With recent developments, the performance of automotive radar has improved significantly. The next generation of 4D radar can achieve imaging capability in the form of high-resolution point clouds. In this context, we believe that the era of deep learning for radar perception has arrived. However, studies on radar deep learning are spread across different tasks, and a holistic overview is lacking. This review paper attempts to provide a big picture of the deep radar perception stack, including signal processing, datasets, labelling, data augmentation, and downstream tasks such as depth and velocity estimation, object detection, and sensor fusion. For these tasks, we focus on explaining how the network structure is adapted to radar domain knowledge. In particular, we summarise three overlooked challenges in deep radar perception, including multi-path effects, uncertainty problems, and adverse weather effects, and present some attempts to solve them.

Over the past decade, the size of neural network models has gradually increased in both breadth and depth, leading to a growing interest in the application of neural network pruning. Unstructured pruning provides fine-grained sparsity and achieves better inference acceleration under specific hardware support. Unstructured Pruning at Initialization (PaI) optimizes the iterative pruning pipeline, but sparse weights increase the risk of underfitting during training. More importantly, almost all PaI algorithms focus only on obtaining the best pruning mask without considering whether the retained weights are suitable for training. Introducing Lipschitz constants during model initialization can reduce the risk of model underfitting and overfitting. As a result, we firstly analyze the impact of Lipschitz initialization on model training and propose the Repaired Initialization (ReI) algorithm for common modules with BatchNorm. Then, we utilize the same idea to repair the weight of unstructured pruned model, and name it Repaired Pruning at Initialization Resilience (RePaIR) algorithm. Extensive experiments and demonstrate that our proposed ReI and RePaIR can improve the training robustness of unpruned and pruned models, respectively, and achieve up to 1.7% accuracy gain with the same sparse pruning mask on TinyImageNet. Furthermore, we provide an improved SynFlow algorithm called Repair SynFlow (ReSynFlow), which employs Lipschitz scaling to overcome the problem of score computation in deeper models. ReSynFlow can effectively improve the maximum compression rate and is suitable for deeper models, with an accuracy improvement of up to 1.3% compared to the SynFlow algorithm on TinyImageNet.

Electrocardiogram (ECG), as a crucial fine-grained cardiac feature, has been successfully recovered from radar signals in the literature, but the performance heavily relies on the high-quality radar signal and numerous radar-ECG pairs for training, restricting the applications in new scenarios due to data scarcity. Therefore, this work focuses on radar-based ECG recovery in new scenarios with limited data and proposes a cardio-focusing and -tracking (CFT) algorithm to precisely track the cardiac location to ensure an efficient acquisition of high-quality radar signals. Furthermore, a transfer learning model (RFcardi) is proposed to extract cardio-related information from the radar signal without ECG ground truth based on the intrinsic sparsity of cardiac features, and only a few synchronous radar-ECG pairs are required to fine-tune the pre-trained model for ECG recovery. The experimental results reveal that the proposed CFT can dynamically identify the cardiac location, and the RFcardi model can effectively generate faithful ECG recoveries after using a small number of radar-ECG pairs for training. The code and dataset will be made available after publication.