Flocking Motion Research Of Complex Multi-agent System

With the development of biology, computer, control science, and artificialintelligence, coordination control problem has drawn increasing attention of manyscholars recently and has become the key subject of the control science. Flockingmotion, as one of the important subject of the coordination control problem, has becomean in-depth hot research studied by scholars. Without centralized control, flockingmotion produces macroscopic synchronous effect by mutual perception, which enhancestheir abilities of searching for food and avoiding predators. In the nature, flockingbehavior exists widely, which is collaborative group behavior in a certain gathering way.In the military, flocking can combat replacing the army in order to reduce casualties.Basing on the system introduces to related research status, this paper mainlyfocuses on the design flocking control protocol and stability analysis of the multi-agentsystem. The paper’s mainly research work is as follows1. Flocking motion of the multi-agent system in the weakly connected network.Based on directed weakly connected network of the second-order agent group, adistributed control protocol with individual local information is proposed. The stabilityof agents’ dynamic behavior is researched by using Lyapunov theory. Simulation resultsshow that all of the agents enable to achieve same velocity when none of the agentsreceive information from leader, and when at least one agent can receive the leader’sinformation, all the agents eventually can converge to the leader’s speed.2. Impulse control of flocking motion of the multi-agent system with disturbanceobserver. Flocking motion of multi-agent system is disturbed by all kinds of interferencefrom the complex environment. Combing with the simple, effective, robust feature ofimpulse control, a disturbance observer-based distributed impulse control protocol isproposed for the multi-agent system. Moreover, a sufficient condition is obtained toensure the flocking motion of multi-agent system with leader.3. Sampling control on collaborative flocking motion of discrete-time system withtime-delays. Collaborative flocking motion’s sampling control of the discrete-timesystem is researched in this paper. For the discrete-time multi-agent system withcommunication delay, a distributed control protocol applying individual localinformation is proposed. Basing on Lyapunov stability theory, an energy function is designed. It is proved theoretically that the system is able to achieve collaborativeflocking motion. Simulation results show that the similarity is small in the time delay ornot and directed network topology or not. But the simulation curves of the single-ratesampling is very similar to the simulation curves of the multi-rate sampling.4. Sampling control of the second order multi-agent system with time-delays.Consistency sampling control of the second order discrete-time multi-agent system withfixed topology network and time-delays is studied. Applying Z transformation,according to the generalized Nyquist criterion, the upper bound of the sampling periodis obtained to ensure the asymptotical convergence of the second order multi-agentsystem with time-delays. The relationship between the multi-sampling period andconvergence time is analyzed in the simulation. We can choose the proper samplingperiod according to different requirements on the system.