Remote control of an omnidirectional mobile robot with time-varying delay and noise attenuation

Abstract

This paper proposes a dead-time compensation control based on the filtered Smith predictor with an integral Linear Quadratic Regulator to remotely control an omnidirectional mobile robot subject to time-varying delay, sampling jitter and measurement noise. The filtered Smith predictor is used to preserve implementation simplicity, attenuate measurement noise and mitigate the performance degradation caused by the time-varying delay and sampling jitter. Implementation simplicity is achieved since the trade-off between robustness and closed-loop performance is handled from experimental tuning approach. Theoretical analysis is provided to show the benefits of the proposed compensation strategy. A set of experimental results with and without network-induced time-varying delay corroborate this analysis. Moreover, the relationship between measurement noise and the undesired time-varying delay effect is theoretically established to emphasize the benefits of the proposed approach. Practical results are presented in order to show that the proposed LQR/filtered Smith predictor strategy can be used in a NCS configuration providing good results with respect to trajectory tracking purposes despite its implementation simplicity.