Global team coordination by local computation

Abstract

Desirability functions are an effective way to define group and individual objectives of a team of cooperating mobile robots. By combining desirability functions, we can identify the individual actions that best satisfy both sets of objectives. Combination, however, is global, posing high demands in terms of communication and computation resources. In this paper, we investigate a technique to perform this combination using local computations. Simulated experiments suggest that, under conditions of spatial locality, team control by local computation achieves the same performance than using a global technique.