Collective Behaviors Of A Heterogeneous Foraging Swarm

While bees and birds often work together in groups to find food, a group of robots can be designed to coordinate their activities to perform missions such as salvage, charting unexplored territories and building mobile sensor networks. In order for such multi-agent systems to succeed it is often critical that they can both maintain cohesive behaviors and appropriately respond to environmental stimuli. Stimulated by the growing interest in biomimicry of the mechanisms of foraging and swarming for use in engineering applications, this thesis focuses on mathematical modeling and analysis of foraging properties of a heterogeneous swarm.Inspired by organized behaviors of honeybee swarms, a heterogeneous swarm, which consists of two kinds of individuals - scouts and normal agents - in terms of different sensing abilities, was constructed by individual-based mathematical models. At first, the swarm was studied in free environment, properties of cohesion and stability were proved when linear-attraction-bounded/unbounded-repulsion functions were used, and effects by various factors were discussed through computer simulations. Then, foraging behaviors of heterogeneous swarm in various environment profiles were analyzed. Firstly, cohesion of swarm applying linear-attraction-bounded-repulsion in environment under certain conditions was proved. Secondly, foraging properties in plane, Gaussian, and multimodal-Gaussian environment were studied. Thirdly, individual stability was proved in some environment profiles. Fourthly, results achieved were extended to heterogeneous swarm using linear-attraction-unbounded -repulsion. After that, simulations tested how environment factors and scout rate could influence collective behaviors, also the priority of short-distance-bounded -attraction function proposed by this thesis was demonstrated in complex environment. At last of the thesis, a multi-robot experiment platform which is under construction was briefly introduced.Results achieved by this thesis showed that heterogeneous swarm is a feasible framework, which is useful in engineering applications of multi-robot systems and mobile sensor networks, etc.