Fault-Tolerate Control for Distance-Only Multi-agent Formation Using Discrete Fourier Transform

Abstract

Multi-agent formation is a mobile robot cluster based on wireless sensor network(WSN), in which each agent has the ability of sensing and computing. It has a wide range of applications in the fields of transportation, aerospace, exploration, etc. The main difficulty of formation control is positioning accuracy and fault tolerance. Existing technology is generally realized by combining distance information and other auxiliary information. However, these technologies have limitations both in cost and generality, so it is infeasible for a large-scale formation system. In this paper, a fault-tolerate formation control method for distance-only multi-agent formation is proposed. The main technique we use is Discrete Fourier Transform (DFT). On the premise of combination in circle and linear motion, the motion parameters of agent are estimated by DFT during discrete time series. On this basis, by monitoring the associated reference points of related targets, each individual in the formation can accurately locate the coordinates of neighbors in its own local coordinate system. In the noise simulation experiment, this method reduces formation distance error by about 20%. The experiment results also show efficiency in non-rigid formation shape control.