Synergy graphs for configuring robot team members

Robots are becoming increasingly modular in their design, allowing different configurations of hardware and software, e.g., different wheels, sensors, and algorithms. We are interested in forming a multi-robot team by configuring each robot (i.e., selecting the different modules) to best fit a task. This general problem is applicable to many domains, such as manufacturing in high-mix low-volume scenarios. In this paper, we formally define the Synergy Graph for Configurable Robots (SGraCR) model, where each robot module is modeled as a vertex in a graph, and we define how to compute the synergy of modules within a single robot, as well as between robots, using the structure of the graph. We define the synergy of a multi-robot team comprised of such configurable robots, and contribute a team formation algorithm that searches a SGraCR to approximate the optimal team. In addition, we contribute a learning algorithm that learns a SGraCR from a small set of training data containing the performance of teams. We evaluate our SGraCR model and algorithm in extensive experiments, both in simulation and with real robots, and compare with competing algorithms.