Method for planning the way of UGV using a modification of dynamic bi-directional RRT algorithm.

Abstract

The path planning problem of unmanned autonomous ground vehicles has always been an acute problem in the field of autonomous ground robotic systems research. From the point of view of the complexity of the tasks during the conduct of combat operations in the urbanized space of the densely built-up city with a constantly changing landscape, the tasks entrusted to the UGV are constantly becoming more difficult, and the scenarios of the use of the UGV demonstrate a diversified trend in the development of new management and control systems. The rapidity of conducting modern military operations in an urbanized space appears as a complex and multifaceted task. Considering the complexity of the UGV movement process, at the current stage of the development of robotic systems, the general trend is to abandon remote control of robotic complexes with the transition to automatic modes, which requires the development and implementation of algorithms for automatic interaction and movement of military mobile robotic systems. To solve the problems of the Rapidly-Exploring Random Tree *Fixed Nodes algorithm regarding its low speed for obtaining track junctions and the impossibility of using it in a dynamic environment when planning a UGV path. To solve the problem of accelerating the acquisition of collision-free paths in real time in two-dimensional space, it is proposed to apply a modified dynamic bidirectional RRT* algorithm with reference nodes. The algorithm is a modification of Rapidly-Exploring Random Tree *Fixed Nodes using a bidirectional greedy search method to speed up and solve the problem of the unidirectional Rapidly-Exploring Random Tree algorithm regarding its slow search speed, as well as the difficulties of decision making in a narrow environment caused by blind random sampling. In the case of dynamic disturbance movement, taking advantage that reference nodes do not require much calculation in planning, in the process of iterative path optimization, the proposed algorithm updates the map information in real time and repairs the damaged output path to complete the dynamic path planning.