We expect that using more communicating robots to explore and map an environment will reduce the time to completion of the task. But this basic question is sometimes neglected: Is the cooperation between more robots really useful?

Here, the authors investigate a scenario in which robots can collaborate locally: a robot sends a request, and the connected robots can decide to stop their activity to answer it. The paper focuses on defining when to collaborate and with whom to collaborate for robots making the exploration of an indoor environment, represented as a graph.

“When” is defined by four collaboration strategies: immediate, postpone, and use heuristics about general or local structure of buildings. The “who” is determined by maximizing a utility value depending on the travel distance and time needed by the robot to complete its current exploration. From the ranked robots, the one selected should also maximize the network connectivity.

A simulation is done, ignoring map, planning, and navigation errors, in randomly generated environments of equal-size rooms; robots communicate to share their maps and build a global map.

The results show the exploration times for two, three, and four robots in case of no collaboration and for the four strategies. Collaboration always increases the time, regardless of the chosen strategy. Time reduces a little for three and four robots, but adding more slows down the exploration due to interference. A detailed analysis of the four strategies finds that the best way to cooperate when cooperation is necessary is using heuristics-based strategies.

One more piece in the scenario of robot cooperation: the assessment is only in simulation and in a reduced problem domain, so it is not the final word about the topic.