Abstract:

We consider the problem of building the map of an unknown environment using multiple mobile robots that have physical limitations arising from dynamics and a limited onboard battery. We consider the setting where the unknown environment has a set of charging stations that the robots must discover and visit often to recharge their battery during the map building process. We propose an iterative approach to solve the resulting energy-constrained multi-robot exploration problem. Our approach uses a combination of frontier-based exploration, graph-based path planning, and multi-robot task assignment. We show that our algorithm admits a computation- ally inexpensive implementation that enables rapid replanning, and propose sufficient conditions for recursive feasibility and finite-time termination. We validate our approach in several Gazebo-based realistic simulations.