Flocking control of multi-agents based on self-adaptively weighting observers driven only by local position measurements

摘要

The most methods to achieve flocking control of multi-agent systems take both position and speed information of all the multi-agents into the control laws. However, various factors and their magnitudes, related to mutual position difference vectors, between the multi-agents are neglected. In this paper, based on distributed observers, flocking control algorithms driven only by local position measurements and with self-adaptively weighting coefficients are proposed. The suggested algorithms are sketched as follows. Firstly, distributed observers are introduced to each individual agent to estimate its speed only by local position measurements of the agents locating in the corresponding neighborhoods. Secondly, by self-adaptively adjusted weighting coefficients to the observers, possible attraction/dispersion force variations caused by position measurement uncertainties are exploited for facilitating flocking formation. Moreover, major properties of the suggested flocking algorithms are analyzed rigorously, and numerical simulations are shown to verify effectiveness of the proposed algorithms.