In this paper, we present an adaptive feedback framework for nonlinear multi-robot systems with unknown execution times and task dependencies. The scheduling issue of periodic tasks with task dependencies is considered. A feedback scheduler with frequency actuator and performance index is presented. Worst-case response time analysis that affects the real-time deadline miss ratio is analyzed. With the given scheduler, the schedulability of periodic tasks that have varied frequencies can be enhanced in the context of real-time control systems. Experiment results show that this adaptive feedback controller can be used in industrial multi-robot coordination and force-guided assembly. With the given adaptive feedback scheduling algorithm, the system performance and the deadline miss ratio can be improved compared with the traditional feedback scheduling algorithms. This feedback framework can achieve high CPU utilization and guarantee a certain deadline miss ratio in real-time control systems without sacrificing the system stability.
展开▼
