Experiment Design for Large Multi-Robot Systems

Multi-robot systems present unique challenges for the experimenter. There are three key components for conducting successful multi-robot experiments: 1) Defining platform-independent metrics for evaluating algorithm performance, 2) determining the ground truth of the robot’s physical configuration and network connectivity, and 3) designing a usable system, with centralized data collection and hands-free operation. These three components require specific hardware and software unique to multi-robot systems. This paper presents a brief survey of the literature on the key challenges facing experiment design for multi-robot systems, and presents the author’s experiences with working with dozens of robots on a daily basis.

McLurkin J
Conference in Seattle, WA