High Speed Intelligent Vehicle Trajectory Planning Considering Dynamic Mobility Of Environmental Vehicles

Path planning is the core of intelligent driving technology research and the key link of realizing intelligent vehicle.At present,most path planning studies regard obstacles as static states.In this paper,considering the dynamic mobility of vehicles with obstacles ahead in the actual driving environment,a high-speed intelligent vehicle trajectory planning method is proposed based on the consideration of the dynamic mobility of vehicles in the environment.The main research work includes:(1)Suppport Vector Machine(SVM)algorithm was used to identify driving intentions of the leading car.The typical driving scenarios faced by intelligent vehicles are analyzed,and the corresponding road environment is built by using the Prescan software.At the same time,according to the smart car sensor configuration scheme,the smart car is equipped with a virtual sensor system in the Prescan to collect the status information of the surrounding vehicles during the driving process.The consistency of the sensor reference system is guaranteed by coordinate transformation calculation.Finally,the collected data are used as the feature vectors of the support vector machine after filtering and de-noising,so as to train the support vector machine model that can recognize the driving intention of the leading vehicle,which lays a foundation for the prediction of the leading vehicle trajectory based on driving intention in the following paper.(2)A trajectory prediction algorithm combining environmental vehicle motion state and driving intention is designed.According to the driving intention of the environment vehicle identified by the support vector machine,the trajectory cluster of the environment vehicle was programmed using the quintic polynomial in Frenet coordinate system.The trajectory constraint function was established based on the safety and comfort of the environment vehicle,and the optimal long-term trajectory of the environment vehicle was selected.At the same time,based on the current state information of the environment vehicle,the short-term driving trajectory of the vehicle ahead is predicted by lossless Kalman filtering.By selecting an appropriate weight function,the predicted long and short trajectories are fused to improve the accuracy of the forecast of environmental vehicle trajectories.(3)Improve the Artificial Potential Field(APF)algorithm applicable to structured roads and dynamic obstacles.At present,the traditional APF algorithm is easy to fall into the local optimal value,does not consider the dynamic obstacles,and is not suitable for structured roads.In this paper,the circular repulsive field of the traditional APF is optimized into an elliptical repulsive field,and the velocity repulsive potential field function related to the velocity of the obstacle is added,considering the road environment constraints of the intelligent vehicle in the process of driving.By analyzing the steering mechanism limitation and comfort constraint of intelligent vehicle,the constraint conditions of maximum turning Angle and its change rate are established.On this basis,aiming at the problem that the obstacles are usually set as static in APF algorithm,which cannot reflect the dynamic mobility of obstacles,according to the predicted trajectory of moving obstacles,the position of obstacles in APF algorithm is dynamically updated to realize the dynamic planning of collision avoidance trajectory.To verify the feasibility of the above dynamic path planning algorithm,a PrescanSimulink co-simulation model was designed,and the Pure Pursuit algorithm was used for trajectory tracking control.The simulation results show that the path planning strategy proposed in this paper can effectively plan the obstacle avoidance path with both safety and comfort on the basis of relatively accurate prediction of the future trajectory of environmental vehicles.