Coverage Control for Multirobot Teams With Heterogeneous Sensing Capabilities

Abstract

This letter investigates how mobile agents with qualitatively different sensing capabilities should be organized in order to effectively cover an area. In particular, by encoding the different capabilities as different density functions in the locational cost, the result is a heterogeneous coverage control problem where the different density functions serve as a way of both abstracting and encapsulating different sensing capabilities. However, different density functions imply that mass is not conserved as the agents move and, as a result, the normal cancellations that occur across boundaries between regions of dominance in the homogeneous case no longer take place when computing the gradient of the locational cost. As a result, new terms are needed if the robots are to execute a descent flow in order to minimize the locational cost, and we show how these additional terms can be formulated as boundary-disagreement terms that are added to the standard Lloyd's algorithm. The results are implemented on real robotic platforms for a number of different use cases.