Consensus filter on networked double-integrator systems

This paper proposes a novel consensus filter to solve the multi-agent consensus problem in the presence of measurement noise, where double-integrator is considered as the dynamic model of the agent. Consensus filter is established by modifying the fundamental consensus algorithm to achieve not only cooperation of the state but attenuation of the noise. The low-pass property of this consensus filter is shown by a specific and fixed topology among the agents. In addition, the tracking stability to a dynamic reference is proven by the linear system theory and matrix theory. Due to the communication constraint, the consensus filter under the switching topology is also considered. The effectiveness of the proposed algorithms is verified by simulation results.