Decentralized Formation Selection Mechanisms Inspired by Foraging Bottlenose Dolphins

Formation control is an important sub-problem in multi-agent robotics and a number of decentralized control strategies have been developed to solve the formation control problem, e.g., [1], [2], [3]. However, there is still no underlying theory that governs the selection of a particular formation. In other words, if an agent has a choice between two formations, which one should it choose? In this paper, we would like to answer this question by drawing inspiration from foraging bottlenose dolphins, Tursiops truncatus. In fact, the networked control community has drawn significant inspiration from interaction-rules in social animals and insects [4], [5], [6]. In particular, the widely used nearest-neighbor-based interaction rules, used for example for formation control [7], [8], [9], consensus [3], [10], and coverage control [11], [12], has a direct biological counterpart, as shown in [4]. Bottlenose dolphins employ an unusual foraging technique known as the horizontal carousal; here the dolphins, after locating a sizeable amount of prey, form a large circle to trap the prey inside that circle. The prey is usually a school of fish and the dolphins slowly tighten the encirclement to restrict the movement of the fish. Eventually, the circle becomes small enough and the dolphins then take turns to charge through the fish while maintaining the integrity of the circle. In this paper, we model the first phase of this method of capturing prey, and offer a choice of two formations for each agent — Large and Small Circle and by using