A deadband-control and prediction co-design approach for networked robotic systems

The cooperation of robotic systems over wireless network results in some challenges, finite network bandwidth, and limited node energy. A novel co-design approach is presented by means of combining deadband-control with dynamic prediction in this paper. To reduce the network traffic and energy consumption, the control packets of "Controlling" robots are only sent if the response error and error changes are larger than the given deadband threshold. Furthermore, the modified Brown third exponential smoothing prediction is introduced to reconstruct the untransmitted values in "Actuating" robots, where parameters are dynamically optimized by a modified Particle Swarm Optimization (PSO) algorithm. Finally, the comparative experimental results indicate that the effectiveness and applicability of the proposed approach are very well.