<![CDATA[A Networked Control System (NCS) is a control system in which actuation and feedback signals are transmitted over a communication network. One of the key challenges in NCS is the presence of random delays introduced by the communication protocol. This study proposes an adaptive approach for tuning the integral gain (Kint) of the Linear Quadratic Integral (LQI) controller, based on both the current delay state and a predicted delay obtained via a Markov chain model, which has not been explored extensively. The proposed method first maps each delay interval to a corresponding Kint, establishing a delay–gain pair. Then, the integral gain is dynamically updated at each control cycle by combining the current (Kint_t) and the predicted Kint for the next time step (Kint_t+1), using weighted coefficients a and b, respectively, as follows: Kint = a* Kint_t + b* Kint_t+1. Experimental validation demonstrates that, with optimal weights a=0.5 and b=0.5, the proposed method significantly improves system performance. Compared to a fixed (static) Kint, it reduces the percentage overshoot from 17.06% to 2.45% and decreases the settling time from 457.6 seconds to 254.06 seconds.]]>
Copyrights © 2025
