Some artificial intelligence algorithms have gained much attention in the rotating machinery fault diagnosis due to their robust nonlinear regression properties. In addition, existing deep learning algorithms are usually dependent on single signal features, which would lead to the loss of some information or incomplete use of the information in the signal. To address this problem, three kinds of popular signal processing methods, including Fast Fourier Transform (FFT), Short-Time Fourier Transform (STFT) and directly slicing one-dimensional data into the two-dimensional matrix, are used to create four different datasets from raw vibration signal as the input data of four enhancement Convolutional Neural Networks (CNN) models. Then, a fuzzy fusion strategy is used to fuse the output of four CNN models that could analyze the importance of each classifier and explore the interaction index between each classifier, which is different from conventional fusion strategies. To show the performance of the proposed model, an artificial fault bearing dataset and a real-world bearing dataset are used to test the feature extraction capability of the model. The good anti-noise and interpretation characteristics of the proposed method are demonstrated as well.

In this article, the issue of the state and unknown input estimation is ironed out for nonlinear systems subject to random sensor failure (RSF) under the dynamic event-triggering mechanism (DETM). To save energy, DETM is utilized to regulate the frequency of data transmissions from the sensor to the estimator, in which each transmission is delivered only when the absolute error is more significant than a predefined threshold. Moreover, RSF is described by a random variable with a certain distribution function. The purpose of this article is to develop a joint state and unknown input estimator such that in the presence of RSF, DETM as well as nonlinearities, the upper bounds on the estimation error covariances of the state and the unknown input are guaranteed and then minimized by parametrizing the estimator gain properly. In addition, a sufficient condition is constructed for the convergence of the designed estimator, and the monotonicity analysis is also established subsequently. Finally, comprehensive illustrative examples are employed to illustrate the validity of the proposed estimation algorithm.