An improved rate-monotonic scheduler and a real-time open control platform for robotic control algorithm verification

The verification of new intelligent robotic control algorithm not only requires the correction of computation but also satisfies the timing constraints. In this paper, an open real-time control environment is designed and implemented for robotic control algorithm verification. With Windows NT´s real time extension RTX, the PC based control system will have good real-time performance for new control algorithm research. The software structure and the task scheduling approach are described in detail and the software comprises of real-time motion control & trajectory planning process, the use´s operating interface process and the inter-process communication module. Considering the uncertainty in computation time of complex intelligent algorithms in robotic systems, the running time of the algorithm is monitored online and an improved rate-monotonic scheduler is presented for schedulability verification. To prove the effectiveness of the open real-time control environment and task scheduler, two examples, a direct driven air-bearing stage for IC packaging and a SCARA type robot for assembly, are introduced. The control performance and task scheduling results show that the open real time control environment and new scheduler are successful in complex control algorithm research and timing verification.