Tracking Control On Multi-agent Systems With Reduced Interaction

With the rapid development of modern computer technology, information technology,network technology and automatic control theory, the study of multi-agent systems hasbecome a key field and quite a number of researches have been adopted into the realsystems. But the environment outside is complicated and unpredictable, there may besensor fault, bad network condition during controlling the real system. So it becomes reallyimportant to reduce the information that the system needs. Tracking control of multi-agentsystems with reduced interaction is to achieve tracking control of multi-agent systems whilethe system could only acquire part of the essential information. The study mainly focuseson Lagrangian systems and wheeled mobile robots systems which are typical nonlinearsystem, systematic study is carried out mainly in the following two aspects:1. Distributed tracking on Lagrangian systems with reduced interaction. The studymainly focuses on the Lagrangian systems. Assumed only a fraction of followers couldinteract with the leader, an adaptive control law has been adopted to achieve consensuscontrol based on the velocity error between agents. Position error has also been added to thecontrol law to achieve velocity consensus between the follower and the leader. Besides, anartificial potential function is used to preserve the network connectivity of the systemduring control process. Finally velocity consensus of the system has been achieved andlarge amount of simulations have been made to illustrate the theoretical results.2. Trajectory tracking of wheeled mobile robots with reduced interaction. The studymainly focuses on the wheeled mobile robot. Assumed follower could not acquire its ownorientation and there are unknown nonlinear disturbances within the angular control input.An observer was built to estimate the trigonometric equation of the error between thefollower’s orientation and the leader’s orientation instead of estimating the error directlyafter analyzing the mathematic equation of the model. An adaptive control theorem isadopted to approach the disturbance and weaken its effect. Finally orientation consensuscould be achieved. Simulations and experiments on real robots have illustrated theeffectiveness of the control law and contrast experiments with different disturbances hasbeen done to explore the influence that they have on controlling real robots.