Azimuth-dependent and flux-conserved force field model: A new control method for robot swarm

In the collective motion control of a robot swarm, physics-based approaches offer unique advantages. A virtual force field is assumed around each robot, and produces attractive or repulsive force to neighbouring robots. Yet physics-based approaches have some limitations, because a robot should have more complicated actions than a physical particle, and the force fields could bring unexpected local balances during foraging and obstacle avoiding. In this paper, we present a novel Azimuth-dependent and Flux-conserved Force Field (ADeFF) model, in which an azimuth-dependent force field is assumed around each robot and a flux-conserved force field around each obstacle and destination area. The azimuth-dependent force field allows the force to be set arbitrarily so that forces more than simply attraction and repulsion can be set to achieve more desired swarm configuration. The flux-conserved force field prevents local balances from happening during swarm foraging. Simulation results show that using the ADeFF model, homogeneous swarm can self-arrange into lattice configurations which are difficult to achieve through previous physics-based approaches. And local balances can be effectively avoided using this new method.