Second-Order Heterogeneous Multi-Agent Target Tracking Without Relative Velocities

The multi-agent target tracking problem has received growing interest in the robotics and controls community in recent years. In particular, distributed target tracking is an abstraction for many potential applications. However, typical results assume that full state information is available. This paper addresses the multi-agent target tracking problem, where only the relative distance between neighbors is known. To yield this result, a novel distributed observer is designed that employs an auxiliary distributed filter. The distributed observer achieves network-to-target regulation by enabling the network of agents to estimate the relative velocities of all agents and the target. The distributed filter/observer structure is motivated by a Lyapunov-based stability analysis, which is provided to ensure that all agents are exponentially regulated to a neighborhood of the target. Comparative simulations are provided to demonstrate the performance of the developed method. The simulation results indicate that six agents modeled by an unknown heterogeneous damped and driven harmonic oscillator can successfully track a target agent. The developed method provides a 67% improvement in the RMS tracking error when compared to a baseline.