Distributed Velocity Control in Cooperative Multi-Agent Moving Source Seeking

This paper uses distributed velocity control to address the agent-based moving source-seeking problem. The source is represented by a scalar field. In cooperative multi-agent seeking, agents communicate with one another using a communication topology to exchange and collect information regarding their positions and the value of the scalar field at any given time. This information is then compiled to create a single overall prediction that is used to locate the moving source that moves in a linear or sinusoidal manner. The velocity at which the agents track the moving source is determined using distributed velocity control that examines two agent velocity types: the gradient velocity and the desired velocity. Formation control is also used to maintain the desired formation of the agents. Finally, a computer simulation is conducted to examine how different agents search for a moving source. In both cases, the agents can find the source at a different seeking time. The results show that the agents under a given desired velocity may perform better than those under gradient velocity for the desired velocity v_d > 1.85 m/s and v_d > 1.9 m/s for linear and sinusoidal moving sources, respectively.