A hybrid systems approach to force-guided robotic assemblies

Assemblies with uncertainties exceeding the assembly tolerance require the use of compliance control combined with search strategies for parts alignment. Automating such assemblies requires a general framework that allows various search strategies to be combined at a higher-level of planning and yet be compatible with low-level compliance control. We examine these issues with a hybrid systems approach. This paper also presents experiments on a specially constructed peg-in-maze assembly abstracting the challenges of example real-world assemblies, which were automated in earlier work.