Decentralized Adaptive Coverage Control of Nonholonomic Mobile Robots*

This paper proposes an algorithm for decentralized adaptive control for coverage of a given convex region using multiple agents with double integrator dynamics. The result is then extended to nonholonomic mobile robots. The density function which describes the event of interest to be sensed by the agents is assumed to be unknown and an adaptation law is derived which can learn the sensory function with time. This work is an extension of the work by Schwager et al. (2007) for mobile sensors with single integrator dynamics. We propose an algorithm which ensures convergence to a near optimal configuration and demonstrate the same using simulations.