Fuzzy logic soft-switch controller for networked control systems

This paper proposed a novel soft-switch controller for networked control systems. This controller is composed of fuzzy logic system and neural networks. One of the most important drawbacks in networked control systems is stochastic time delay which causes instability in control system. Proposed controller can overcome large delays by applying suitable control signal softly. Ten neural networks are designed based on related network time delay ranges, then using a TSK fuzzy logic system, the proper weights for outputs of each neural network are calculated based on online estimated network time delay. By summation of multiplied obtained weights and outputs of neural networks, control signal of soft-switch controller is generated. Comparison of simulation results between soft-switch method and two other common methods shows the proposed method improves the system performance especially in large delays such as 450ms. For example in network time delays over 400ms the Integral of Time-weighted Absolute Error, ITAE, of soft-switch controller is reduced to 0.78 and 0.11 rather than ITAE of Smith predictor and common fuzzy logic controller, respectively.