Understanding human intentions and actions through egocentric videos is important on the path to embodied artificial intelligence. As a branch of egocentric vision techniques, hand trajectory prediction plays a vital role in comprehending human motion patterns, benefiting downstream tasks in extended reality and robot manipulation. However, capturing high-level human intentions consistent with reasonable temporal causality is challenging when only egocentric videos are available. This difficulty is exacerbated under camera egomotion interference and the absence of affordance labels to explicitly guide the optimization of hand waypoint distribution. In this work, we propose a novel hand trajectory prediction method dubbed MADiff, which forecasts future hand waypoints with diffusion models. The devised denoising operation in the latent space is achieved by our proposed motion-aware Mamba, where the camera wearer's egomotion is integrated to achieve motion-driven selective scan (MDSS). To discern the relationship between hands and scenarios without explicit affordance supervision, we leverage a foundation model that fuses visual and language features to capture high-level semantics from video clips. Comprehensive experiments conducted on five public datasets with the existing and our new evaluation metrics demonstrate that MADiff predicts comparably reasonable hand trajectories compared to the state-of-the-art baselines.

Multi-Layer Perceptron (MLP) models are the foundation of contemporary point cloud processing. However, their complex network architectures obscure the source of their strength and limit the application of these models. In this article, we develop a two-stage abstraction and refinement (ABS-REF) view for modular feature extraction in point cloud processing. This view elucidates that whereas the early models focused on ABS stages, the more recent techniques devise sophisticated REF stages to attain performance advantages. Then, we propose a High-dimensional Positional Encoding (HPE) module to explicitly utilize intrinsic positional information, extending the "positional encoding" concept from Transformer literature. HPE can be readily deployed in MLP-based architectures and is compatible with transformer-based methods. Within our ABS-REF view, we rethink local aggregation in MLP-based methods and propose replacing time-consuming local MLP operations, which are used to capture local relationships among neighbors. Instead, we use non-local MLPs for efficient non-local information updates, combined with the proposed HPE for effective local information representation. We leverage our modules to develop HPENets, a suite of MLP networks that follow the ABS-REF paradigm, incorporating a scalable HPE-based REF stage. Extensive experiments on seven public datasets across four different tasks show that HPENets deliver a strong balance between efficiency and effectiveness. Notably, HPENet surpasses PointNeXt, a strong MLP-based counterpart, by 1.1% mAcc, 4.0% mIoU, 1.8% mIoU and 0.2% Cls. mIoU, with only 50.0%, 21.5%, 23.1%, 44.4% of FLOPs on ScanObjectNN, S3DIS, ScanNet, and ShapeNetPart, respectively.

Outdoor loop closure detection is essential for correcting odometry drift and constructing a globally consistent map. Semantic-graph-based approaches effectively model object-level topology and achieve strong loop closure performance; however, their effectiveness degrades in background-dominated scenes with few distinctive objects, and establishing accurate injective node correspondences remains challenging. In contrast, structural descriptor methods, though offering stronger environmental generality through spatial-distribution modeling, remain susceptible to LiDAR noise and the discriminative power of point-level features. These limitations motivate the need for a more robust method that combines adaptability with enhanced descriptive power. We propose a novel loop-closure detection framework, SAGE, that integrates highly adaptable point-cloud shape-distribution features and generally reliable semantic graph topology, adaptively combining their similarity measures to improve detection performance. Specifically, we design a semantic graph matching module with dual constraints, local graph feature consistency and global spatial consistency, to achieve more accurate injective node correspondences. In addition, we extract point-cloud shape-distribution features and introduce a fusion mechanism that integrates them with the semantic graph module, assessing reliability and adaptively weighting their contributions. Extensive loop closure detection and pose estimation experiments on various datasets demonstrate that SAGE achieves superior performance over strong baselines. We provide the code at https://github.com/SAGE-11/SAGE.

Multi-robot collaboration is becoming increasingly critical and presents significant challenges in modern robotics, especially for building a globally consistent, accurate map. Traditional multi-robot pose graph optimization (PGO) methods ensure basic global consistency but ignore the geometric structure of the map, and only use loop closures as constraints between pose nodes, leading to divergence and blurring in overlapping regions. To address this issue, we propose LEMON-Mapping, a loop-enhanced framework for large-scale, multi-session point cloud fusion and optimization. We re-examine the role of loops for multi-robot mapping and introduce three key innovations. First, we develop a robust loop processing mechanism that rejects outliers and a loop recall strategy to recover mistakenly removed but valid loops. Second, we introduce spatial bundle adjustment for multi-robot maps, reducing divergence and eliminating blurring in overlaps. Third, we design a PGO-based approach that leverages refined bundle adjustment constraints to propagate local accuracy to the entire map. We validate LEMON-Mapping on several public datasets and a self-collected dataset. The experimental results show superior mapping accuracy and global consistency of our framework compared to traditional merging methods. Scalability experiments also demonstrate its strong capability to handle scenarios involving numerous robots.

Reliable localization under Global Positioning System-denied or visually degraded conditions remains a fundamental challenge for autonomous systems. Vision- and Light Detection and Ranging (LiDAR)-based approaches often degrade in low illumination, adverse weather, or appearance-changing environments, as they rely on stable surface texture or geometry. In contrast, ground-penetrating radar (GPR) captures subsurface electromagnetic reflections that remain relatively stable across lighting, seasonal, and weather variations, making it a promising complementary sensing modality for long-term localization. However, spatial variability in subsurface dielectric properties induces fluctuations in electromagnetic wave velocity, leading to geometric distortions in GPR echoes and unstable feature extraction. To address this challenge, we propose the Velocity-Invariant Feature Transform (VIFT), a physics-guided self-supervised learning framework for GPR keypoint detection and description. VIFT explicitly models wave-velocity-induced distortions through a continuous velocity space parameterized by a Beta distribution, and leverages velocity-conditioned wavefield migration as physically consistent data augmentation. A Siamese network is trained with velocity-consistency supervision to jointly learn repeatable keypoint score maps and discriminative local descriptors from unlabeled real GPR scans. To further enhance robustness, sparsity-aware, dispersion, distinctiveness, and orthogonality losses are incorporated to improve repeatability, spatial coverage, and descriptor discriminability. Extensive experiments on public benchmarks and large-scale real-world GPR datasets demonstrate that VIFT consistently outperforms traditional handcrafted methods and recent learning-based Vison and GPR methods, achieving a 5–10% improvement in keypoint repeatability over state-of-the-art methods, particularly under extremely sparse keypoint sampling regimes, while also improving matching accuracy and registration robustness under diverse subsurface conditions.

Ground penetrating radar (GPR) based localization has gained significant recognition in robotics due to its ability to detect stable subsurface features, offering advantages in environments where traditional sensors like cameras and LiDAR may struggle. However, existing methods are primarily focused on small-scale place recognition (PR), leaving the challenges of PR in large-scale maps unaddressed. These challenges include the inherent sparsity of underground features and the variability in underground dielectric constants, which complicate robust localization. In this work, we investigate the geometric relationship between GPR echo sequences and underground scenes, leveraging the robustness of directional features to inform our network design. We introduce learnable Gabor filters for the precise extraction of directional responses, coupled with a direction-aware attention mechanism for effective geometric encoding. To further enhance performance, we incorporate a shift-invariant unit and a multi-scale aggregation strategy to better accommodate variations in dielectric constants. Experiments conducted on public datasets demonstrate that our proposed EDENet not only surpasses existing solutions in terms of PR performance but also offers advantages in model size and computational efficiency.

The task of occupancy forecasting (OCF) involves utilizing past and present perception data to predict future occupancy states of autonomous vehicle surrounding environments, which is critical for downstream tasks such as obstacle avoidance and path planning. Existing 3D OCF approaches struggle to predict plausible spatial details for movable objects and suffer from slow inference speeds due to neglecting the bias and uneven distribution of changing occupancy states in both space and time. In this paper, we propose a novel spatiotemporal decoupling vision-based paradigm to explicitly tackle the bias and achieve both effective and efficient 3D OCF. To tackle spatial bias in empty areas, we introduce a novel spatial representation that decouples the conventional dense 3D format into 2D bird’s-eye view (BEV) occupancy with corresponding height values, enabling 3D OCF derived only from 2D predictions thus enhancing efficiency. To reduce temporal bias on static voxels, we design temporal decoupling to improve end-to-end OCF by temporally associating instances via predicted flows. We develop an efficient multi-head network EfficientOCF to achieve 3D OCF with our devised spatiotemporally decoupled representation. A new metric, conditional IoU (C-IoU), is also introduced to provide a robust 3D OCF performance assessment, especially in datasets with missing or incomplete annotations. The experimental results demonstrate that EfficientOCF surpasses existing baseline methods on accuracy and efficiency, achieving state-of-the-art performance with a fast inference time of 82.33 ms with a single GPU. Our code is released at: https://github.com/BIT-XJY/EfficientOCF.

Semantic segmentation is a key technique that enables mobile robots to understand and navigate surrounding environments autonomously. However, most existing works focus on segmenting known objects, overlooking the identification of unknown classes, which is common in real-world applications. In this paper, we propose a feature-oriented framework for open-set semantic segmentation on LiDAR data, capable of identifying unknown objects while retaining the ability to classify known ones. We design a decomposed dual-decoder network to simultaneously perform closed-set semantic segmentation and generate distinctive features for unknown objects. The network is trained with multi-objective loss functions to capture the characteristics of known and unknown objects. Using the extracted features, we introduce an anomaly detection mechanism to identify unknown objects. By integrating the results of close-set semantic segmentation and anomaly detection, we achieve effective feature-driven LiDAR open-set semantic segmentation. Evaluations on both SemanticKITTI and nuScenes datasets demonstrate that our proposed framework significantly outperforms state-of-the-art methods. The source code will be made publicly available at https://github.com/nubot-nudt/DOSS.

Place recognition is an important technique for autonomous mobile robotic applications. While single-modal sensor-based approaches have shown satisfactory performance, cross-modal place recognition remains underexplored due to the challenge of bridging the cross-modal heterogeneity gap. In this work, we introduce an instance-aware cross-modal place recognition approach, named InsCMPR. We design a novel instance-aware modality alignment module, which aligns multi-modal data at both pixel-level and instance-level by leveraging a pre-trained vision foundation model SAM. Then a novel dual-branch hybrid Mamba-Transformer network is proposed to efficiently enhance the distinctiveness of the produced descriptors by integrating global features with local instance features. Experimental results on the KITTI, NCLT, and HAOMO datasets show that our proposed methods achieve state-of-the-art performance while operating in real time. We will open source the implementation of our method at: https://github.com/nubot-nudt/InsCMPR.

While millimeter-wave radars are widely used in robotics and autonomous driving, extrinsic calibration with other sensors remains challenging due to the sparsity and uncertainty of radar point clouds. In this paper, we propose a novel deep feature-matching-based online extrinsic calibration approach for a 4D millimeter-wave radar and 3D LiDAR system. We formulate the calibration problem as a crossmodal point cloud registration task, initiating with keypointlevel matching followed by dense matching refinement. Efficient yet powerful neural networks are employed to extract prior keypoint matches, which are then expanded to surrounding regions, establishing dense point correspondences. Our approach effectively leverages the majority of the information from millimeter-wave radar, mitigating the impact of radar point cloud sparsity. We evaluate our approach on two datasets, and experimental results demonstrate that it outperforms state-of-the-art baseline methods and achieves an average improvement of 66.96% in calibration success rate, while reducing translational error and rotational error by 23.84% and 30.31%, respectively. Our implementation will be made open-source at https://github.com/nubot-nudt/RLCNet.

Place recognition is an important component for autonomous robot navigation. Many existing LiDAR-based place recognition methods encode the structural information of 3D LiDAR data into 2D image representations. However, most of these intermediates only exploit the projection in a single view, ignoring a great amount of useful information. In this paper, a compact fusion-view image representation of LiDAR point cloud is proposed to extract important structural information from different views. Our proposed method generates such fusion-view images using the corresponding geometric information among points highlighting the edges of objects. It then extracts texture features encoding the shapes and layouts of scene elements into global descriptors, where regional features are designed to adapt local discrepancies caused by seasonal changes. Extensive experiments on the Oxford RobotCar, NCLT, UTBM datasets and our cross-season dataset validate the proposed method and demonstrate its superior generalization performance under different LiDAR sensors and season shifts. Moreover, our proposed method can operate online with a single CPU, making it suitable for resource-limited real robot platforms.

Point cloud prediction (PCP) aims to forecast future 3D point clouds of scenes by leveraging sequential historical LiDAR scans, offering a promising avenue to enhance the perceptual capabilities of autonomous systems. However, existing methods mostly adopt an end-to-end approach without explicitly modeling moving instances, limiting their effectiveness in dynamic real-world environments. In this paper, we propose IMPNet, a novel instance motion-aware network for future point cloud scene prediction. Unlike prior works, IMPNet explicitly incorporates motion and instance-level information to enhance PCP accuracy. Specifically, we extract appearance and motion features from range images and residual images using a dual-branch convolutional network and fuse them via a motion attention block. Our framework further integrates a motion head for identifying moving objects and an instance-assisted training strategy to improve instance-wise point cloud predictions. Extensive experiments on multiple datasets demonstrate that our proposed network achieves state-of-the-art (SOTA) performance in PCP with superior predictive accuracy and robust generalization across diverse driving scenarios. Our method has been released at https://github.com/nubot-nudt/IMPNet.

In this paper, we introduce a novel image-goal navigation approach, named RFSG. Our focus lies in leveraging the fine-grained connections between goals, observations, and the environment within limited image data, all the while keeping the navigation architecture simple and lightweight. To this end, we propose the spatial-channel attention mechanism, enabling the network to learn the importance of multi-dimensional features to fuse the goal and observation features. In addition, a self-distillation mechanism is incorporated to further enhance the feature representation capabilities. Given that the navigation task needs surrounding environmental information for more efficient navigation, we propose an image scene graph to establish feature associations at both the image and object levels, effectively encoding the surrounding scene information. Cross-scene performance validation was conducted on the Gibson and HM3D datasets, and the proposed method achieved state-of-the-art results among mainstream methods, with a speed of up to 53.5 frames per second on an RTX3080. This contributes to the realization of end-to-end image-goal navigation in real-world scenarios. The implementation and model of our method have been released at: https://github.com/nubot-nudt/RFSG.

As lunar exploration missions grow increasingly complex, ensuring safe and autonomous rover-based surface exploration has become one of the key challenges in lunar exploration tasks. In this work, we have developed a lunar surface simulation system called the Lunar Exploration Simulator System (LESS) and the LunarSeg dataset, which provides RGB-D data for lunar obstacle segmentation that includes both positive and negative obstacles. Additionally, we propose a novel two-stage segmentation network called LuSeg. Through contrastive learning, it enforces semantic consistency between the RGB encoder from Stage I and the depth encoder from Stage II. Experimental results on our proposed LunarSeg dataset and additional public real-world NPO road obstacle dataset demonstrate that LuSeg achieves state-of-the-art segmentation performance for both positive and negative obstacles while maintaining a high inference speed of approximately 57 Hz. We have released the implementation of our LESS system, LunarSeg dataset, and the code of LuSeg at: https://github.com/nubot-nudt/LuSeg.

LiDAR-based place recognition (LPR) is a key component for autonomous driving, and its resilience to environmental corruption is critical for safety in high-stakes applications. While state-of-the-art (SOTA) LPR methods perform well in clean weather, they still struggle with weather-induced corruption commonly encountered in driving scenarios. To tackle this, we propose ResLPRNet, a novel LiDAR data restoration network that largely enhances LPR performance under adverse weather by restoring corrupted LiDAR scans using a wavelet transform-based network. ResLPRNet is efficient, lightweight and can be integrated plug-and-play with pretrained LPR models without substantial additional computational cost. Given the lack of LPR datasets under adverse weather, we introduce ResLPR, a novel benchmark that examines SOTA LPR methods under a wide range of LiDAR distortions induced by severe snow, fog, and rain conditions. Experiments on our proposed WeatherKITTI and WeatherNCLT datasets demonstrate the resilience and notable gains achieved by using our restoration method with multiple LPR approaches in challenging weather scenarios. Our code and benchmark are publicly available here: https://github.com/nubot-nudt/ResLPR.

Place recognition is a fundamental technology of high relevance for autonomous robot navigation. Existing methods encounter significant challenges arising from scene variations (e.g., illumination changes, dynamic objects), view-point shifts, and difficulties in data fusion and alignment. These factors often lead to a substantial drop in recognition recall, which is typically addressed in the literature by training deep neural networks to learn invariant feature representations. In this paper, we propose MRMT-PR, a novel multi-scale reverse-view Mamba-Transformer architecture for LiDAR-based place recognition that uses a single-frame point cloud as its input. Our MRMT-PR framework consists of a multi-scale reverse-view preprocessing module for LiDAR point clouds, a Mamba-Transformer feature encoder, and a global feature fusion module. This architecture effectively mitigates the impact of perspective and illumination variations, enhances the global representational capacity of LiDAR features, and significantly improves recognition robustness under challenging conditions such as viewpoint changes and long-term localization. Experiments conducted on NCLT dataset with challenging scenarios demonstrate that MRMT-PR outperforms existing LiDAR-based place recognition baselines in terms of overall performance.

Predicting hand motion is critical for understanding human intentions and bridging the action space between human movements and robot manipulations. Existing hand trajectory prediction (HTP) methods forecast the future hand waypoints in 3D space conditioned on past egocentric observations. However, such models are only designed to accommodate 2D egocentric video inputs. There is a lack of awareness of multimodal environmental information from both 2D and 3D observations, hindering the further improvement of 3D HTP performance. In addition, these models overlook the synergy between hand movements and headset camera egomotion, either predicting hand trajectories in isolation or encoding egomotion only from past frames. To address these limitations, we propose novel diffusion models (MMTwin) for multimodal 3D hand trajectory prediction. MMTwin is designed to absorb multi-modal information as input encompassing 2D RGB images, 3D point clouds, past hand waypoints, and text prompt. Besides, two latent diffusion models, the egomotion diffusion and the HTP diffusion as twins, are integrated into MMTwin to predict camera egomotion and future hand trajectories concurrently. We propose a novel hybrid Mamba-Transformer module as the denoising model of the HTP diffusion to better fuse multimodal features. The experimental results on three publicly available datasets and our self-recorded data demonstrate that our proposed MMTwin can predict plausible future 3D hand trajectories compared to the state-of-the-art baselines, and generalizes well to unseen environments. The code and pretrained models will be released at https://github.com/IRMVLab/MMTwin.

Understanding how humans would behave during hand-object interaction (HOI) is vital for applications in service robot manipulation and extended reality. To achieve this, some recent works simultaneously forecast hand trajectories and object affordances on human egocentric videos. The joint prediction serves as a comprehensive representation of future HOI in 2D space, indicating potential human motion and motivation. However, the existing approaches mostly adopt the autoregressive paradigm, which lacks bidirectional constraints within the holistic future sequence, and accumulates errors along the time axis. Meanwhile, they overlook the effect of camera egomotion on first-person view predictions. To address these limitations, we propose a novel diffusion-based HOI prediction method, namely Diff-IP2D, to forecast future hand trajectories and object affordances with bidirectional constraints in an iterative non-autoregressive manner on egocentric videos. Motion features are further integrated into the conditional denoising process to enable Diff-IP2D aware of the camera wearer’s dynamics for more accurate interaction prediction. Extensive experiments demonstrate that Diff-IP2D significantly outperforms the state-of-the-art baselines on both the off-the-shelf and our newly proposed evaluation metrics. This highlights the efficacy of leveraging a generative paradigm for 2D HOI prediction. The code and the video have been released at https://github.com/IRMVLab/Diff-IP2D.

Forecasting how human hands would move around target objects on egocentric videos can provide prior knowledge to enhance the path planning capabilities of service robots and assistive wearable devices. During the hand-object interaction process, head movements always occur concurrently to provide observations for the interaction scene from different egocentric views. Although some prior works have successfully integrated head motion information into hand trajectory prediction (HTP), they basically overlook the multi-view consistency (MVC) inherent in headset camera egomotion. We argue that multi-view consistency reveals geometric and semantic relationships during hand-object interaction, and can be regarded as additional supervision signals for predicting more realistic hand trajectories. Therefore, in this work, we propose a novel learning scheme dubbed EER to improve diffusion-based 2D hand trajectory prediction methods, which involves exploiting the geometric consistency, enhancing the multi-canvas consistency, and reconstructing the semantic consistency inherent in MVC. The experimental results show that our proposed EER scheme significantly improves the prediction accuracy of existing diffusion-based 2D HTP methods on the publicly available datasets. We will release the code as open-source at https://github.com/IRMVLab/EER-HTP.

Articulated tracked robots face significant challenges in maintaining stable locomotion over uneven terrain due to unknown contact points between tracks and ground, which are critical for dynamic control. Unlike legged robots, where contact locations can be predicted, tracked systems require real-time adaptation to varying terrains. This paper presents C-TRAC, a terrain-adaptive control framework that integrates reinforcement learning with a contact-modeling variational autoencoder (C-VAE) to enable robust obstacle traversal. We first train a C-VAE in simulation to reconstruct high-fidelity contact information (position and binary probability) from noisy sensor measurements. This model learns a latent representation of terrain contacts, capturing complex interactions between the robot’s kinematics and environment. Subsequently, we employ an asymmetric Soft Actor-Critic (SAC) algorithm to optimize a control policy that leverages the predicted contact data for adaptive track control during locomotion. Extensive experiments validate C-TRAC in both simulated and real-world scenarios. In benchmark tests against state-of-the-art (SOTA) methods using RoboCup Rescue Robot League environments, our approach achieves superior obstacle traversal speed (up to 66.67% faster on 45◦staircase) and stability (up to 47.53% more stable on the oblique terrace) compared to contact-agnostic RL baselines and model-based methods. Notably, zero-shot sim-to-real transfer demonstrates consistent performance in unstructured outdoor ruins, also confirming the framework’s practicality.

Multimodal 3D object detectors leverage the strengths of both geometry-aware LiDAR point clouds and semantically rich RGB images to enhance detection performance. However, the inherent heterogeneity between these modalities, including unbalanced convergence and modal misalignment, poses significant challenges. Meanwhile, the large size of the detection-oriented feature also constrains existing fusion strategies to capture long-range dependencies for the 3D detection tasks. In this work, we introduce a fast yet effective multimodal 3D object detector, incorporating our proposed Instance-level Contrastive Distillation (ICD) framework and Cross Linear Attention Fusion Module (CLFM). ICD aligns instance-level image features with LiDAR representations through object-aware contrastive distillation, ensuring fine-grained cross-modal consistency. Meanwhile, CLFM presents an efficient and scalable fusion strategy that enhances cross-modal global interactions within sizable multimodal BEV features. Extensive experiments on the KITTI and nuScenes 3D object detection benchmarks demonstrate the effectiveness of our methods. Notably, our 3D object detector outperforms state-of-the-art (SOTA) methods while achieving superior efficiency. The implementation of our method has been released as open-source at: https://github.com/nubot-nudt/ICD-Fusion.

Accurate and robust simultaneous localization and mapping (SLAM) is crucial for autonomous mobile systems, typically achieved by leveraging the geometric features of the environment. Incorporating semantics provides a richer scene representation that not only enhances localization accuracy in SLAM but also enables advanced cognitive functionalities for downstream navigation and planning tasks. Existing pointwise semantic LiDAR SLAM methods often suffer from poor efficiency and generalization, making them less robust in diverse real-world scenarios. In this paper, we propose a semantic graph-enhanced SLAM framework, named SG-SLAM, which effectively leverages the geometric, semantic, and topological characteristics inherent in environmental structures. The semantic graph serves as a fundamental component that facilitates critical functionalities of SLAM, including robust relocalization during odometry failures, accurate loop closing, and semantic graph map construction. Our method employs a dual-threaded architecture, with one thread dedicated to online odometry and relocalization, while the other handles loop closure, pose graph optimization, and map update. This design enables our method to operate in real time and generate globally consistent semantic graph maps and point cloud maps. We extensively evaluate our method across the KITTI, MulRAN, and Apollo datasets, and the results demonstrate its superiority compared to state-of-the-art methods. Our method has been released at https://github.com/nubot-nudt/SG-SLAM.

Localization is one of the core parts of modern robotics. Classic localization methods typically follow the retrieve-then-register paradigm, achieving remarkable success. Recently, the emergence of end-to-end localization approaches has offered distinct advantages, including a streamlined system architecture and the elimination of the need to store extensive map data. Although these methods have demonstrated promising results, current end-to-end localization approaches still face limitations in robustness and accuracy. Bird’s-Eye-View (BEV) image is one of the most widely adopted data representations in autonomous driving. It significantly reduces data complexity while preserving spatial structure and scale consistency, making it an ideal representation for localization tasks. However, research on BEV-based end-to-end localization remains notably insufficient. To fill this gap, we propose BEVDiffLoc, a novel framework that formulates LiDAR localization as a conditional generation of poses. Leveraging the properties of BEV, we first introduce a specific data augmentation method to significantly enhance the diversity of input data. Then, the Maximum Feature Aggregation Module and Vision Transformer are employed to learn robust features while maintaining robustness against significant rotational view variations. Finally, we incorporate a diffusion model that iteratively refines the learned features to recover the absolute pose. Extensive experiments on the Oxford Radar RobotCar and NCLT datasets demonstrate that BEVDiffLoc outperforms the baseline methods. Our code is available at https://github.com/nubot-nudt/BEVDiffLoc.

Semantic segmentation and stereo matching, respectively analogous to the ventral and dorsal streams in our human brain, are two key components of autonomous driving perception systems. Addressing these two tasks with separate networks is no longer the mainstream direction in developing computer vision algorithms, particularly with the recent advances in large vision models and embodied artificial intelligence. The trend is shifting towards combining them within a joint learning framework, especially emphasizing feature sharing between the two tasks. The major contributions of this study lie in comprehensively tightening the coupling between semantic segmentation and stereo matching. Specifically, this study makes three key contributions: (1) a tightly coupled, gated feature fusion strategy, (2) a hierarchical deep supervision strategy, and (3) a coupling tightening loss function. The combined use of these technical contributions results in TiCoSS, a state-of-the-art joint learning framework that simultaneously tackles semantic segmentation and stereo matching. Through extensive experiments on the KITTI, vKITTI2, and Cityscapes datasets, along with both qualitative and quantitative analyses, we validate the effectiveness of our developed strategies and loss function. Our approach demonstrates superior performance compared to prior arts, with a notable increase in mean intersection over union by over 9%.

4D LiDAR semantic segmentation classifies the semantic category of each LiDAR point and detects whether it is dynamic, a critical ability for tasks like obstacle avoidance and autonomous navigation. Existing approaches often rely on computationally heavy 4D convolutions or recursive networks, which result in poor real-time performance. In this paper, we introduce SegNet4D, a novel real-time 4D semantic segmentation network, offering both efficiency and strong semantic understanding. SegNet4D addresses 4D segmentation as two tasks: single-scan semantic segmentation and moving object segmentation, each tackled by a separate network head. Both results are combined in a motion-semantic fusion module to achieve comprehensive 4D segmentation. Additionally, instance information is extracted from the current scan and exploited for instance-wise segmentation consistency. Extensive experiments on the SemanticKITTI and nuScenes datasets demonstrate that our method outperforms the state-of-the-art in both 4D semantic segmentation and moving object segmentation. Through detailed runtime analysis, our method shows greater efficiency, enabling real-time operation. Besides, its effectiveness and efficiency have also been validated on a real-world robotic platform. The implementation of our method has been released at https: //github.com/nubot-nudt/SegNet4D.

Visual place recognition (VPR) has played an essential role in simultaneous localization and mapping-based mobile robotics and autonomous driving in the past decade, which can identify previously visited places by matching the current observed view against a view database for global localization and loop closure. However, existing VPR methods always suffer fromfalse place recognitiondue to the following issues: insensitive spatial–temporal embedding extraction, lack of multiview descriptor for matching, and misconsideration of false views for model optimization. To address these issues, a novel robust framework calledVPR under variational views (VPR-VV)is proposed. VPR-VV is integrated with: a sequence encoder to extract robust spatial–temporal features from a view sequence, then a hierarchical view retrieval module is employed for multiview feature descriptor aggregation, and a novel enhanced ranking feedback average precision loss with the normalized discounted cumulative gain metric is designed for model optimization. As a result, VPR-VV can significantly enhance the accuracy and robustness of VPR for robot localization under variational views. Experiments with ablation studies are conducted on various challenging indoor and outdoor datasets, and our framework’s superiority is demonstrated: VPR-VV outperforms state-of-the-art (SOTA) methods by up to 9.4% in recall@1, and real-time inference is achieved without additional memory or computational overhead.

Autonomous surface vehicles (ASVs) are widely employed in various marine industrial tasks. Regardless of the specific operation, one common process is docking, particularly the ASV's return to port. Therefore, the docking process is crucial to ASV operations in industrial applications. Docking space perception (DSP) provides essential environmental information, such as the docking space, driving space, and the shore. However, only a few studies have utilized marks for cooperative docking space perception, and no research has explored noncooperative perception. This article proposes a novel docking space bird's eye view (DSBEV) network to segment docking space on the bird's eye view (BEV) plane for the DSP task. DSBEV does not require any marks and can operate directly in port environments. DSBEV network includes an undistortion-splat module and an attention-water module, designed to enhance the DSP performance. A new watershore loss is introduced to generate the docking space utilizing the shape information. Due to the lack of the docking perception dataset, we construct a water surface platform to create a docking space dataset. The experimental results demonstrate that our DSBEV network achieved an average IoU of 76.8%, attaining state-of-the-art performance. Module analyses and ablation studies are conducted to validate the effectiveness of our designed modules.

Category-level object pose estimation and tracking has achieved impressive progress in computer vision, augmented reality, and robotics. Existing methods either estimate the object states from a single observation or only track the 6-DoF pose of a single object. In this paper, we focus on category-level multi-object 9-Dimensional (9D) state tracking from the point cloud stream. We propose a novel 9D state estimation network to estimate the 6-DoF pose and 3D size of each instance in the scene. It uses our devised multi-scale global attention and object-level local attention modules to obtain representative latent features to estimate the 9D state of each object in the current observation. We then integrate our network estimation into a Kalman filter to combine previous states with the current estimates and achieve multi-object 9D state tracking. Experiment results on two public datasets show that our method achieves state-of-the-art performance on both category-level multi-object state estimation and pose tracking tasks. Furthermore, we directly apply the pre-trained model of our method to our air-ground robot system with multiple moving objects. Experiments on our collected real-world dataset show our method's strong generalization ability and real-time pose tracking performance.

Global localization using onboard perception sensors, such as cameras and light detection and ranging (LiDAR) sensors, is crucial in autonomous driving and robotics applications when Global Positioning System (GPS) signals are unreliable. Most approaches achieve global localization by sequential place recognition (PR) and pose estimation (PE). Some methods train separate models for each task, while others employ a single model with dual heads, trained jointly with separate task-specific losses. However, the accuracy of localization heavily depends on the success of PR, which often fails in scenarios with significant changes in viewpoint or environmental appearance. Consequently, this renders the final PE of localization ineffective. To address this, we introduce a new paradigm,PR-by-PE localization, which bypasses the need for separate PR by directly deriving it from PE. We propose RING#, an end-to-endPR-by-PE localizationnetwork that operates in the bird's-eye-view (BEV) space, compatible with both vision and LiDAR sensors. RING# incorporates a novel design that learns two equivariant representations from BEV features, enabling globally convergent and computationally efficient PE. Comprehensive experiments on the north campus long-term vision and LiDAR (NCLT) and Oxford datasets show that RING# outperforms state-of-the-art methods in both vision and LiDAR modalities, validating the effectiveness of the proposed approach.

This article introduces BEVPlace++, a novel, fast, and robust LiDAR global localization method for unmanned ground vehicles. It uses lightweight convolutional neural networks (CNNs) on Bird's Eye View (BEV) image-like representations of LiDAR data to achieve accurate global localization through place recognition, followed by 3-DoF pose estimation. Our detailed analyses reveal an interesting fact that CNNs are inherently effective at extracting distinctive features from LiDAR BEV images. Remarkably, keypoints of two BEV images with large translations can be effectively matched using CNN-extracted features. Building on this insight, we design a Rotation Equivariant Module (REM) to obtain distinctive features while enhancing robustness to rotational changes. A Rotation Equivariant and Invariant Network (REIN) is then developed by cascading REM and a descriptor generator, NetVLAD, to sequentially generate rotation equivariant local features and rotation invariant global descriptors. The global descriptors are used first to achieve robust place recognition, and then local features are used for accurate pose estimation. Experimental results on seven public datasets and our UGV platform demonstrate that BEVPlace++, even when trained on a small dataset (3000 frames of KITTI) only with place labels, generalizes well to unseen environments, performs consistently across different days and years, and adapts to various types of LiDAR scanners. BEVPlace++ achieves state-of-theart performance in multiple tasks, including place recognition, loop closure detection, and global localization. Additionally, BEVPlace++ is lightweight, runs in real-time, and does not require accurate pose supervision, making it highly convenient for deployment. The source codes are publicly available at https://github.com/zjuluolun/BEVPlace2.

Moving object segmentation (MOS) and Ego velocity estimation (EVE) are vital capabilities for mobile systems to achieve full autonomy. Several approaches have attempted to achieve MOSEVE using a LiDAR sensor. However, LiDAR sensors are typically expensive and susceptible to adverse weather conditions. Instead, millimeter-wave radar (MWR) has gained popularity in robotics and autonomous driving for real applications due to its cost-effectiveness and resilience to bad weather. Nonetheless, publicly available MOSEVE datasets and approaches using radar data are limited. Some existing methods adopt point convolutional networks from LiDAR-based approaches, ignoring the specific artifacts and the valuable radial velocity information of radar measurements, leading to suboptimal performance. In this paper, we propose a novel transformer network that effectively addresses the sparsity and noise issues and leverages the radial velocity measurements of radar points using our devised radar self- and cross-attention mechanisms. Based on that, our method achieves accurate EVE of the robot and performs MOS using only radar data simultaneously. To thoroughly evaluate the MOSEVE performance of our method, we annotated the radar points in the public View-of-Delft (VoD) dataset and additionally constructed a new radar dataset in various environments. The experimental results demonstrate the superiority of our approach over existing state-of-the-art methods. The code is available at https://github.com/ORCAUboat/RadarMOSEVE.

Understanding how the surrounding environment changes is crucial for performing downstream tasks safely and reliably in autonomous driving applications. Recent occupancy estimation techniques using only camera images as input can provide dense occupancy representations of large-scale scenes based on the current observation. However, they are mostly limited to representing the current 3D space and do not consider the future state of surrounding objects along the time axis. To extend camera-only occupancy estimation into spatiotemporal prediction, we propose Cam4DOcc, a new benchmark for camera-only 4D occupancy forecasting, evaluating the surrounding scene changes in a near future. We build our benchmark based on multiple publicly available datasets, including nuScenes, nuScenes-Occupancy, and Lyft-Level5, which provides sequential occupancy states of general movable and static objects, as well as their 3D backward centripetal flow. To establish this benchmark for future research with comprehensive comparisons, we introduce four baseline types from diverse camera-based perception and prediction implementations, including a static-world occupancy model, voxelization of point cloud prediction, 2D-3D instance-based prediction, and our proposed novel end-to-end 4D occupancy forecasting network. Furthermore, the standardized evaluation protocol for preset multiple tasks is also provided to compare the performance of all the proposed baselines on present and future occupancy estimation with respect to objects of interest in autonomous driving scenarios. The dataset and our implementation of all four baselines in the proposed Cam4DOcc benchmark are released as open source at https://github.com/haomo-ai/Cam4DOcc.

Moving object segmentation (MOS) provides a reliable solution for detecting traffic participants and thus is of great interest in the autonomous driving field. Dynamic capture is always critical in the MOS problem. Previous methods capture motion features from the range images directly. Differently, we argue that the residual maps provide greater potential for motion information, while range images contain rich semantic guidance. Based on this intuition, we propose MF-MOS, a novel motion-focused model with a dual-branch structure for LiDAR moving object segmentation. Novelly, we decouple the spatial-temporal information by capturing the motion from residual maps and generating semantic features from range images, which are used as movable object guidance for the motion branch. Our straightforward yet distinctive solution can make the most use of both range images and residual maps, thus greatly improving the performance of the LiDAR-based MOS task. Remarkably, our MF-MOS achieved a leading IoU of 76.7% on the MOS leaderboard of the SemanticKITTI dataset upon submission, demonstrating the current state-of-the-art performance. The implementation of our MF-MOS has been released at https://github.com/SCNU-RISLAB/MF-MOS.

High-definition (HD) semantic map generation of the environment is an essential component of autonomous driving. Existing methods have achieved good performance in this task by fusing different sensor modalities, such as LiDAR and camera. However, current works are based on raw data or network feature-level fusion and only consider short-range HD map generation, limiting their deployment to realistic autonomous driving applications. In this paper, we focus on the task of building the HD maps in both short ranges, i.e., within 30m, and also predicting long-range HD maps up to 90m, which is required by downstream path planning and control tasks to improve the smoothness and safety of autonomous driving. To this end, we propose a novel network named SuperFusion, exploiting the fusion of LiDAR and camera data at multiple levels. We use LiDAR depth to improve image depth estimation and use image features to guide long-range LiDAR feature prediction. We benchmark our SuperFusion on the nuScenes dataset and a self-recorded dataset and show that it outperforms the state-of-the-art baseline methods with large margins on all intervals. Additionally, we apply the generated HD map to a downstream path planning task, demonstrating that the long-range HD maps predicted by our method can lead to better path planning for autonomous vehicles. Our code and self-recorded dataset have been released at https://github.com/haomo-ai/SuperFusion.

The millimeter-wave radar sensor maintains stable performance under adverse environmental conditions, making it a promising solution for all-weather perception tasks, such as outdoor mobile robotics. However, the radar point clouds are relatively sparse and contain massive ghost points, which greatly limits the development of mmWave radar technology. In this paper, we propose a novel point cloud super-resolution approach for 3D mmWave radar data, named Radar-diffusion. Our approach employs the diffusion model defined by mean-reverting stochastic differential equations (SDE). Using our proposed new objective function with supervision from corresponding LiDAR point clouds, our approach efficiently handles radar ghost points and enhances the sparse mmWave radar point clouds to dense LiDAR-like point clouds. We evaluate our approach on two different datasets, and the experimental results show that our method outperforms the state-of-the-art baseline methods in 3D radar super-resolution tasks. Furthermore, we demonstrate that our enhanced radar point cloud is capable of downstream radar point-based registration tasks.

Visual place recognition (VPR) plays a pivotal role in autonomous exploration and navigation of mobile robots within complex outdoor environments. While cost-effective and easily deployed, camera sensors are sensitive to lighting and weather changes, and even slight image alterations can greatly affect VPR efficiency and precision. Existing methods overcome this by exploiting powerful yet large networks, leading to significant consumption of computational resources. In this paper, we propose a high-performance teacher and lightweight student distillation framework called TSCM. It exploits our devised cross-metric knowledge distillation to narrow the performance gap between the teacher and student models, maintaining superior performance while enabling minimal computational load during deployment. We conduct comprehensive evaluations on large-scale datasets, namely Pittsburgh30k and Pittsburgh250k. Experimental results demonstrate the superiority of our method over baseline models in terms of recognition accuracy and model parameter efficiency. Moreover, our ablation studies show that the proposed knowledge distillation technique surpasses other counterparts. The code of our method has been released at https://github.com/nubot-nudt/TSCM.

In recent years, object-oriented simultaneous localization and mapping (SLAM) has attracted increasing attention due to its ability to provide high-level semantic information while maintaining computational efficiency. Some researchers have attempted to enhance localization accuracy by integrating the modeled object residuals into bundle adjustment. However, few have demonstrated better results than feature-based visual SLAM systems, as the generic coarse object models, such as cuboids or ellipsoids, are less accurate than feature points. In this paper, we propose a Visual Object Odometry and Mapping framework VOOM using high-level objects and low-level points as the hierarchical landmarks in a coarse-to-fine manner instead of directly using object residuals in bundle adjustment. Firstly, we introduce an improved observation model and a novel data association method for dual quadrics, employed to represent physical objects. It facilitates the creation of a 3D map that closely reflects reality. Next, we use object information to enhance the data association of feature points and consequently update the map. In the visual object odometry backend, the updated map is employed to further optimize the camera pose and the objects. Meanwhile, local bundle adjustment is performed utilizing the objects and points-based covisibility graphs in our visual object mapping process. Experiments show that VOOM outperforms both object-oriented SLAM and feature points SLAM systems such as ORB-SLAM2 in terms of localization. The implementation of our method is available at https://github.com/yutongwangBIT/VOOM.git.

The reliance on accurate camera poses is a significant barrier to the widespread deployment of Neural Radiance Fields (NeRF) models for 3D reconstruction and SLAM tasks. The existing method introduces monocular depth priors to jointly optimize the camera poses and NeRF, which fails to fully exploit the depth priors and neglects the impact of their inherent noise. In this paper, we propose Truncated Depth NeRF (TD-NeRF), a novel approach that enables training NeRF from unknown camera poses - by jointly optimizing learnable parameters of the radiance field and camera poses. Our approach explicitly utilizes monocular depth priors through three key advancements: 1) we propose a novel depth-based ray sampling strategy based on the truncated normal distribution, which improves the convergence speed and accuracy of pose estimation; 2) to circumvent local minima and refine depth geometry, we introduce a coarse-to-fine training strategy that progressively improves the depth precision; 3) we propose a more robust inter-frame point constraint that enhances robustness against depth noise during training. The experimental results on three datasets demonstrate that TD-NeRF achieves superior performance in the joint optimization of camera pose and NeRF, surpassing prior works, and generates more accurate depth geometry. The implementation of our method has been released at https://github.com/nubot-nudt/TD-NeRF.

Place recognition is an important task for robots and autonomous cars to localize themselves and close loops in pre-built maps. While single-modal sensor-based methods have shown satisfactory performance, cross-modal place recognition that retrieving images from a point-cloud database remains a challenging problem. Current cross-modal methods transform images into 3D points using depth estimation for modality conversion, which are usually computationally intensive and need expensive labeled data for depth supervision. In this work, we introduce a fast and lightweight framework to encode images and point clouds into place-distinctive descriptors. We propose an effective Field of View (FoV) transformation module to convert point clouds into an analogous modality as images. This module eliminates the necessity for depth estimation and helps subsequent modules achieve real-time performance. We further design a non-negative factorization-based encoder to extract mutually consistent semantic features between point clouds and images. This encoder yields more distinctive global descriptors for retrieval. Experimental results on the KITTI dataset show that our proposed methods achieve state-of-the-art performance while running in real time. Additional evaluation on the HAOMO dataset covering a 17 km trajectory further shows the practical generalization capabilities. We have released the implementation of our methods as open source at: https://github.com/haomo-ai/ModaLink.git.

Fusion-based place recognition is an emerging technique jointly utilizing multi-modal perception data, to recognize previously visited places in GPS-denied scenarios for robots and autonomous vehicles. Recent fusion-based place recognition methods combine multi-modal features in implicit manners. While achieving remarkable results, they do not explicitly consider what the individual modality affords in the fusion system. Therefore, the benefit of multi-modal feature fusion may not be fully explored. In this paper, we propose a novel fusion-based network, dubbed EINet, to achieve explicit interaction of the two modalities. EINet uses LiDAR ranges to supervise more robust vision features for long time spans, and simultaneously uses camera RGB data to improve the discrimination of LiDAR point clouds. In addition, we develop a new benchmark for the place recognition task based on the nuScenes dataset. To establish this benchmark for future research with comprehensive comparisons, we introduce both supervised and self-supervised training schemes alongside evaluation protocols. We conduct extensive experiments on the proposed benchmark, and the experimental results show that our EINet exhibits better recognition performance as well as solid generalization ability compared to the state-of-the-art fusion-based place recognition approaches. Our open-source code and benchmark are released at: https://github.com/BIT-XJY/EINet.

LiDAR-based place recognition (LPR) is one of the most crucial components of autonomous vehicles to identify previously visited places in GPS-denied environments. Most existing LPR methods use mundane representations of the input point cloud without considering different views, which may not fully exploit the information from LiDAR sensors. In this article, we propose across-viewtransformer-based network, dubbed CVTNet, to fuse the range image views and bird's eye views generated from the LiDAR data. It extracts correlations within the views using intratransformers and between the two different views using intertransformers. Based on that, our proposed CVTNet generates a yaw-angle-invariant global descriptor for each laser scan end-to-end online and retrieves previously seen places by descriptor matching between the current query scan and the prebuilt database. We evaluate our approach on three datasets collected with different sensor setups and environmental conditions. The experimental results show that our method outperforms the state-of-the-art LPR methods with strong robustness to viewpoint changes and long-time spans. Furthermore, our approach has better real-time performance that can run faster than the typical LiDAR frame rate does.

LiDAR-based scene flow estimation (SFE) and moving object segmentation (MOS) are important tasks with broad-ranging applications in autonomous driving, such as traffic surveillance, motion analysis, obstacle avoidance, etc. Most existing works address SFE and MOS separately, ignoring the underlying shared geometric constraints and their inherent correlation. This article rethinks LiDAR-based SFE and MOS tasks, providing our key insight that jointly addressing them can tackle challenges in both tasks, and their solutions can reinforce one another to improve the performance of both. Based on this insight, we introduce a novel framework that exploits shared geometric constraints by explicitly partitioning the scene into static and moving regions and subsequently estimating flow differently for these regions. A lightweight and interpretable neural network dubbed SFEMOS is proposed. It employs an encoder and two specially designed head modules for each task, achieving MOS without relying on prior poses and online point-wise flow estimation for 360-degree point clouds. Due to the absence of public datasets for concurrently evaluating both tasks, we generate ground truth flow data using MOS labels from SemanticKITTI. Additionally, we establish a new dataset using a rotational LiDAR mounted on our own autonomous vehicle. Evaluation results on both datasets validate the superior performance of our proposed SFEMOS. Our dataset and label generation method are released athttps://github.com/nubot-nudt/SFEMOS.

Loop closing and relocalization are crucial techniques to establish reliable and robust long-term SLAM by addressing pose estimation drift and degeneration. This article begins by formulating loop closing and relocalization within a unified framework. Then, we propose a novel multi-head network LCR-Net to tackle both tasks effectively. It exploits novel feature extraction and a pose-aware attention mechanism to precisely estimate similarities and 6-DoF poses between pairs of LiDAR scans. In the end, we integrate our LCR-Net into a SLAM system and achieve robust and accurate online LiDAR SLAM in outdoor driving environments. We thoroughly evaluate our LCR-Net through three setups derived from loop closing and relocalization, including candidate retrieval, closed-loop point cloud registration, and continuous relocalization using multiple datasets. The results demonstrate that LCR-Net excels in all three tasks, surpassing the state-of-the-art methods and exhibiting a remarkable generalization ability. Notably, our LCR-Net outperforms baseline methods without using a time-consuming robust pose estimator, rendering it suitable for online SLAM applications. To our best knowledge, the integration of LCR-Net yields the first LiDAR SLAM with the capability of deep loop closing and relocalization. The implementation of our methods is open-sourced athttps://github.com/nubot-nudt/LCR-Net.

Semantic perception is a core building block in autonomous driving, since it provides information about the drivable space and location of other traffic participants. For learning-based perception, often a large amount of diverse training data is necessary to achieve high performance. Data labeling is usually a bottleneck for developing such methods, especially for dense prediction tasks, e.g., semantic segmentation or panoptic segmentation. For 3D Li-DAR data, the annotation process demands even more effort than for images. Especially in autonomous driving, point clouds are sparse, and objects appearance depends on its distance from the sensor, making it harder to acquire large amounts of labeled training data. This paper aims at taking an alternative path proposing a self-supervised representation learning method for 3D LiDAR data. Our approach exploits the vehicle motion to match objects across time viewed in different scans. We then train a model to maximize the point-wise feature similarities from points of the associated object in different scans, which enables to learn a consistent representation across time. The experimental results show that our approach performs better than previous state-of-the-art self-supervised representation learning methods when fine-tuning to different downstream tasks. We furthermore show that with only 10% of labeled data, a network pre-trained with our approach can achieve better performance than the same network trained from scratch with all labels for semantic segmentation on SemanticKITTI.11Code: https://github.com/PRBonn/TARL

Simultaneously odometry and mapping using LiDAR data is an important task for mobile systems to achieve full autonomy in large-scale environments. However, most existing LiDAR-based methods prioritize tracking quality over reconstruction quality. Although the recently developed neural radiance fields (NeRF) have shown promising advances in implicit reconstruction for indoor environments, the problem of simultaneous odometry and mapping for large-scale scenarios using incremental LiDAR data remains unexplored. To bridge this gap, in this paper, we propose a novel NeRF-based LiDAR odometry and mapping approach, NeRF-LOAM, consisting of three modules neural odometry, neural mapping, and mesh reconstruction. All these modules utilize our proposed neural signed distance function, which separates LiDAR points into ground and non-ground points to reduce Z-axis drift, optimizes odometry and voxel embeddings concurrently, and in the end generates dense smooth mesh maps of the environment. Moreover, this joint optimization allows our NeRF-LOAM to be pre-trained free and exhibit strong generalization abilities when applied to different environments. Extensive evaluations on three publicly available datasets demonstrate that our approach achieves state-of-the-art odometry and mapping performance, as well as a strong generalization in large-scale environments utilizing LiDAR data. Furthermore, we perform multiple ablation studies to validate the effectiveness of our network design. The implementation of our approach will be made available at https://github.com/JunyuanDeng/NeRF-LOAM.

A distinctive representation of image patches in form of features is a key component of many computer vision and robotics tasks, such as image matching, image retrieval, and visual localization. State-of-the-art descriptors, from hand-crafted descriptors such as SIFT to learned ones such as HardNet, are usually high-dimensional; 128 dimensions or even more. The higher the dimensionality, the larger the memory consumption and computational time for approaches using such descriptors. In this paper, we investigate multi-layer perceptrons (MLPs) to extract low-dimensional but high-quality descriptors. We thoroughly analyze our method in unsuper-vised, self-supervised, and supervised settings, and evaluate the dimensionality reduction results on four representative descriptors. We consider different applications, including visual localization, patch verification, image matching and retrieval. The experiments show that our lightweight MLPs trained using supervised method achieve better dimensionality reduction than PCA. The lower-dimensional descriptors generated by our approach outperform the original higher-dimensional descriptors in downstream tasks, especially for the hand-crafted ones. The code is available at https://github.com/PRBonn/descriptor-dr.

Accurately perceiving instances and predicting their future motion are key tasks for autonomous vehicles, enabling them to navigate safely in complex urban traffic. While bird’s-eye view (BEV) representations are commonplace in perception for autonomous driving, their potential in a motion prediction setting is less explored. Existing approaches for BEV instance prediction from surround cameras rely on a multi-task auto-regressive setup coupled with complex post-processing to predict future instances in a spatio-temporally consistent manner. In this paper, we depart from this paradigm and propose an efficient novel end-to-end framework named PowerBEV, which differs in several design choices aimed at reducing the inherent redundancy in previous methods. First, rather than predicting the future in an auto-regressive fashion, PowerBEV uses a parallel, multi-scale module built from lightweight 2D convolutional networks. Second, we show that segmentation and centripetal backward flow are sufficient for prediction, simplifying previous multi-task objectives by eliminating redundant output modalities. Building on this output representation, we propose a simple, flow warping-based post-processing approach which produces more stable instance associations across time. Through this lightweight yet powerful design, PowerBEV outperforms state-of-the-art baselines on the NuScenes Dataset and poses an alternative paradigm for BEV instance prediction. We made our code publicly available at: https://github.com/EdwardLeeLPZ/PowerBEV.

Open-world Instance Segmentation (OIS) is a challenging task that aims to accurately segment every object instance appearing in the current observation, regardless of whether these instances have been labeled in the training set. This is important for safety-critical applications such as robust autonomous navigation. In this paper, we present a flexible and effective OIS framework for LiDAR point cloud that can accurately segment both known and unknown instances (i.e., seen and unseen instance categories during training). It first identifies points belonging to known classes and removes the back-ground by leveraging close-set panoptic segmentation networks. Then, we propose a novel ellipsoidal clustering method that is more adapted to the characteristic of LiDAR scans and allows precise segmentation of unknown instances. Furthermore, a diffuse searching method is proposed to handle the common over-segmentation problem presented in the known instances. With the combination of these techniques, we are able to achieve accurate segmentation for both known and unknown instances. We evaluated our method on the SemanticKITTI open-world LiDAR instance segmentation dataset. The experimental results suggest that it outperforms current state-of-the-art methods, especially with a 10.0% improvement in association quality. The source code of our method will be publicly available at https://github.com/nubot-nudt/ElC-OIS.

Autonomous robotic tasks require actively perceiving the environment to achieve application-specific goals. In this paper, we address the problem of positioning an RGB camera to collect the most informative images to represent an unknown scene, given a limited measurement budget. We propose a novel mapless planning framework to iteratively plan the next best camera view based on collected image measurements. A key aspect of our approach is a new technique for uncertainty estimation in image-based neural rendering, which guides measurement acquisition at the most uncertain view among view candidates, thus maximising the information value during data collection. By incrementally adding new measurements into our image collection, our approach efficiently explores an unknown scene in a mapless manner. We show that our uncertainty estimation is generalisable and valuable for view planning in unknown scenes. Our planning experiments using synthetic and real-world data verify that our uncertainty-guided approach finds informative images leading to more accurate scene representations when compared against baselines.

Fast and accurate path planning is important for ground robots to achieve safe and efficient autonomous navigation in unstructured outdoor environments. However, most existing methods exploiting either 2D or 2.5D maps struggle to balance the efficiency and safety for ground robots navigating in such challenging scenarios. In this paper, we propose a novel hybrid map representation by fusing a 2D grid and a 2.5D digital elevation map. Based on it, a novel path planning method is proposed, which considers the robot poses during traversability estimation. By doing so, our method explicitly takes safety as a planning constraint enabling robots to navigate unstructured environments smoothly. The proposed approach has been evaluated on both simulated datasets and a real robot platform. The experimental results demonstrate the efficiency and effectiveness of the proposed method. Compared to state-of-the-art baseline methods, the proposed approach consistently generates safer and easier paths for the robot in different unstructured outdoor environments. The implementation of our method is publicly available at https://github.com/nubot-nudt/T-Hybrid-planner.

Identifying moving objects is a crucial capability for autonomous navigation, consistent map generation, and future trajectory prediction of objects. In this paper, we propose a novel network that addresses the challenge of segmenting moving objects in 3D LiDAR scans. Our approach not only predicts point-wise moving labels but also detects instance information of main traffic participants. Such a design helps determine which instances are actually moving and which ones are temporarily static in the current scene. Our method exploits a sequence of point clouds as input and quantifies them into 4D voxels. We use 4D sparse convolutions to extract motion features from the 4D voxels and inject them into the current scan. Then, we extract spatio-temporal features from the current scan for instance detection and feature fusion. Finally, we design an upsample fusion module to output point-wise labels by fusing the spatio-temporal features and predicted instance information. We evaluated our approach on the LiDAR-MOS benchmark based on SemanticKITTI and achieved better moving object segmentation performance compared to state-of-the-art methods, demonstrating the effectiveness of our approach in integrating instance information for moving object segmentation. Furthermore, our method shows superior performance on the Apollo dataset with a pre-trained model on SemanticKITTI, indicating that our method generalizes well in different scenes. The code and pre-trained models of our method will be released at https://github.com/nubot-nudt/InsMOS.

Point cloud registration is an important task in robotics and autonomous driving to estimate the ego-motion of the vehicle. Recent advances following the coarse-to-fine manner show promising potential in point cloud registration. However, existing methods rely on good superpoint correspondences, which are hard to be obtained reliably and efficiently, thus resulting in less robust and accurate point cloud registration. In this paper, we propose a novel network, named RDMNet, to find dense point correspondences coarse-to-fine and improve final pose estimation based on such reliable correspondences. Our RDMNet uses a devised 3D-RoFormer mechanism to first extract distinctive superpoints and generates reliable superpoints matches between two point clouds. The proposed 3D-RoFormer fuses 3D position information into the transformer network, efficiently exploiting point clouds’ contextual and geometric information to generate robust superpoint correspondences. RDMNet then propagates the sparse superpoints matches to dense point matches using the neighborhood information for accurate point cloud registration. We extensively evaluate our method on multiple datasets from different environments. The experimental results demonstrate that our method outperforms existing state-of-the-art approaches in all tested datasets with a strong generalization ability.

Accurate moving object segmentation is an es-sential task for autonomous driving. It can provide effective information for many downstream tasks, such as collision avoidance, path planning, and static map construction. How to effectively exploit the spatial-temporal information is a critical question for 3D LiDAR moving object segmentation (LiDAR-MOS). In this work, we propose a novel deep neural network exploiting both spatial-temporal information and different representation modalities of LiDAR scans to improve LiDAR-MOS performance. Specifically, we first use a range image-based dual-branch structure to separately deal with spatial and temporal information that can be obtained from sequential LiDAR scans, and later combine them using motion-guided attention modules. We also use a point refinement module via 3D sparse convolution to fuse the information from both LiDAR range image and point cloud representations and reduce the artifacts on the borders of the objects. We verify the effectiveness of our proposed approach on the LiDAR-MOS benchmark of SemanticKITTI. Our method outperforms the state-of-the-art methods significantly in terms of LiDAR-MOS IoU. Benefiting from the devised coarse-to-fine architecture, our method operates online at sensor frame rate. Code is available at: https://github.com/haomo-ai/MotionSeg3D.

Robots that are supposed to interact with or manipulate objects in the world must be able to track the poses of objects in their sensor data. Thus, Detecting and tracking the 6-DoF poses of targeted objects is important for aerial manipulation and is still in the early stage due to the high dynamics and limited onboard capacity of such systems. In this paper, we propose ICK-Track, a novel method for onboard category-level object 6-DoF pose tracking that can be applied to aerial manipulation without using any pre-defined object CAD models. It first utilizes a semi-supervised video segmentation to detect objects in the eye-in-hand RGB-D camera stream to segment the 3D points of objects. Then, canonical keypoints are extracted using iterative farthest point sampling. We propose a novel inter-frame consistent keypoints generation network to generate the corresponding keypoint pairs, which are used together with ICP to estimate the pose changes of objects for tracking. Experimental results show that our method is more robust to viewpoint changes and runs faster than the state-of-the-art methods on category-level pose tracking. We further test our proposed method on a real aerial manipulator. A demo video showing the use of our method on a real aerial manipulator and the implementation of our method are available at: https://github.com/S-JingTao/ICK-Track.

Robust and accurate, map-based localization is crucial for autonomous mobile systems. In this paper, we exploit range images generated from 3D LiDAR scans to address the problem of localizing mobile robots or autonomous cars in a map of a large-scale outdoor environment represented by a triangular mesh. We use the Poisson surface reconstruction to generate the mesh-based map representation. Based on the range images generated from the current LiDAR scan and the synthetic rendered views from the mesh-based map, we propose a new observation model and integrate it into a Monte Carlo localization framework, which achieves better localization performance and generalizes well to different environments. We test the proposed localization approach on multiple datasets collected in different environments with different LiDAR scanners. The experimental results show that our method can reliably and accurately localize a mobile system in different environments and operate online at the LiDAR sensor frame rate to track the vehicle pose.

Robust and reliable ego-motion is a key component of most autonomous mobile systems. Many odometry estimation methods have been developed using different sensors such as cameras or LiDARs. In this work, we present a resilient approach that exploits the redundancy of multiple odometry algorithms using a 3D LiDAR scanner and a monocular camera to provide reliable state estimation for autonomous vehicles. Our system utilizes a stack of odometry algorithms that run in parallel. It chooses from them the most promising pose estimation considering sanity checks using dynamic and kinematic constraints of the vehicle as well as a score computed between the current LiDAR scan and a locally built point cloud map. In this way, our method can exploit the advantages of different existing ego-motion estimating approaches. We evaluate our method on the KITTI Odometry dataset. The experimental results suggest that our approach is resilient to failure cases and achieves an overall better performance than individual odometry methods employed by our system.

Accurately localizing in and mapping an environment are essential building blocks of most autonomous systems. In this paper, we present a novel approach for LiDAR odometry and mapping, focusing on improving the mapping quality and at the same time estimating the pose of the vehicle. Our approach performs frame-to-mesh ICP, but in contrast to other SLAM approaches, we represent the map as a triangle mesh computed via Poisson surface reconstruction. We perform the surface reconstruction in a sliding window fashion over a sequence of past scans. In this way, we obtain accurate local maps that are well suited for registration and can also be combined into a global map. This enables us to build a 3D map showing more geometric details than common mapping approaches relying on a truncated signed distance function or surfels. Our experimental evaluation shows quantitatively and qualitatively that our maps offer higher geometric accuracies than these other map representations. We also show that our maps are compact and can be used for LiDAR-based odometry estimation with a novel ray-casting-based data association.

The ability to localise is key for robot navigation. We describe an efficient method for vision-based localisation, which combines sequential Monte Carlo tracking with matching ground-level images to 2-D cartographic maps such as OpenStreetMaps. The matching is based on a learned embedded space representation linking images and map tiles, encoding the common semantic information present in both and providing potential for invariance to changing conditions. Moreover, the compactness of 2-D maps supports scalability. This contrasts with the majority of previous approaches based on matching with single-shot geo-referenced images or 3-D reconstructions. We present experiments using the StreetLearn and Oxford RobotCar datasets and demonstrate that the method is highly effective, giving high accuracy and fast convergence.

Localization is a crucial capability for mobile robots and autonomous cars. In this paper, we address learning an observation model for Monte-Carlo localization using 3D LiDAR data. We propose a novel, neural network-based observation model that computes the expected overlap of two 3D LiDAR scans. The model predicts the overlap and yaw angle offset between the current sensor reading and virtual frames generated from a pre-built map. We integrate this observation model into a Monte-Carlo localization framework and tested it on urban datasets collected with a car in different seasons. The experiments presented in this paper illustrate that our method can reliably localize a vehicle in typical urban environments. We furthermore provide comparisons to a beam-endpoint and a histogram-based method indicating a superior global localization performance of our method with fewer particles.

Reliable and accurate localization and mapping are key components of most autonomous systems. Besides geometric information about the mapped environment, the semantics plays an important role to enable intelligent navigation behaviors. In most realistic environments, this task is particularly complicated due to dynamics caused by moving objects, which can corrupt the mapping step or derail localization. In this paper, we propose an extension of a recently published surfel-based mapping approach exploiting three-dimensional laser range scans by integrating semantic information to facilitate the mapping process. The semantic information is efficiently extracted by a fully convolutional neural network and rendered on a spherical projection of the laser range data. This computed semantic segmentation results in point-wise labels for the whole scan, allowing us to build a semantically-enriched map with labeled surfels. This semantic map enables us to reliably filter moving objects, but also improve the projective scan matching via semantic constraints. Our experimental evaluation on challenging highways sequences from KITTI dataset with very few static structures and a large amount of moving cars shows the advantage of our semantic SLAM approach in comparison to a purely geometric, state-of-the-art approach.