Abstract
Ramdane Hedjar and Messaoud Bounkhel. Wireless model predictive control: Application to water-level system. Advances in Mechanical Engineering 2016; 8(4): 1-13. DOI: 10.1177/1687814016643639
The Advances in Mechanical Engineering Editorial Office became aware that the peer review process organised by the Guest Editors of the Special Collection
Based on the feedback from the new independent referees, the following corrections are to be made to the above-referenced article.
With reference to the control signal, the authors did not explain why the simulation does not oscillate and the experiment does.
The authors wish to note that the simulation deals with an ideal scenario, without noise or uncertainties. However, the experiment deals with a physical system, which includes noise and uncertainties, even though modeling and identification of the mathematical model of the water-level system has been conducted.
The authors also note that the delay in networked control systems is a random process, citing the difference between figures 5 and 9 as an example. It is also known in control theory that an advantage of integral action is to eliminate the steady state error. However, its main drawback is the amplification of the control signal with noise, in comparison with the case without integral action. In addition, the authors note that the output of the controller is saturated between 0v and 10v. These factors account for large differences between simulations and real-time execution of the algorithm.
In the article, figure 9 shows the control signal without integral action, with small oscillations due to noise, uncertainties, and random delay. Figure 11 depicts the control signal when integral action is included in the control loop. In figure 11, integral action has amplified the control signal and the controller has saturated the output signal between 0v and 10v.