In this paper, the adaptive fault-tolerant control (FTC) problem is addressed for the stochastic nonlinear systems with multiple faults. The multiple faults, including the actuator and abrupt system faults, are first discussed in the same theoretical framework. The unknown nonlinearities are approximated by multi-dimensional Taylor networks (MTNs). By taking advantage of backstepping technique and finite-time control, the actual control law and virtual control signals are constructed, and then a novel adaptive FTC scheme based on the finite-time control method is proposed. The proposed controller guarantees that the closed-loop system is semi-global finite-time stable in probability (SGFSP) and the tracking error converges to a small neighborhood around the origin in the finite-time. Lastly, three examples are given to illustrate the effectiveness of the proposed scheme. Note to Practitioners —This research is motivated by the fact that actuators faults exist widely in real applications, which often degrade the control accuracy of the system and even result in the system instability. So far, the actuator and abrupt system faults of the stochastic nonlinear system have not yet been considered under the same theoretical framework. Therefore, this study designs a new MTN-based adaptive finite-time FTC scheme, which can ensure that the closed-loop system is SGFSP. The proposed control scheme has excellent practical value.

This paper is concerned with the tracking control for a class of switched nonlinear systems subject to prescribed performance. Firstly, a barrier function and a normalized function are introduced to achieve prescribed performance control, which ensures that the tracking error evolves within prescribed boundary. Then, multi-dimensional Taylor network provides a new approach for how to estimate the nonlinearity arising from the backstepping control process. Significantly, the proposed multi-switching-based adaptive controller realizes that all signals in the closed-loop system are globally uniformly ultimately bounded while ensuring asymptotic tracking. Different from most existing network-approximation-based control strategies, the developed method in this paper is not only independent of the initial state, but also can achieve global stability. Finally, it is easy to verify the effectiveness of the proposed control method through two simulations. Note to Practitioners —This research is motivated by the fact that the control ideas of many practical engineering systems can be provided by making a profound study on switched nonlinear systems. However, most of the existing results focused on semi-global stability of systems. Therefore, this study devotes to develop a novel adaptive prescribed performance control strategy, which can not only ensure global stability of closed-loop system but also realize the asymptotic tracking of the switched nonlinear systems. It is worth noting that the proposed control approach has great significance for many practical systems, such as circuit systems and single-link inverted pendulum systems.

Estimating the ocean subsurface thermal structure (OSTS) based on multisource sea surface data in the western Pacific Ocean is of great significance for studying ocean dynamics and El Niño phenomenon, but it is challenging to accurately estimate the OSTS from sea surface parameters in the area. This paper proposed an improved neural network model to estimate the OSTS from 0–2000 m from multisource sea surface data including sea surface temperature (SST), sea surface salinity (SSS), sea surface height (SSH), and sea surface wind (SSW). In the model experiment, the rasterized monthly average data from 2005–2015 and 2016 were selected as the training and testing set, respectively. The results showed that the sea surface parameters selected in the paper had a positive effect on the estimation process, and the average RMSE value of the ocean subsurface temperature (OST) estimated by the proposed model was 0.55 °C. Moreover, there were pronounced seasonal variation signals in the upper layers (the upper 200 m), however, this signal gradually diminished with increasing depth. Compared with known estimation models such as the random forest (RF), the multiple linear regression (MLR), and the extreme gradient boosting (XGBoost), the proposed model outperformed these models under the data conditions of the paper. This research can provide an advanced artificial intelligence technique for estimating subsurface thermohaline structure in major sea areas.