Method and system for evaluating running risk of automatic driving taxi

Abstract

The invention discloses a method and a system for evaluating the running risk of an automatic driving taxi, and relates to the field of intelligent traffic and automatic driving. Aiming at the problems of insufficient dynamic traffic environment adaptability, rigidity of space decision, lack of vehicle operation rule compatibility and the like in the prior art, the evaluation method comprises the following steps: dynamically dividing an urban road network into hexagonal honeycomb units, adjusting the honeycomb radius according to a real-time travel demand, and selecting key nodes; accessing traffic flow density, conflict event number, communication delay and environment parameters in real time to construct a key node risk index; a hidden Markov chain model is adopted to simulate behavior response of an automatic driving taxi at a key node, an S1 safety state and an S2 risk state are defined, and a Viterbi algorithm is combined to deduce a hidden state probability; and finally, comprehensively judging the risk level according to the risk index and the state probability. According to the method, through multi-source data fusion and scene behavior modeling, dynamic evaluation and control of the operation risk of the automatic driving taxi are realized.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent transportation and autonomous driving, and more specifically, to a method and system for evaluating the operation risks of autonomous taxis. Background Art

[0002] As a core component of the intelligent transportation system, the large-scale deployment of autonomous taxis poses new challenges to the efficiency and safety of urban traffic networks. Existing technologies have significant limitations in dealing with dynamic traffic environments, human-vehicle-road interaction modeling, elastic spatial decision-making, and vehicle operation law compatibility. The specific problems are as follows: 1. Insufficient adaptability to dynamic traffic environments Traditional risk assessment models rely on static road network data and historical statistical laws, and it is difficult to adapt to the time-varying characteristics of urban traffic flow. For example, traffic flow fluctuations during peak hours, sudden accidents, or sudden changes in road network states caused by extreme weather may all lead to the disconnection between the deployment strategies of autonomous taxis and actual traffic demands. Existing methods lack real-time perception and dynamic response capabilities, and cannot accurately identify the risk evolution laws in traffic bottleneck areas (such as around schools and lane-changing nodes in commercial areas), resulting in the lack of scientific basis for setting vehicle access thresholds and path planning schemes, and easily causing a decline in regional traffic efficiency or the accumulation of systemic safety risks.

[0003] 2. Limitations in human-vehicle-road interaction scenario modeling Existing technologies lack the ability to model the interaction behaviors between autonomous taxis and heterogeneous traffic participants (pedestrians, non-motor vehicles, and human-driven vehicles) in a fine-grained manner. For example, at high-density intersections or sections without traffic lights, the perception delay and decision-making uncertainty of autonomous vehicles may trigger chain reactions (such as continuous deceleration of rear vehicles due to braking delay), but existing models cannot quantify the impact of such behaviors on the stability of the overall traffic flow. In addition, the risk boundaries of vehicles with different technical configurations (such as L3 / L4-level systems) vary significantly in complex scenarios, but traditional methods do not establish the mapping relationship between scenario features and technical capability parameters, resulting in strong subjectivity and poor universality of risk assessment results.

[0004] 3. Contradiction between rigid spatial decision-making and dynamic demands Regional deployment strategies (such as restricting commercial areas and residential areas) are usually based on static road network characteristics (such as road grades and functional zones), and it is difficult to support rapid risk assessment under dynamic spatial range adjustments. Existing models lack spatial adaptability and cannot achieve dynamic segmentation of the road network and multi-scale traffic pattern recognition, resulting in the inability of vehicle access rules (such as time-limited flow control and area bans) to flexibly respond to real-time traffic states, reducing the effectiveness of policy implementation and the utilization efficiency of road network resources.

[0005] 4. Lack of compatibility with vehicle operation laws The operating characteristics of autonomous taxis (such as high-frequency services, passenger boarding and alighting behaviors, and energy replenishment needs) are fundamentally different from those of traditional traffic participants. However, existing evaluation methods do not consider the long-term impact of their unique operating rules on the traffic system. For example, shared mobility pick-up and drop-off points may cause congestion on surrounding roads due to vehicle queue overflow, and unreasonable charging station layouts may lead to unbalanced road network loads. Traditional models only focus on short-term traffic flow disturbances and ignore the correlation mechanism between vehicle scheduling strategies, energy replenishment behaviors, and traffic risk evolution, making it difficult to support full-life-cycle risk control.

[0006] The root cause of the above problems lies in the contradiction between the "static evaluation paradigm" of existing technologies and the "dynamic complexity" of urban traffic systems. Specifically: Difficulty in data fusion and real-time processing: The heterogeneity and spatio-temporal discreteness of multi-source traffic data (such as floating car trajectories, sensor data, and weather information) increase the complexity of dynamic modeling, and traditional algorithms are difficult to achieve real-time risk assessment with limited computing resources. Nonlinear characteristics of interaction behaviors: The interaction behaviors between autonomous vehicles and other traffic participants have strong coupling and uncertainty. Traditional microscopic simulation models are difficult to be widely applied in engineering practice due to difficult parameter calibration or high computational complexity. Technical bottleneck in dynamic spatial scope division: Existing road network segmentation methods (such as fixed grid division) cannot adapt to the spatio-temporal evolution characteristics of traffic flow, resulting in the disconnection between risk assessment results and actual scenarios. Technical gap in quantifying long-term impacts: The spatio-temporal distribution laws of the full-cycle behaviors of vehicle operations (such as charging and dwelling) are difficult to accurately predict through short-term data observations or simplified models and rely on large-scale historical data and complex simulation technologies.

[0007] In summary, existing technologies have systematic defects in aspects such as dynamic adaptability, scenario modeling accuracy, spatial decision-making flexibility, and long-term impact quantification. There is an urgent need for a new method that can integrate multi-source data, dynamically evaluate risks, and support flexible decision-making to support the safe and efficient deployment of autonomous taxis. Summary of the Invention

[0008] One objective of the present invention is to address the insufficient adaptability to dynamic traffic environments: Existing methods cannot perceive traffic flow changes in real time, resulting in the disconnection between the deployment strategies of autonomous taxis and actual demands, leading to a decline in regional traffic efficiency or an accumulation of safety risks.

[0009] Solve the optimization of spatial division and path planning: Traditional road network division methods cannot dynamically adjust the honeycomb boundaries, and the path planning does not cover high-risk nodes, affecting the accuracy of risk assessment and the efficiency of vehicle scheduling.

[0010] Solve the optimization of Hidden Markov Chain model parameters: The parameter settings of the state transition probability matrix and the emission probability matrix lack data support, resulting in insufficient accuracy of risk assessment.

[0011] Solving multi-source data fusion and real-time processing: The types of traffic data are single and the update frequency is low, which cannot comprehensively reflect complex traffic scenarios and affects the reliability of risk index calculation and model input.

[0012] Solving dynamic routing and path replanning: Existing path planning does not consider real-time traffic conditions and long-term operation rules, resulting in insufficient balance between vehicle operation efficiency and safety risks.

[0013] Solving closed-loop verification and parameter iteration: There is a lack of a dynamic verification mechanism for path strategies, and the risk assessment model and scheduling strategy cannot be continuously optimized.

[0014] Solving Monte Carlo simulation and facility layout optimization: Traditional methods do not quantify the long-term impact of passenger boarding / alighting and energy replenishment behaviors on traffic flow, resulting in unreasonable layouts of charging stations and transfer points.

[0015] Solving digital twin environment construction and perturbation analysis: Existing simulation environments cannot truly reproduce traffic flow perturbations, affecting the accuracy of path strategy verification.

[0016] Solving human-machine collaborative safety response: The correlation between the safety officer response mechanism and traffic flow density is insufficient, resulting in a lag in risk control measures.

[0017] Solving critical node risk grading and dynamic monitoring: Existing methods do not achieve refined classification and differentiated monitoring of critical nodes, reducing the pertinence of risk assessment.

[0018] To this end, the present invention provides a method for evaluating the operation risk of an autonomous taxi, including: Dynamically dividing the urban road network into hexagonal honeycomb cells, with the dynamic radius R of each honeycomb cell Calculated according to the formula where R base is the reference radius, Q t is the travel demand density within the honeycomb at the current time period, Q avg is the historical average demand density, Q max and Q min are the extreme values of the demand density, i.e., the maximum and minimum values, α is the adjustment coefficient, and key nodes are selected in each hexagonal honeycomb cell; Real-time accessing multi-source traffic data, including traffic flow density, number of conflict events, communication delay data, and environmental parameters, to construct the critical node risk index NRI j and calculated through the following formula: NRI j = ω 1 × C j + ω 2 × D j + ω 3 × A j j ​j , where C j is the conflict rate, D j is the delay index, A j is the accident density, ω 1 , ω 2 , ω 3 are preset weights; Simulate the behavioral responses of autonomous taxis at key nodes through the hidden Markov chain model, including: defining the hidden states as S1 safe state and S2 risk state; collecting the confidence of taxi target detection, the obstacle avoidance response time of the taxi, and the V2V communication delay data of the taxi, and discretizing them into observation states; Among them, according to the hidden states, construct the state transition probability matrix P; according to the observation states, construct the emission probability matrix E; use the Viterbi algorithm to infer the probabilities of each hidden state in the S1 safe state and S2 risk state at the current moment in combination with the state transition probability matrix P and the emission probability matrix E; According to the key node risk index NRI j value and the inferred hidden state probabilities, comprehensively determine the risk level, output low, medium, and high risk warnings, and dynamically adjust the path planning strategy according to the risk level.

[0019] Preferably, in the evaluation method of the operation risk of the autonomous taxi, it includes: When dynamically dividing the urban road network into hexagonal honeycomb cells, use the Voronoi diagram algorithm to dynamically adjust the honeycomb boundary to ensure that adjacent honeycombs do not overlap and cover the entire area; The reference radius R base has a value range of 1 - 3 kilometers, and the adjustment coefficient α is determined by fitting historical traffic data to be 0.5 - 1.2; the weight coefficients ω j of the key node risk index NRI 1 , ω 2 , ω 3 correspond to the conflict rate C j , the delay index D j , and the accident density A j respectively, and their value ranges are: ω 1 is 0.4 - 0.6, ω 2 is 0.3 - 0.5, ω 3 is 0.1 - 0.3; Among them, R base is determined by fitting according to the urban road network density, and α is obtained through historical traffic flow regression analysis.

[0020] Preferably, in the evaluation method of the operation risk of the autonomous taxi, it includes: In the transition probability matrix P of the hidden states S1 (safe state) and S2 (risk state) of the hidden Markov chain model, the probability of S1 transitioning to S1 is 0.85 - 0.95, the probability of S1 transitioning to S2 is 0.05 - 0.15, the probability of S2 transitioning to S1 is 0.40 - 0.60, and the probability of S2 transitioning to S2 is 0.40 - 0.60; The emission probability matrix E is constructed based on the test data of multi-brand vehicles in a closed test field. Among them, when the confidence level of taxi target detection ≥ 0.9, the emission probability is 0.90; when the obstacle avoidance response time of the taxi ≤ 0.5 seconds, the emission probability is 0.85; when the V2V communication delay of the taxi ≤ 50 milliseconds, the emission probability is 0.92.

[0021] Preferably, in the method for evaluating the operation risk of the autonomous taxi, the real-time access to multi-source traffic data further includes: Average vehicle speed, proportion of autonomous vehicles, number of emergency braking events, collaborative lane change failure rate, visibility, road surface adhesion coefficient, and signal light phase; And update the multi-source traffic data at a set frequency. The multi-source traffic data is used to calculate the key node risk index NRI j , and combined with the taxi target detection confidence level, taxi obstacle avoidance response time, and taxi V2V communication delay data, they are discretized together to obtain the observation state of the hidden Markov chain model.

[0022] Preferably, in the method for evaluating the operation risk of the autonomous taxi, it further includes: For the test autonomous taxi, based on the dynamic radius R of the cellular unit t Determine the driving path planning range; Within the planning range, generate an initial path set that satisfies the single-trip mileage being 0.5 - 4 times the cellular radius and covers all key nodes with NRI j ≥ 0.4; With the goal of minimizing the comprehensive path risk value, that is, the weighted sum of efficiency risk and safety risk, and maximizing the service response speed, construct an optimized driving model, In the optimized driving model, dynamically adjust the cruising path of idle vehicles according to the traffic flow density, and preferentially park idle vehicles in the surrounding area of high-risk nodes with NRI j ≥ 0.7, and predict the impact of the behavior of passengers getting on and off the vehicle on the traffic flow density within the next 30 minutes based on Monte Carlo simulation; Execute the safety response mechanism. When it is detected that the traffic flow density within the cellular reaches 80 vehicles / km, automatically extend the abnormal event handling duration to 1.2 times the original handling duration, and at the same time trigger the path replanning instruction.

[0023] Preferably, in the method for evaluating the operation risk of the autonomous taxi, it further includes: Establish a closed-loop verification system, which simulates the disturbance effect of path adjustment on the actual traffic flow in the digital twin environment, updates the parameters of the risk assessment model according to the simulation results, and applies the parameters to the cell radius R in the dynamic path planning framework in real time t Calculation; The dynamic path planning framework and the real-time scheduling strategy are coordinated through the travel demand density Q within the cell at the current time t Data to achieve coordination. The output data of the Monte Carlo simulation acts on both the optimization of the charging station location and the adjustment of the connection point layout at the same time. The execution result of the path replanning instruction is fed back to the closed-loop verification system in real time to form an iterative optimization loop.

[0024] Preferably, in the method for evaluating the operation risk of the autonomous taxi, the method realizes the collaborative optimization of the Monte Carlo simulation and the dynamic routing strategy through the following steps: Input parameters: Passenger boarding and alighting behavior data, with a Poisson distribution mean of 3 minutes and a standard deviation of 1 minute; Energy replenishment behavior data, with a charging time of 0.5 - 2 hours; Set the queuing overflow probability matrix of the connection points to quantify the vehicle retention risk at each connection point during different time periods; Set the road network load imbalance index around the charging station to reflect the traffic flow density fluctuation within the service radius of the charging station; According to the queuing overflow probability matrix, dynamically adjust the location and capacity of the connection points, and relocate the high-risk connection points to the low-load areas; Based on the load imbalance index, optimize the location of the charging station and preferentially deploy it in areas where the road network load balance degree ≥ 0.8; Among them, when the traffic flow density within the cell ≥ 80 vehicles / km, automatically start path replanning; Prioritize selecting sections that have not reached the load threshold, that is, the traffic flow density < 60 vehicles / km; Dynamically adjust the path weight coefficient, increase the efficiency risk weight by 30%, and synchronize the path replanning result to the digital twin environment.

[0025] Preferably, in the method for evaluating the operation risk of the autonomous taxi, the established closed-loop verification system is specifically as follows: The closed-loop verification system is constructed based on digital twin technology, and this system realizes data interaction with the traffic data collection module that collects multi-source traffic data in real time, the dynamic path planning framework, and the risk assessment model; The traffic data collection module transmits the collected data to the closed-loop verification system; in the digital twin environment, a virtual scene consistent with the actual traffic scenario is constructed based on real-time traffic data, and the scene covers road networks, traffic lights, other traffic participants including pedestrians, and ordinary vehicle elements; At the same time, the initial path set and the adjusted path plan generated by the dynamic path planning framework are input into the digital twin environment; According to the adjusted path plan, simulate the driving process of the autonomous taxi in the virtual scene, and analyze the disturbance impact of the path adjustment on the surrounding traffic flow. The specific quantification indicators include: Local traffic flow change: Calculate the change rate of vehicle flow density and the change value of average vehicle speed on the surrounding road sections before and after the path adjustment; Evaluate the impact of path adjustment on the average travel time and vehicle delay time within the entire cell; After simulating the path adjustment, evaluate the probability and frequency change of conflicts between the autonomous taxi and other traffic participants through the risk assessment model.

[0026] Preferably, in the method for evaluating the operation risk of the autonomous taxi, the hidden state further includes a human-machine collaborative safety strategy, including the following steps: It is clear that the average response delay of the safety officer is 30 seconds, and the average handling duration for abnormal events is 90 seconds; Establish the correlation between the handling duration and the traffic flow density: when the traffic flow density within the cell is greater than or equal to 80 vehicles per kilometer, the handling duration is increased by 20% on the original basis; The safety risk will increase due to the changes in the response delay and handling duration, and the increased amplitude is positively correlated with the adjustment of the handling duration; When the increased amplitude of the safety risk reaches or exceeds the threshold of 30%, or the increased amplitude of the efficiency risk reaches or exceeds another threshold of 20%, reduce the upper limit of the access density of the autonomous taxi within the cell, adjust it to 70% of the original, and increase the weight of the safety risk by 30%. At the same time, push a warning message to the remote monitoring center.

[0027] Preferably, in the method for evaluating the operation risk of the autonomous taxi, the key node risk index NRI j The construction and risk grading method are as follows: Real-time access to multi-source data such as the movement trajectory, accident records, and weather information of the test autonomous taxi, combined with the traffic flow density, number of conflict events, communication delay data, and environmental parameters; Fuse the above data to extract the conflict rate, delay index, and accident density of the key nodes; Calculate the key node risk index NRI based on the conflict rate, delay index, and accident density j According to the risk index NRI j The key nodes are divided into three categories: Type I nodes, NRI j ≥0.7: Implement real-time monitoring and focus on analyzing its disturbance impact on the regional traffic flow; Type II nodes, 0.4 ≤ NRI j <0.7: Conduct periodic evaluations and dynamically adjust the monitoring frequency; Type III nodes, NRI j <0.4: Implement basic monitoring and regularly update the basic data.

[0028] The technical problems solved by the present invention also include: 1. The problem that the existing risk assessment system for autonomous driving taxis is difficult to dynamically adapt to the complex changes of the urban road network. Traditional methods use fixed area division or simple grid models, which cannot dynamically adjust the coverage range and granularity of road network units according to real-time travel demands, resulting in insufficient risk assessment accuracy, especially in accurately locating risk areas during peak hours or sudden traffic events.

[0029] 2. The problem of insufficient multi-source traffic data fusion and risk quantification ability. Existing systems lack multi-dimensional real-time fusion analysis of traffic flow density, conflict events, communication delays, and environmental parameters, resulting in one-sided calculation of risk indices and inability to accurately reflect the comprehensive risk status of key nodes.

[0030] 3. The problem of low reliability in predicting the behavior response of autonomous driving vehicles. Existing models are difficult to simulate the real-time behavior changes of vehicles in a dynamic road network, especially the insufficient discretization processing of key parameters such as target detection, obstacle avoidance response, and communication delay, resulting in large deviations in risk state inference.

[0031] 4. The problem of poor dynamic coordination between risk warning and path planning. Traditional methods rely on static risk level division and cannot dynamically adjust path strategies in combination with real-time hidden state probabilities, resulting in warning delays or disconnection between planning strategies and real road conditions.

[0032] Therefore, the present invention also provides an assessment system for the operation risk of autonomous driving taxis, including: A road network dynamic division module for dynamically dividing the urban road network into hexagonal honeycomb units according to real-time travel demands, where the dynamic radius R of each honeycomb unit t is calculated by the formula R base is the reference radius, Q t is the travel demand density within the honeycomb at the current time period, Q avg is the historical average demand density, Q max and Qmin is the extreme value of demand density, α is the adjustment coefficient, and key nodes are selected in each hexagonal honeycomb cell; A multi-source data access module for obtaining traffic flow density, the number of conflict events, communication delay data, and environmental parameters in real time; A key node risk assessment module for calculating the key node risk index NRI according to the multi-source data j , where NRI j is calculated by the formula NRI j =ω 1 ×C j +ω 2 ×D j +ω 3 ×A j where C j is the conflict rate, D j is the delay index, A j is the accident density, and ω 1 , ω 2 , ω 3 are preset weights; A behavior response simulation module that includes a hidden Markov chain model for defining hidden states S1 (safe state) and S2 (risk state), discretizing the target detection confidence, obstacle avoidance response time, and V2V communication delay data of the autonomous driving taxi into observation states, constructing a state transition probability matrix P and an emission probability matrix E, and inferring the probabilities of each hidden state at the current moment through the Viterbi algorithm in combination with matrices P and E; A comprehensive risk decision-making module for comprehensively determining low, medium, and high risk levels according to the key node risk index NRI j and the hidden state probabilities, outputting a warning signal and dynamically adjusting the path planning strategy of the autonomous driving taxi.

[0033] The present invention has at least the following beneficial effects: 1. Dynamic risk assessment and path optimization: Through dynamic division of honeycomb cells and multi-source data fusion, real-time quantification of traffic risks and dynamic adjustment of path strategies are achieved, improving the scientificity and safety of the deployment of autonomous driving taxis.

[0034] 2. Spatial division and parameter optimization: The Voronoi diagram algorithm ensures dynamic adaptation of honeycomb boundaries, and the scientific setting of parameter ranges enhances the universality of the model for different urban road networks.

[0035] 3. Probability model parameter calibration: Based on the state transition and emission probability matrices of measured data, the prediction accuracy of the hidden Markov chain model for vehicle behavior is improved, and the risk misjudgment rate is reduced.

[0036] 4. Multi-dimensional data support: Expand data types and update frequencies to comprehensively reflect traffic conditions, providing a richer basis for risk index calculation and model input.

[0037] 5. Dynamic routing and safety response: Route planning covers high-risk nodes, combines Monte Carlo simulation to predict long-term impacts, balances efficiency and safety; Path replanning triggered by traffic flow density improves emergency response capabilities.

[0038] 6. Closed-loop verification and continuous optimization: The digital twin environment verifies the actual effects of path strategies, forming a "data - decision - verification" closed loop to ensure the continuous evolution of strategies.

[0039] 7. Quantify long-term impacts and optimize facilities: Quantify the long-term impacts of occupant behavior and energy replenishment through Monte Carlo simulation, optimize the layout of charging stations and transfer points, and reduce the imbalance of road network load.

[0040] 8. Perturbation analysis and precise verification: The digital twin environment reproduces traffic flow perturbations, quantifies the actual impacts of path adjustments, and provides data support for strategy optimization.

[0041] 9. Human-machine collaborative safety enhancement: Associate the response of safety officers with traffic flow density, dynamically adjust the handling duration and access rules, and improve the risk control ability in complex scenarios.

[0042] 10. Differentiated monitoring of key nodes: Dynamically classify nodes through risk grading, optimize the allocation of monitoring resources, and improve the pertinence and efficiency of risk assessment.

[0043] 11. Dynamic road network division improves the adaptability of risk assessment: Through the dynamic radius adjustment mechanism of hexagonal honeycomb cells, it can adaptively scale the road network coverage according to the real-time travel demand density, solving the problem of lagged response of traditional fixed grid division to traffic flow mutation scenarios. Adjust the size of honeycomb cells by combining the ratio of historical and current demand densities, reduce the honeycomb radius during peak hours to improve risk positioning accuracy, and expand the radius during low-demand hours to reduce computational load, achieving the optimal balance between resource allocation and risk assessment efficiency.

[0044] 12. Multi-source data fusion enhances the comprehensiveness of risk quantification: By integrating multi-dimensional real-time data such as traffic flow density, number of conflict events, communication delay, and environmental parameters, and constructing a key node risk index (NRI j ) based on preset weights, it solves the problem of one-sidedness in traditional single-index evaluation. The weighted calculation of conflict rate (C j ), delay index (D j ), and accident density (A j ) in the formula can comprehensively quantify risks from three dimensions of traffic conflicts, traffic efficiency, and safety accidents, significantly improving the scientificity and reliability of key node risk determination.

[0045] 13. Hidden Markov Model Optimization for Behavioral Response Prediction: By defining hidden states (safe / risky) and discretizing observed states (object detection confidence, obstacle avoidance response time, communication delay), combined with the state transition matrix and emission probability matrix, the Viterbi algorithm is used to dynamically infer the hidden state probability. This model can effectively capture the behavioral uncertainty of autonomous vehicles in complex road networks. Especially through discretization, it reduces noise interference, improves the real-time performance and accuracy of risk state prediction, and provides a reliable basis for dynamic path planning.

[0046] 14. Comprehensive Risk Decision-making to Enhance Path Planning Synergy: Based on the dual decision-making mechanism of NRI j and the hidden state probability, the static risk index is combined with the dynamic behavioral response prediction to achieve multi-level early warnings for low, medium, and high risk levels. By dynamically adjusting the path planning strategy, it can automatically avoid potential risk areas at high-conflict nodes and switch to redundant communication links when the communication delay is high, ensuring the optimal operation safety and efficiency of autonomous taxis in complex urban road networks. Description of the Drawings

[0047] Figure 1 : Schematic diagram of the ant climbing pole theory; Figure 2 : Schematic diagram of the "area - cellular - key node" three-level architecture; Figure 3 : Model diagram of the cellular operation of autonomous taxis; Figure 4 : Schematic diagram of the road risk model of autonomous taxis at key nodes; Figure 5 : Flowchart for formulating the deployment parameters of autonomous taxis. Detailed Implementation Manner

[0048] The following further elaborates on the present invention in conjunction with the drawings, so that those skilled in the art can implement it with reference to the text of the specification.

[0049] As Figures 1-5 shown: Figure 1 is the schematic diagram of the ant climbing pole theory, Figure 1Based on the bionics principle of ants climbing a pole, the theoretical basis for the dynamic division of honeycomb units is explained. When a large number of ants randomly crawl on the pole and turn back after meeting, their paths are ultimately equivalent to dividing the pole into two independent regions, and the number of ants in each region tends to be balanced. This phenomenon is mapped to the deployment strategy of autonomous taxis, indicating that within dynamically adjusted honeycomb units, the randomness of vehicle operation paths can be equivalent to risk assessment within a fixed region, thus achieving spatial decoupling. Based on the ant colony algorithm, honeycomb units are dynamically generated centered around transportation hubs, and the Voronoi diagram algorithm is used to ensure that there are no overlapping boundaries and the entire area is covered, supporting the elastic calculation formula of the honeycomb radius R t of the elastic calculation formula which provides a spatial division basis for subsequent critical node identification and risk assessment.

[0050] Figure 2 It is a schematic diagram of the three-level architecture of "region - honeycomb - critical node". Figure 2 It shows the hierarchical architecture of global risk assessment: the bottom layer is the critical node layer, which realizes local risk early warning by real-time monitoring of NRI j values (such as a type of node with NRI j ≥ 0.7); the middle layer is the honeycomb layer, which aggregates the peak risks of nodes and calculates the admission capacity, dynamically adjusting the vehicle density (such as path planning covering nodes with NRI j ≥ 0.4); the top layer is the region layer, which quantifies the cascading effect between honeycombs through a risk propagation model and generates time-sharing and zoning strategies (such as peak-hour traffic restrictions). This architecture strengthens the "barrel effect" theory, that is, the regional risk is determined by the peak risk of critical nodes, supports the application of the hidden Markov chain model and the Viterbi algorithm, and realizes the dynamic determination of risk levels through the state transition probability matrix P and the emission probability matrix E.

[0051] Figure 3 It is a model diagram of the honeycomb operation of autonomous taxis. Figure 3 It depicts the operation rules of autonomous taxis within honeycomb units: vehicles start from transportation hubs (such as subway stations, charging stations), and their paths cover a range of 0.5 - 4 times the honeycomb radius to ensure service coverage of high-risk nodes (NRI j ≥ 0.4). The stop time for passengers to get on and off follows a Poisson distribution (with a mean of 2 - 5 minutes), and when the battery level is below the threshold, the vehicle automatically returns to the charging station (the charging time is 0.5 - 2 hours), and the empty vehicle dispatching is limited within the honeycomb based on reinforcement learning. This model is closely related to Monte Carlo simulation. By predicting the impact of passenger behavior and energy replenishment requirements on traffic flow density, it optimizes the location of charging stations and the layout of transfer points, balances the long-term operation rules and short-term traffic conditions, and achieves the goal of minimizing the comprehensive risk value of the path.

[0052] Figure 4Schematic diagram of the road risk model for autonomous driving taxis at key nodes. Figure 4 A mathematical model for the risk evolution of key nodes is constructed. The inputs include the initial risk X of the node f , the static features θ, and the number C of autonomous driving vehicles max , and the output is the comprehensive risk Y after deployment f (coupling of efficiency and safety risks). This model is verified through microscopic simulation experiments and combines the NRI j grading method (real-time monitoring of type-I nodes, periodic assessment of type-II nodes) to quantify the impact of technical configurations (such as L3 / L4-level systems) on the risk boundary. For example, when a vehicle detects a pedestrian intrusion, the probability of successful braking is 92% (parameter of the emission probability matrix E). False judgment or delayed response will trigger the intervention of the safety officer, and the disposal duration and access density are adjusted through the human-machine collaborative safety strategy to achieve dynamic risk control.

[0053] Figure 5 Flowchart for formulating deployment parameters of autonomous driving taxis. Figure 5 It shows a closed-loop process of "evaluation - decision - verification": First, the area is divided based on GIS to generate a honeycomb structure, and spatio-temporal strategies (such as operating hours, human-machine ratio) are configured; the perturbation impact of the path strategy is verified through a digital twin environment (closed-loop verification system), and the local traffic flow changes (change rate of traffic flow density, change value of average vehicle speed) and overall efficiency indicators (average passing time, delay time) are quantified; finally, the parameters are iteratively optimized according to the simulation results, such as adjusting the honeycomb radius R t The α coefficient (value range 0.5 - 1.2) or the path weight coefficient (efficiency risk weight increased by 30%) in the calculation. This process supports the collaborative optimization of Monte Carlo simulation and dynamic routing, ensures scientific decision-making on the layout of charging stations and the capacity of transfer points, and finally forms a vehicle deployment plan that meets the regional risk control.

[0054] According to an embodiment of the present invention, when dynamically dividing the urban road network into hexagonal honeycomb units, the reference radius R base can be determined by fitting according to the urban road network density, and the value range is 1 - 3 kilometers (for example, 1 kilometer is taken in high-density urban areas and 3 kilometers is taken in suburban areas). The adjustment coefficient α is determined through historical traffic flow regression analysis, and the value range is 0.5 - 1.2 (such as α = 1.0 during the morning peak period and α = 0.6 during the off-peak period). This process can use a GIS geographic information system (such as ArcGIS) to integrate high-precision map data, and combine the Voronoi diagram algorithm library (such as the scipy library in Python) to dynamically adjust the honeycomb boundary to ensure that adjacent honeycombs do not overlap and cover the entire area. The road network data is stored in the cloud server and transmitted to the in-vehicle terminal in real time through the edge computing node. The honeycomb division result is displayed through the in-vehicle navigation system, and the driver can view the distribution of key nodes within the current honeycomb through the human-machine interaction interface.

[0055] When accessing multi-source traffic data in real time, floating car trajectory data can be collected through a GPS module (such as Trimble BD992) installed on the test vehicle, the number of conflict events is identified by an in-vehicle camera (such as Mobileye EyeQ5) combined with computer vision algorithms, and V2V communication delay data is obtained through a DSRC module (such as Cohda Wireless MK5). The data update frequency is set to 1 minute, and real-time data stream transmission is achieved through the Apache Kafka message queue, and SparkStreaming is used for distributed computing. The weight coefficients of the key node risk index NRI j are: ω 1 is 0.4 - 0.6 (corresponding to the conflict rate), ω 2 is 0.3 - 0.5 (corresponding to the delay index), ω 3 is 0.1 - 0.3 (corresponding to the accident density), and the calculation results are pushed to the control center through the 5G network after being completed on the edge server.

[0056] In the state transition probability matrix P of the hidden Markov chain model, the probability of S1 transitioning to S1 is 0.85 - 0.95 (such as 0.90), the probability of S1 transitioning to S2 is 0.05 - 0.15 (such as 0.10), the probability of S2 transitioning to S1 is 0.40 - 0.60 (such as 0.50), and the probability of S2 transitioning to S2 is 0.40 - 0.60 (such as 0.50). The emission probability matrix E is constructed based on the test data of multi-brand vehicles in a closed test field. When the target detection confidence level ≥ 0.9, the emission probability is 0.90; when the obstacle avoidance response time ≤ 0.5 seconds, it is 0.85; when the V2V communication delay ≤ 50 milliseconds, it is 0.92. The data is stored in a MySQL database, and the transition probability matrix P is calculated through the matrix operation of the NumPy library in Python. When determining the risk level, the low-risk threshold is NRI j <0.4 and the probability of S1 ≥ 0.85, the medium risk is 0.4 ≤ NRI j <0.7 or the probability of S2 ≥ 0.20, the high risk is NRI j ≥ 0.7 or the probability of S2 ≥ 0.50. When the risk level is high, the system automatically reduces the vehicle access density within the honeycomb to 70% of the original upper limit, and pushes a new path through the in-vehicle terminal (such as avoiding nodes with NRI j ≥ 0.7). Through actual measurement and verification, this method increases the regional average vehicle speed by 15% and reduces the conflict events by 22%, effectively balancing traffic efficiency and safety risks.

[0057] According to another embodiment of the present invention, when dynamically dividing the urban road network into hexagonal honeycomb cells, the reference radius R baseThe value of t is determined according to the urban road network density, and the specific range is from 1.5 kilometers to 2.5 kilometers. The adjustment coefficient α is obtained through historical traffic flow regression analysis, and the value range is from 0.7 to 1.0. The real-time travel demand density Q t is collected by multi-source traffic sensors deployed at road intersections and main roads. For example, Siemens Sicore traffic flow detectors are installed on the top of street lamp poles or traffic signal poles. The outer shell is made of anti-ultraviolet ABS plastic and fixed by bolts. The computing unit selects the NVIDIA Jetson AGX Xavier embedded system and is installed in a waterproof metal box. The box is fixed in the roadside control cabinet through a bracket. During operation, the sensor uploads vehicle trajectory data to the computing unit every 5 minutes, and the computing unit dynamically adjusts the honeycomb radius R t according to the formula and updates the honeycomb boundary through the Voronoi diagram algorithm. For example, when Q t increases by 20% compared with the historical average value, R

[0058] extends from 2 kilometers to 2.8 kilometers to cover the newly added congested area. j The calculation of the key node risk index NRI j relies on the fusion of multi-source traffic data. The conflict rate C j is collected by video surveillance devices deployed at intersections. For example, Hikvision DS-2CD2346G2-IU cameras are installed at a height of 6 meters, fixed to the surveillance pole through a universal bracket, and the outer shell is made of aluminum alloy. The delay index D j is calculated based on the difference between the real-time vehicle speed and the historical average vehicle speed. The vehicle speed data is provided by GPS floating cars, and the threshold is set to trigger an alarm when the vehicle speed drops by more than 30%. The accident density A j is obtained from the traffic management department's database, and the statistical period is 30 days. Data processing uses Dell's PowerEdgeR750 servers deployed in the cloud data center. The cabinet is equipped with redundant power supplies and cooling systems. During operation, the camera captures traffic flow videos in real time, and the edge computing node extracts the number of conflict events per hour. If the number of conflict events at a certain node is ≥ 3 times per hour, then C j = 0.8; combined with a 40% drop in real-time vehicle speed (D j = 0.6) and an accident density of 0.3, the calculated NRI

[0059] The hidden states of the Hidden Markov Model include S1 (safe) and S2 (risk), and the observed states are generated by discretizing vehicle sensor data. The target detection confidence threshold is set to ≥0.85, corresponding to an emission probability of 0.9; the obstacle avoidance response time threshold is ≤0.6 seconds, corresponding to an emission probability of 0.8; the V2V communication delay threshold is ≤60 milliseconds, corresponding to an emission probability of 0.9. The in-vehicle computing unit uses the Aurora Innovation autonomous driving domain controller, which is integrated into the vehicle's central control system. The housing is made of magnesium alloy and is fixed to the trunk by bolts; the V2X communication module selects the Huawei MH5000-31 module, which supports 5G and DSRC protocols and is installed in the vehicle's pre-installed T-Box. During operation, the vehicle continuously collects the target detection confidence (such as 0.92), the obstacle avoidance response time (such as 0.55 seconds), and the V2V delay (such as 45 milliseconds). After inputting into the model, the probability of the current state being S1 is calculated as 0.88 through the Viterbi algorithm. Combining NRI j =0.73, the system comprehensively determines it as a high-risk level, triggers a route replanning instruction, and guides the vehicle to a section with a traffic flow density lower than 60 vehicles per kilometer.

[0060] According to another embodiment of the present invention, when the dynamic road network division adopts the Voronoi diagram algorithm, commercial GIS software (such as ArcGIS) can be used in combination with real-time traffic data interfaces. The reference radius R base ranges from 1 km (applicable to high-density urban areas) to 3 km (applicable to suburban areas), and the adjustment coefficient α is fitted from 0.5 (off-peak hours) to 1.2 (peak hours) according to historical traffic data. The selection of key nodes can be combined with floating car trajectory data, and transportation hubs (such as subway stations, commercial areas) can be used as initial seed points to identify high-conflict areas in the road network through the Dijkstra algorithm. In terms of equipment, lidar (such as Velodyne HDL-64E) and cameras (such as Basler acA2040-90um) deployed on road infrastructure can be used to continuously collect traffic flow data.

[0061] The reference radius R base needs to call urban road network density data. For example, the density value can be calculated by obtaining the road length and regional area through OpenStreetMap. The fitting process of the adjustment coefficient α can use SPSS software for multiple linear regression analysis, and the independent variables include the traffic flow, accident rate, and weather data in historical periods. The weight coefficient ω 1 takes 0.5 (dominated by the conflict rate), ω 2 takes 0.4 (the delay index is the second), ω 3Take 0.1 (accident density assistance), which is calculated by the Analytic Hierarchy Process (AHP). In terms of equipment, the Huawei Atlas 500 intelligent edge computing platform can be used for real-time data processing with a memory configuration of 16GB DDR4.

[0062] Key node risk index NRI j The calculation of NRI needs to access the traffic management platform data in real time, including the traffic flow density obtained through the V2X communication module (such as the Huawei 5G vehicle terminal MH5000) and the number of conflict events counted by the video analysis system (such as Hikvision iDS-9600). The data update frequency is set to 10Hz to ensure dynamics. When calculating, the conflict rate C j is obtained by dividing the number of emergency braking events per unit time by the total traffic volume, and the delay index D j uses the ratio of the actual travel time to the free flow time, and the accident density A j is the number of historical accidents divided by the node area. In terms of equipment, NVIDIA Jetson AGX Orin can be used as an edge computing node with TensorRT accelerated inference.

[0063] By dynamically adjusting the cellular boundary through the Voronoi algorithm, the road network segmentation accuracy can be improved by more than 30%, effectively reducing the boundary overlap problem caused by traditional fixed grid division. The scientific setting of the parameter value range improves the model adaptability. In tests of different city scales, the error rate of the reference radius R base is controlled within 15%. The optimized combination of weight coefficients makes the correlation coefficient between the NRI j calculation result and the actual accident rate reach 0.82, improving the prediction accuracy by 45% compared with the traditional single-index method. This implementation provides a reliable basis for spatial division and risk quantification for the dynamic deployment of autonomous driving taxis.

[0064] According to another embodiment of the present invention, the probability of the hidden state S1 transitioning to S1 is set to 0.90 (typical value), S1 transitioning to S2 is 0.10, S2 transitioning to S1 is 0.50, and S2 transitioning to S2 is 0.50. These parameters are based on the statistical data measured in a closed test field. For example, in 5000 urban road tests, the safety state retention rate is 88.7%. In terms of equipment, the CompactRIO system of National Instruments can be used for data acquisition, and LabVIEW software is used to record vehicle state transition events. This system can be installed in the trunk of an autonomous driving taxi to obtain real-time data of the vehicle control unit through the CAN bus interface.

[0065] When the object detection confidence is ≥ 0.9, the emission probability is set to 0.90. This threshold can be obtained by training the YOLOv5s model on the Cityscapes dataset. The emission probability for obstacle avoidance response time ≤ 0.5 seconds is 0.85, and the data is from the test of the fusion perception system of lidar (such as Ouster OS1-64) and millimeter-wave radar (such as Bosch MRR440). The emission probability for V2V communication delay ≤ 50 ms is 0.92, and the test is carried out using the MK5 OBU device of Cohda Wireless under the 802.11p protocol. These devices are installed on the front bumper (lidar), side mirror (millimeter-wave radar), and roof (OBU antenna) of the vehicle respectively.

[0066] When initializing the hidden Markov chain model, the state probability vector is set to [0.9, 0.1]. Observation data is received every 500 ms and input into the model after discretization. The Viterbi algorithm runs on the NVIDIA Jetson AGX Orin platform and accelerates inference through TensorRT. Parameter calibration uses the Baum-Welch algorithm and is iteratively optimized based on 100 hours of actual road test data. In terms of equipment, Renesas R-Car H3 can be used as the in-vehicle computing platform to run the state estimation program with the Linux operating system.

[0067] The probability matrix based on the measured data makes the state prediction accuracy reach 92.3%, which is 18 percentage points higher than the traditional empirical value method. The precise threshold setting of the emission probability controls the false alarm rate below 6.8%, meeting the requirements of the ISO 26262 functional safety standard. In the mixed traffic scenario test, the response delay of the model to the risk state is shortened to 120 ms, effectively supporting the emergency decision-making of the autonomous driving system. This implementation provides a reliable probability modeling basis for dynamic risk assessment and improves the behavior prediction ability of autonomous driving taxis in complex scenarios.

[0068] According to another embodiment of the present invention, the new added data includes average vehicle speed (0 - 120 km / h), proportion of autonomous driving vehicles (10% - 80%), number of emergency braking events (0 - 5 times / minute), cooperative lane change failure rate (0.1% - 5%), visibility (50 m - 2000 m), road surface adhesion coefficient (0.1 - 1.0), signal light phase (cycle 60 - 180 seconds). Among them, the visibility threshold is set as: low (< 200 m), medium (200 - 500 m), high (> 500 m). In terms of equipment, the Bosch BMP280 barometric pressure sensor can be used to measure visibility and is installed inside the vehicle's front windshield; the Honeywell MLX90640 infrared thermal imager is used to detect the road surface temperature and calculate the adhesion coefficient, and is installed at the front end of the chassis.

[0069] Multi-source data is transmitted in real time to the edge computing node through a 5G communication module (such as Huawei MH5000), and the update frequencies are set to 10 Hz (high-frequency data) and 1 Hz (low-frequency data). The average vehicle speed is obtained through a GPS module (such as u-blox NEO-M8N), which is installed on the roof; the proportion of autonomous vehicles is obtained by receiving the identity information of surrounding vehicles through V2X communication, and the OBU device is installed at the rear of the vehicle. The collaborative lane change failure rate is recorded by an in-vehicle camera (such as Sony IMX490) for lane change events and analyzing failure cases, and the camera is installed on the rearview mirror housing.

[0070] The multi-source data is fused through the Kalman filtering algorithm to generate a data stream with a unified timestamp. The conflict rate C j is calculated as the number of emergency braking events divided by the total traffic volume, and the delay index D j adopts the ratio of travel time to free flow time, and the accident density A j is the number of historical accidents divided by the node area. The observation state discretization uses K-means clustering to map continuous data to discrete values from 0 to 2 (for example, visibility 0 = low, 1 = medium, 2 = high). In terms of equipment, Renesas R-Car H3 can be used as an in-vehicle computing platform to run data processing programs with the Linux system.

[0071] Expanding the data type increases the calculation dimension of the risk index by 40%, and the correlation with the actual accident rate is increased to 0.85. The high-frequency data update (10 Hz) shortens the model response delay to 80 ms, which is 3 times faster than the traditional method. The discretization process reduces the data dimension by 60% and improves the inference efficiency of the hidden Markov chain model by 25%. In the heavy rain weather test, the addition of visibility data increases the risk warning accuracy rate from 72% to 89%, effectively enhancing the risk identification ability in complex environments. This implementation provides more comprehensive data support for dynamic risk assessment and improves the multi-dimensional perception ability of the model for the traffic system.

[0072] According to another embodiment of the present invention, in terms of determining the dynamic path planning range, the radius R of the cellular unit t adopts a reference value of 1 - 3 kilometers and is dynamically adjusted in combination with the real-time travel demand density Q t For example, during the morning rush hour, when Q t exceeds 120% of the historical average, R tIt can be shortened to 0.8 times the reference value. The path planning range is set to 0.5 - 4 times the honeycomb radius, and the specific multiple is adjusted in real time according to the traffic flow state. The Velodyne HDL-64E lidar is used for environmental perception, and the HERE high-precision map is combined to achieve dynamic division of the honeycomb boundary. The vehicle-mounted GPS module obtains the position information in real time, and calculates the current honeycomb coverage range through the Voronoi diagram algorithm to ensure that adjacent honeycombs do not overlap and the whole area is covered.

[0073] The initial path set is generated based on the NVIDIA DRIVE PX2 computing platform, and the real-time road conditions are obtained by combining the TomTom traffic data API. The path planning needs to cover all NRI j key nodes with ≥0.4, and the single-trip mileage is controlled within 1.5 - 3 times the honeycomb radius. For example, in a honeycomb cell with a honeycomb radius of 2 km, the path length is set to 3 - 6 km. The Dijkstra algorithm is used to generate candidate paths, and the path combination with a balance of efficiency risk and safety risk is selected through multi-objective optimization. The cruising path of idle vehicles is dynamically adjusted according to the traffic flow density and is preferentially deployed within 500 meters around NRI j ≥0.7 high-risk nodes.

[0074] In the construction of the optimized driving model, the initial weight of the efficiency risk is set to 0.6, and the safety risk weight is 0.4. When the detected traffic flow density reaches 80 vehicles / km, the path replanning mechanism is automatically triggered, and the section with a traffic flow density <60 vehicles / km is selected. The Mobileye EyeQ4 chip is used for driving decision-making, and V2X data interaction is realized through the 4G / 5G communication module. Monte Carlo simulation predicts the passenger boarding and alighting behavior within the next 30 minutes, and the input parameters include the mean of the Poisson distribution of 3 minutes and the standard deviation of 1 minute. When the traffic flow density exceeds the standard, the abnormal event handling duration is extended from 90 seconds to 108 seconds, and at the same time, the safety risk weight is increased to 0.52. The path replanning response time is controlled within 200 ms to ensure real-time requirements.

[0075] Through dynamic path planning, the coverage ratio of high-risk nodes is increased to 92%, and the average service response time is shortened by 18%. The Monte Carlo simulation prediction accuracy reaches 85%, and the charging station location error is controlled within 1.2 km. The traffic flow density trigger mechanism reduces the path replanning response time by 40%, and the accident rate is reduced by 23% after the abnormal event handling duration is extended. This solution significantly improves the safety performance in complex traffic scenarios while maintaining the vehicle operation efficiency.

[0076] According to another embodiment of the present invention, the construction of a closed-loop verification system is based on digital twin technology. The system can interact with traffic data collection modules, dynamic path planning frameworks, and risk assessment models. The traffic data collection module can use Huawei Road-X perception equipment to collect data such as traffic density and the number of conflict events in real time at a frequency of 5Hz. In the digital twin environment, a virtual scene is constructed with the help of the Unity3D engine. The scene contains the road network provided by the HERE high-precision map and pedestrians, ordinary vehicles and other elements generated by the CARLA simulation platform. The initial path set and the adjusted path plan generated by the dynamic path planning framework will be synchronously input into the virtual scene to simulate the driving process of the self-driving taxi.

[0077] The parameter update mechanism operates according to a certain process. First, the digital twin environment calculates indicators such as the rate of change of traffic density in the surrounding sections before and after the path adjustment (the threshold is set to ±15%) and the average speed change value (the threshold is set to ±8km / h). Then, these indicators are fed back to the risk assessment model, and the model will adjust the state transition probability matrix P of the hidden Markov chain accordingly. For example, the probability of transitioning from S1 to S2 is dynamically adjusted by ±20% based on the original 0.05-0.15. At the same time, the output data of the Monte Carlo simulation will be used to optimize the site selection of charging stations, and its load balancing threshold is set to 0.8, and the capacity of the docking point will be adjusted. The upper limit of the capacity of a single docking point is 5 vehicles.

[0078] The collaborative optimization strategy is implemented in three aspects. First, the dynamic path planning framework will t Data adjustment cell radius R t , the adjustment range is 0.8-1.2 times the baseline value. Second, when the traffic density reaches 80 vehicles / km, the system will automatically start path replanning, giving priority to sections with a traffic density of less than 60 vehicles / km, and increasing the efficiency risk weight from 0.6 to 0.78. Third, the results of path replanning will be fed back to the closed-loop verification system in real time, forming an optimization loop with an iteration cycle of 15 minutes. After actual measurement, the system can shorten the average travel time in the region by 12% and reduce the optimization frequency of path strategies by 40%.

[0079] In this way, the simulation accuracy of the digital twin environment reached 92%, and the error with the actual traffic flow was controlled within ±10%. The error in the site selection of charging stations was less than 1.5 kilometers, and the queue overflow rate at the docking point was reduced by 65%. The parameter update frequency of the risk assessment model was increased to once per minute, and the response speed of the system was increased by 30%. The stability indicators of regional traffic flow (such as speed variance) improved by 22%, and the overall performance of the system was effectively improved.

[0080] According to another embodiment of the present invention, in terms of setting the input parameters of Monte Carlo simulation, the data of the behavior of passengers getting on and off the vehicle adopts a Poisson distribution model, with the mean set to 3 minutes and the standard deviation to 1 minute. In the data of the energy replenishment behavior, the charging time is set to 0.5 - 2 hours, and the specific time is dynamically adjusted according to the charging pile power (such as 7kW slow charging or 120kW fast charging). The floating car data API of TomTom can be used to obtain samples of the behavior of passengers getting on and off the vehicle, and the charging time statistics can be obtained through the charging pile database of the State Grid. The data input frequency is set to be updated once a minute to ensure the timeliness of the simulation.

[0081] The optimization strategies for the transfer points and charging stations are implemented through two matrices. The matrix of the overflow probability of the transfer point queue adopts a 3×3 grid division, and the maximum capacity of each grid is set to 5 vehicles. When the queuing vehicles exceed 80% of the capacity for 15 consecutive minutes, the migration mechanism is triggered. When calculating the unbalanced index of the road network around the charging station, the service radius is set to 1.5 kilometers. When the traffic flow density fluctuation of a certain road section exceeds ±25% of the historical average, it is marked as an unbalanced area. The Huawei Atlas 500 intelligent small station can be used for real-time analysis of traffic flow data, and combined with Esri ArcGIS for geographic information processing.

[0082] The implementation of the path replanning strategy is divided into three steps. First, set the traffic flow density threshold to 80 vehicles per kilometer. When the detected value reaches the threshold, the path replanning algorithm is automatically started, and the road sections with a traffic flow density less than 60 vehicles per kilometer are preferentially selected. Second, dynamically adjust the path weight coefficient, increase the efficiency risk weight from 0.6 to 0.78, and correspondingly reduce the safety risk weight. Finally, synchronize the adjusted path plan to the digital twin environment for verification, and the verification period is set to 15 minutes. The path planning module of Baidu Apollo can be used to implement the algorithm deployment, and the data synchronization is completed through the 5G communication module.

[0083] In this way, the overflow rate of the transfer point queue is reduced by 65%, and the average residence time of a single node is shortened from 8 minutes to 3 minutes; the location error of the charging station is controlled within 1.2 kilometers, and the unbalanced index of the surrounding road network is improved by 38%; the response time of path replanning is shortened from 45 seconds of the traditional method to 18 seconds; the prediction accuracy of Monte Carlo simulation reaches 85%, and the prediction error of the behavior of passengers getting on and off the vehicle is less than ±1.5 minutes.

[0084] According to another embodiment of the present invention, in terms of the architecture design of the closed-loop verification system, Huawei Atlas 500 is used as an edge computing node and deployed at the intersection of transportation hubs and main roads to collect data such as traffic flow density and the number of conflict events in real time at a frequency of 5Hz. The digital twin environment is built based on the Unity 3D engine, integrating HERE high-precision map data, and the road network accuracy reaches the 0.1-meter level. The pedestrian behavior model in the virtual scene is trained using the CARLA open-source dataset, and the dynamic parameters of ordinary vehicles come from the TomTom floating car database. The path plan generated by the dynamic path planning framework is synchronized to the digital twin environment through the 5G communication module, and the data transmission delay is controlled within 20ms.

[0085] In terms of setting the quantitative indicators for perturbation analysis, the local traffic flow change rate threshold is set to ±15%, and the average vehicle speed change threshold is ±8km / h. The average travel time at the cellular unit level is calculated based on the weighted average of the travel times of all road segments in the road network, and the delay time is calculated using the Webster formula. The conflict probability calculation of the risk assessment model combines the state transition probability of the hidden Markov chain, and an early warning is triggered when the probability of the S2 state exceeds 0.6. The simulation time step is set to 0.1 second, and the single simulation duration covers the typical cycle of traffic flow changes (such as 15 minutes).

[0086] The implementation of the verification process is divided into three stages. First, the traffic data collection module transmits the real-time data to the cloud platform, and after preprocessing by Kalman filtering, it is input into the digital twin environment. Secondly, the initial path and the adjusted path generated by the dynamic path planning framework are respectively subjected to 100 Monte Carlo simulations to statistically obtain the confidence interval of the perturbation index. Finally, the parameters of the risk assessment model are adjusted according to the simulation results. For example, the weight coefficient ω j of the key node risk index NRI 1 is dynamically calibrated from 0.5±0.1. The entire verification cycle is set to 15 minutes, and after the parameters are updated, they are synchronized to the autonomous driving vehicle terminal.

[0087] As a result, the simulation error of the digital twin environment is controlled within ±10%, and the matching degree with the actual traffic flow reaches 92%; the prediction accuracy of the impact of path adjustment on the surrounding traffic flow density is increased to 88%; the prediction error of the average travel time of the cellular unit is reduced from 25% of the traditional method to 12%; the parameter update frequency of the risk assessment model is increased to 1 time per minute, and the system response speed is increased by 30%.

[0088] According to another embodiment of the present invention, in terms of parameter setting of the human-machine collaborative safety strategy, the average value of the safety officer's response delay is set to 30 seconds, which is determined by collecting historical data through the on-vehicle event recorder and performing normal distribution fitting. The initial value of the abnormal event handling duration is set to 90 seconds, and when the traffic flow density in the cell reaches 80 vehicles per kilometer, the handling duration is extended to 108 seconds. The Mobileye EyeQ5 chip can be used to monitor the traffic flow density in real time, and the data is transmitted to the remote monitoring center through the 4G communication module. Parameter calibration is completed through tests in a closed test field. The test vehicle is an L4-level autonomous driving taxi equipped with the Bosch ESP system for braking control.

[0089] The implementation steps of the risk dynamic adjustment mechanism are as follows. First, when the traffic flow density exceeds the threshold, the system automatically triggers the mechanism for extending the handling duration, and the extension amplitude is 20% of the original duration. Second, the increase amplitude of the safety risk is calculated through the hidden Markov chain model. When it exceeds the 30% threshold, the upper limit of the admission density is reduced from 15 vehicles per kilometer to 10.5 vehicles per kilometer. At the same time, the safety risk weight coefficient is increased from 0.5 to 0.65, and the efficiency risk weight is correspondingly reduced. NVIDIA DRIVE Sim can be used for risk propagation simulation to verify the impact of weight adjustment on the system stability. The data update frequency is set to once per second to ensure real-time requirements.

[0090] The implementation of the warning and response process is divided into three links. First, the traffic flow density is monitored in real time through the Huawei RoadEye lidar, and a warning is triggered when the detected value continuously exceeds the threshold for 5 minutes. Second, the system automatically generates an admission density adjustment plan and pushes it to all autonomous driving vehicle terminals through the V2X communication module. Finally, the adjusted parameters are synchronized to the digital twin environment for verification, and the verification period is 30 minutes. The TomTom traffic index API can be used to obtain regional traffic flow data, and spatial analysis is performed in combination with Esri ArcGIS. The measured data shows that this mechanism can reduce the intervention frequency of the safety officer by 40% and shorten the emergency response time to within 15 seconds.

[0091] Thus, the misjudgment rate of safety risks is reduced by 28%, and the risk assessment accuracy rate is increased to 91%; the mechanism for extending the abnormal event handling duration reduces the accident rate by 23%; the dynamic adjustment of the admission density improves the road network load balance degree by 35%; after the risk weight is optimized, the response speed of the system to sudden traffic events is increased by 40%.

[0092] According to another embodiment of the present invention, in terms of multi-source data access, a test self-driving taxi is equipped with a lidar (such as Velodyne HDL-32E) and a camera (such as Sony IMX490), and motion trajectory data is collected at a frequency of 10 Hz. Accident records are obtained through an in-vehicle EDR black box, and weather information comes from the meteorological bureau's API interface. Traffic flow data is collected in real time by geomagnetic sensors (such as 3M™ 7910) and microwave radars (such as Siemens SIRESPOT) deployed at intersections, with an update frequency of 5 Hz. The visibility in environmental parameters is monitored by a lidar visibility meter (such as Vaisala FD12P) installed on a street lamp pole, and the road surface adhesion coefficient is obtained through a friction coefficient sensor (such as Scantronix MC5) embedded in the road surface.

[0093] The data fusion processing flow is as follows. First, the Kalman filter algorithm is used to denoise the vehicle trajectory data, and the positioning accuracy is controlled within 0.3 meters. Second, the accident records, weather information, and real-time traffic flow data are fused through a BP neural network to extract three core indicators: the conflict rate (unit: times / hour), the delay index (dimensionless), and the accident density (times / km²). The key node risk index NRI j The calculation formula for is NRI j =0.5×C j +0.3×D j +0.2×A j , where C j Calculates the number of conflict events at intersections through video analysis, D j Is calculated based on the Webster delay formula, and A j Is interpolated from the historical accident database.

[0094] The risk grading and monitoring strategy implementation steps are as follows. When NRI j ≥0.7, trigger the monitoring mechanism for type-I nodes, and conduct real-time video analysis through edge computing nodes (such as Huawei Atlas 500) deployed at key locations to focus on monitoring high-risk behaviors such as pedestrians crossing the street and vehicle lane changes. Type-II nodes adopt a rolling assessment with a period of 15 minutes, and use the support vector machine algorithm to predict the risk level in the future period. Type-III nodes update the basic information through crowdsourced data, and the data collection period is set to 24 hours. The monitoring resource allocation ratio is 50% for type-I nodes, 30% for type-II nodes, and 20% for type-III nodes, and the deployment density of cameras and sensors is optimized through the genetic algorithm.

[0095] As a result, the recognition accuracy of key nodes reaches 94%, and the false alarm rate is lower than 5%; the real-time monitoring of type-I nodes shortens the accident response time by 25%; the operation and maintenance cost is reduced by 30% after the optimized allocation of monitoring resources; the risk grading strategy improves the overall traffic efficiency of the road network by 18%.

[0096] The operation risk assessment system of the self-driving taxi of the present invention realizes accurate risk perception and dynamic path optimization through dynamic road network division, multi-dimensional data fusion and intelligent behavior prediction. The specific implementation process is as follows: The first step: Dynamic road network division and key node identification: The system obtains the travel demand density data of each region of the city in real time, including vehicle distribution heat maps, order request volumes and historical demand averages. Based on a dynamic adjustment algorithm, the urban road network is divided into hexagonal honeycomb cells, and the coverage range of each honeycomb automatically scales with the demand density: when the real-time demand is higher than the historical average level, the honeycomb radius shrinks to improve the risk positioning accuracy; when the demand is lower than the average value, the radius expands to reduce the computational load. Within each honeycomb cell, by analyzing historical accident data and real-time traffic flow trajectories, areas with frequent vehicle lane changes, long signal waiting times or high accident incidences are selected as key nodes, which serve as the core areas for subsequent risk assessment.

[0097] The second step: Multi-source data fusion and risk index calculation: The system accesses and integrates multi-dimensional traffic data in real time: the traffic flow density is collected through roadside sensing devices and in-vehicle sensors, multi-modal fusion technology is used to detect vehicle hard braking, lane change conflicts and pedestrian intrusion events, and the conflict frequency within a unit time is counted; the communication delay data between the vehicle and the cloud is monitored synchronously, and the impact of the environment on traffic is evaluated in combination with meteorological parameters. Based on the above data, three core indicators are calculated for each key node: 1. Conflict rate: The ratio of the number of conflict events to the total number of passing vehicles, reflecting the traffic conflict risk; 2. Delay index: The average deviation between the actual travel time of the vehicle and the theoretical free flow time, representing the loss of traffic efficiency; 3. Accident density: The spatial distribution statistics of the number of historical accidents around the node; A comprehensive risk index is generated through weighted fusion of the above indicators. The weight allocation gives priority to the real-time conflict rate, followed by the delay index and the historical accident density. The system updates the risk level every five minutes, divides it into high, medium and low three-level thresholds, and triggers the corresponding early warning mechanism.

[0098] The third step: Behavior prediction and dynamic path decision-making: Based on the hidden Markov model, a vehicle behavior response prediction framework is constructed. The object detection confidence, obstacle avoidance response speed and communication delay data are discretized into three-level observation states, and two hidden states of "safe" and "risk" are defined. The state transition probability and the observation state distribution law are trained through historical data, and the dynamic programming algorithm is used to infer the probability that the vehicle is in a risk state in real time. Combining the key node risk index and the behavior prediction result, a hierarchical decision is made: When the node risk index reaches the high-risk threshold and the vehicle behavior risk probability continues to increase, immediately send a detour instruction to surrounding vehicles and update the no-go area map; If either the node risk index or the behavior risk probability exceeds the standard alone, limit the vehicle speed and activate the redundant communication link; For low-risk nodes, maintain regular monitoring and periodically refresh the evaluation data. Decision instructions are sent in real time through the vehicle-road collaborative network, synchronously optimizing the risk heat map of the cloud road network and achieving global path planning collaboration.

[0099] This system reduces the positioning error of traditional fixed grids through dynamic road network division, and the risk identification accuracy during peak hours is improved by more than 40%; the multi-dimensional risk assessment mechanism controls the false alarm rate within 7%, and the reliability is significantly enhanced compared with the single-index method; the behavior prediction model has an identification delay of less than 2 seconds for sudden risks, and the path dynamic adjustment enables the detour rate of vehicles in high-risk areas to exceed 95%, and the traffic efficiency is improved by about 20%.

[0100] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and embodiments. It can be fully applied to various fields suitable for the present invention. For those familiar with the field, additional modifications can be easily made. Therefore, without departing from the general concept defined by the claims and the equivalent scope, the present invention is not limited to the specific details and the illustrated examples here.

Claims

1. A method for assessing the operating risk of an autonomous taxi, characterized in that: include: Dynamically divide the urban road network into hexagonal honeycomb units, and the dynamic radius R of each honeycomb unit t According to the real-time travel demand, the formula calculate, Where R base is the base radius, Q t is the travel demand density within the cell in the current period, Q avg is the historical average demand density, Q max With Q min is the extreme value of demand density, α is the adjustment coefficient, and key nodes are selected in each hexagonal honeycomb unit; Real-time access to multi-source traffic data, including traffic density, number of conflict events, communication delay data and environmental parameters, to build the key node risk index NRI j and by the formula NRI j =ω1×C j +ω2×D j +ω3×A j Calculate, where C j is the conflict rate, D j is the delay index, A j is the accident density, ω1, ω2, ω3 are the preset weights; The behavior response of the self-driving taxi at key nodes is simulated through the hidden Markov chain model, including: defining the hidden state as S1 safe state and S2 risk state; collecting taxi target detection confidence, taxi obstacle avoidance response time and taxi V2V communication delay data, and discretizing them into observation states; Among them, according to the hidden state, the state transition probability matrix P is constructed; according to the observed state, the emission probability matrix E is constructed; the Viterbi algorithm is used to combine the state transition probability matrix P and the emission probability matrix E to infer the probability of each hidden state in the S1 safe state and the S2 risk state at the current moment; According to the key risk index NRI j The value and the inferred hidden state probability are used to comprehensively determine the risk level, output low, medium, and high risk warnings, and dynamically adjust the path planning strategy according to the risk level.

2. The method for assessing the operating risk of an autonomous taxi according to claim 1, wherein: include: When the urban road network is dynamically divided into hexagonal honeycomb units, the Voronoi diagram algorithm is used to dynamically adjust the honeycomb boundaries to ensure that adjacent honeycombs do not overlap and cover the entire area; The reference radius R base The value range is 1-3 kilometers, and the adjustment coefficient α is determined to be 0.5-1.2 based on the historical traffic data fitting; the key node risk index NRI j The weight coefficients ω1, ω2, and ω3 correspond to the conflict rate C j , Delay Index D j 、Accident density A j , whose value range is: ω1 is 0.4-0.6, ω2 is 0.3-0.5, ω3 is 0.1-0.3; Among them, R base It is determined by fitting the urban road network density, and α is obtained through historical traffic flow regression analysis.

3. The method for assessing the operating risk of an autonomous taxi according to claim 1, wherein: include: In the transition probability matrix P of the hidden state S1 safety state and S2 risk state of the hidden Markov chain model, the probability of S1 transitioning to S1 is 0.85-0.95, the probability of S1 transitioning to S2 is 0.05-0.15, the probability of S2 transitioning to S1 is 0.40-0.60, and the probability of S2 transitioning to S2 is 0.40-0.60; The emission probability matrix E is constructed based on multi-brand vehicle test data in a closed test field, where the emission probability is 0.90 when the taxi target detection confidence is ≥0.9, the emission probability is 0.85 when the taxi obstacle avoidance response time is ≤0.5 seconds, and the emission probability is 0.92 when the taxi V2V communication delay is ≤50 milliseconds.

4. The method for assessing the operating risk of an autonomous taxi according to claim 1, characterized in that: The real-time access to multi-source traffic data further includes: Average vehicle speed, percentage of autonomous vehicles, number of emergency braking events, coordinated lane change failure rate, visibility, road adhesion coefficient, and signal light phase; And update the multi-source traffic data at a set frequency, and the multi-source traffic data is used to calculate the key node risk index NRI j , and the taxi target detection confidence, taxi obstacle avoidance response time and taxi V2V communication delay data are discretized together to obtain the observation state of the hidden Markov chain model.

5. The method for assessing the operating risk of an autonomous taxi according to claim 1, characterized in that: Also includes: For test autonomous taxis, the dynamic radius R of the cellular unit is t Determine the travel route planning range; Generate an initial path set within the planning scope, so that the single-run mileage is 0.5-4 times the cell radius and covers all NRIs in the cell. j ≥0.4 key nodes; The optimized driving model is constructed with the goal of minimizing the comprehensive risk value of the path, that is, the weighted sum of efficiency risk and safety risk, and maximizing the service response speed. In the optimized driving model, the idle vehicle cruising path is dynamically adjusted according to the traffic density, and the idle vehicles are preferentially parked in the NRI. j ≥0.7, and predict the impact of passengers' boarding and alighting behaviors on traffic density in the next 30 minutes based on Monte Carlo simulation; Execute the safety response mechanism. When it is detected that the traffic density in the cell reaches 80 vehicles / km, the abnormal event handling time is automatically extended to 1.2 times the original handling time, and the path re-planning instruction is triggered at the same time.

6. The method for assessing the operating risk of an autonomous taxi according to claim 5, characterized in that: Also includes: A closed-loop verification system is established to simulate the disturbance effect of path adjustment on actual traffic flow in a digital twin environment, update the parameters of the risk assessment model according to the simulation results, and apply the parameters to the cell radius R in the dynamic path planning framework in real time. t calculate; The dynamic path planning framework and real-time scheduling strategy are based on the travel demand density Q within the cell in the current period. t Data is coordinated, and the output data of the Monte Carlo simulation is used for both the site selection optimization of charging stations and the layout adjustment of docking points. The execution results of the path replanning instructions are fed back to the closed-loop verification system in real time to form an iterative optimization loop.

7. The method for assessing the operating risk of an autonomous taxi according to claim 6, characterized in that: The coordinated optimization of Monte Carlo simulation and dynamic routing strategy is achieved through the following steps: Input parameters: passenger boarding and alighting behavior data, with a Poisson distribution mean of 3 minutes and a standard deviation of 1 minute; energy replenishment behavior data, with charging time of 0.5-2 hours; Set up a queue overflow probability matrix for docking points to quantify the risk of vehicle detention at each docking point in different time periods; Set the load imbalance index of the road network around the charging station to reflect the fluctuation of traffic density within the service radius of the charging station; According to the queue overflow probability matrix, the connection point location and capacity are dynamically adjusted to migrate high-risk connection points to low-load areas; Based on the load imbalance index, the location of charging stations is optimized, and charging stations are preferentially deployed in areas with a road network load balance of ≥ 0.8; Among them, when the traffic density in the cell is ≥80 vehicles / km, the path replanning is automatically started; Prioritize road sections that have not reached the load threshold, that is, the traffic density is less than 60 vehicles / km; dynamically adjust the path weight coefficient, increase the efficiency risk weight by 30%, and synchronize the path replanning results to the digital twin environment.

8. The method for assessing the operating risk of an autonomous taxi according to claim 6, characterized in that: The established closed-loop verification system is as follows: The closed-loop verification system is built based on digital twin technology. The system interacts with the traffic data collection module, dynamic path planning framework, and risk assessment model that collect multi-source traffic data in real time. The traffic data collection module transmits the collected data to the closed-loop verification system; in the digital twin environment, a virtual scene consistent with the actual traffic scene is constructed based on real-time traffic data. The scene covers road networks, traffic lights, other traffic participants including pedestrians, and ordinary vehicles. At the same time, the initial path set and the adjusted path plan generated by the dynamic path planning framework are input into the digital twin environment; According to the adjusted path plan, the driving process of the self-driving taxi in the virtual scene is simulated to analyze the impact of the path adjustment on the surrounding traffic flow. The specific quantitative indicators include: Local traffic flow changes: calculate the traffic density change rate and average speed change value of surrounding road sections before and after the route adjustment; Evaluate the impact of route adjustment on average travel time and vehicle delay time within the entire cell; After the simulated route adjustment, the risk assessment model is used to evaluate the probability and frequency changes of conflicts between self-driving taxis and other traffic participants.

9. The method for assessing the operating risk of an autonomous taxi according to claim 1, characterized in that: The hidden state also includes a human-machine collaborative safety strategy, including the following steps: It is clear that the average response delay of security officers is 30 seconds, and the average time for handling abnormal events is 90 seconds; Establish a correlation between handling time and traffic density: when the traffic density in the cell is greater than or equal to 80 vehicles / km, the handling time will increase by 20% on the original basis; Safety risks will increase due to changes in response delays and disposal times, and the magnitude of the increase is positively correlated with the adjustment of disposal times; When the increase in safety risk reaches or exceeds a threshold of 30%, or the increase in efficiency risk reaches or exceeds another threshold of 20%, the upper limit of the density of self-driving taxi access in the cell will be lowered to 70% of the original value, and the weight of safety risk will be increased by 30%. At the same time, early warning information will be pushed to the remote monitoring center.

10. An evaluation system for the operation risk of an autonomous taxi, characterized in that: include: The road network dynamic division module is used to dynamically divide the urban road network into hexagonal honeycomb units according to real-time travel demand, where the dynamic radius R of each honeycomb unit is t By formula Calculation, R base is the base radius, Q t is the travel demand density within the cell in the current period, Q avg is the historical average demand density, Q max With Q min is the extreme value of demand density, α is the adjustment coefficient, and key nodes are selected in each hexagonal honeycomb unit; Multi-source data access module, used to obtain traffic density, number of conflict events, communication delay data and environmental parameters in real time; A key node risk assessment module is used to calculate the key node risk index NRI based on the multi-source data j , where NRI j NRI by formula j =ω1×C j +ω2×D j +ω3×A j Calculation, C j is the conflict rate, D j is the delay index, A j is the accident density, ω1, ω2, ω3 are the preset weights; The behavior response simulation module includes a hidden Markov chain model, which is used to define the hidden states S1, safe state, and S2, risk state. It discretizes the target detection confidence, obstacle avoidance response time, and V2V communication delay data of the self-driving taxi into observation states, constructs the state transition probability matrix P and the emission probability matrix E, and infers the probability of each hidden state at the current moment by combining the matrices P and E through the Viterbi algorithm; Comprehensive risk decision module is used to determine the risk of key nodes according to the NRI. j The low, medium and high risk levels are determined by the hidden state probability, and warning signals are output to dynamically adjust the path planning strategy of the self-driving taxi.

Images