A Novel Robust Predictive Control System Over Imperfect Networks

Abstract

Abstract-This paper aims to study on feedback control for a networked system with both uncertain delays, packet dropouts and disturbances. Here, a so-called robust predictive control (RPC) approach is designed as follows: 1-delays and packet dropouts are accurately detected online by a network problem detector (NPD); 2-a so-called PI-based neural network grey model (PINNGM) is developed in a general form for a capable of forecasting accurately in advance the network problems and the effects of disturbances on the system performance; 3-using the PINNGM outputs, a small adaptive buffer (SAB) is optimally generated on the remote side to deal with the large delays and/or packet dropouts and, therefore, simplify the control design; 4-based on the PINNGM and SAB, an adaptive sampling-based integral state feedback controller (ASISFC) is simply constructed to compensate the small delays and disturbances. Thus, the steady-state control performance is achieved with fast response, high adaptability and robustness. Case studies are finally provided to evaluate the effectiveness of the proposed approach.Index Terms-Networked control system, time delay, state feedback control, predictor, neural network, buffer. I. INTRODUCTIONETWORKED control systems (NCSs) are spatially distributed systems, in which actuators and sensors at the plant side are connected to a controller at the remote side. Due to having many advantages over the traditional control schemes via point-to-point wiring, NCSs have been deploying worldwide in various applications, such as power grids, transportation systems, teleoperation or remote systems.However for practical uses of NCSs, the most important issues are network-included random time-varying delays and packet dropouts which deteriorate the control performances and, easily cause the instability [1], [5], [25], and [29]. Generally, time delays can be classified into three components: computation delays at the controller and communication delays at the forward and backward channels.Many studies in literature addressed stability of NCSs with delay problem only [1]- [6]. Here, the controller designs were depended on the assumptions in which the time delay was constant [1], bounded [2]- [5] or had a probability distribution function [6]. Other researchers focused on analyzing time delay problem at communication channels to evaluate its effects on the system performances as well as to compensate these undesirable effects [7] and [8]. For NCSs with large delays, a new control concept based on variable sampling periods using neural network or prediction theories has been recently adopted [9]- [11]. Nevertheless, the observation of real delay data to train and construct the NCSs was not appropriately discussed in these studies. To solve this problem, an advanced variable sampling period control concept has been developed for systems containing random delays [29]. The effectiveness of this concept in accurately detecting and predicting the delays without requiring a training process was proved through real-time ...

Stochastic stabilisation of wireless networked control systems with lossy multi‐packet transmission

“…there exists matrix Z ∈ ℝ m × m such that PB = BZ and Z = VΣ −1 P 1 ΣV T , where P 1 ∈ ℝ m × m and P 2 ∈ ℝ (n− m) × (n− m) . Theorem 2: The closed-loop system in (15) with the controller in (14) is stochastic stable, if there exists scalars γ 1 , γ 2 > 0 and a positive definite matrix P ∈ ℝ n× n with the structure as in (32) and a matrix M with an appropriate dimension such that (see (33a))…”

“…To overcome these difficulties, tremendous efforts have been made by the scientists and engineers in the field, and various control theories and techniques have been applied to NCSs, e.g. robust control theory [7][8][9], approaches based on Markov jump systems [10][11][12], predictive control [13][14][15], and joint design schemes between communication and control [16][17][18], to name a few.…”

Effects of Communication Signal Delay on the Power Grid: A Review

“…Because of this feature, NN-based control can be used to improve the resilience of control frameworks and address communication link weaknesses. The use of NN-based wide-area damping controllers to adjust for wide range of communication delays was investigated in [104]. Wang S. et al [71], neural network approximate-based feedback adaptive quantized control protocol based on k-filter observer is proposed to reduce the estimated error and external disturbances resulted from multi-machine excitation system.…”

Discrete‐time higher‐order sliding mode control of systems with unmatched uncertainty

“…Disturbance forecasting is a well-established concept and a frequently used technique in systems with smooth disturbances. [40][41][42][43] There are many forecasting techniques, such as discrete Kalman filter, weighted moving average, and double exponential smoothing. 44 The disturbance terms are represented as time-series structure and are utilized to forecast future values of disturbances.…”

“…Disturbance forecasting is a well‐established concept and a frequently used technique in systems with smooth disturbances . There are many forecasting techniques, such as discrete Kalman filter, weighted moving average, and double exponential smoothing .…”