Recent multimodal fusion methods, integrating images with LiDAR point clouds, have shown promise in scene flow estimation. However, the fusion of 4D millimeter wave radar and LiDAR remains unexplored. Unlike LiDAR, radar is cheaper, more robust in various weather conditions and can detect point-wise velocity, making it a valuable complement to LiDAR. However, radar inputs pose challenges due to noise, low resolution, and sparsity. Moreover, there is currently no dataset that combines LiDAR and radar data specifically for scene flow estimation. To address this gap, we construct a Radar-LiDAR scene flow dataset based on a public real-world automotive dataset. We propose an effective preprocessing strategy for radar denoising and scene flow label generation, deriving more reliable flow ground truth for radar points out of the object boundaries. Additionally, we introduce RaLiFlow, the first joint scene flow learning framework for 4D radar and LiDAR, which achieves effective radar-LiDAR fusion through a novel Dynamic-aware Bidirectional Cross-modal Fusion (DBCF) module and a carefully designed set of loss functions. The DBCF module integrates dynamic cues from radar into the local cross-attention mechanism, enabling the propagation of contextual information across modalities. Meanwhile, the proposed loss functions mitigate the adverse effects of unreliable radar data during training and enhance the instance-level consistency in scene flow predictions from both modalities, particularly for dynamic foreground areas. Extensive experiments on the repurposed scene flow dataset demonstrate that our method outperforms existing LiDAR-based and radar-based single-modal methods by a significant margin.

4D radar has received significant attention in autonomous driving thanks to its robustness under adverse weathers. Due to the sparse points and noisy measurements of the 4D radar, most of the research finish the 3D object detection task by integrating images from camera and perform modality fusion in BEV space. However, the potential of the radar and the fusion mechanism is still largely unexplored, hindering the performance improvement. In this study, we propose a cross-view two-stage fusion network called CVFusion. In the first stage, we design a radar guided iterative (RGIter) BEV fusion module to generate high-recall 3D proposal boxes. In the second stage, we aggregate features from multiple heterogeneous views including points, image, and BEV for each proposal. These comprehensive instance level features greatly help refine the proposals and generate high-quality predictions. Extensive experiments on public datasets show that our method outperforms the previous state-of-the-art methods by a large margin, with 9.10% and 3.68% mAP improvements on View-of-Delft (VoD) and TJ4DRadSet, respectively. Our code will be made publicly available.

Despite the great success of deep learning in stereo matching, recovering accurate disparity maps is still challenging. Currently, L1 and cross-entropy are the two most widely used losses for stereo network training. Compared with the former, the latter usually performs better thanks to its probability modeling and direct supervision to the cost volume. However, how to accurately model the stereo ground-truth for cross-entropy loss remains largely under-explored. Existing works simply assume that the ground-truth distributions are uni-modal, which ignores the fact that most of the edge pixels can be multimodal. In this paper, a novel adaptive multimodal cross-entropy loss (ADL) is proposed to guide the networks to learn different distribution patterns for each pixel. Moreover, we optimize the disparity estimator to further alleviate the bleeding or mis-alignment artifacts in inference. Extensive experimental results show that our method is generic and can help classic stereo networks regain state-of-the-art performance. In particular, GANet with our method ranks 1ston both the KITTI 2015 and 2012 benchmarks among the published methods. Meanwhile, excellent synthetic-to-realistic generalization performance can be achieved by simply replacing the traditional loss with ours. Code is available at https://github.com/xxxupeng/AdL.

Wireless link scheduling in D2D communication systems aims at maximizing the weighted sum rate of D2D pairs by determining which subset of D2D pairs should be activated. However, it is a non-convex combinatorial optimization problem, which is generally NP-hard and difficult to achieve the optimal solution. Inspired by the recent attempt of introducing machine learning and graph embedding to reach the general goal, we propose an efficient method to solve the weighted sum rate maximization problem. We first model the system as a one-nearest neighbor graph, in which each D2D pair is a node and the strongest interference link for each node is an edge. Then we compute the feature vectors of both weights and nodes by graph embedding and use the feature vectors as the input of a subsequent multi-layer classifier. The parameters of classifier and graph embedding are trained jointly in a supervised manner. Simulation result shows that the proposed method can obtain near-optimal performance with only hundreds of training samples and is capable to be generalized to more complicated scenarios.