A Modification of PI Controller Based on an Online Adaptive Approach to Use in Networked Control Systems for Remote Surgery

Abstract

Background: Networked control systems (NCSs) are used to control industrial and medical plants via data communication networks. These systems have many wide applications in a broad range of areas such as remote surgery, industrial and space sciences. Two important challenging problems in these systems are stochastic time delays and packet dropouts. Classic proportional- integral controllers due to their simple inherent design and implementation have many applications in controlling industrial and medical plants. However, these simple controllers do not have high performance in NCS because of communication networks induced time-varying delays and so this causes instability in NCS. In this paper, an adaptive proportional-integral controller is proposed using an online estimation of network time delay technique in a node application layer. The coefficients of this new controller change according to the values of estimated time delays online. Therefore, the proposed controller causes stability in NCS loop. The performance of the proposed method is simulated for a DC motor that can be used in remote surgery. The simulation results show the proposed controller is better at least about 1000 times according to IAE performance index rather than a classic proportional-integral controller. Also, the results of practical implementation show that the proposed controller causes the stability of NCS. Objective: The study intends to analyze and design an online adaptive approach that can stabilize networked control systems having important applications such as remote surgery. Materials and Methods: The article proposes an online adaptive proportional-integral controller that can be used in NCS and has applications such as remote surgery. The coefficients of the newly proposed controller are changed online based on the estimation of network time delay. In the proposed controller firstly, the variable time delay value is estimated online, then the coefficients of the PI controller is updated based on this estimated value. Results: The proposed method causes the controller to generate and transmit the suitable control signal according to different and random conditions of the network. Adaption of coefficients compensates time-varying delay effect on system performance and causes increasing stability that is necessary for medical applications. Conclusion: The proposed system performs better than the traditional approach in terms of measuring the average value of the error, recall, and ITAE. According to simulation and practical results, when the network average time delay is about 40ms, the performance index for an online adaptive PI controller is equal to 4.4236, and value for a classic PI controller is 4409. Thus, the performance of an online adaptive PI controller has been improved about 1000 times rather than a classic one. Therefore, the proposed controller in real network time delay has proper performance and keeps the stability of the control loop.