No abstract

Classifications based on deep learning have been widely applied in the estimation of the direction of arrival (DOA) of signal. Due to the limited number of classes, the classification of DOA cannot satisfy the required prediction accuracy of signals from random azimuth in real applications. This paper presents a Centroid Optimization of deep neural network classification (CO-DNNC) to improve the estimation accuracy of DOA. CO-DNNC includes signal preprocessing, classification network, and Centroid Optimization. The DNN classification network adopts a convolutional neural network, including convolutional layers and fully connected layers. The Centroid Optimization takes the classified labels as the coordinates and calculates the azimuth of received signal according to the probabilities of the Softmax output. The experimental results show that CO-DNNC is capable of acquiring precise and accurate estimation of DOA, especially in the cases of low SNRs. In addition, CO-DNNC requires lower numbers of classes under the same condition of prediction accuracy and SNR, which reduces the complexity of the DNN network and saves training and processing time.

No abstract

The disturbance of cable force is an important aspect that should be taken into account for precision motion systems involved in CNC machine tools, wafer scanners, etc. In this paper, unlike paying attention to the cable force loaded on long stroke in previous works, the disturbance between long and short stroke had been analyzed in details. To minimize the disturbance force loaded on short stroke of motion system, a disturbance model identification and the corresponding compensation approach method was proposed. A two-step identification in frequency domain was developed such that the parameters of stiffness and damping between long and short stroke can be estimated. Then a state feedback method using ESO and estimated parameters had be applied to compensate the disturbance, which improved the servo performance of short stroke significantly. Finally, the effectiveness of the proposed method is illustrated through numerical simulation.

LMS algorithm is a kind of classic adaptive algorithms. Although it has the virtue of simple operation, it also shows the defects of relatively slow convergence and big steady state errors in low SNR. To remedy these defects, this paper put forward a new variable steps adaptive LMS algorithm. In the transient state, the learning rate increases slowly with the iteration times which accelerate the convergence rate of LMS algorithm. In the steady state, the learning rate decreases gradually with the iteration times which guarantee the convergence accuracy of LMS algorithm. After this improved algorithm is applied in the design of adaptive wavetrap, the simulation results show that it can not only effectively ease up the conflicts between convergence rates and steady state errors, but also improve the performance of wavetrap in real-time trapping.