This paper focuses on the problem of adaptive neural networks (NNs) tracking control for a class of completely nonaffine switched pure-feedback uncertain nonlinear systems with switched reference model. A sufficient and necessary condition for the control problem to be solvable is derived by exploiting the common Lyapunov function (CLF) method, backstepping, input-to-state stability analysis, and the small-gain technique. Also, a small-gain technique-based adaptive NN control scheme is provided to avoid the designed difficulty caused by the construction of an overall CLF for the switched closed-loop system, which is usually required when studying the switched pure-feedback system. Adaptive NN controllers of individual subsystems are constructed to guarantee that all of the signals in the closed-loop system are semi-globally uniformly ultimately bounded under arbitrary switchings, and the tracking error converges to a small neighborhood of the origin. Two examples, which include a continuously stirred tank reactor system, are presented to demonstrate the effectiveness of the proposed design approach.

Summary This paper investigates the problem of adaptive output‐feedback neural network (NN) control for a class of switched pure‐feedback uncertain nonlinear systems. A switched observer is first constructed to estimate the unmeasurable states. Next, with the help of an NN to approximate the unknown nonlinear terms, a switched small‐gain technique‐based adaptive output‐feedback NN control scheme is developed by exploiting the backstepping recursive design scheme, input‐to‐state stability analysis, the common Lyapunov function method, and the average dwell time (ADT) method. In the recursive design, the difficulty of constructing an overall Lyapunov function for the switched closed‐loop system is dealt with by decomposing the switched closed‐loop system into two interconnected switched systems and constructing two Lyapunov functions for two interconnected switched systems, respectively. The proposed controllers for individual subsystems guarantee that all signals in the closed‐loop system are semiglobally, uniformly, and ultimately bounded under a class of switching signals with ADT, and finally, two examples illustrate the effectiveness of theoretical results, which include a switched RLC circuit system.

The irrigation decision-making system based on Knowledge-based Engineering (KBE) can accurately predict water requirements and realize smart irrigation. Recurrent neural network(RNN) model have recently showed state-of-the-art performance in this system. This paper deals with the problem of long-term rainfall forecasting based on this network which predicts target rainfalls based on contextual information. A novel recurrent neural network with long short term memory(LSTM) is put for model sequence process for forecasting rainfall. Back-propagation through time(BPTT) algorithm is described for updating recurrent network’s weights. Extensive empirical comparison with three networks, Feed-forward neural network (FNN), Wavelet neural network(WNN) and Auto-regressive Integrated Moving Average(ARIMA), are also provided at various numbers of parameters and configurations. Simulation results demonstrate that the recurrent model with LSTM, trained by the suggested methods, outperforms the others networks.