Research On Multiple Optimization Strategies For RRT Path Planning Algorithms

In the complex and dangerous working environment,the mobile robot can not only ensure the personal safety of workers,but also create higher economic benefits.Path planning,which supports mobile robots in their tasks,is the process of finding a set of conflictfree paths in a given workspace and connecting starting points.Among many related strategies,the sampling-based algorithm Rapidly-exploring Random Tree(RRT)is widely used because of the low complexity of calculation and the fact that it does not have to model the surrounding environment,but this algorithm also has many disadvantages,such as the randomness of searching for paths,the paths obtained are not optimal or sub-optimal,there are many path inflection points,and the paths are not smooth enough.The algorithm is not smooth enough.Therefore,this article uses mobile robots as research subjects,aiming at the problems of RRT algorithm,various optimisation strategies are proposed and the main research is as follows.(1)To address the problem that the traditional RRT algorithm does not consider path safety and smoothness,a Safe and Smooth Rapidly-exploring Random Tree(SS-RRT*)algorithm is proposed.Firstly,the Quick-RRT* algorithm combines the proposed node safety strategy and path safety strategy to obtain an initial path.Then,by using a smoothing strategy based on improved Bezier curves,the initial path is smoothed to reduce path costs,improve the quality of the path at turning points,and achieve path smoothing.At last,the validity of the suggested method is proven by the modelling.(2)To tackle the problems of slow convergence of RRT* and non-optimal or suboptimal paths obtained,an improved bidirectional RRT* algorithm,Bi-RRT* is proposed by combining the greedy strategy of Bidirectional RRT with the idea of node backtracking and improving the traditional collision detection algorithm based on fork product.Firstly,the initial path is obtained by the basic algorithm framework combined with the improved collision detection algorithm;then after the post-processing procedures: path optimization strategy and path smoothing strategy,the indexes of path cost reduction and path quality improvement are achieved respectively.From the Matlab simulation and Gazebo simulation results,it can be seen that the Bi-RRT* method is effective in decreasing the planning time and increasing the quality of the path.(3)To address the problems of high path cost of traditional RRT algorithm and large computation when Bessel curve is used for path smoothing,a path planning algorithm based on circular arc rounding method named CAF-RRT* is proposed and applied to global path planning.The algorithm firstly expands the map to ensure the path safety.Then the proposed algorithm quickly obtains the initial path with lower path cost.Finally,the path segmentation is smoothed by using the circular arc rounding method to satisfy the smoothing requirement.The results of both the Matlab simulation and the Turtle Bot3 simulation indicate that the approach can significantly decrease the path cost of the algorithm and enhance the path’s quality.