Abstract: The continuing drop of the price-to-performance ratio of embedded systems, wireless networking, and high-speed processors in recent years has fueled the growing interest in multi-robot systems. As the number of robots increase, it is useful to think in terms of a swarming paradigm where capabilities are expressed by populations rather than super-capable individuals, as seen in the group dynamics of biological swarms. This is especially relevant in complex tasks where it is difficult to provide robots with specific instructions a priori due to incomplete knowledge of the environment, e.g. interplanetary exploration, or where mission criteria require capabilities that are varied in both quantity and difficulty, e.g. urban search and rescue. In this talk, I will present some of the challenges towards the development and deployment of actual robotic swarms for real world applications. Specifically, I consider the effects of communication on teams of robots and propose methods to maintain the integrity of the communication network. I will show how reactive controllers can be used to maintain end-to-end communication links and consider the differences between monitoring point-to-point signal strength versus data throughput. I will further demonstrate how these controllers, combined with more deliberative planning, can be used to coordinate a robot swarm to generate a two-dimensional geometric pattern while maintaining specified relative distance constraints. This is relevant for applications such as perimeter surveillance, environmental boundary tracking, and cooperative manipulation.
- This event has passed.