Dynamic anti-windup compensation for multi-agent systems with input saturation

In this paper, the multi-agent consensus problem is considered for a group of double integrator systems with input saturation. For each agent, velocity and position information of its neighbors is used to control itself to achieve consensus. Therefore, signal communication between different agents plays a significant role in the success of distributed control. However, due to control signal amplitude limits, velocity, position or both velocity and position produced control signals could saturate. Therefore, in this paper a dynamic anti-windup compensation architecture is employed to tackle the challenge of saturated control signals. Simulation results are provided to verify the effectiveness of the proposed anti-windup compensation for the multi-agent systems.