Multi-Beam LiDAR (MBL) sensors sample the real-world with discrete 3D point clouds (PC) and have become a major and essential 3D sensing capability for autonomous robots. To ensure an accurate point sampling on surfaces, high-resolution MBL sensors (e.g., Ouster OS0-128) are commonly used to collect dense point clouds for robot tasks, including object detection and tracking, simultaneous localization and mapping (SLAM), in applications such as autonomous driving vehicles (ADVs). However, the high cost and large volume/weight/energy consumption of such sensors limit their usage in broader applications such as UAV/UGV swarms with small-scale agents with limited payload. Existing studies on Super-Resolution (SR) upsampling of the PC from low-resolution MBL have not considered the geometry semantics of the scenes, thus resulting in less optimal SR points for downstream subtasks (e.g., SLAM). Thus, this paper proposes SGSR-Net, a structure semantics-guided MBL Super-Resolution network. SGSR-Net takes the low-resolution range images of the MBL sensors as input and produces dense and structure-aware Super-Resolution point cloud from those sparse measurements through a vertical spatial and channel attention-enhanced CNN model coupling with guided Monte Carlo filtering, for indoor LiDAR-SLAM applications. The SGSR-Net is validated using datasets collected by a UGV equipped with multiple MBL sensors. The results demonstrate that the proposed CG-LSR (CASE Attention Guided Encoder-Decoder LiDAR Super-Resolution Network) reduces the MAE of the SR points by 12.4% down to 0.177 m when compared with the state-of-the-art (SOTA) method [1}, [2], [3], [4]. The indoor SLAM results with SR-points produced by SGSR-Net show that the mean and RMSE of the absolute pose error (APE) are decreased by 27% and 30%, down to 0.849 m and 0.902 m, respectively, which significantly improve the indoor-SLAM performance and stability of SOTA LiDAR-SLAM systems (i.e. LeGO-LOAM [5], [6], [7], [8]).

Given the prominence of 3D sensors in recent years, 3D point clouds are worthy to be further investigated for environment perception and scene understanding. Learning accurate local and global contexts in point clouds is pivotal for semantic segmentation, and neighbor aggregation and Transformers have achieved notable success in local and global perception in point cloud analysis, respectively. Nevertheless, studying each independently is far from the optimal solution for comprehensive feature learning. To address this, we take a novel step towards investigating and integrating the structures of neighbor aggregation and Transformers. In this paper, we introduce Point Neighbor Aggregation with Transformer (PointNAT), a conceptually straightforward and effective approach aiming to enhance the performance of 3D point cloud semantic segmentation. PointNAT consists of a Neighbor Aggregation Block (NAB) for local perception, a Point Transformer Block (PTB) for global modeling, and a Hybrid Block to connect NABs and PTBs. NABs effectively learn complex local features at varying scales through an improved neighbor aggregation operation and a multi-head mechanism. PTBs efficiently perform global attention using a small set of learnable key points. Hybrid Blocks serve as high-and-low frequency signal hybridizers, merging the strengths of these two blocks by adaptively assigning hybrid weights to local and global contexts. We have evaluated the performance of PointNAT with state-of-the-art networks on several benchmarks, including S3DIS, Toronto3D, and SensatUrban. PointNAT achieves mIoU scores of 77.8%, 84.7%, and 65.2% in these three dataset, respectively. Furthermore, it outperforms the baseline approach PointNeXt by 3.0%, 1.3%, and 4.2%, respectively, while utilizing only 59.9% of the parameters and 15.2% of the FLOPs. The results demonstrate PointNAT’s superior ability in accurately segmenting large-scale 3D point cloud scenes, emphasizing its potential to advance environment perception and scene understanding. Our code is available at https://github.com/zeng-ziyin/PointNAT.

Autonomous driving technology has achieved significant breakthroughs in open scenarios, enabling the deployment of excellent positioning, detection, and navigation algorithms on passenger vehicles. However, there has been limited research attention devoted to autonomous driving for specialized vehicles in non-open scenarios. This manuscript introduces a perception system designed for heavy-duty mining transportation trucks operating in open-pit mines, which are typical of non-open scenarios. The system comprises four independent algorithms: high-precision fusion positioning, multi-task 2D detection, 9 Degrees of Freedom (9 DoF) 3D head, and autonomous navigation technology. Experimental verification demonstrates the effectiveness of these methods in addressing the challenges posed by mining environments, ultimately leading to enhanced safety and efficiency for trucks. This research outcome, through the comprehensive examination of positioning, detection, and navigation, aims to address the challenges encountered by mining trucks during operations. Its significance lies in enhancing automation levels in mining scenarios.

Lane information plays an important role in high-definition (HD) maps because it provides invaluable road information for autonomous vehicles. The most widely used lane extraction method for HD maps is based on mobile mapping systems, and it is prohibitively costly and time-consuming. In this study, we propose a novel approach for lane information extraction based on crowdsourcing vehicles equipped with monocular cameras and global navigation satellite system devices for recording road images and position data. First, we propose a lane mask propagation network to detect the lane markings in images, which are then projected from a perspective space into a three-dimensional space in accordance with the position data. Second, we propose a data management method to store lane information in the cloud data center. An improved density-based spatial clustering of applications with noise clustering algorithm and a gradual fitting algorithm are used to remove the outliers and improve the lane data accuracy. The proposed method is quantitatively evaluated against a real-world HD map produced by a mobile mapping vehicle. The experimental results show that more than 80% of the extracted lane markings meet the accuracy requirements of HD maps. In conclusion, our method can be used as a low-cost and efficient approach for updating the lane information in HD maps for autonomous vehicles.