Path Planning for Shepherding a Swarm in a Cluttered Environment using Differential Evolution

Shepherding involves herding a swarm of agents (\\emph{sheep}) by another a\ncontrol agent (\\emph{sheepdog}) towards a goal. Multiple approaches have been\ndocumented in the literature to model this behaviour. In this paper, we present\na modification to a well-known shepherding approach, and show, via simulation,\nthat this modification improves shepherding efficacy. We then argue that given\ncomplexity arising from obstacles laden environments, path planning approaches\ncould further enhance this model. To validate this hypothesis, we present a\n2-stage evolutionary-based path planning algorithm for shepherding a swarm of\nagents in 2D environments. In the first stage, the algorithm attempts to find\nthe best path for the sheepdog to move from its initial location to a strategic\ndriving location behind the sheep. In the second stage, it calculates and\noptimises a path for the sheep. It does so by using \\emph{way points} on that\npath as the sequential sub-goals for the sheepdog to aim towards. The proposed\nalgorithm is evaluated in obstacle laden environments via simulation with\nfurther improvements achieved.\n