Cooperative multi-robot patrol in an indoor infrastructure

Abstract

Multi-robot patrol (MRP)is essentially a collective decision-making problem, where mobile robotic units must coordinate their actions effectively in order to schedule visits to every critical point of the environment. The problem is commonly addressed using centralized planners with global knowledge and/or calculating a priori routes for all robots before the beginning of the mission. However, distributed strategies for MRP have very interesting advantages, such as allowing the team to adapt to changes in the system, the possibility to add or remove patrol units during the mission, and leading to trajectories that are much harder to predict by an external observer. In this work, we present a distributed strategy to solve the patrolling problem in a real world indoor environment, where each autonomous agent decides its actions locally and adapts to the system’s needs using distributed communication. Experimental results show the ability of the team to coordinate so as to visit every important point of the environment. Furthermore, the approach is able to scale to an arbitrary number of robots as well as overcome communication failures and robot faults.