Research On Path Planning Algorithm Of Mobile Robot In Dynamic Environment

In recent years,the level of intelligence and automation of mobile robots has been constantly improved,and it has been widely used in agriculture,logistics,medical care,military and other aspects,which has greatly improved the work efficiency and quality.Path planning as a key part of mobile robots to perform tasks safely and efficiently,has important research significance.The purpose is to find a reliable obstacle avoidance route for the robot from the starting point to the task point.There are often dynamic obstacles in the actual task scenario,which can pose a safety threat to the robot if it is unable to adopt a timely strategy for avoidance.Therefore,the path planning problem in dynamic environments needs to be studied.The main research of this paper includes:Firstly,the principles and characteristics of several commonly used environmental modeling methods are introduced,and the merits and demerits are compared and analyzed.The grid method with accurate location description,high precision and easy to update and maintain is selected as the environmental modeling method.Path planning is divided into static programming for known environment and dynamic programming for unknown environment and its principles are illustrated.The basic theories and advantages and disadvantages of several commonly used algorithms are analyzed,and A*algorithm with short time and strong optimization ability and rolling window method with more flexibility are selected as the basic algorithms.Secondly,in view of the static obstacle avoidance path planned by the traditional A*algorithm,there are many turns,easy collision and sub-optimal length,and so on,an improved A*algorithm is proposed.In order to reduce redundant turning points,shorten the path length and improve the path security,the algorithm is improved by dynamically adjusting the strategies of searching neighborhood and screening security nodes by collision detection.Based on this,the optimal path between two points is generated by smoothing the optimized path corners with a third-order Bezier curve with continuous curvature.Then,aiming at the dynamic obstacle avoidance problem of single task point,a dynamic obstacle avoidance algorithm based on rolling window is proposed to avoid unknown moving obstacles in real time.First,a mathematical model is built to predict the movement path of obstacles through the position and speed information detected in the window to determine whether there is a collision.Then according to the collision situation to develop collision avoidance strategy,timely and effective to eliminate the conflict,to ensure the safety of the robot.Finally,the multi-task point path planning problem is studied,and a hybrid algorithm is proposed to realize static and dynamic obstacle avoidance.Based on the global planning problem,the improved A*algorithm is used to plan the obstacle avoidance path between any two points and generate the distance matrix,which is then combined with the ant colony algorithm to clarify the priority and generate the global path of multiple task points to achieve static obstacle avoidance.Then the dynamic obstacle avoidance algorithm is invoked to eliminate the collision risk and achieve dynamic obstacle avoidance,so that the robot can safely reach all the task points.Simulation results show that the hybrid algorithm is effective.