A progressive surrogate gradient learning for memristive spiking neural network

In recent years, spiking neural networks(SNNs) have received increasing attention of research in the field of artificial intelligence due to their high biological plausibility, low energy consumption, and abundant spatio-temporal information.However, the non-differential spike activity makes SNNs more difficult to train in supervised training. Most existing methods focusing on introducing an approximated derivative to replace it, while they are often based on static surrogate functions. In this paper, we propose a progressive surrogate gradient learning for backpropagation of SNNs, which is able to approximate the step function gradually and to reduce information loss. Furthermore, memristor cross arrays are used for speeding up calculation and reducing system energy consumption for their hardware advantage. The proposed algorithm is evaluated on both static and neuromorphic datasets using fully connected and convolutional network architecture, and the experimental results indicate that our approach has a high performance compared with previous research.

Beamforming analysis based on CSB sin-FDA

This paper studies the adaptive beamforming algorithm based on the frequency diverse array(FDA)array where the interference is located at the same angle(but different range)with the target.We take the cross subarray-based FDA with sinusoidal frequency offset(CSB sin-FDA)as the receiving array instead of the basic FDA.The sampling covariance matrix under insufficient snapshot can be corrected by the automatic diagonal loading method.On the basis of decomposing the mismatched steering vector error into a vertical component and a parallel one,this paper searches the vertical component of the error by the quadratic constraint method.The numerical simulation verifies that the beamformer based on the CSB sin-FDA can effectively hold the mainlobe at the target position when the snapshot is insufficient or the steering vector is mismatched.

A Study on The Multi-Antenna Geometrical Depolarization Channel Modeling

The traditional geometrical depolarization model that single transmitter to single receiver provides a simple method of polarization channel modeling. It can obtain the geometrical depolarization effect of each path if known the antenna configuration, the polarization field radiation pattern and the spatial distribution of scatters. With the development of communication technology, information transmission spectrum is more and more scarce. The original model provides only a single channel polarization state, so the information will be limited that the polarization state carries in the polarization modulation. The research is so significance that how to carries polarization modulation information by using multi-antenna polarization state. However, the present study shows that have no depolarization effect model for multi-antenna systems. In this paper, we propose a multi-antenna geometrical depolarization model. On the basis of a single antenna to calculate the depolarization effect of the model, and through simulation to analysis the main factors that influence the depolarization effect. This article provides a multi-antenna geometrical depolarization channel modeling that can applied to large-scale array antenna, and to some extent increase the speed of information transmission.

Three-dimensional localization of near-field and strictly noncircular sources using steering vector decomposition

The three-dimensional localization problem for noncircular sources in near-field with a centro-symmetric cross array is rarely studied.In this paper,we propose an algorithm with improved estimation performance.We decompose the multiple parameters of the steering vector in a specific order so that it can be converted into the products of several matrices,and each of the matrices includes only one parameter.On this basis,each parameter to be resolved can be estimated by performing a one-dimensional spatial spectral search.Although the computational complexity of the proposed algorithm is several times that of our previous algorithm,the estimation performance,including its error and resolution,with respect to the direction of arrival,is improved,and the range estimation performance can be maintained.The superiority of the proposed algorithm is verified by simulation results.