Coordination of Multirobot Systems Under Temporal Constraints

Abstract:

Multirobot systems have great potential to change our lives by increasing efficiency or decreasing costs in many applications, ranging from warehouse logistics to construction. They can also replace humans in dangerous scenarios, such as the Fukushima nuclear disaster cleanup. However, teleoperating robots in these scenarios would severely limit their capabilities due to communication and reaction delays. Furthermore, ensuring that the overall behavior of the system is safe and correct for a large number of robots is challenging without a principled solution approach.

Ideally, multirobot systems should be able to plan and execute autonomously and be robust to certain external factors, such as synchronization errors and failing robots. Moreover, these systems should be able to scale to large numbers, as the effectiveness of particular tasks might depend directly on these criteria. Motivated by these goals, we introduce a framework for multirobot coordination to accomplish complex tasks autonomously. This framework consists of (i) a formalism to specify multirobot tasks, (ii) algorithms to synthesize paths that collectively satisfy these tasks, and (iii) methods to deal with synchronization errors.

Chair: Professor Necmiye Ozay

Remote Access: https://bluejeans.com/819684505