The Intelligent Vehicle Simulation Platform Development For Multi-vehicle Cooperative Sensing Trajectories

With the rise of the smart car industry and the advent of the era of big data Internet of Vehicles,smart cars and the derived smart car traffic data have begun to enter people’s field of vision.However,smart cars are still in the early stages of development.Only a small number of companies start to operate smart cars on certain road sections,and cannot obtain the proportion of various smart cars and the traffic data of smart cars in various traffic scenarios.On the other hand,smart car simulation is still in its infancy.They often focus on vehicle landing function testing,but ignore the microscopic realistic traffic flow simulation and the information sharing between cars,which leads to the general vehicle traffic behavior in the simulation process.Rough,unable to test the Internet of Vehicles communication and other issues.Therefore,it is very necessary to design a smart car traffic simulation platform for multi-smart cars to collect trajectories,and to build a smart car co-simulation framework with a complete process from scenario to algorithm testing in a simulation environment.It can support the construction of a variety of traffic scenarios,meet the repeated testing of special traffic scenarios,break away from the continuous use time of sensors and vehicles in reality,and improve the progress of smart car development and the exploration of smart car data performance.This article combines Gazebo and Sumo two simulation software to build a smart car traffic simulation platform.The specific implementation steps are as follows: First,the system architecture of the intelligent vehicle traffic simulation platform is established,and the functional modules are implemented and integrated in the software,including the road environment module,the micro-traffic flow module,the vehicle dynamics module,the sensor module,etc.;then this paper independently developed The communication module is used to realize the data flow between different simulation software to ensure the synchronization of the simulation;the front-end module is developed to modify the simulation parameters and enhance the visualization;the smart car algorithm module is developed to realize the coordinated perception of the trajectory of multiple smart cars based on the original lidar data.Among them,the smart car algorithm module is a system independent of the simulation platform,and the overall process of the algorithm module is designed according to the data characteristics of the smart car.First,the bicycle environment perception algorithm is used to obtain the road trajectory based on the original laser radar data,and then the same vehicle trajectory collected by different smart cars is matched based on the DS evidence theory.Finally,the Kalman filter algorithm is used to complete the multi-vehicle trajectory fusion to obtain multiple intelligences.The microscopic trajectory of the traffic after the cooperative perception of vehicles.The development of the simulation platform can be used to test and verify the functions of smart cars,and contribute to the research of smart car simulation and Internet of Vehicles.