Research On Path Planning And Control Methods Of Low-speed Unmanned Vehicles In Limited Areas

With the rapid development of autonomous driving technology,autonomous driving has not only been applied to cars on highways,but has also been applied in various sub-scenarios,such as parks,ports,mining areas,and scenic spots.It has become an important carrier for scenarios such as "smart logistics" and "smart mining area",which improves operation efficiency,saves labor costs,and improves safety.Low-speed unmanned vehicles have a simpler structure and are easier to control than high-speed passenger vehicles.This paper designs path planning,obstacle avoidance and trajectory tracking algorithms for low-speed unmanned vehicles in parks,ports,and mining areas,and verifies the effectiveness of the algorithms through simulation experiments.The first chapter analyzes the research status of unmanned vehicles and lowspeed unmanned vehicles at home and abroad,and then analyzes the research status of global path planning algorithm,local trajectory planning algorithm and unmanned vehicle trajectory tracking control algorithm.The second chapter analyzes the hardware principles of the perception module,planning module and control module,analyzes the software architecture of low-speed unmanned vehicles.The various coordinate systems adopted by the unmanned vehicle are determined.The kinematic model of the unmanned vehicle is established.The third chapter mainly analyzes the principles of A~* global path planning algorithm and VFH series obstacle avoidance algorithm,and respectively carries out simulation experiments.The simulation proves that the A~* global path planning algorithm is capable of simple path planning tasks and the VFH+ algorithm can effectively avoid static obstacles in the environment without stopping.After the unmanned vehicle obtains the local trajectory,it needs to track the trajectory.Chapter 4 designs pure pursuit and model predictive control algorithms,and simulates the two algorithms respectively.The pure pursuit algorithm is simple to implement and robust.Simulations prove that the pure pursuit algorithm can track the global path well;the model predictive control algorithm has a large amount of calculation and model predictive control algorithm is hard to implement.Compared with the pure pursuit algorithm,model predictive control algorithm has smaller lateral tracking error and heading angle error.Chapter 5 customized the ROS navigation stack,replacing the local trajectory planning function package with trajectory tracking control algorithm function package.Combined with the Stage Simulator,the global path planning algorithm and the trajectory tracking control algorithm are jointly simulated,which verifies the feasibility of the algorithms.Finally,the whole process simulation experiment is performed.The global path planning algorithm uses the A~* algorithm,the local obstacle avoidance algorithm uses the VFH+ algorithm,and the trajectory tracking control algorithm uses the pure pursuit algorithm.The simulation results show that the algorithms can work well.