DYNAMIC BIOINSPIRED NEURAL NETWORK FOR MULTI-ROBOT FORMATION CONTROL IN UNKNOWN ENVIRONMENTS

Jianjun Ni, Xiaofang Yang, Junfeng Chen, and Simon X. Yang

Keywords

Formation control, multirobot system, bioinspired neural network, SOM neural network

Abstract

Formation control is a challenging and critical issue in robotics, which is a typical multi-robot cooperation problem. In this study, the formation control task is divided into calculating the expected locations in a required formation, assigning these locations among the robots in the system, and navigating each robot to its expected location. A novel approach based on a dynamic bioinspired neural network is proposed for real-time formation control of multi-robots, where the formation task can change during the moving process of the robots to the destination, the environment is large and unknown. In this paper, a dynamic template is introduced firstly to reduce the computational time of the formation method and make it possible for real-time formation control in large scale environment. The virtual targets obtained by the leader–follower based formation model is assigned to each robot based on the self-organizing map neural network. And a bioinspired neural network is employed to navigate these robots to their expected virtual targets with obstacles avoiding automatically. The proposed approach can deal with various situations such as some robots break down, and the formation task is changing dynamically. The simulation results show that the approach in this paper is capable of achieving the formation task in real time efficiently.

Important Links:



Go Back