Realtime Hybrid Task-Based Control for Robots and Machine Tools

This paper presents work in the field of hard realtime robotics and machine control. We analyse the requirements of a hybrid realtime control task specification allowing the integration of discrete and continuous control tasks. We propose an application independent task structure providing data flow consistency under simulataneous access by different control layers. We provide an execution flow mechanism to guarantee execution time determinism, yet allowing flexibility to react to a changing environment. We use state machines for process monitoring and a thread-safe realtime event system to communicate changes. The tasks can be distributed over a network and communicate using interfaces or manipulate streams of data in the loop. The presented task structure is applied to a real world example.