Using the Swarm Leader Concept for Multi-Agent Obstacle Avoidance

Abstract

The understanding of the operating principles and stability of natural swarms has recently proven to be a useful tool for the design and control of artificial robotic agents. In previous work [1], an algorithm that successfully enhances the navigation performance of a swarm of robots using the swarm leader concept was introduced. An extension to that work is presented in this paper. The real implantation of some applications mentioned in [1] is presented. The new topic in this paper is to use the simulations along with the experimental implementation using a swarming behaviour test-bed, which has been built from scratch to test the swarming behaviour for multi-agent systems. The main objectives are to design and fabricate a mobile robot test-bed, using lego kits as a basis to test the behaviour of 5-6 interacting mobile robots, and to develop a model to simulate the mobile robots‟ behaviour. These objectives help in designing a series of experiments to explore collective behaviour in swarm robotics with non-complicated way of communication between swarm members. Anarena has been installed to perform a range of experiments with the swarm. Through the use of signalling between the agents, such experiments can be developed to achieve high-level goals such as cooperative search tasks and assembly. Experimental and simulation results match in a way that confirms the efficiency of the algorithm as well as its applicability.