The synthetic aperture radar (SAR) for ocean surveillance missions requires low latency and light weight inference. This paper proposes a novel small-size Searched Binary Network (SBNN), with network architecture search (NAS) for ship classification with SAR. In SBNN, convolution operations are modified by binarization technologies. Both input feature maps and weights are quantized into 1-bit in most of the convolution computation, which significantly decreases the overall computational complexity. In addition, we propose a patch shift processing, which can adjust feature maps with learnable parameters at spatial level. This process enhances the performance by reducing the information irrelevant to the targets. Experimental results on the OpenSARShip dataset show the proposed SBNN outperforms both binary neural networks from computer vision and CNN-based SAR ship classification methods. In particular, SBNN shows a great advantage in computational complexity.
Ship classification technology using synthetic aperture radar (SAR) has become a research hotspot. Many deep-learning-based methods have been proposed with handcrafted models or using transplanted computer vision networks. However, most of these methods are designed for graphics processing unit (GPU) platforms, leading to limited scope for application. This paper proposes a novel mini-size searched convolutional Metaformer (SCM) for classifying SAR ships. Firstly, a network architecture searching (NAS) algorithm with progressive data augmentation is proposed to find an efficient baseline convolutional network. Then, a transformer classifier is employed to improve the spatial awareness capability. Moreover, a ConvFormer cell is proposed by filling the searched normal convolutional cell into a Metaformer block. This novel cell architecture further improves the feature-extracting capability. Experimental results obtained show that the proposed SCM provides the best accuracy with only weights, achieving a good trade-off between performance and model size.
Greedy pursuit algorithms are widely used in sparse signal processing for their computational efficiency. However, research on the reconstruction error properties is far from comprehensive. This paper derives the statistical properties of reconstruction error in greedy pursuit algorithms, including probability density function (PDF), expectation, and covariance. The reconstruction error follows a mixture distribution, which is composed of multiple multivariate random variables and weighted by the probability of support sets. The multivariate random variable obeys a truncated distribution that results from restricting the noise domain. The validity of the derivations is verified by using the orthogonal matching pursuit (OMP) algorithm as an example.
A novel framework for a low-cost coding digital receiving array based on machine learning (ML-CDRA) is proposed in this paper. The received full-array signals are encoded into a few radio frequency (RF) channels, and decoded by an artificial neural network in real-time. The encoding and decoding networks are studied in detail, including the implementation of the encoding network, the loss function and the complexity of the decoding network. A generalized form of loss function is presented by constraint with maximum likelihood, signal sparsity, and noise. Moreover, a feasible loss function is given as an example and the derivations for back propagation are successively derived. In addition, a real-time processing implementation architecture for ML-CDRA is presented based on the commercial chips. It is possible to implement by adding an additional FPGA on the hardware basis of full-channel DRA. ML-CDRA requires fewer RF channels than the traditional full-channel array, while maintaining a similar digital beamforming (DBF) performance. This provides a practical solution to the typical problems in the existing low-cost DBF systems, such as synchronization, moving target compensation, and being disabled at a low signal-to-noise ratio. The performance of ML-CDRA is evaluated in simulations.
For a large-scale array, the subarray technology can save the radio frequency (RF) channels. However, the use of subarray technology often tends to reduce the beamforming performance of the array. In some specific scenarios, it is necessary to use simultaneous multi-beam technology for target search. Exploring methods to simultaneously achieve multi-beamforming while reducing radio frequency channels is a worthwhile research endeavor. This paper introduces an innovative application of hybrid beamforming structure to the realm of beam synthesis and proposes an algorithm for simultaneous multi-beam pattern synthesis in subarrays. This method weights the received signal of the array twice. Initially, the analog signal received by each antenna unit undergoes the first weighted, accomplished by incorporating a phase shifter after each antenna. After the initial weighted, the analog signals proceed into their respective subarray channels, where they undergo a second step of weighted after being converted into digital signals. The second step of weighted involves digital beamforming of the subarray signals, utilizing multiple sets of weights, to achieve simultaneous multibeam. The algorithm can obtain an element weighted vector and a subarray weighted matrix. Based on the alternating projection (AP) algorithm, the property of the Kronecker product, and the definition of the Khatri-Rao product, the optimal element weighted vector and subarray weighted matrix can be obtained by solving the 2-norm minimization problem.