Anisotropic active Lagrangian particle swarm control in a meandering jet

Abstract

Underwater robot swarming is of great importance to the development of novel strategies for oceanographic research, oceanic meteorology, and surveillance. Nevertheless, it has not drawn as much attention as its aerial or ground counterparts. In this paper, we investigate the mobility of a robot swarm in dominant and dynamically changing ocean currents. An active Lagrangian particle swarm (ALPS) consists of mobile agents with local sensing and communication capabilities. Motion of these agents is mostly dictated by background flows. Limited self-propulsion is applied intermittently in order to maintain swarm connectivity and avoid inter-agent collisions. To preserve the predictability of the swarm motion through ocean general circulation models, a physicomimetic swarm control law is proposed such that pair-wise self-propulsion is applied analogously to internal forces among atoms or molecules. Interactions among neighboring agents are weighted by local agent densities, which are calculated through a compact smoothing kernel function. The ALPS' movement is simulated in a meandering jet. Results show that swarm connectivity can be effectively maintained by controlling a small portion of the agents with the proposed swarm control method even when large background flow variations present.