Research On Path Planning Of Mobile Robot Based On RRT Algorithm

With the steady development of today’s technological level and the continuous expansion of the robot industry,the number of intelligent robots appearing around people is also increasing.Nowadays,robots can not only meet our daily work and life needs,such as cleaning the ground,transporting goods,and navigating,but also help humans complete high-quality,high standard,and even high-risk tasks.As the most critical technology in the field of mobile robots,the solution of path planning problem is undoubtedly the key content of the majority of robotics scholars.Therefore,this article focuses on the research of sampling based path planning methods for mobile robots,and provides a detailed introduction to the principles and processes of traditional RRT algorithm,RRT Connect algorithm,and RRT * algorithm.We also write programs for these three algorithms and conduct simulation experiments in designed scenarios.Analyze their advantages and disadvantages one by one,and compare their performance to identify their main characteristics,laying the foundation for the design of subsequent improved algorithms.The main innovation points and workload of the thesis are as follows:1、Designed path planning for mobile robots based on variable probability constraints.In order to solve the problem of unclear path search direction and poor stability of random tree growth in the RRT Connect algorithm,a mobile robot path planning based on variable probability constraints was designed.Firstly,a target bias mechanism was introduced into the random sampling process of the RRT Connect algorithm,and a variable probability target bias sampling function was designed.Used to guide the growth direction of random trees,reduce the occurrence of lost paths,improve search efficiency,and balance search paths.Furthermore,Bézier curve is introduced for smoothing,which can optimize the path length to a certain extent while optimizing the path path.To verify the feasibility of the improved strategy,this thesis conducts comparative experiments with two classical sampling based algorithms on the Pycharm platform in two scenes.The experimental results verify the feasibility of the variable probability target bias sampling function and the effectiveness of Bézier curve processing.2、Designed a path backtracking planning for mobile robots based on dual tree elliptical constraints.To solve the problem of RRT * algorithm not being able to efficiently search for near optimal paths,a path backtracking planning for mobile robots based on elliptical constraints was designed.Firstly,the target bias mechanism from the previous design scheme is added to the sampling process,and the idea of dual tree elliptical sampling constraint is introduced.The purpose is to accelerate the convergence of the first search path,quickly generate an elliptical sampling set,and enable the random tree to search for new paths within this elliptical subset.Secondly,an adaptive dynamic step size and path backtracking processing strategy are adopted to optimize the searched paths and shorten the total length of the paths.To verify the effectiveness of the improved algorithm,comparative experiments were conducted on the Pycharm platform with three sampling based algorithms in two types of scenarios.The results confirmed the effective improvement of the search efficiency of the improved algorithm on the first path,and it also played a certain role in search time and iteration times,ensuring that the searched path was a near optimal path.3、Simulation experiments were conducted on mobile robots based on the ROS experimental platform.Research and build an ROS based experimental platform for learning in the Ubuntu virtual machine environment.After configuring the system environment,use Gazebo simulation software to generate the required 3D simulation map for the experiment.By combining SLAM technology with mobile robot simultaneous positioning and map environment building functions,the LiDAR sensor carried by the mobile robot itself is simulated.Scan the information data of the generated map using the SLAM based Gmapping algorithm and build a visual map.Finally,autonomous cruise exploration experiments were conducted on the built map of the mobile robot car,verifying the feasibility of the algorithm and laying the foundation for future physical experiments.This thesis proposes different improvement methods for two algorithms and conducts simulation comparative experiments in multiple scenarios.The results show that the improved algorithm in this article has advantages in multiple evaluation indicators;Finally,a simulation exploration experiment was conducted on the ROS experimental platform to verify the feasibility of the proposed improvement method.