Pilot-free optimal control of linear systems over OFDM networks via control-aided channel prediction

Designing optimal controllers for wireless networked control systems (WNCS) typically requires real-time channel state information (CSI), which is difficult to obtain in the presence of time-varying wireless channels. In this work, we propose a pilot-free framework for optimal control over wireless networks, where a linear dynamic plant communicates with a remote controller over an Orthogonal Frequency Division

Multiplexing (OFDM) link. The CSI is predicted online from plant states and control inputs using Kalman filtering, thereby eliminating the need for pilot transmissions. Based on the predicted CSI, the control policy is derived using the Bellman principle. To address the resulting curse of dimensionality, we approximate the optimal control solution via a coupled algebraic Riccati equation (CARE), which is solved using a stochastic approximation (SA) algorithm. We establish joint system stability and provide sufficient conditions for the existence, uniqueness, and convergence of the approximate control

solution. Simulation results demonstrate that the proposed framework achieves improved control stability performance and channel prediction accuracy compared with widely used

baseline methods.