Research And System Design Of Path Planning And Trajectory Tracking Methods For Unmanned Vehicles

With the development of unmanned technology,a variety of unmanned vehicles are sold on the ground,and most of unmanned systems are equipped with industrial personal computer(IPC)as computing platforms,but IPCs are expensive and large in size,and the adoption of less expensive embedded computing platforms is the trend of development in the coming years.If the unmanned system changes the computing platform,it will bring two problems.However,on the one hand,due to the high computational complexity of the traditional RRT path planning algorithm and the limited computing resources of the embedded platform,the real-time performance of the system can not be effectively guaranteed.On the other hand,the traditional controller based on model predictive control(MPC)has complex kinematic models and constraints,which will occupy too much computational resources and finally affect the accuracy of trajectory tracking.In this regard,this paper analyzes and designs the path planning and trajectory tracking system according to the system requirements,and this system is applied to the low-speed scenario in the park and can plan path in real time and track trajectory accurately.The research topic of this paper comes from the Guangxi innovation-driven development special fund project "Research on Intelligent Driving Technology of Electric Sightseeing Vehicles",and the main research contents are as follows.(1)In order to improve the real-time planning of the unmanned system based on the embedded platform,the RRT path planning method based on the state grid is proposed.Firstly,the state grid algorithm is used to generate the search extension domain that guides the RRT algorithm,which guides the search direction and reduces the computational complexity of RRT algorithm path,and then a random sampling point generation function is introduced to limit the area of the sampled nodes.Whenever a new node is generated,the Euclidean distance between its previous node is calculated to judge whether it is the nearest node to remove the node with too large distance between two points.Next,the excessive steering paths are pruned and the path further becomes smooth by the B-sample curve function to improve the real-time performance of unmanned vehicle path planning.Finally,the embedded platform-based unmanned system is built for path planning simulation and experimental vehicle testing.The experimental data results show that compared with the traditional RRT algorithm,the total path length of the random search tree is reduced by 14%,the average path length is reduced by 16% after pruning,and the average time spent is reduced by 64%.However,compared to the RRT and artificial potential field algorithm,the total path length of the random search tree is reduced by 8%,the average path length after pruning is reduced by 6%,and the average time spent is reduced by 33%.Compared with the previous two algorithms,the proposed algorithm not only reduces the computational complexity,but also has a smaller number of iterations,reaches the target point faster,and ensures the continuity of the path curvature.(2)For the problem of unmanned vehicle tracking desired trajectory and the accuracy of autonomous steering control are not high,a trajectory tracking controller of MPC based on Ackermann steering model is proposed in this paper.Firstly,Ackermann steering kinematic model and its constraints are constructed to solve objective function,and then tracking trajectory path points are performed 5 times polynomial fitting.Next,according to system requirements,the controller is processed by linear simplification to reduce the computational resource occupation,which can improve the accuracy of unmanned vehicle trajectory tracking.Finally,the experimental vehicle based on embedded unmanned vehicle driving system is built,and the experimental sensing layer is equipped with Velodyne16 LIDAR,industrial camera and IMU combined with inertial guidance.The embedded computing platform is selected to process the data to realize effective path planning and software and hardware design of trajectory tracking,and the system is verified by functional simulation and real vehicle test.The experimental data results show that the MPC controller based on Ackermann model has high accuracy in trajectory tracking.The trajectory tracking accuracy reaches 99.8% at a vehicle speed of 30km/h with continuous transverse yaw angle and acceleration curvature when the 5th polynomial curve fitting is used in this paper.