Discrete exponentially stable prediction-based controller under time delay with application to unmanned helicopters

Abstract

This paper presents a discrete exponentially stable prediction-based controller (DEPC) for discrete time-delay aerial systems subject to significant time delays and external disturbances. Unlike conventional predictors that could not consider future disturbances, the DEPC integrates an exponential stability criterion into the disturbance prediction process, facilitating accurate recursive state prediction. This strategy ensures that the predicted state exponentially converges to the future state, enabling robust tracking and superior disturbance rejection for inherently unstable platforms. The effectiveness of the DEPC is validated through both numerical simulations and comprehensive flight tests on an unmanned helicopter. The DEPC maintains stability under time delays of up to 0.3s in both indoor and outdoor environments. The control performance shows a maximum reduction of 48% in root-mean-square tracking error compared to existing prediction-based controllers, particularly during sinusoidal input trajectory. The results confirm the feasibility and performance of the DEPC for direct implementation on highly agile aerial platforms, even in the presence of significant time delays.