Intention trajectory collision prediction and multi-source anti-counterfeiting vehicle collaborative verification method and system

Abstract

The application discloses a kind of vehicle collaborative verification method and system of intention trajectory collision prediction and multi-source anti-fake, it is related to automatic driving and vehicle-road collaborative technology field.The application dynamically constructs virtual conflict space-time domain containing intention trajectory using vehicle-mounted edge computing unit, virtually calculates collision time in real time to determine physical causality and assess avoidance feasibility.System executes two-way collaborative operation after generating consent to let go instruction, and the physical conflict resolution is confirmed locally by edge computing unit.Subsequently, cloud server introduces roadside unit dynamic signature or radar feature point cloud hash value based on space-time synchronous snapshot for anti-fake evidence, and after completing traffic entropy verification and asynchronous settlement, collaborative completion notification instruction is issued to the interaction parties.The application realizes closed-loop control of end-side calculation, cloud verification and two-way notification, effectively filters false brushing behavior, and ensures the causality authenticity and system credibility of right-of-way allocation transaction.

Description

Technical Field

[0001] This invention relates to the fields of intelligent transportation systems (ITS), vehicle-to-everything (V2X) security, and driving behavior analysis, specifically to a vehicle collaborative verification method and system for intention trajectory collision prediction and multi-source anti-counterfeiting. Background Technology

[0002] In current vehicle-to-everything (V2X) traffic yielding scenarios, the existing judgment logic has serious flaws: First, the causal relationship is reversed. The simple "rear vehicle request + in front vehicle slows down" judgment model cannot distinguish whether the in front vehicle's slowing down is due to "looking at a phone / obstacle ahead" or is a genuine proactive yielding to the following vehicle. Second, anti-spoofing verification methods are weak. Malicious users can easily use software simulators to forge the GPS tracks of two vehicles to fraudulently obtain right-of-way points. Third, the system generally lacks a closed-loop two-way collaborative notification mechanism and a mechanism to suppress the impact of macro-traffic flow (such as phantom traffic jams). Summary of the Invention

[0003] 1. The technical problem to be solved by this invention: Existing technologies suffer from several technical defects in right-of-way yielding determination, including reversed causality (inability to distinguish between proactive yielding and passive deceleration), one-way interaction protocol (lack of collaborative confirmation and two-way feedback), and susceptibility to interference from analog signals in highly dynamic environments.

[0004] 2. Technical Solution: A vehicle collaborative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting is applied to a vehicle yielding behavior verification system including a cloud server, distributed storage nodes or settlement center, roadside unit, and vehicle terminal. The method includes: Step S1: Motion Vector Acquisition and Intended Trajectory Prediction: In response to the right-of-way interaction request initiated by the vehicle terminal of the first vehicle, the vehicle terminal of the second vehicle acquires its own driving status data in real time through on-board sensors, and receives the driving status data of the first vehicle and the preset target overtaking trajectory vector through the V2X communication link; the vehicle terminal of the second vehicle predicts the virtual motion trajectory of the first vehicle traveling according to the target overtaking trajectory vector based on the vehicle dynamics model, and predicts the inertial motion trajectory of the second vehicle maintaining its current state. Step S2: Virtual Conflict Construction and Cooperative Avoidance Decision: The vehicle terminal of the second vehicle uses the on-board edge computing unit to calculate the intersection point of the virtual motion trajectory of the first vehicle and the inertial motion trajectory of the second vehicle, and dynamically calculates the virtual estimated collision time based on the intersection point; it determines whether the virtual estimated collision time is less than a preset safety threshold; if it is less than the preset safety threshold, it determines that there is a real physical right-of-way conflict; after confirming the existence of a conflict, the vehicle terminal of the second vehicle combines the front and side environmental data obtained by its own sensors to assess whether there are safe avoidance conditions; if there are safe avoidance conditions and the second vehicle reaches a yielding decision, it generates a yielding instruction containing a preset yielding strategy, and feeds the yielding instruction back to the first vehicle through the V2X communication link, wherein the preset yielding strategy includes a deceleration yielding feature sequence or a lane change avoidance feature vector; if there are no safe avoidance conditions or the virtual estimated collision time is greater than or equal to the preset safety threshold, it executes silent interception or denial of the request; Step S3: Two-way collaborative execution and local resolution calculation: In response to receiving the yielding instruction, the vehicle terminal of the first vehicle executes the corresponding acceleration overtaking or lane change maneuver according to the preset yielding strategy; simultaneously, during the yielding behavior monitoring task activation period, the vehicle terminal of the second vehicle continuously receives high-frequency real-time driving status data of the first vehicle through the V2X communication link, tracks its own acceleration decay rate and lateral displacement envelope in real time, and continuously couples with the real-time driving status of the first vehicle to recalculate the real-time estimated collision time; if the vehicle terminal of the second vehicle calculates and determines that the real-time estimated collision time rises above the preset safety threshold, and receives the overtaking completion signal from the vehicle terminal of the first vehicle, then the local determination that the physical conflict resolution is effective is made, and the interaction log containing trajectory changes and feature data is reported to the cloud server; Step S4: Cloud-based multi-source verification and two-way closed-loop notification: The cloud server receives the interaction log and obtains the real-time dynamic signature broadcast by the nearby roadside unit; the cloud server verifies the consistency between the interaction log and the real-time dynamic signature, and after successful verification, packages the two to generate a courtesy verification certificate, and locks the courtesy verification certificate through an asymmetric digital signature; the cloud server triggers the distributed storage node or settlement center to perform asynchronous settlement of points assets, and synchronously generates a collaborative completion notification instruction containing the settlement result; the cloud server sends the collaborative completion notification instruction to the vehicle terminal of the first vehicle and the vehicle terminal of the second vehicle respectively, so as to trigger the corresponding vehicle terminal to execute the successful feedback of the internal human-machine interaction interface and the output of the external visual interaction status signal.

[0005] Preferably, the preset safety threshold is dynamically corrected through the following algorithm mechanism: the edge computing unit of the second vehicle obtains in real time the road surface adhesion coefficient of the current road segment, the braking response delay time of the self-driving chassis system, and the dynamic obstacle spatial complexity factor of the forward vision; based on the road surface adhesion coefficient and the braking response delay time, the basic braking safety time under physical limit state is calculated; the basic braking safety time and the spatial complexity factor are nonlinearly weighted and fused to dynamically output the real-time updated safety threshold.

[0006] Preferably, the step of assessing whether safe avoidance conditions are met by combining the frontal and lateral environmental data acquired by its own sensors includes: the edge computing unit of the second vehicle constructing a local spatiotemporal occupancy grid matrix with the vehicle's center of gravity as the origin; mapping the frontal and lateral environmental data to the local spatiotemporal occupancy grid matrix, extracting the relative velocity vector and static boundary coordinates of potential obstacles; calculating the limit longitudinal acceleration attenuation or limit lateral yaw rate necessary for deceleration or lane change avoidance according to a preset yielding strategy; if the calculated limit longitudinal acceleration attenuation or limit lateral yaw rate exceeds the physical execution boundary of the second vehicle's chassis dynamics or a preset occupant comfort envelope, then it is determined that safe avoidance conditions are not met.

[0007] Preferably, the process of generating the courtesy verification credential in step S4 further includes: If it is determined that there is no signal coverage from the roadside unit in the current road segment, the vehicle terminal of the second vehicle automatically switches to radar feature notarization mode. When the second vehicle generates the consent to yield instruction, it embeds a synchronization timestamp based on a precise time protocol and sends it to the first vehicle. When the physical time reaches the synchronization timestamp, the vehicle terminals of the first vehicle and the second vehicle synchronously lock the radar reflection feature cloud of a specific static landmark scanned by the vehicle-mounted sensors in the current frame, and after performing vehicle motion compensation, extract the hash value of the radar reflection feature cloud based on the spatial distribution feature matrix. The vehicle terminal of the second vehicle reports the hash value to the cloud server along with the interaction log; The cloud server embeds the hash value into the interaction log to generate the courtesy verification credential, and the hash value is used by the cloud server to subsequently verify the authenticity of the vehicle's physical location.

[0008] Preferably, the method further includes a local traffic entropy verification algorithm mechanism, including: The cloud server obtains the rate of change of the average speed of traffic flow and the standard deviation of vehicle distance in a local road segment before and after the occurrence of the yielding behavior. If the cloud server determines that the yielding behavior causes the average speed of traffic flow in the local road segment to decrease and exceed the preset congestion threshold, then it determines that the yielding behavior causes a local traffic entropy increase, and determines that the yielding behavior is invalid or deducts the evaluation weight of the second vehicle. If the average speed of traffic flow within the local road segment remains stable or increases and the standard deviation of vehicle distance decreases, the cloud server generates a full-amount points asset settlement voucher and triggers the distributed storage node or settlement center to execute the settlement.

[0009] This invention also proposes a vehicle cooperative verification system for realizing a vehicle cooperative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting, characterized by comprising: Conflict Calculation Module: An edge computing unit deployed on the vehicle terminal, used to predict the intended trajectory in real time and calculate the risk of multi-vehicle trajectory intersection and the virtual estimated collision time value; Causal determination engine: An edge computing unit deployed on the vehicle terminal, used to evaluate avoidance conditions and execute collaborative avoidance decisions and two-way handshake protocols based on the virtual predicted collision time threshold; V2X Notary Interface: Used to communicate with roadside units and cloud servers to obtain and verify unforgeable spatiotemporal existence proofs; Utility assessment unit: Deployed on a cloud server, it is used to calculate the traffic entropy impact factor of micro-level collaborative driving behavior on macro-level traffic flow and trigger two-way closed-loop notification instructions.

[0010] 3. Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: This invention achieves causal attribute locking of physical driving behavior by establishing a closed-loop architecture of edge computing, cloud verification, and two-way notification. It completely eliminates the vulnerability of false positives in passive deceleration by utilizing virtual TTC based on intent trajectory prediction; it achieves absolute anti-spoofing in highly dynamic environments with zero-latency intervention through spatiotemporal synchronous snapshots and radar feature hash mapping; and it ensures the credibility of the right-of-way allocation system while improving overall traffic efficiency through a traffic entropy verification mechanism. Attached Figure Description

[0011] Figure 1 This is a flowchart of the first embodiment of the vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting proposed in this invention. Figure 2 This is a schematic diagram of the vehicle collaborative verification system proposed in this invention; Figure 3 This is a flowchart of the core algorithm for notarization of virtual TTC closed loop and spatiotemporal synchronization features proposed in this invention. Detailed Implementation

[0012] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0013] As attached Figure 1 As shown, in the first embodiment of the vehicle collaborative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting proposed in this invention, this embodiment is applied to a vehicle yielding behavior verification system including a cloud server, distributed storage nodes or settlement centers, roadside units, and vehicle terminals. This embodiment includes the following steps: Step S1, Motion Vector Acquisition and Intended Trajectory Prediction: In response to the right-of-way interaction request initiated by the vehicle terminal of the first vehicle, the vehicle terminal of the second vehicle acquires its own driving status data in real time through on-board sensors, and receives the driving status data of the first vehicle and the preset target overtaking trajectory vector through the V2X communication link; the vehicle terminal of the second vehicle predicts the virtual motion trajectory of the first vehicle traveling according to the target overtaking trajectory vector based on the vehicle dynamics model, and predicts the inertial motion trajectory of the second vehicle maintaining its current state.

[0014] It should be noted that the vehicle dynamics model can adopt the industry-standard single-track vehicle kinematic model or a multi-degree-of-freedom dynamics model that includes tire lateral slip characteristics. The edge computing unit can adaptively select the appropriate model based on the underlying computing resources of the current onboard chip and the real-time vehicle speed. This invention does not impose strict restrictions on this basic model, but focuses on ensuring low latency and high reliability of subsequent intention trajectory prediction and projection, and minimizing the limitation of protection scope caused by hard-coded single models.

[0015] Step S2, Virtual Conflict Construction and Cooperative Avoidance Decision: The vehicle terminal of the second vehicle uses the on-board edge computing unit to calculate the intersection point of the virtual motion trajectory of the first vehicle and the inertial motion trajectory of the second vehicle, and dynamically calculates the virtual estimated collision time based on the intersection point; it determines whether the virtual estimated collision time is less than a preset safety threshold; if it is less than the preset safety threshold, it determines that there is a real physical right-of-way conflict; after confirming the existence of a conflict, the vehicle terminal of the second vehicle combines the front and side environmental data obtained by its own sensors to assess whether there are safe avoidance conditions; if there are safe avoidance conditions and the second vehicle reaches a yielding decision, it generates a yielding instruction containing a preset yielding strategy, and feeds the yielding instruction back to the first vehicle through the V2X communication link, wherein the preset yielding strategy includes a deceleration yielding feature sequence or a lane change avoidance feature vector; if there are no safe avoidance conditions or the virtual estimated collision time is greater than or equal to the preset safety threshold, it executes silent interception or denial of the request.

[0016] In the algorithm execution of step S2, the preset safety threshold is not a fixed constant. The edge computing unit of the second vehicle extracts the road surface adhesion coefficient of the current road segment and the braking response delay time of the autonomous vehicle's drive-by-wire chassis system (such as brake air pressure / hydraulic pressure build-up delay) at high frequency, and calculates the basic braking safety time of the chassis under physical limit conditions. Subsequently, the spatial complexity factor of dynamic obstacles within the field of view, quantified by the forward-facing camera and millimeter-wave radar, is introduced to nonlinearly weight and adjust the basic braking safety time, outputting a real-time adaptive safety threshold. This establishes a multi-dimensional adaptive threshold judgment defense line in a highly dynamic physical environment, comprehensively intercepting any cheating driving attempts to bypass the safety causal judgment by deceiving static parameters.

[0017] Furthermore, when assessing whether safe avoidance conditions are met, the edge computing unit constructs a local spatiotemporal occupancy grid matrix with the vehicle's center of gravity as the origin at the bottom layer, and projects and maps the frontal and lateral environmental data into the matrix in real time to calculate the relative velocity vector of potential obstacles. If the limit longitudinal acceleration attenuation or limit lateral yaw rate calculated according to the strategy, which is necessary to perform deceleration, yielding, or lane changing to avoid obstacles, exceeds the physical execution boundary of the second vehicle drive-by-wire chassis dynamics (such as being limited by the current tire lateral force limit), or breaks the preset occupant comfort envelope, the algorithm bottom layer will directly circuit-break and determine that "safe avoidance conditions are not met," and then execute the silent interception or rejection request described in step S2 to ensure that right-of-way allocation is deeply coupled with the vehicle chassis's ultimate safety limits.

[0018] Step S3, Two-way collaborative execution and local resolution calculation: In response to receiving the yielding instruction, the vehicle terminal of the first vehicle executes the corresponding acceleration overtaking or lane change maneuver according to the preset yielding strategy; simultaneously, during the yielding behavior monitoring task activation period, the vehicle terminal of the second vehicle continuously receives high-frequency real-time driving status data of the first vehicle through the V2X communication link, tracks its own acceleration decay rate and lateral displacement envelope in real time, and continuously couples with the real-time driving status of the first vehicle to recalculate the real-time estimated collision time; if the vehicle terminal of the second vehicle calculates and determines that the real-time estimated collision time rises above the preset safety threshold, and receives the overtaking completion signaling feedback from the vehicle terminal of the first vehicle, then the local determination that the physical conflict resolution is effective is made, and the interaction log containing trajectory changes and feature data is reported to the cloud server.

[0019] Step S4, Cloud-based Multi-Source Verification and Two-Way Closed-Loop Notification: The cloud server receives the interaction log and obtains the real-time dynamic signature broadcast by the nearby roadside unit; the cloud server verifies the consistency between the interaction log and the real-time dynamic signature, and after successful verification, packages the two to generate a courtesy verification credential, and locks the courtesy verification credential through an asymmetric digital signature; the cloud server triggers the distributed storage node or settlement center to perform asynchronous settlement of points assets, and synchronously generates a collaborative completion notification instruction containing the settlement result; the cloud server sends the collaborative completion notification instruction to the vehicle terminal of the first vehicle and the vehicle terminal of the second vehicle respectively, so as to trigger the corresponding vehicle terminal to execute the successful feedback of the internal human-machine interaction interface and the output of the external visual interaction status signal.

[0020] In the second embodiment of the vehicle collaborative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting proposed in this invention, based on the first embodiment, the process of generating the courtesy verification certificate in step S4 further includes the following steps: Step S21: If it is determined that there is no signal coverage of the roadside unit in the current road segment, the vehicle terminal of the second vehicle automatically switches to radar feature notarization mode.

[0021] Step S22: When the second vehicle generates the consent to yield instruction, it embeds a synchronization timestamp based on a precise time protocol and sends it to the first vehicle; when the physical time reaches the synchronization timestamp, the vehicle terminals of the first vehicle and the second vehicle synchronously lock the specific static landmark radar reflection feature point cloud scanned by the vehicle-mounted sensors in the current frame, and after performing vehicle motion compensation, extract the hash value of the radar reflection feature point cloud based on the spatial distribution feature matrix.

[0022] Step S23: The vehicle terminal of the second vehicle reports the hash value to the cloud server along with the interaction log.

[0023] Step S24: The cloud server embeds the hash value into the interaction log to generate the courtesy verification credential. The hash value is used by the cloud server to subsequently verify the authenticity of the vehicle's physical location.

[0024] In the third embodiment of the vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting proposed in this invention, based on the first embodiment, the method further includes a local traffic entropy verification algorithm mechanism, comprising the following steps: Step S31: The cloud server obtains the rate of change of the average speed of traffic flow and the standard deviation of vehicle distance in the local road segment before and after the yielding behavior occurs.

[0025] Step S32: If the cloud server determines that the yielding behavior causes the average speed of the traffic flow in the local road segment to decrease and exceed the preset congestion threshold, then it determines that the yielding behavior causes a local traffic entropy increase, and determines that the yielding behavior is invalid or deducts the evaluation weight of the second vehicle.

[0026] Step S33: If the average speed of traffic flow in the local road segment remains stable or increases and the standard deviation of vehicle distance decreases, the cloud server generates a full-amount points asset settlement voucher and triggers the distributed storage node or settlement center to execute the settlement.

[0027] In the fourth embodiment of the vehicle collaborative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting proposed in this invention, based on the first embodiment, the preset safety threshold is dynamically corrected through the following algorithm mechanism: the edge computing unit of the second vehicle obtains in real time the road surface adhesion coefficient of the current road segment, the braking response delay time of the self-driving chassis system, and the dynamic obstacle spatial complexity factor of the forward vision; based on the road surface adhesion coefficient and the braking response delay time, the basic braking safety time under the physical limit state is calculated; the basic braking safety time and the spatial complexity factor are nonlinearly weighted and fused to dynamically output the real-time updated safety threshold.

[0028] In the fifth embodiment of the vehicle collaborative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting proposed in this invention, based on the first embodiment, the assessment of whether safe avoidance conditions are met by combining the frontal and lateral environmental data acquired by its own sensors specifically includes the following: Step S51: The edge computing unit of the second vehicle constructs a local spatiotemporal occupancy grid matrix with the vehicle's centroid as the origin.

[0029] Step S52: Map the front and side environmental data to the local spatiotemporal occupancy grid matrix to extract the relative velocity vector and static boundary coordinates of potential obstacles.

[0030] Step S53: Based on the preset yielding strategy, calculate the limit longitudinal acceleration attenuation or the limit lateral yaw rate necessary to perform deceleration yielding or lane changing to avoid a collision.

[0031] Step S54: If the calculated limit longitudinal acceleration attenuation or the limit lateral yaw rate exceeds the physical execution boundary of the second vehicle chassis dynamics or the preset occupant comfort envelope, then it is determined that there are no conditions for safe avoidance.

[0032] Example 1: Causality Verification Edge Algorithm Based on Intent-Driven and TTC (Preventing False Positives and Collaborative Closed Loop): Combination Figures 1-3As shown, this embodiment details the edge calculation mechanism for Virtual Predicted Time of Collision (TTC). In a highway ramp merging scenario, the first vehicle (the requesting vehicle) sends a "target overtaking trajectory vector" containing the target merging acceleration and the coordinates of the target's entry point. The onboard edge computing unit of the second vehicle (the yielding vehicle) places this vector and its own inertial trajectory vector into the same spatiotemporal coordinate system.

[0033] Define the intention trajectory vector of the first vehicle as P. 1(t) The inertial trajectory vector of the second vehicle is P. 2(t) Let the relative position vector of the two vehicles at time t0 be D(t0), and the relative velocity vector be V. rel (t0). The basic triggering solution model for virtual TTC is: TTC virtual (t0) = - [ D(t0) · Vrel(t0) ] / ||V rel (t0)||2 It should be noted that the above TTC solution model only applies to D(t)·V rel When D(t) < 0 (i.e., indicating that the first and second vehicles are spatially close to each other), it is activated and considered valid; if D(t) · V rel (t)≥0 (i.e., the two vehicles are relatively stationary or far apart), the edge computing unit will directly determine that TTC tends to infinity. (This eliminates the need to trigger collision warnings and avoidance decisions.)

[0034] After generating the yielding instruction and activating the yielding behavior monitoring task, the system enters the high-frequency closed-loop tracking stage. At this time, the edge computing unit of the second vehicle collects the instantaneous acceleration vector a1(t) of the first vehicle and the instantaneous acceleration vector a2(t) of the second vehicle in real time at a preset high-frequency sampling rate (such as 10Hz-50Hz), and continuously integrates and updates the relative motion state. The real-time relative velocity vector is calculated as follows: V rel (t) = V rel (t0) + The real-time relative displacement is calculated as follows: D(t) = D(t0) + In the formula, Let be the integration time variable (i.e., the time integration pseudo-variable), representing any intermediate instant within the high-frequency closed-loop tracking time period from the initial time t0 to the current time t. Introducing... This is to mathematically distinguish it from the current time t at the upper limit of integration, and to characterize the continuous integral accumulation evolution process of the vehicle's acceleration and velocity over a very short time window.

[0035] The real-time computational model for closed-loop resolution has been updated as follows: TTC realtime (t) = - [ D(t) · V rel (t) ] / ||V rel (t)|| 2 Explanation of the technical essence and physical significance: Static formula TTC virtual (t0) is used to determine at time t0 whether there is an initial physical right-of-way conflict (i.e., spatiotemporal cross-validation of intent shadows); while the dynamic closed-loop formula TTC realtime (t) constitutes the core criterion for the entire monitoring task. The system monitors TTC. realtime The first derivative of (t) with time is greater than zero (indicating that the danger is moving away), and this ultimately confirms the TTC. realtime (t)≥T safe When the (safety threshold) is reached, a rigorous mathematical closed loop is formed at the bottom layer of the algorithm, thereby accurately verifying the true effectiveness of the action.

[0036] Example 2: Spatiotemporal snapshot synchronization and radar feature notarization algorithm in high dynamic environments (anti-cheating): Combination Figure 3 As shown, to address the pain point of open parking lots lacking roadside unit (RSU) coverage and being prone to GPS-simulated score manipulation, the system invokes the vehicle-mounted radar to perform environmental feature verification. Considering that the sampling time difference when the vehicle is traveling at high speed can cause the anti-counterfeiting comparison to crash, this embodiment introduces a V2V-based nanosecond-level spatiotemporal synchronization snapshot mechanism.

[0037] When the second vehicle sends its consent to yield instruction, it embeds a future synchronization timestamp T. sync When T is reached sync At that time, the bottom layer of the edge perception architecture of both vehicles is interrupted, and the static landmark radar reflection feature point cloud of the current frame is locked simultaneously.

[0038] Define the set of local static landmark cloud points at the lock time Tsync as C={p1,p2,...,p n}, where the four-dimensional spatial feature matrix of a single point cloud is p i =[x i ,y i ,z i ,I i ] T (I) i(Reflection intensity). The edge computing unit first calculates the geometric center p of the point cloud set. bar And construct the spatial covariance matrix M cov : System Extraction Matrix M cov The non-zero eigenvalue combination is used to reduce its dimension to a one-dimensional environmental feature vector E, followed by irreversible hashing to generate a multi-source evidence chain: Hash evidence = SHA-256( E || Timestamp || Vehicle ID ) Explanation of the technical essence and physical significance: This mechanism extracts a unique, non-replicable "3D environmental fingerprint" of the road. Directly performing hash calculations on the massive point cloud set C would lead to computational overload and susceptibility to noise interference. This invention utilizes the spatial covariance matrix M... cov The dimensionality of tens of thousands of discrete point clouds is reduced to extract the essential topological feature vector E of their geometric distribution. Even if cheaters can deceive GPS positioning, they cannot fabricate three-dimensional radar reflection data with the same spatial topological feature matrix out of thin air. The cloud server only asynchronously compares the hash digest generated based on this topological feature afterward, ensuring zero-latency intervention for physical risk avoidance and completely eliminating the possibility of pure software simulation for score manipulation.

[0039] Example 3: Macroscopic Traffic Entropy Suppression Algorithm (Solving Ghost Traffic Congestion Negative Externalities): This system incorporates macroscopic dynamic utility assessment. In congested urban areas, if a second vehicle brakes suddenly to gain yield points, it will trigger a chain reaction in the traffic flow behind. The cloud server tracks the "traffic entropy" of local road segments in real time: if it detects a sharp drop in the average speed of the traffic flow behind and a dramatic fluctuation in the standard deviation of vehicle distance (a surge in disorder) shortly after yielding, the cloud-based utility assessment unit determines that this action has generated serious negative externalities. The system then triggers a dynamic penalty factor, reducing the point weight of the transaction or deeming it invalid, thereby guiding drivers to adopt a smooth "zipper-style alternating passage" at the algorithmic level. After verification and settlement (or circuit breaker) are completed, the cloud sends a "collaborative completion notification instruction" to trigger two-way HMI feedback, declaring the agreement complete.

[0040] Example 4: Vehicle Collaborative Verification System Architecture: Combination Figure 2 As shown, the present invention also provides a vehicle collaborative verification system for implementing the collaborative verification methods in embodiments one to three above. This system physically spans both edge and cloud infrastructure, specifically including: Conflict Calculation Module: An edge computing unit deployed on the vehicle terminal, used to predict the intended trajectory in real time and calculate the risk of intersection of multiple vehicle trajectories and the virtual estimated collision time value.

[0041] Causal determination engine: An edge computing unit deployed on the vehicle terminal, used to evaluate avoidance conditions and execute collaborative avoidance decisions and two-way handshake protocols based on the virtual predicted collision time threshold.

[0042] V2X Notary Interface: Used to communicate with roadside units and cloud servers to obtain and verify unforgeable proof of spatiotemporal existence.

[0043] Utility assessment unit: Deployed on a cloud server, it is used to calculate the traffic entropy impact factor of micro-level collaborative driving behavior on macro-level traffic flow and trigger two-way closed-loop notification instructions.

[0044] The causal determination engine and conflict calculation module deployed on the terminals of the first and second vehicles utilize on-board edge computing power to achieve high-frequency V2V handshake and real-time virtual TTC closed-loop tracking, ensuring millisecond-level risk avoidance response and physical causal verification. The V2X notary interface deployed on the vehicle terminal is used to connect with the roadside unit (RSU) to extract the synchronization timestamp and dynamic signature; The utility evaluation unit, deployed on a cloud server, is used to perform traffic entropy reduction calculations and macro-level right-of-way scheduling, and to work with distributed storage nodes or settlement centers to complete hash storage and asynchronous settlement of multi-source anti-counterfeiting evidence chains.

[0045] Algorithm equivalent substitution declaration: Those skilled in the art should understand that the virtual predicted collision time (TTC) calculation model (such as the division model and calculus tracking model based on relative displacement and relative velocity vectors) and the hash function for environmental feature anti-counterfeiting (such as SHA-256) described in this embodiment are merely preferred mathematical expressions of the present invention. Without departing from the core concept of the present invention, using other integral evolution equations, Kalman filter prediction models, or neural network spatiotemporal prediction models to replace TTC calculation, or using other asymmetric encryption algorithms or national cryptographic algorithms to replace hash functions, to achieve the same technical effects of "causal determination based on intent trajectory" and "anti-tampering and rights confirmation based on physical features," will inevitably fall within the scope of protection of the present invention.

[0046] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0047] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting, characterized in that, A vehicle yielding behavior verification system applied to cloud servers, distributed storage nodes or settlement centers, roadside units, and vehicle terminals, the method comprising: Step S1: Motion Vector Acquisition and Intended Trajectory Prediction: In response to the right-of-way interaction request initiated by the vehicle terminal of the first vehicle, the vehicle terminal of the second vehicle acquires its own driving status data in real time through on-board sensors, and receives the driving status data of the first vehicle and the preset target overtaking trajectory vector through the V2X communication link; the vehicle terminal of the second vehicle predicts the virtual motion trajectory of the first vehicle traveling according to the target overtaking trajectory vector based on the vehicle dynamics model, and predicts the inertial motion trajectory of the second vehicle maintaining its current state. Step S2: Virtual Conflict Construction and Cooperative Avoidance Decision: The vehicle terminal of the second vehicle uses the on-board edge computing unit to calculate the intersection point of the virtual motion trajectory of the first vehicle and the inertial motion trajectory of the second vehicle, and dynamically calculates the virtual estimated collision time based on the intersection point; it determines whether the virtual estimated collision time is less than a preset safety threshold; if it is less than the preset safety threshold, it determines that there is a real physical right-of-way conflict; after confirming the existence of a conflict, the vehicle terminal of the second vehicle combines the front and side environmental data obtained by its own sensors to assess whether there are safe avoidance conditions; if there are safe avoidance conditions and the second vehicle reaches a yielding decision, it generates a yielding instruction containing a preset yielding strategy, and feeds the yielding instruction back to the first vehicle through the V2X communication link, wherein the preset yielding strategy includes a deceleration yielding feature sequence or a lane change avoidance feature vector; if there are no safe avoidance conditions or the virtual estimated collision time is greater than or equal to the preset safety threshold, it executes silent interception or denial of the request; Step S3: Two-way collaborative execution and local resolution calculation: In response to receiving the yielding instruction, the vehicle terminal of the first vehicle executes the corresponding acceleration overtaking or lane change maneuver according to the preset yielding strategy; simultaneously, during the yielding behavior monitoring task activation period, the vehicle terminal of the second vehicle continuously receives high-frequency real-time driving status data of the first vehicle through the V2X communication link, tracks its own acceleration decay rate and lateral displacement envelope in real time, and continuously couples with the real-time driving status of the first vehicle to recalculate the real-time estimated collision time; if the vehicle terminal of the second vehicle calculates and determines that the real-time estimated collision time rises above the preset safety threshold, and receives the overtaking completion signal from the vehicle terminal of the first vehicle, then the local determination that the physical conflict resolution is effective is made, and the interaction log containing trajectory changes and feature data is reported to the cloud server; Step S4: Cloud-based multi-source verification and two-way closed-loop notification: The cloud server receives the interaction log and obtains the real-time dynamic signature broadcast by the nearby roadside unit; the cloud server verifies the consistency between the interaction log and the real-time dynamic signature, and after successful verification, packages the two to generate a courtesy verification certificate, and locks the courtesy verification certificate through an asymmetric digital signature; the cloud server triggers the distributed storage node or settlement center to perform asynchronous settlement of points assets, and synchronously generates a collaborative completion notification instruction containing the settlement result; the cloud server sends the collaborative completion notification instruction to the vehicle terminal of the first vehicle and the vehicle terminal of the second vehicle respectively, so as to trigger the corresponding vehicle terminal to execute the successful feedback of the internal human-machine interaction interface and the output of the external visual interaction status signal.

2. The vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting as described in claim 1, characterized in that, The preset safety threshold is dynamically corrected through the following algorithm mechanism: the edge computing unit of the second vehicle obtains the road surface adhesion coefficient of the current road segment, the braking response delay time of the self-driving chassis system, and the dynamic obstacle spatial complexity factor of the forward vision in real time. Based on the road surface adhesion coefficient and the braking response delay time, the basic braking safety time under the physical limit state is calculated. The basic braking safety time and the space complexity factor are nonlinearly weighted and fused to dynamically output the safety threshold that is updated in real time.

3. The vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting as described in claim 1, characterized in that, The assessment of whether safe avoidance conditions are met, based on the frontal and lateral environmental data acquired by the vehicle's own sensors, includes: the edge computing unit of the second vehicle constructing a local spatiotemporal occupancy grid matrix with the vehicle's center of gravity as the origin; mapping the frontal and lateral environmental data to the local spatiotemporal occupancy grid matrix, extracting the relative velocity vector and static boundary coordinates of potential obstacles; calculating the limit longitudinal acceleration attenuation or limit lateral yaw rate necessary for deceleration, yielding, or lane changing based on a preset yielding strategy; if the calculated limit longitudinal acceleration attenuation or limit lateral yaw rate exceeds the physical execution boundary of the second vehicle's chassis dynamics or a preset occupant comfort envelope, then it is determined that safe avoidance conditions are not met.

4. The vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting as described in claim 1, characterized in that, The process of generating the courtesy verification credential in step S4 further includes: If it is determined that there is no signal coverage from the roadside unit in the current road segment, the vehicle terminal of the second vehicle automatically switches to radar feature notarization mode. When the second vehicle generates the consent to yield instruction, it embeds a synchronization timestamp based on a precise time protocol and sends it to the first vehicle. When the physical time reaches the synchronization timestamp, the vehicle terminals of the first vehicle and the second vehicle synchronously lock the radar reflection feature cloud of a specific static landmark scanned by the vehicle-mounted sensors in the current frame, and after performing vehicle motion compensation, extract the hash value of the radar reflection feature cloud based on the spatial distribution feature matrix. The vehicle terminal of the second vehicle reports the hash value to the cloud server along with the interaction log; The cloud server embeds the hash value into the interaction log to generate the courtesy verification credential, and the hash value is used by the cloud server to subsequently verify the authenticity of the vehicle's physical location.

5. The vehicle collaborative verification method for intent trajectory collision prediction and multi-source anti-counterfeiting as described in claim 1, characterized in that, The method also includes a local traffic entropy verification algorithm mechanism, including: The cloud server obtains the rate of change of the average speed of traffic flow and the standard deviation of vehicle distance in a local road segment before and after the occurrence of the yielding behavior. If the cloud server determines that the yielding behavior causes the average speed of traffic flow in the local road segment to decrease and exceed the preset congestion threshold, then it determines that the yielding behavior causes a local traffic entropy increase, and determines that the yielding behavior is invalid or deducts the evaluation weight of the second vehicle. If the average speed of traffic flow within the local road segment remains stable or increases and the standard deviation of vehicle distance decreases, the cloud server generates a full-amount points asset settlement voucher and triggers the distributed storage node or settlement center to execute the settlement.

6. A vehicle collaborative verification system, used to implement the vehicle collaborative verification method for intention trajectory collision prediction and multi-source anti-counterfeiting as described in any one of claims 1-4, characterized in that, include: Conflict Calculation Module: An edge computing unit deployed on the vehicle terminal, used to predict the intended trajectory in real time and calculate the risk of multi-vehicle trajectory intersection and the virtual estimated collision time value; Causal determination engine: An edge computing unit deployed on the vehicle terminal, used to evaluate avoidance conditions and execute collaborative avoidance decisions and two-way handshake protocols based on the virtual predicted collision time threshold; V2X Notary Interface: Used to communicate with roadside units and cloud servers to obtain and verify unforgeable spatiotemporal existence proofs; Utility assessment unit: Deployed on a cloud server, it is used to calculate the traffic entropy impact factor of micro-level collaborative driving behavior on macro-level traffic flow and trigger two-way closed-loop notification instructions.

Images