摘要

In consensus protocols of multi-agent system, we can often find state differences between some agents and a reference agent. However, we find that taking the relative states value may not be best. Conversely, using the differences between some agents states and several times state of the reference agent can achieve quickly convergence by choosing a appropriate parameter r. Therefore, in this paper a new consensus protocol taking r times state of reference agent is proposed. Based on this idea, two convergence protocols of single-integrator kinematics are considered, and some sufficient conditions on consensus protocols of double-integrator dynamics are derived. Finally, numerical examples illustrate these theoretical results.