34 results about "Social force model" patented technology

The invention belongs to the field of artificial intelligence and intelligent traffic systems, particularly relates to a path control system combining multi-modal perception and dynamic trajectory prediction, and aims to solve the problems of insufficient perception fusion, low trajectory prediction precision and control response lag in a complex dynamic environment. The system comprises a multi-modal perception fusion unit, a dynamic interaction modeling unit, a space-time coupling prediction unit, a risk field construction unit and an adaptive path generation unit. Multi-source sensor data are fused through confidence coefficient weighting, an interaction weight matrix under an attention mechanism is constructed, a future trajectory is predicted in combination with individual dynamics and a social force model, a four-dimensional space-time risk field is generated, and a minimum risk path is solved based on an improved fast marching algorithm. The system realizes sensing-prediction-control closed-loop cooperation, the end-to-end delay is less than 250 milliseconds, and the driving safety and comfort in a complex traffic scene are improved.

The invention is suitable for the technical field of crowd simulation and emergency evacuation, and provides an emotional infection and social force model coupling-based crowd evacuation method, which comprises the following steps of: constructing an emotional infection model based on an SEIRS framework, dividing individual emotional states into a susceptible state, an exposed state, an infected state and a recovery state, the infection rate and the recovery rate are dynamically adjusted through a feedback function; establishing a social force model, and embedding an emotion value into calculation of a target driving force and a repulsive force between individuals to realize real-time adjustment of the motion behavior by the emotion; global path planning and an ant colony pheromone mechanism are fused, a density perception cost function and pheromone guiding force are introduced, and dynamic blockage avoidance is achieved; a bidirectional feedback mechanism is designed, and local path reconstruction is triggered when congestion is detected. According to the method, bidirectional coupling of emotion propagation and behavior decision is realized, the authenticity and adaptability of evacuation simulation are improved, and a scientific basis can be provided for emergency management and evacuation strategy optimization in public places.

The invention provides a traffic organization method and system based on urban traffic transfer efficiency evaluation, and the method comprises the steps: building a building model based on a pre-obtained urban traffic building space, setting an initial traffic organization scheme of traffic nodes, simulating the movement behaviors of passengers at each traffic node through a social force model, and carrying out the simulation of the passenger movement behaviors at each traffic node; obtaining passenger flow simulation data; setting an initial weight for the factor set based on the evaluation model to obtain a comment set; and dynamically adjusting the initial weight of the factor set based on the passenger flow simulation data and the evaluation model to obtain the target weight of the factor set, calculating the comprehensive association degree in combination with the comment set, and obtaining the urban traffic transfer efficiency evaluation result according to the comprehensive association degree to adjust the initial traffic organization scheme. And obtaining a target traffic organization scheme to organize urban traffic. According to the method, different passenger flow scenes are simulated to dynamically adjust the weight of the evaluation index, the urban traffic transfer efficiency level under the multi-factor interaction influence is quantified, and the reliability of a traffic organization scheme is improved.

The invention relates to the field of artificial intelligence, and discloses a charging robot intelligent scheduling and control method based on visual perception. According to the method, a double-layer architecture integrating global planning and local decision making is constructed; a central scheduling system generates a global conflict-free space-time path; each robot constructs a local three-dimensional semantic map in real time through a visual sensor, and predicts a dynamic target trajectory by using a long and short term memory network fused with a social force model; and when a collision risk is detected, triggering a decentralized negotiation protocol, autonomously re-planning a local obstacle avoidance path according to the dynamic priority, and synchronizing the state to a central system. The system comprises a global path planning module, a visual perception module, a trajectory prediction module, a risk decision module, a conflict negotiation module, a local re-planning module and a state synchronization module. According to the method, millisecond-level local response and global collaborative optimization are realized, and the safety, the task continuity and the system expandability are remarkably improved.

The application discloses a kind of electric bus active suspension control method and system, method includes according to bus parameter and motor distribution establishes the half vehicle dynamics model of bus, constructs the state space equation of model, obtains the active suspension model suitable for bus;Considering the mass distribution change and centroid position change caused by bus passengers getting on and off and walking in the car, a social force model is established;And through the social force model, the relationship between the number of passengers on the bus and the front and rear distribution of passengers is obtained according to the particle swarm optimization algorithm, the passenger distribution law is obtained, and the active suspension model suitable for mass distribution change and centroid position change is established after the range of uncertain parameters is determined;For the direct interference of rear motor vibration, PID control is designed to adapt to the vibration of motor under different working conditions;By deriving the theorem of H2 / H ∞ control, the active suspension feedback control law when the mass distribution changes and the centroid position changes is obtained.The application can effectively improve the comfort of bus.

The invention relates to the technical field of digital twinning, in particular to a layout generation optimization method and system for a digital twinning body of a space facility, and the method comprises the steps: constructing a two-dimensional data matrix for each layout scheme of the space facility through employing the evaluation indexes of a preset number of channels in a discretization grid; constructing a social force model based on the two-dimensional data matrix, and determining a layout cost index; analyzing evaluation indexes about efficiency safety, energy consumption and a flexible channel in the two-dimensional data matrix to obtain a multi-dimensional evaluation vector, obtaining a dynamic weight vector by adopting a Bayesian network, obtaining a close degree score by utilizing a decision analysis method, obtaining scene adaptation robustness in combination with an element numerical value dispersion degree of the dynamic weight vector, and determining a fitness function, so as to obtain the scene adaptation robustness. And adopting a particle swarm optimization algorithm to obtain an optimal layout scheme. The invention aims to enhance the comprehensive performance of the space facility layout scheme.

The invention belongs to the technical field of building emergency evacuation path planning, and particularly provides an intelligent building-oriented emergency evacuation path dynamic planning method and related equipment thereof, and the method mainly comprises the steps: obtaining building real-time monitoring data, carrying out the preprocessing of the real-time monitoring data, and obtaining building security situation data; performing threat diffusion simulation and region division on the building security situation data to obtain safety region division data; and performing personnel cluster analysis and social force model prediction based on the building security situation data to obtain personnel evacuation behavior prediction data and evacuation bottleneck early warning data. According to the invention, accurate adaptation of the evacuation path with the on-site threat diffusion trend and the real-time change of personnel flow is realized, personnel are prevented from falling into an unavoided threat area, local congestion is reduced, and the overall evacuation time consumption is effectively shortened.

The application discloses an indoor dense crowd emergency evacuation method considering a fragment scattering effect, belongs to the technical field of emergency evacuation, and comprises the following steps: a group intelligence algorithm and a fragment scattering effect social force model are used to calculate different optimal evacuation exit schemes corresponding to a sudden event and partition crowds under different conditions; site, partition crowds and sudden event information are taken as input, and the optimal evacuation exit scheme is taken as output; all the input and output are used to train a decision tree model to obtain a real-time emergency evacuation decision model; indoor partition crowd information under a sudden event is collected, and the information is input into the real-time emergency evacuation decision model together with the site and sudden event information to obtain the optimal evacuation exit scheme; and the optimal target exit of each partition in the optimal evacuation exit scheme is transmitted to a drone of each partition, and the drone uses time-sharing broadcasting to guide the partition crowds to evacuate to the optimal exit.

The application provides a simulation method and device for subway passenger behavior and electronic equipment, and belongs to the field of data processing. The method comprises the following steps: performing grid processing on a subway model scene to obtain a grid map; determining a target moving path in the grid map according to an initial position and a preset destination of a target passenger; obtaining a current state of the target passenger, wherein the current state is used to indicate a boarding process; determining a behavior decision of the target passenger through a preset behavior tree according to the current state, wherein the behavior decision is used to indicate a next state of the target passenger; and simulating the behavior of the target passenger according to the behavior decision, the target moving path and a social force model. The simulation accuracy and authenticity can be improved.

The application relates to a pedestrian motion trajectory evolution method and system based on a social force model, and the method comprises the following steps: setting a simulation environment, obtaining the initial position of pedestrians from a video; obtaining the actual force of the pedestrians at the collision moment and the non-collision moment according to a pedestrian dynamics model, updating the speed of the pedestrians according to the speed change of the pedestrians with time, and then updating the position of the pedestrians; obtaining the pedestrian motion trajectory according to the position of the pedestrians at each moment; wherein the construction process of the pedestrian dynamics model is as follows: obtaining the interaction force between the pedestrians and other individuals and the interaction force between the pedestrians and walls based on the social force model; constructing the social force, constructing the physical force according to the contact or collision action between the pedestrians and other individuals, and constructing the speed change of the pedestrians with time; and constructing the pedestrian dynamics model. Compared with the prior art, the application can more accurately reflect the actual force condition between the pedestrians under different situations, and the dynamics analysis is more in line with the actual situation.

To accurately predict the position of a moving body.SOLUTION: The information processing apparatus includes a controller. In the prediction of the position of the first moving object based on an SFM (SocialForceModel), when it is determined that the plurality of second moving objects are a moving object group in which the plurality of second moving objects are moving together in the same direction, the control unit adjusts a repulsive force received by the first moving object from the plurality of second moving objects to be a repulsive force received from one second moving object.SELECTED DRAWING: Figure 1

The invention discloses an electric bicycle intrusion behavior simulation method based on reinforcement learning parameter self-adaption, and the method comprises the following steps: obtaining environment parameters, initializing a reinforcement learning agent, and enabling the motion space of the reinforcement learning agent to comprise the hyper-parameter adjustment amount of a social force model; iteratively training the reinforcement learning agent until a convergence condition is reached to obtain a simulation model, acquiring all vehicle states and updating a state space by the reinforcement learning agent in each iteration, then calculating a corresponding action space, and meanwhile, calculating an electric bicycle intrusion decision probability according to the state space, and the social force model calculates the social force of the electric bicycle according to the intrusion decision probability of the electric bicycle after adjusting hyper-parameters in the action space, and updates the state of the electric bicycle. According to the method, through deep fusion of reinforcement learning and a social force model, a complete simulation-evaluation-learning closed-loop optimization framework is constructed to realize parameter adaptation, decision dynamics and multi-scene adaptation.

The present application relates to the technical field of intelligent livestock monitoring, and discloses a flock density dynamic monitoring method based on multi-camera fusion, comprising the following steps: S1, acquiring a flock dynamic image sequence through a plurality of cameras, modeling the interaction force, expected speed and target migration trend between flock individuals based on a social force model, generating group behavior dynamics characteristic data, and dynamically detecting the in-out flow of sheep in the boundary area of the camera field of view, and outputting density distribution data after boundary compensation; S2, the group behavior dynamics characteristic data generated in S1 is input into a space-time joint deep learning compensation model to obtain the density distribution data of the flock in the camera field of view. Through the group behavior dynamics modeling and space-time joint deep learning compensation technical scheme, the present application achieves the technical effect of continuous and stable density estimation in a dynamic scene, and solves the problem of boundary in-out occlusion and significant density jump caused by the rapid movement of the flock, compared with the scheme of relying on fixed camera image splicing and target detection in the prior art.

The invention relates to the related technical field of scene generation, in particular to a scene generation method based on a physical guidance diffusion model, and the method comprises the steps: processing an automatic driving data set, segmenting the data set into a training set, a verification set and a test set, and defining an intelligent agent track to be composed of a state sequence and an action sequence; capturing the time, space and map-level dependency relationship of the trajectory by using a space-time Transform; a track is generated through forward diffusion noise adding and backward diffusion denoising processes, a social strength model containing self-driving force, repulsive force and attraction is injected into the backward diffusion process, and track adjustment is guided to conform to physical laws and human-like driving behaviors. According to the method, the problem that the scene is not real due to the fact that physical constraints are ignored or only post correction is carried out in the prior art is solved, and the generated scene has authenticity and feasibility and can be effectively used for automatic driving system test evaluation.

The invention relates to the technical field of intelligent animal husbandry monitoring, and discloses a sheep flock density dynamic monitoring method based on multi-camera fusion, and the method comprises the steps: S1, collecting a sheep flock dynamic image sequence through multiple cameras, and carrying out the modeling of the interaction force, expected speed and target migration trend among sheep flock individuals based on a social force model, generating group behavior dynamic characteristic data, performing dynamic detection on sheep access flow in a view field boundary area of a camera, and outputting density distribution data after boundary compensation; and S2, performing dynamic feature data on the group behaviors generated in the step S1. According to the method, the technical effect of continuous and stable density estimation in a dynamic scene is achieved through the technical scheme of group behavior dynamics modeling and spatio-temporal joint deep learning compensation; the defects of boundary in-out shielding and obvious density jump caused by rapid movement of the sheep flock are overcome.

This invention relates to a method and system for assessing crowd collision risk in a sloping micro-road network scenario. The method includes the following steps: in the target sloping micro-road network scenario, a multi-view video acquisition network covering the main movement channel of the crowd is used to acquire a crowd movement video stream; pedestrian visual information is extracted from the crowd movement video stream using a target detection algorithm; the pedestrian visual information is input into an improved social force model after local parameter calibration to obtain the total force experienced by each pedestrian in the target sloping micro-road network scenario; a global-level dynamic risk index is calculated based on the total force experienced by each pedestrian; based on the global-level dynamic risk index, the crowd collision risk in the target sloping micro-road network scenario is assessed in real time using preset multi-level risk judgment rules to obtain the assessment result. Compared with the prior art, this invention improves the accuracy, reliability, and practical value of the crowd collision risk assessment results.

This invention discloses a civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity. The method divides the cabin into seating, aisle, and exit areas and sets independent target point sets. A two-stage screening technique is employed to determine effective action edges, avoiding abnormal repulsive forces at aisle entrances. Differentiated social force model parameters are established to address mechanical deadlock and lateral oscillation issues. A priority passage mechanism triggered by a waiting threshold is used to overcome the problem of permanent waiting. The kinetic energy is kept constant during area transitions through kinetic energy conservation and transformation, addressing the issue of acceleration lag during turns. An aisle boundary constraint protection mechanism is set up to correct positions and velocities exceeding the boundaries in real time, eliminating lateral oscillations. Verified through personnel evacuation tests under different operating conditions, the average error between the simulation results and the actual evacuation time is controlled within 10%, significantly improving the realism and reliability of civil aircraft emergency evacuation simulations and providing a scientific calculation tool for airworthiness certification and cabin safety assessment.

The application discloses an intelligent intersection pedestrian protection system based on trajectory prediction and intention recognition, and through deployment of a perception layer, an edge computing layer and a communication layer, a perception network of multi-source data fusion is constructed to realize accurate detection and continuous tracking of intersection pedestrians and vehicles; a multi-agent trajectory prediction model is established based on a social force model, target driving force, pedestrian interaction and vehicle avoidance factors are comprehensively considered, and future motion trajectories of pedestrians and vehicles are generated; a pedestrian's crossing intention is identified by analyzing the pedestrian's multi-modal behavior characteristics; a collision risk level is evaluated based on the trajectory prediction and intention recognition results; and a differentiated active intervention measure is taken by an execution layer according to the risk level, including signal timing optimization, sound and light warning and vehicle braking instruction sending, so that intelligent active protection of the vulnerable groups at the intersection is realized. The application realizes a paradigm shift from passive protection to active protection and provides all-round and intelligent safety protection for the vulnerable groups at the intersection.

The application relates to a non-motor vehicle track reconstruction method in a motor-non-motor shared space, and the method steps comprise the following: extracting an initial track through a CV algorithm; adopting an LSTM to predict and complete a track missing due to occlusion, and matching the track before and after occlusion through a Hungarian algorithm to complete a complete track; adopting a three-layer cyclic social force particle filtering algorithm constructed by integrating a social force model representing non-motor vehicle driving behavior into a particle filtering framework to denoise the completed complete track to obtain a reconstructed track. Compared with the prior art, the application can effectively improve the performance of a video detector in acquiring high-precision track data, provide more fine perception data for a holographic intersection and a traffic control system, and has great application value in the industry.

The application provides an improved social force model parameter tuning method, device and equipment and a storage medium, and relates to the technical field of virtual group simulation. The method comprises the following steps: obtaining a virtual group simulation target scene and evaluation indexes determined according to the target scene; setting multi-objective particle swarm optimization algorithm parameters by taking interaction force parameters A, n, n', gamma and lambda in the improved social force model as decision variables and taking the evaluation indexes as a target function; setting virtual group simulation experiment parameters for the target scene; and completing evaluation of particles in the multi-objective particle swarm optimization algorithm through virtual group simulation experiment of the target scene to obtain calibrated values of the A, n, n', gamma and lambda under the optimized target scene. The method can quickly obtain optimal model parameters under different simulation scenes, and virtual group simulation using the optimized model parameters can improve the motion interaction effect of the virtual group.

The application is suitable for the field of crowd simulation and emergency evacuation, and provides a crowd evacuation method based on emotion contagion and social force model coupling, which comprises the following steps: constructing an emotion contagion model based on SEIRS framework, dividing individual emotion state into susceptible state, exposed state, infected state and recovered state, and dynamically adjusting infection rate and recovery rate through feedback function; establishing a social force model, embedding emotion value into target driving force and individual repulsive force calculation, realizing real-time adjustment of emotion on movement behavior; fusing global path planning and ant colony pheromone mechanism, introducing density perception cost function and pheromone guiding force, realizing dynamic congestion avoidance; designing a bidirectional feedback mechanism, triggering local path reconstruction when congestion is detected. The application realizes the bidirectional coupling of emotion transmission and behavior decision, improves the reality and adaptability of evacuation simulation, and can provide a scientific basis for public place emergency management and evacuation strategy optimization.

The invention discloses a pedestrian trajectory prediction method based on intention assistance in a man-machine collision avoidance scene. The method comprises the following steps: S1, carrying out mathematical definition on pedestrian intention prediction and trajectory prediction; s2, decomposing a pedestrian intention prediction problem into two sub-problems of short-term intention direction prediction and long-term intention end point prediction; s3, predicting the intention direction of the pedestrian based on a social force model; s4, performing pedestrian intention end point prediction based on a naive Bayes method; s5, modeling a pedestrian motion track based on a Gaussian process; and S6, pedestrian trajectory prediction and trajectory correction are carried out based on Gaussian process regression, so that the pedestrian intention and the historical trajectory are fused to predict a future trajectory. In this way, the reliability of the prediction result is improved, the mobile robot can well avoid moving pedestrians, and the risk of man-machine collision is reduced.

The invention discloses an automatic driving vehicle full-speed-domain lane changing method based on longitudinal and transverse coupling, which is suitable for acceleration and deceleration lane changing scenes of common roads and ramps and comprises the following steps: 1, collecting states of a vehicle and surrounding vehicles in real time, and constructing virtual boundary vehicles in a sparse scene; 2, constructing a longitudinal merging feasible region according to a comfortable and safe time headway, introducing a lane changing urgency factor into a ramp scene to dynamically compress a safety boundary, and determining an optimal merging point according to a minimum longitudinal error; 3, calculating a longitudinal acceleration by using a nonlinear longitudinal controller; 4, generating a transverse reference point in combination with the longitudinal merging progress; calculating transverse acceleration by adopting an improved social force model; 5, introducing a tire friction circle theory to carry out dynamic coupling correction, and dynamically cutting the transverse acceleration according to a longitudinal priority principle; and 6, repeating the steps until a lane changing target is achieved. According to the method, full-speed-range lane changing control can be unified, so that the driving stability and safety can be effectively improved.

This invention discloses an emergency evacuation method based on the dynamic coupling of cellular automata and social force models. The method first obtains the coordinates of pedestrians within the evacuation space and constructs a Voronoi diagram. Next, it calculates the area of ​​Voronoi cells with physical boundary constraints and introduces a pedestrian radius *r* correction term for lower bound constraints. It then calculates the local density within the evacuation space and introduces an exponential smoothing method to correct the local density. Finally, it designs the weight function of the SFM model; calculates the weights of the CA model based on the SFM model weights; and weights the SFM acceleration and the CA expected velocity correction term to obtain the final acceleration of each individual. Based on the final acceleration, it updates the individual's velocity and position. This invention can better reflect the realism of the microscopic behavior of evacuated individuals and achieve higher evacuation efficiency.

The application relates to the technical field of unmanned aerial vehicle trajectory planning, and provides a multi-unmanned aerial vehicle trajectory optimization method in post-disaster emergency communication and related equipment, which comprises the following steps: modeling the movement state of each ground user equipment based on the current position of all ground user equipments, obtaining a user social force model of each ground user equipment, and constructing a reward function according to the user social force model; constructing an unmanned aerial vehicle trajectory constraint condition according to the user social force model of all ground user equipments and the takeoff airport position of all unmanned aerial vehicles; optimizing the trajectory of the unmanned aerial vehicle based on the reward function under the constraint of the unmanned aerial vehicle trajectory constraint condition, obtaining the running trajectory of the unmanned aerial vehicle in a communication time period; and controlling the movement of the unmanned aerial vehicle according to the running trajectory of the unmanned aerial vehicle, so that the emergency communication of the ground user equipment is realized. The method can improve the rationality of the unmanned aerial vehicle trajectory in post-disaster emergency communication.

The application discloses a kind of based on correction order parameter's unmanned aerial vehicle cluster motion consistency evaluation method, this method includes: step one: establish unmanned aerial vehicle dynamics model;Step two: unmanned aerial vehicle dynamics model is mapped as particle model, realize the conversion between unmanned aerial vehicle dynamics model and particle model;Step three: unmanned aerial vehicle individual control law based on social force model;Step four: generate cluster steady-state motion and transition state data of transformation;Step five: calculate individual correction order parameter.The application can improve the evaluation mode of unmanned aerial vehicle cluster motion consistency, establish more scientific order parameter evaluation index for the possible surrounding, consistent turning and other motion modes of unmanned aerial vehicle cluster, improve the evaluation performance of unmanned aerial vehicle cluster consistent motion.

The invention relates to the technical field of automatic driving, and provides an automatic driving vehicle control method, device and equipment based on a social force model, and the method comprises the steps: recognizing the information of an environment object main body included in obtained environment data, and dividing the environment object main body into a static object main body and a maneuvering object main body; after the information of the environment object main body is tracked and a traffic scene is constructed, future behavior information of each maneuvering object main body is predicted based on a social force model; determining a plurality of planned paths of the autonomous vehicle in the traffic scene according to the driving information of the autonomous vehicle, the future behavior information and the information of the tracked static object main body; and determining an optimal path and generating a control instruction of the automatic driving vehicle. The method provided by the invention not only can accurately predict the future behavior information of the non-motor vehicle, but also can adapt to complex and changing traffic scenes, has higher adaptability and flexibility, and can process more unknown and abnormal conditions.

The invention discloses an indoor dense crowd emergency evacuation method considering a fragment scattering effect, and belongs to the technical field of emergency evacuation, and the method comprises the steps: employing a swarm intelligence algorithm and a fragment scattering effect social force model to calculate different optimal evacuation exit schemes corresponding to emergencies and subarea crowds under different conditions; the site, the subarea crowd and the emergency event information serve as input, the optimal evacuation exit scheme serves as output, all the input and the output are adopted to train the decision tree model, and a real-time emergency evacuation decision model is obtained; collecting crowd information of each partition in a room under emergencies, and inputting the crowd information, the site information and the emergencies information into a real-time emergency evacuation decision model to obtain an optimal evacuation exit scheme; and the optimal target exit of each partition in the optimal evacuation exit scheme is transmitted to the unmanned aerial vehicle of each partition, and the unmanned aerial vehicle guides the crowd of the partition to evacuate to the optimal exit by adopting time-sharing broadcast. The scheme has the advantages that the planned evacuation path is accurate, and the field evacuation efficiency is high.

The application discloses a sports event crowd gathering risk prevention and control system based on situation awareness, and relates to the technical field of intelligent security and protection, comprising: a dynamic stress modeling unit, which is used for establishing a social force model, converting instantaneous velocity vectors in a crowd motion vector field into interactive stress parameters, and calculating local physical pressure; a risk evolution assessment unit, which is used for calculating local pressure gradient based on the interactive stress, identifying kinetic energy accumulation areas with opposite vector directions through a hedging vector detection algorithm, and determining a risk level; and a closed-loop guidance feedback unit, which is used for automatically solving evacuation guidance vectors based on the spatial distribution of the pressure gradient, and driving an external pointing terminal to perform dynamic guidance. High-density crowds are regarded as compressible non-Newtonian fluids, and physical quantities such as velocity vector divergence and pressure gradient are introduced, so that advanced prediction and closed-loop intervention of crowd hedging risks are realized. The application can be widely applied to risk prevention and control in high-density crowd gathering scenes such as large-scale sports events.

The invention provides an intersection mixed traffic optimization method and system, and the method comprises the steps: firstly recognizing pedestrians and bicycles in a traffic flow, and extracting microcosmic parameters, such as a track and flow; then, a high-fidelity multi-agent simulation model fusing a social force model and an intelligent driver model is constructed to reproduce high-density mixed traffic flow features; on the basis, a key optimization factor is selected to carry out an orthogonal experiment so as to achieve the goals of minimum overall delay of the intersection, optimal traffic efficiency and highest traffic safety. And finally, performing multi-target optimization on the multi-agent simulation model by constructing a comprehensive evaluation index to obtain a model optimal scheme. According to the method, the authenticity and pertinence of traffic simulation are improved, the traffic efficiency can be effectively optimized, the delay duration and the traffic conflict rate are reduced, and scientific and reliable decision support is provided for intersection planning design and management optimization in a mixed traffic scene.