Deep-learning-based radar imaging is developed with distributed frequency modulated continuous waveform multiple-input multiple-output (FMCW MIMO) radars in which a deep-learning approach based on the convolutional neural network (CNN) is proposed to achieve radar images robust to adverse circumstances. Differently from the existing deeplearning methods applied to radar object recognition, the deramped radar signal is exploited as the input of the proposed deep CNN (DCNN) without any processing related to the spectrogram transform and the subspace decomposition. To effectively train the proposed DCNN, the received signal is reformulated in terms of the reflection gain values in the (azimuth, range) patches in the image region of interest such that the output vector of the DCNN is composed of the reflection gain values in the associated patches. Furthermore, to overcome the limitations on the amount of training data and training time, the transfer learning approach is effectively applied to the distributed FMCW MIMO radar imaging. The proposed radar imaging is assessed with synthetic simulation data. Specifically, by transferring the pretrained DCNN model for a given reference radar to other distributed radars, the distributed radars can save about 52.4 % in training time compared with a DCNN having the same architecture but without transfer learning.
In this paper, efficient gradient updating strategies are developed for the federated learning when distributed clients are connected to the server via a wireless backhaul link. Specifically, a common convolutional neural network (CNN) module is shared for all the distributed clients and it is trained through the federated learning over wireless backhaul connected to the main server. However, during the training phase, local gradients need to be transferred from multiple clients to the server over wireless backhaul link and can be distorted due to wireless channel fading. To overcome it, an efficient gradient updating method is proposed, in which the gradients are combined such that the effective SNR is maximized at the server. In addition, when the backhaul links for all clients have small channel gain simultaneously, the server may have severely distorted gradient vectors. Accordingly, we also propose a binary gradient updating strategy based on thresholding in which the round associated with all channels having small channel gains is excluded from federated learning. Because each client has limited transmission power, it is effective to allocate more power on the channel slots carrying specific important information, rather than allocating power equally to all channel resources (equivalently, slots). Accordingly, we also propose an adaptive power allocation method, in which each client allocates its transmit power proportionally to the magnitude of the gradient information. This is because, when training a deep learning model, the gradient elements with large values imply the large change of weight to decrease the loss function.
When micro‐Doppler (MD) radars are distributed, a federated learning strategy over wireless backhaul links is developed for motion classification. Specifically to identify the human motion, a common convolutional neural network (CNN) model is shared for all the distributed radars (i.e. clients) and it is trained through the federated learning strategy over wireless backhaul connected to the main server. In the proposed system, a main bottleneck is the estimation of local gradients for CNN training at the server, which are transferred from distributed radars over the wireless backhaul link. To overcome it, a deep learning (DL) aided gradient estimation algorithm is proposed, in which the deep neural networks (DNNs) for encoding local gradient vectors at the distributed radars and the DNN for decoding (i.e. estimating) them at the server are jointly trained in an end‐to‐end autoencoder‐based learning strategy. To avoid the inter‐client interference over the wireless backhaul link, the DNN structure for the gradient estimation algorithm with the orthogonal multiple access is proposed, in which the proposed DNN effectively learns the encoding/decoding at the transceiver over wireless backhaul. By exploiting the experimental data measured through the USRP‐based MD radars, the authors validate the motion classification performance of the proposed federated learning strategy and DL aided gradient estimation over the wireless backhaul link.
In this paper, we propose a cooperative linear discriminant analysis (LDA)-based motion classification algorithm for distributed micro-Doppler (MD) radars which are connected to a data fusion center through the limited backhaul. Due to the limited backhaul, each radar cannot report the high-dimensional data of a multi-aspect angle MD signature to the fusion center. Instead, at each radar, the dimensionality of the MD signature is reduced by using the LDA algorithm and the dimensionally-reduced MD signature can be collected at the data fusion center. To further reduce the burden of backhaul, we also propose the softmax processing method in which the distances of the sensed MD signatures from the centers of clusters for all motion candidates are computed at each radar. The output of the softmax process at each radar is quantized through the pyramid vector quantization with a finite number of bits and is reported to the data fusion center. To improve the classification performance at the fusion center, the channel resources of the backhaul are adaptively allocated based on the classification separability at each radar. The proposed classification performance was assessed with synthetic simulation data as well as experimental data measured through the USRP-based MD radar.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.