Feature extraction plays an important role in the field of fault diagnosis of analog circuits. How to effectively extract fault features is crucial to diagnostic accuracy. The components tolerance and circuit nonlinearities of analog circuits can cause some part overlapping of primal signal among different component faults in time domain and frequency domain. Currently, the existing method aims at wavelet features, statistical property features, conventional frequency features and conventional time-domain features. There is no decoupling ability for the feature extraction methods mentioned above. To solve the problem, a new fault features extraction method is proposed. The diagnostic results are compared with those from other methods. Firstly, it is proposed to use the statistical property features of transformed signals by the fractional Fourier transform in the optimal fractional order domain as fault features, such as range, mean, standard deviation, skewness, kurtosis, entropy, median, the third central moment, and centroid. And then, KPCA is used to reduce the dimensionality of candidate features so as to obtain the optimal features. Next, normalization is applied to rescale input features. Finally, extracted features are trained by SVM to diagnose faulty components in analog circuits. The simulation results show that compared with traditional methods, the proposed method is quite efficient to improve diagnostic accuracy.

Many accidents happen under shunting mode when the speed of a train is below 45 km/h. In this mode, train attendants observe the railway condition ahead using the traditional manual method and tell the observation results to the driver in order to avoid danger. To address this problem, an automatic object detection system based on convolutional neural network (CNN) is proposed to detect objects ahead in shunting mode, which is called Feature Fusion Refine neural network (FR-Net). It consists of three connected modules, i.e., the depthwise-pointwise convolution, the coarse detection module, and the object detection module. Depth-wise-pointwise convolutions are used to improve the detection in real time. The coarse detection module coarsely refine the locations and sizes of prior anchors to provide better initialization for the subsequent module and also reduces search space for the classification, whereas the object detection module aims to regress accurate object locations and predict the class labels for the prior anchors. The experimental results on the railway traffic dataset show that FR-Net achieves 0.8953 mAP with 72.3 FPS performance on a machine with a GeForce GTX1080Ti with the input size of 320 × 320 pixels. The results imply that FR-Net takes a good tradeoff both on effectiveness and real time performance. The proposed method can meet the needs of practical application in shunting mode.

No abstract

No abstract