Key points are not available for this paper at this time.
Within the field of robotics, especially in the multi-agent systems research, there are multiple solutions or proposals based on centralized systems that facilitate group control. How-ever, this centralization is a potential vulnerability to failures of the major system and a limitation in scaling or reducing the group members, as it is with swarm systems. Indeed, for swarms, having the same control system in each agent facilitates their development and scalability. The development of swarm-type robotic systems is still an emerging field full of challenges. One of these is the agent ability to avoid obstacles in unknown environ-ments. Therefore, to solve it, a decentralized control architecture based on subsumption is proposed in this work. This architecture applies the VFH + algorithm for obstacle avoidance and agent inter-collisions. A virtual test-bench for multi-robotic systems investigation is developed to evaluate the application of swarm behaviors and the avoidance algorithm. It integrates MATLAB development tools, ROS-GAZEBO and the TurtleBot3 robotic platform. In this virtual environment, several tests are executed to verify the operation of the proposed control architecture and the flexibility and scalability of the swarm system.
Ortega et al. (Thu,) studied this question.