Affiliation:
1. MIREA – Russian Technological University
Abstract
Objectives. The article substantiates the relevance of automatic docking of autonomous mobile robots. Specific examples show that the implementation of the automatic docking functions of autonomous robots reveals the potential for creating multi-agent systems with a transformable structure. The aim of the work is to develop means for automatic docking of autonomous mobile robots in complex scenarios and an uncertain environment.Methods. The proposed approach to automating autonomous mobile robot docking is reduced to a modification of the counter-growth rapidly-exploring random tree (RRT) method. It is based on the parallel execution of a decentralized route planning algorithm with mutual coordination of distributed computing processes. The effectiveness of the complex of algorithmic and software tools developed was evaluated using computer and natural simulation methods. The final series of full-scale experiments was carried out on the example of JetBot AI kit Nvidia platforms for automatic docking of autonomous robots. This was performed using the means and methods of intelligent control, visual navigation, technical vision and wireless network communication.Results. The study analyzed the features of automatic docking as one of the tasks of group control of autonomous robots. This is part of multi-agent systems, capable of reconfiguring structures for purposeful changes to the existing set of functional properties and application possibilities. The study also proposes a decentralized modification of the counter-growth RRT method. This allows the movements of autonomous mobile robots in the course of their mutual approach and subsequent docking to be planned. A set of software-algorithmic tools was developed to automate the docking of autonomous robots. A series of model and full-scale experiments were carried out to confirm the effectiveness of the approach developed herein.Conclusions. The modification presented herein of the counter-growth RRT method, traditionally used for planning the movements of manipulators and mobile platforms, is complementary to the tasks it resolves. This enables the docking of autonomous robots to be automated. The results obtained open up the potential for universal schedulers with extended functionality for autonomous robot control systems to be designed.