Optimization-Driven Control and Organization of a Robot Swarm for Cooperative Transportation

Abstract The cooperation of autonomous robotic systems, forming a robotic network by communicating with each other, brings about the potential to improve both performance and robustness in robotics applications, but introduces additional complexity in control design. In this context, this paper examines a cooperative transportation task that can serve as a test bed for distributed control algorithms, but may also inspire applications in logistics. In the proposed approach, the task solution is split into an organization scheme, determining formations useful for the current situation, as well as a distributed controller maintaining the formation. Both of these rely on the solution of optimization problems, with the latter drawing from distributed model predictive control theory. The optimization problems are formulated in a way that they can be derived automatically from the intrinsic properties of the considered scenario, e.g. the number of participating robots and the shape of the transported object. Therefore, no human intervention or scenario-specific design steps are necessary while still being able to cater to a wide range of scenarios. Challenging simulation scenarios illustrate the performance achievable with the proposed scheme.