Distributed Strategies for Dynamic Coverage with Limited Sensing Capabilities

TL;DR

This paper introduces a distributed robotic coverage algorithm that operates without metric information, focusing on optimal deployment, limited communication, and event clustering in obstacle-rich environments.

Contribution

It presents a novel distributed approach for robotic coverage and event clustering that accounts for physical limitations and minimizes communication links.

Findings

Derived a lower bound on the number of agents needed based on topological assumptions.

Achieved minimized communication links while maintaining a connected network.

Successfully identified and directed agents towards the event's most intense point.

Abstract

In this work, it is presented the development of a novel distributed algorithm performing robotic coverage, clustering and dispatch around an event in static-obstacle structured environments without relying on metric information. Specifically, the aim is to account for the trade-off between local communication given by bearing visibility sensors installed on each agent involved, optimal deployment in closed unknown scenarios and focus of a group of agents on one point of interest. The particular targets of this study can be summarized as 1. the computation, under certain topological assumptions, of a lower bound for the number of required agents, which are provided by a realistic geometric model (e.g. a round shape) to emphasize physical limitations; 2. the minimization of the number of nodes and links maintaining a distributed approach over a connected communication graph; 3. the identification of an activation cluster around an event with a radial decreasing intensity, sensed by each agent; 4. the attempt to send the agents belonging to the cluster towards the most intense point in the scenario by minimizing a weighted isoperimetric functional.