Learning‐based adaptive feedback control for tracking optimisation in wireless sensor actuator networking systems

Abstract

In wireless sensor and actuator networks (WSAN), sensors can have much longer lifetime as mobile sink nodes move through the network and collect data from static sensor nodes. Applying mobile sinks in WSAN systems has faced challenges in sink path planning and data gathering scheduling. Many studies have focused on priori-trail planning for mobile sinks; however, they have not considered tracking performance degradation during operation due to system nonlinearities and uncertain WSAN environments. Thus, two learning-based adaptive feedback control methods, adaptive neural-network-based feedback control (ANNFC) and adaptive neural-network-based feedback control with integrator model (ANNFC-IM), are proposed to improve and optimise tracking performance in unknown and uncertain WSAN environments. The proposed algorithms enable a mobile sink to adjust its flying trajectory with online learning by applying a neural-network-based control method. The performance of the proposed algorithm is validated by theoretical analysis and simulation results. The results demonstrate that the proposed methods significantly outperform existing methods in terms of tracking error, system stability, and convergence.