Research On Motion Planning And Tracking Control Of Nonholonomic Constrained Mobile Robot

Patrol mobile robot is an indispensable role in the process of modern industrial production,which realizes a high degree of autonomy and intelligence in the process of production inspection with low cost.the design of its efficient and autonomous navigation system is the key to the realization of intelligence.In the industrial environment,the layout of road equipment is complex,and dynamic objects such as new static objects or pedestrians often appear on the path.At the same time,the four-wheel configuration commonly used by patrol mobile robots has the characteristics of nonholonomic constraints,and its controllable degree of freedom is less than the degree of freedom of space motion,and can not move to any direction of space.The above two aspects of patrol process characteristics make the design of autonomous navigation system more complex.This paper will study and design the autonomous navigation system of patrol mobile robot under the application background of the above complex environment constraint characteristics and nonholonomic motion constraint characteristics of mobile robot.Among them,this paper will mainly carry out research from three aspects: improving the global planning ability of the robot navigation system,the path tracking control ability and the local planning ability in the multi-constraint environment.First of all,in terms of global planning,in order to improve planning efficiency and at the same time generate a global path that meets the motion constraints,this article will analyze the non-holonomic constraint characteristics of the environmental scene and the mobile robot motion model and improve the hybrid A~* algorithm in three aspects: the expansion mode of the child nodes,the setting of the heuristic function and the cost function.Through experimental test comparison,it reflects the superiority of the improved hybrid A~* algorithm designed in this paper in planning efficiency and planning results.In terms of path tracking control,in order to overcome the slip disturbance caused by the incomplete motion constraint of the inspection robot and improve the dynamic tracking performance of the robot,this paper first uses successive linearization to linearize the robot's nonlinear motion model to obtain the error model and the prediction model.Then,based on the model predictive control strategy MPC,the state and control variables are rolled and optimized according to the optimization goal,and the prediction time domain is dynamically solved.The robot optimizes the sequence of control variables.Through simulation and experimental analysis,the path tracking controller based on model predictive control designed in this paper can control the inspection robot to track the motion stably and has a certain ability to resist slip disturbance,and the path tracking accuracy reaches centimeter level.In terms of local dynamic programming,this paper models the new obstacle constraints in the unknown environment and uses approximate convexity to simplify its mathematical expression.Based on the model predictive control strategy,the dynamic planning process is combined with the tracking process.The robot motion model constraints and environmental obstacle constraints are integrated,and the local dynamic programming problem is described as a mathematical optimization problem.The MPC local planner can dynamically process robot constraints and environmental obstacle constraints in real time,and construct optimization problems based on optimization goals to dynamically solve the optimization state and control variables,so as to achieve obstacle avoidance motion at a small cost while tracking the reference path as much as possible.Finally,a non-holonomic constrained mobile robot experimental platform was built based on the robot development framework ROS system.Subsequently,a number of scenarios were designed to carry out the navigation performance experiment of path tracking and dynamic obstacle avoidance.Through actual navigation experiment test and analysis,the superiority of the tracking controller and local planner based on model predictive control designed in this paper is verified.