highest weight. For example, the gravity weight of the company is assigned as 0.9. Based on the vehicle's current position and the gravity point, a probability vector field representing the probability of the vehicle's short-term driving direction is generated. This probability vector field is then used as a logically equivalent future trajectory input to maintain the product recommendation service. This dynamic robustness maintenance mechanism also continuously monitors whether there are re-anchoring events where the vehicle trajectory is re-anchored to a new or existing navigation route. Once a re-anchoring event is detected, the system will automatically exit the implicit trajectory inference mode and resume the execution of steps a to d. Through this logical closed loop of deviation capture-circuit abandonment-mode switching-re-re-restoration recovery, this method can adapt to route changes in real driving environments.

[0032] Meanwhile, to effectively prevent privacy risks associated with inferring user identity through historical gravity points, this invention adheres to the following multi-layered privacy protection principles in the design and operation of the implicit trajectory deduction mode: First, the principle of localized data processing. For example, the storage, clustering analysis, and identification of historical gravity points of vehicles are all completed in a closed loop on the vehicle's edge computing nodes. The raw trajectory data (Raw GPS Data) containing the user's precise geographical location is always stored locally and is not uploaded to any cloud server, physically isolating the risk of leakage of raw sensitive data. Second, the anonymization and labeling of gravity point data. After identifying historical gravity points (such as home addresses and company addresses) locally, the system does not directly use their actual geographic coordinates. Instead, it anonymizes them; for example, the home address is abstracted as the internal code gravity point A, and the company address is abstracted as gravity point B. In all subsequent calculations and weight allocations, these anonymous tags, which contain no geographic information, are used. The output results are probabilistic and abstract: when interacting with external systems, the vehicle terminal does not upload specific gravity point coordinates or their anonymous tags, but rather a probability vector field representing the vehicle's short-term driving direction, generated based on the gravity weights of these gravity points. This vector field is purely mathematical data, representing only the probability distribution of driving direction; for example, within the next 5 minutes, the probability of driving north is 70%, and the probability of driving east is 20%. External systems (including recommendation engines) cannot extract any specific, sensitive geographic endpoint from this abstract vector field. User authorization and controllability design: This function is designed as an opt-in mode. Before first use, the user is clearly informed of its working principle and data processing method, and user authorization is obtained. Simultaneously, the system provides a visual management interface, allowing users to view, delete, or manually correct historical gravity points automatically identified by the system at any time, ensuring that users have ultimate control over their personal data. These are all extended implementation methods known to those skilled in the art.

[0033] In some implementations, to improve the accuracy and practical feasibility of recommendations, the following steps are also included: to predict and avoid service point congestion caused by demand convergence, before step d, the future trajectory lines of multiple vehicles pointing to the same merchant and with overlapping time windows are aggregated to calculate an expected encounter traffic density characterizing future passenger flow density. Its calculation follows the rules: ,in, In the time window The future trajectory lines all point to the total number of vehicles belonging to that merchant. For the first The system calculates the expected encounter probability between a vehicle and the merchant, compares the calculated expected encounter traffic density with the service capacity in the merchant's supply information to determine an encounter quality level, such as smooth, saturated, or congestion warning. Based on this level, the system adjusts the content of the guidance information pushed to the vehicle. To maintain the real-time accuracy of the service capacity in the merchant's supply information, the method also includes a dynamic modulation mechanism for supply-side service capacity. When an expected encounter event is confirmed as a real transaction, the merchant system sends an encounter confirmation signal to the cloud. Based on this signal, the cloud deducts the service capacity count value in the merchant's corresponding supply information in real time, and... The mechanism calculates the expected probability of service capacity after deduction, forming a closed-loop processing loop that adjusts feedback based on actual transaction consumption. In addition, to explore the user's business intentions, the guidance information pushed to the vehicle can be designed as a binary timing opportunity containing at least two alternative encounter events that differ in their core value propositions. The system receives the behavioral feedback signal generated by the user's choice of the binary timing opportunity and builds a personalized profile for the user locally based on the accumulated signal, which includes time sensitivity and price sensitivity weighting factors. When generating expected encounter events, the weighting factors in the profile can be used to sort or filter the events.

[0034] Example 1: In a highway scenario during a holiday return peak, a vehicle A, with its navigation destination set, travels approximately 50 kilometers to a large service area X. This service area is the last major stop before entering the city ahead. According to historical traffic data, this service area frequently experiences congestion at its entrance and long queues at its internal commercial facilities during this time period. At this point, the driver of vehicle A faces a parking decision dilemma. Meanwhile, the operator of the coffee shop S within the service area cannot predict the customer flow within the next hour and can only rely on experience-based inventory preparation and reactive service. In this scenario, the method disclosed in this invention generates a future trajectory line on the edge computing node of vehicle A, based on its navigation destination, containing a series of future spatiotemporal points and their corresponding time arrival probability distributions. At the same time, coffee shop S has published supply information through its terminal, which is valid for the next 2 hours and has a service capacity of 5 customers per minute. After receiving the future trajectory line of vehicle A and the supply information of coffee shop S, the system cloud node calculates that the expected encounter probability between vehicle A and coffee shop S is 92%. The generation of this calculation result triggers the system to assess the macro-level passenger flow status of service area X.

[0035] Specifically, the cloud node aggregates the expected encounter probability data of all other vehicles whose future trajectory lines also point to service area X and arrive within a similar time window, and calculates the probability based on the expected encounter traffic density rules. The system determined that during the expected arrival time of vehicle A, coffee shop S would experience an expected traffic density of 8 customers per minute, exceeding its service capacity of 5 customers per minute. Based on this, the system classified the expected encounter as a congestion warning. Accordingly, the guidance information sent to vehicle A was configured as a binary timing opportunity with two differentiated options: Option 1, indicating that service area X coffee shop S was expected to be reached, but there might be a queue with an estimated waiting time of approximately 20 minutes; Option 2, based on simultaneous calculations of alternative routes, indicated that... The current route continues for 18 kilometers to coffee shop T in service area Y, with an expected encounter probability of 88%. There is no queue at the shop, but the total journey will be extended by about 10 minutes. The driver of vehicle A, weighing the time cost, chooses option two. Ultimately, vehicle A's navigation route is automatically updated to service area Y, where the driver receives service without waiting. Meanwhile, coffee shop S in service area X, through the system's heat map of future passenger flow density based on all potential passengers, including vehicle A, has pre-configured personnel and materials to cope with the upcoming passenger flow peak.

[0036] Example 2: To objectively verify the actual effectiveness of the technical solution of this invention in addressing the dynamic supply-demand mismatch problem in a mobile environment, this example constructs a simulation test environment based on multi-agent micro-traffic flow. This test environment simulates a 200-kilometer-long two-way highway with three service areas. Service area X, located at kilometer 150, is designated as a core commercial node, and the service capacity of its coffee shop S is set to 5 units per minute. The total test duration is set to 8 hours to cover at least one complete passenger flow tidal cycle, including off-peak, peak, and sub-peak periods. A total of 5000 vehicles with randomly generated navigation destinations and service preferences, such as a potential demand for coffee, are simulated. Vehicles; the experiment was conducted in two groups for comparison: the control group, whose vehicle recommendation service adopted the traditional method based on real-time geolocation, that is, when a vehicle enters a service area X within a preset 5-kilometer radius, service information is pushed to it; the experimental group, whose vehicle and merchant node behavior fully followed the complete technical solution disclosed in this invention, which includes future trajectory prediction, expected traffic density calculation, and dynamic guidance; both groups operated under the same traffic flow generation model, road network topology, and merchant service capacity settings to eliminate interference from irrelevant variables; during the experiment, the system recorded and statistically analyzed key performance indicators such as vehicle queue length, average waiting time, transaction success rate, and service resource utilization rate in each service area at a 1-second interval.

[0037] During the third hour of the test run, when the first return peak began, the simulation platform observed that a large number of vehicles in the control group with a need for coffee arrived at the 5-kilometer recommended trigger circle of service area X and received the recommended information in a similar time period. This caused a high degree of convergence in vehicle decisions in a short period of time, and long queues quickly formed at the entrance ramp of service area X and in front of coffee shop S. In the test group, because the system calculated the expected encounter probability of each vehicle heading to coffee shop S in advance and aggregated to generate the expected encounter traffic density, when this density exceeded the service capacity of coffee shop S, the system dynamically adjusted the guidance information for subsequent vehicles. For example, it provided differentiated opportunity options such as going to the next service area or including waiting time warnings. Some vehicles chose to divert based on this information, which suppressed and smoothed the peak passenger flow to service area X.

[0038] A comparison of key performance indicators (KPIs) between the two groups during peak hours revealed a clear difference in operational efficiency. In terms of service experience, the average waiting time in the control group was 22.1 minutes, while in the experimental group it was 4.1 minutes. Correspondingly, the transaction abandonment rate due to long waiting times was 41.5% in the control group and 5.8% in the experimental group. Regarding business efficiency, the success rate of transactions via guidance information was 89.7% in the experimental group, compared to 45.2% in the control group. These differences stem from the technical solution adopted by the experimental group, which, through anticipated encounters… The calculation of traffic density allows for the prediction of future congestion to be placed at the forefront of the recommendation and decision-making process for individual vehicles, thereby transforming disordered individual decision-making behavior into collective behavior that is coordinated and regulated at the system level. In addition, the experimental data also recorded that the idle time of coffee shop S in the experimental group was 11.3%, while that in the control group was 2.3%. The reason for this phenomenon is that, in order to ensure the service quality of service area X, the experimental group's system actively guided a portion of the expected passenger flow exceeding its carrying capacity to other service nodes. This was a system adjustment behavior that sacrificed the maximum throughput capacity of a single node in exchange for the overall stable operation of the road network service system.

[0039] Example 3: This example combines Figures 1 to 3 This paper describes the implementation of an IoT-based method for recommending car-related products. Figure 1As shown, this process involves four core entities: vehicles, an IoT platform, merchants, and a recommendation engine. It begins with the vehicle uploading information including its navigation destination to the IoT platform. The recommendation engine then generates a future trajectory line containing a series of future spatiotemporal points and their corresponding arrival probability distributions. Simultaneously, merchants publish supply information, including service content and effective spatiotemporal windows, through the IoT platform. This supply information is then transmitted to the recommendation engine. Based on the received vehicle future trajectory line and merchant supply information, the recommendation engine performs spatiotemporal matching calculations to determine the expected probability of their encounter and generate an expected encounter event. Finally, based on the probability of this expected encounter event, the system pushes information bidirectionally to both the vehicle and the merchant through the IoT platform. The vehicle receives an opportunity reminder including booking or navigation detour suggestions, while the merchant receives a future passenger flow density heatmap for passenger flow prediction.

[0040] like Figure 2 As shown, the diagram begins with two parallel initial data input steps: acquiring the future trajectory of vehicles and acquiring merchant supply information. The former outputs a data structure containing the probability distribution of arrival times, while the latter outputs data containing service content and effective spatiotemporal windows. These two data streams converge into the core module for calculating the expected probability of spatiotemporal encounters. This module aims to perform spatiotemporal dimension calculations on the information from both supply and demand sides, producing one or more expected encounter events. Subsequently, the system enters the decision-making stage for pushing guidance information. This stage makes decisions based on the probability of the generated expected encounter events, distributing guidance information to two target objects: one is to push opportunity reminders to vehicles, which contain specific booking or navigation detour suggestions; the other is to push passenger flow heat maps to merchants, which aim to present future passenger flow density to assist merchants in making inventory preparation decisions.

[0041] like Figure 3 As shown in the figure, this diagram illustrates the core functions provided by the vehicle-related product recommendation system of this invention from a user interaction perspective. The diagram clearly defines two types of external interactive entities: vehicles / drivers and merchants. For vehicles / drivers, they can interact with the system to receive personalized opportunity reminders and obtain dynamic route suggestions. For merchants, they can use the system to perform operations such as publishing supply information, receiving future customer flow warnings, and updating real-time service capabilities. This diagram clearly defines the service boundaries and core value provided by the system to both supply and demand users.

[0042] Example 4: To ensure consistent deployment of the method across different hardware platforms, engineering calibration and parameter configuration are required for its dynamic robustness maintenance mechanism and implicit trajectory extrapolation mode. This configuration process includes establishing a baseline threshold for determining trajectory deviation events and determining historical gravity points and activation rules for the implicit trajectory extrapolation mode, creating a data-driven setting method. The calibration process first targets the baseline threshold in the dynamic robustness maintenance mechanism. This threshold is set to balance the sensitivity to identifying genuine deviation intentions with the tolerance to normal driving behavior and signal noise. The calibration process collects GPS trajectory data for a specific vehicle model for more than 30 consecutive minutes on various roads during normal driving, lane changes, and cornering operations. By analyzing this dataset, the maximum lateral error caused by signal drift during straight-line driving is statistically determined. And the maximum lateral displacement during standard lane change maneuvers. To avoid misclassifying the above two situations as trajectory deviation events, a baseline threshold is set. The value of follows The calculation is performed according to the rules, where, A safety factor greater than 1, such as 1.5, is used to provide a buffer zone when a certain vehicle model's measured safety margin is lower than the required level. It is 5 meters. If it is 3.5 meters, then its The threshold is set at 7.5 meters, thus providing a quantifiable basis for calculation.

[0043] Subsequently, a configuration is performed for the implicit trajectory extrapolation mode. This configuration transforms the user's historical driving habits into a computable short-term intent prediction model. The configuration process begins with offline mining of historical gravity points. This process employs a density clustering algorithm applied to a dataset containing at least 90 days of local vehicle historical trajectories. In this algorithm, the neighborhood radius... With minimum sample size The parameters are set to 200 meters and 5, respectively, to aggregate parking points in the same geographical area and to filter parking point clusters with a certain access frequency as candidate historical gravity points. After the historical gravity points are determined, the system further establishes a context-based weighted activation rule base, which is composed of a series of IF-THEN logics. One rule is defined as follows: IF the current time is between 7:00 AM and 9:00 AM on a weekday, THEN sets the weight factor of historical gravity points labeled "Company" to 0.9, sets the weight factor of historical gravity points labeled "Home" to 0.1, and sets the weight of all other gravity points to 0. After the above calibration and configuration procedures, when the method of the present invention is deployed on a specific vehicle platform, the key parameters and logical rules on which its dynamic robustness maintenance function and implicit trajectory inference function depend have a reproducible engineering setting basis.

[0044] Example 5: When deploying the method of the present invention, its personalized profiling mechanism needs to be initially configured and its iterative logic set to handle the initial recommendation of new users and subsequent preference adaptation. This configuration process is used to generate basic preference weights for new user profiles. In the offline stage, the system performs cluster analysis on historical behavior feedback signals from anonymous user groups to identify several user preference profile prototypes, such as time-priority, price-sensitive, and balanced types, and calculates the corresponding average time sensitivity and price sensitivity weight factors for each prototype. When a new user first uses this method, the system sets the initial weight factor of its personalized profile to the value of the origin of the balanced profile, which constitutes the benchmark starting point for its personalized iteration.

[0045] During user interaction with the system, the weighting factors of the personalized profile follow a clear iterative update procedure; when a user chooses an efficiency-oriented option in a binary timing opportunity, their time sensitivity weight... Perform a gain update, while the price sensitivity weight... Then a decay update is performed, and vice versa; the weight update follows the exponential moving average algorithm, and its update rule is as follows: ,in, For the updated weights, The weights before the update. For the target values ​​corresponding to this selection, when selecting the efficiency option, the target value for time sensitivity is 1, and the target value for price sensitivity is 0. A fixed learning rate is used to control the impact of a single action on the overall profile. Through the above-mentioned initial value setting and iterative update procedure for personalized profiles, the recommendation function in the method of this invention has the ability to provide initial recommendations based on group statistics for new users and to make personalized adjustments for existing users based on their behavioral feedback.

[0046] Meanwhile, to verify the actual effectiveness of the personalized profiling mechanism, this invention also conducted a specific simulation experiment to quantify and compare the recommendation accuracy before and after profile optimization. This experiment generated 1000 simulated driver users with different real preferences, such as time-priority, price-sensitive, and balanced preferences, and randomly divided them into two groups: a control group with a fixed initial profile and an experimental group following the weight update procedure described in Example 5. During the simulation period, the system pushed 20 binary timing opportunities to all users and recorded the number of accurate recommendations that matched the system's preferred recommendation with the user's actual choice. The experimental results showed that the average recommendation accuracy of the control group remained stable at 51.2% throughout the period, while the accuracy of the experimental group increased from a similar initial level to 87.9% as the user profile was continuously optimized. This significant improvement of over 35 percentage points verifies that the personalized profile dynamic iteration mechanism adopted in this invention has significant and quantifiable technical effects in improving user experience and service matching efficiency.

[0047] Example 6: Before connecting a new service area containing multiple commercial facilities to the IoT recommendation system of this invention, to ensure that the system has a unified standard and reliable operation in evaluating the service status and counting resources of all merchants in the service area, a set of configuration procedures for the front-end business logic and fault tolerance mechanism needs to be executed. This procedure first quantifies the encounter quality level determination logic to map the continuously changing expected encounter traffic density into discrete service status levels. The core of the determination logic is to calculate the expected encounter traffic density. Service capabilities of merchant broadcasting ratio And based on this ratio, a three-level state division is performed: when When the value is less than 0.8, the system marks the encounter quality level of the corresponding expected event as smooth; when... When it is in the range of 0.8 to 1.2, it is marked as saturated; when If the value is greater than 1.2, it will be marked as a congestion warning.

[0048] Furthermore, to address the issue of inaccurate service capacity counts due to missing encounter confirmation signals caused by user cancellations or network communication delays, this procedure is configured with a fault-tolerant logic based on state and clock. When an anticipated encounter event is confirmed by a user through a booking operation, the system deducts one unit from the merchant's corresponding service capacity count and marks that unit as a time-sensitive reservation. The duration of this time-sensitive period is set based on the estimated arrival time calculated from the user's vehicle's future trajectory line, plus a fixed grace period. If the system successfully receives the final encounter confirmation signal uploaded by the merchant and associated with the booking before the end of the time-sensitive period, the reservation status is cleared. Conversely, if no final confirmation signal is received by the end of the period, the system determines that the booking is invalid and automatically performs a rollback operation, restoring the merchant's service capacity count value to one unit. After completing the above configuration procedure, the server logic of the newly accessed service area has a unified congestion level judgment standard and a resource counting mechanism with state rollback capability, providing a standardized data processing foundation for subsequent online operation.

[0049] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0050] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for recommending car-related products based on the Internet of Things, characterized in that, The method includes the following steps: Step a, obtain the future trajectory line generated by the vehicle based on its navigation destination. The future trajectory line is a data structure that contains a series of future spatiotemporal points and the time arrival probability distribution of the corresponding future spatiotemporal points. Step b: Obtain supply information published by the merchant, including service content and effective time and space window; Step c: Calculate the expected probability of the vehicle's future trajectory line encountering the merchant's supply information in time and space, so as to generate at least one expected encounter event. Step d: Based on the probability of the expected event, push guidance information to the vehicle or merchant; the method also includes a dynamic robustness maintenance mechanism, which performs the following operations: continuously compare the real-time position of the vehicle with the future trajectory line, and determine that a trajectory deviation event has occurred when the lateral offset distance between the real-time position of the vehicle and the future trajectory line exceeds the baseline threshold. In response to a trajectory deviation event, all calculation results of expected encounter events based on the future trajectory line are discarded; and an implicit trajectory inference mode is activated. The implicit trajectory inference mode is based on the historical gravity points determined by cluster analysis of the vehicle's local historical trajectory data, and selects the historical gravity points with the highest weights in combination with the current time and date context to generate a probability vector field that represents the probability of the vehicle's short-term driving direction, so as to maintain the product recommendation service using the probability vector field.

2. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The dynamic robustness maintenance mechanism also performs the following operations: after activating the implicit trajectory extrapolation mode, it continuously monitors whether there are re-anchoring events where the vehicle trajectory is re-anchored to the new or original navigation route, and in response to the re-anchoring event, it automatically resumes the execution of steps a to d.

3. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, Following step d, the method further includes: aggregating future trajectory lines of multiple vehicles pointing to the same merchant and with overlapping time windows to calculate an expected encounter traffic density characterizing future passenger flow density; comparing the expected encounter traffic density with the merchant's service capacity to determine an encounter quality level; and adjusting the content of the guidance information pushed to the vehicles in step d based on the encounter quality level.

4. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The time arrival probability distribution in step a is dynamically calculated and updated based on the vehicle's current speed and real-time road conditions using a normal distribution model.

5. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The calculation of the expected probability in step c involves calculating the probability that each discrete spatiotemporal point on the future trajectory falls within the merchant's effective spatiotemporal window, and then summing the probabilities of each point by weight.

6. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The guidance information pushed to the vehicle in step d includes the expected probability of encountering the vehicle and the opportunity reminder for booking or navigation detour suggestions.

7. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The guidance information pushed to merchants in step d is a heat map of future passenger flow density generated based on the expected encounter probability of multiple vehicles.

8. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The method also includes a dynamic modulation mechanism for supply-side service capacity, which performs the following operations: when an expected event is confirmed as a real transaction, it receives an event confirmation signal associated with the real transaction; based on the event confirmation signal, it deducts the service capacity count value in the supply information corresponding to the merchant in real time. Furthermore, based on the service capacity count value after deduction, the subsequent expected probability is calculated.

9. The method for recommending vehicle-related products based on the Internet of Things according to claim 1, characterized in that, The guidance information pushed to the vehicle in step d is a binary timing opportunity containing at least two alternative encounter events that differ in their core value propositions; the method also includes: receiving behavioral feedback signals generated by the user's selection behavior of the binary timing opportunity; constructing a personalized profile for the user based on the accumulated behavioral feedback signals, including time sensitivity and price sensitivity weight factors; and using the weight factors in the personalized profile to sort or filter the subsequently generated expected encounter events.

10. A method for recommending vehicle-related products based on the Internet of Things according to claim 3, characterized in that, The calculation of expected traffic density follows these rules: ,in, The expected traffic density represents the number of vehicles expected to be encountered per unit time. In the time window The future trajectory lines all point to the total number of vehicles belonging to the merchant. For the first The expected probability of a vehicle encountering a business.

Images