Research On Path Planning And Trajectory Tracking Of Patrol Robot With Reconnaissance

Patrol robots are mostly used in communities,factories and other environments to replace mankind to perform tedious security inspection tasks.However,most of the existing patrol robots have technical problems.For example,the path planning part occupies a lot of computing resources,dynamic planning is liable to fall into local oscillations,and the trajectory tracking is poor in real-time.These problems make patrol robots unable to cope with tasks such as sudden reconnaissance.In response to the above problems,This paper develops a prototype of Patrol Robot with Reconnaissance(PRR)and carries out research on key technologies for path planning and trajectory tracking to improve the robot's autonomous ability when facing multiple security inspection environments.And it also provides a theoretical basis for the robots to perform tasks autonomously in the security and inspection scene.First of all,this paper analyzes the requirements of the current security and inspection scene,determines the performance requirements for motion,perception,planning and interaction that the robot should have,designs the PRR from the four subsystems of drive control,sensor communication,main processing and human-computer interaction,and finally develops a PRR prototype.The robot motion experiment proves that the robot prototype meets the basic motion requirements,and provides a basic platform for subsequent experiments and applications.Secondly,in order to solve the problems of existing robots in path planning,such as poor real-time performance and resource occupation,the related algorithms are applied and improved.Among them,Jump Point Search(JPS)algorithm is used for path planning in known map scenarios,and its effectiveness is proved through simulation and comparison.At the same time,in view of the characteristics of the unknown map,this paper proposes a Dynamic Jump Point Search(DJPS)path planning algorithm and corresponding planning strategies suitable for dynamic environments,which can plan dynamic paths in real-time under changing situations and avoid local oscillation problems.Simulation and analysis prove that the algorithm can improve the real-time and stability of robot path planning under the unknown map.Subsequently,in order to solve the trajectory tracking problem of the robot,a PRR kinematic model is established and an Model Predictive Control(MPC)controller suitable for the robot is designed.The effectiveness of the motion controller in trajectory tracking is proved by simulation.At the same time,to further improve the operation speed of the MPC algorithm,the Alternating Direction Method of Multipliers(ADMM)algorithm framework is used to optimize the quadratic programming part of the MPC algorithm.Through simulation and comparison,it is proved that the fusion algorithm can effectively reduce the calculation time of the quadratic programming part and improve the real-time performance of the robot trajectory tracking control.Finally,a joint experimental platform of PRR,ROS,and MATLAB is established,and robot path planning and trajectory tracking experiments are designed in indoor and outdoor environments.By analyzing and comparing the actual motion data and expected motion data of the robot,it is verified that the PRR prototype,the path planning algorithm and trajectory tracking algorithm developed in this paper meet the application requirements in the security and inspection scene.