Simultaneous Localization And Mapping Research Articles

Building the Future of Transportation: A Comprehensive Survey on AV Perception, Localization, and Mapping

Autonomous vehicles (AVs) depend on perception, localization, and mapping to interpret their surroundings and navigate safely. This paper reviews existing methodologies and best practices in these domains, focusing on object detection, object tracking, localization techniques, and environmental mapping strategies. In the perception module, we analyze state-of-the-art object detection frameworks, such as You Only Look Once version 8 (YOLOv8), and object tracking algorithms like ByteTrack and BoT-SORT (Boosted SORT). We assess their real-time performance, robustness to occlusions, and suitability for complex urban environments. We examine different approaches for localization, including Light Detection and Ranging (LiDAR)-based localization, camera-based localization, and sensor fusion techniques. These methods enhance positional accuracy, particularly in scenarios where Global Positioning System (GPS) signals are unreliable or unavailable. The mapping section explores Simultaneous Localization and Mapping (SLAM) techniques and high-definition (HD) maps, discussing their role in creating detailed, real-time environmental representations that enable autonomous navigation. Additionally, we present insights from our testing, evaluating the effectiveness of different perception, localization, and mapping methods in real-world conditions. By summarizing key advancements, challenges, and practical considerations, this paper provides a reference for researchers and developers working on autonomous vehicle perception, localization, and mapping.

LiDAR, IMU, and camera fusion for simultaneous localization and mapping: a systematic review

Simultaneous Localization and Mapping (SLAM) is a crucial technology for intelligent unnamed systems to estimate their motion and reconstruct unknown environments. However, the SLAM systems with merely one sensor have poor robustness and stability due to the defects in the sensor itself. Recent studies have demonstrated that SLAM systems with multiple sensors, mainly consisting of LiDAR, camera, and IMU, achieve better performance due to the mutual compensation of different sensors. This paper investigates recent progress on multi-sensor fusion SLAM. The review includes a systematic analysis of the advantages and disadvantages of different sensors and the imperative of multi-sensor solutions. It categorizes multi-sensor fusion SLAM systems into four main types by the fused sensors: LiDAR-IMU SLAM, Visual-IMU SLAM, LiDAR-Visual SLAM, and LiDAR-IMU-Visual SLAM, with detailed analysis and discussions of their pipelines and principles. Meanwhile, the paper surveys commonly used datasets and introduces evaluation metrics. Finally, it concludes with a summary of the existing challenges and future opportunities for multi-sensor fusion SLAM.

20 Years of Integrated Exploration

The autonomous construction of environment maps with the help of mobile robots is an important problem in modern robotics; because practically all tasks performed by robots require a representation of the working environment. Many solutions have been proposed to solve this problem known as SLAM (Simultaneous Localization and Mapping). The inclusion of a motion planner to the classical SLAM problem gives way to a new approach known as “Integrated Exploration”, in which a robot gradually builds a map while simultaneously localizing itself and making local decisions on where to go to maximize the acquisition of map information. In this paper we will analyze the proposals that have been developed in the last 20 years in this area, having as a primary interest to show the advances that have been made in the area of motion planning and the challenges presented by its integration and coordination with the SLAM problem.

Faster than Fast: Accelerating Oriented FAST Feature Detection on Low-end Embedded GPUs

The visual-based SLAM (Simultaneous Localization and Mapping) is a technology widely used in applications such as robotic navigation and virtual reality, which primarily focuses on detecting feature points from visual images to construct an unknown environmental map and simultaneously determines its own location. It usually imposes stringent requirements on hardware power consumption, processing speed and accuracy. Currently, the ORB (Oriented FAST and Rotated BRIEF)-based SLAM systems have exhibited superior performance in terms of processing speed and robustness. However, they still fall short of meeting the demands for real-time processing on mobile platforms. This limitation is primarily due to the time-consuming Oriented FAST calculations accounting for approximately half of the entire SLAM system. This paper presents two methods to accelerate the Oriented FAST feature detection on low-end embedded GPUs. These methods optimize the most time-consuming steps in Oriented FAST feature detection: FAST feature point detection and Harris corner detection, which is achieved by implementing a binary-level encoding strategy to determine candidate points quickly and a separable Harris detection strategy with efficient low-level GPU hardware-specific instructions. Extensive experiments on a Jetson TX2 embedded GPU demonstrate an average speedup of over 7.3 times compared to widely used OpenCV with GPU support. This significant improvement highlights its effectiveness and potential for real-time applications in mobile and resource-constrained environments.

DMS-SLAM: Semantic Visual SLAM Based on Deep Mask Segmentation in Dynamic Environments

Abstract Visual Simultaneous Localization and Mapping (VSLAM), which is a core technology for intelligent mobile robots, is typically based on static environment assumptions. However, it encounters challenges in dynamic scenes such as accurately distinguishing between static and dynamic feature points and high computational costs. These issues lead to significant deviations in camera tracking and introduce ghosting in mapping. To address these issues, we propose DMS-SLAM, a fast semantic visual SLAM based on deep mask segmentation. First, semantic segmentation is performed using the YOLOv8s-seg network, followed by a mask correction algorithm based on depth consistency and a missed detection compensation algorithm based on relative displacement invariance, ensuring precise segmentation of dynamic objects within the semantic thread. Then, the segmented objects are classified according to dynamic probability, and the motion of potential dynamic objects is determined through a depth mask association algorithm. Dynamic feature points are eliminated by utilizing epipolar geometric constraints to improve localization accuracy. Finally, local point clouds with dynamic objects removed are stitched together to construct a global dense point cloud map, presenting a static scene free of ghosting. Experiments on the TUM RGB-D dataset and in real-world scenarios show that the method enhances absolute trajectory accuracy by an average of 96.19% and relative pose accuracy by 94.27% over the ORB-SLAM2 system in highly dynamic environments. Compared to other dynamic semantic visual SLAM algorithms, DMS-SLAM improves operational efficiency by over 51.78%, significantly reducing runtime and generating clear and accurate global 3D point cloud maps, which fully demonstrates its superiority in terms of accuracy, stability and efficiency.

R2SCAT-LPR: Rotation-Robust Network with Self- and Cross-Attention Transformers for LiDAR-Based Place Recognition

LiDAR-based place recognition (LPR) is crucial for the navigation and localization of autonomous vehicles and mobile robots in large-scale outdoor environments and plays a critical role in loop closure detection for simultaneous localization and mapping (SLAM). Existing LPR methods, which utilize 2D bird’s-eye view (BEV) projections of 3D point clouds, achieve competitive performance in efficiency and recognition accuracy. However, these methods often struggle with capturing global contextual information and maintaining robustness to viewpoint variations. To address these challenges, we propose R2SCAT-LPR, a novel, transformer-based model that leverages self-attention and cross-attention mechanisms to extract rotation-robust place feature descriptors from BEV images. R2SCAT-LPR consists of three core modules: (1) R2MPFE, which employs weight-shared cascaded multi-head self-attention (MHSA) to extract multi-level spatial contextual patch features from both the original BEV image and its randomly rotated counterpart; (2) DSCA, which integrates dual-branch self-attention and multi-head cross-attention (MHCA) to capture intrinsic correspondences between multi-level patch features before and after rotation, enhancing the extraction of rotation-robust local features; and (3) a combined NetVLAD module, which aggregates patch features from both the original feature space and the rotated interaction space into a compact and viewpoint-robust global descriptor. Extensive experiments conducted on the KITTI and NCLT datasets validate the effectiveness of the proposed model, demonstrating its robustness to rotation variations and its generalization ability across diverse scenes and LiDAR sensors types. Furthermore, we evaluate the generalization performance and computational efficiency of R2SCAT-LPR on our self-constructed OffRoad-LPR dataset for off-road autonomous driving, verifying its deployability on resource-constrained platforms.

LiDAR Based SLAM: A Comparative Evaluation with LeGO-LOAM and HDL-Graph SLAM

Simultaneous localization and mapping (SLAM) is an area of research that is experiencing rapid advancements, with a significant impact on improving the navigation and perception capabilities of vehicles, thereby enabling their safe operation in complex environments. This study presents a comprehensive overview of the recent developments in SLAM and conducts a comparative evaluation of two widely employed SLAM methods. The evaluation is based on rigorous performance analysis using the KITTI dataset, which is one of the most popular benchmark datasets for evaluating SLAM algorithms. The evaluation focuses on two essential metrics: absolute trajectory error (ATE) and relative pose error (RPE), which provides valuable insights into the accuracy and consistency of pose estimation over time. By quantifying the deviation between estimated trajectories and ground truth data, ATE sheds light on the global accuracy of the SLAM system, while RPE examines the local error and the system's ability to maintain reliable pose estimates within sequential frames. A thorough discussion is provided on the advantages, distinctive features, and performance characteristics of each technique, thereby offering valuable insights to propel future research in this area.

Semantic SLAM system for mobile robots based on large visual model in complex environments

Simultaneous localization and mapping (SLAM) plays an important role in many fields, one of which is to help unmanned devices such as drones, self-driving cars and intelligent robots to achieve precise positioning and mapping. However, when facing complex or changing surroundings, especially when healthcare robots face a large number of mobile healthcare workers and patients in wards, the hospital environment is relatively complex, and the traditional positioning and mapping methods based on geometric features, such as points and lines, are not able to achieve accurate positioning and mapping results for healthcare robots. This paper mainly focuses on the characteristics of complex dynamic environment, and proposes a method to obtain semantic information of surrounding ring and dynamic point culling strategy for robot localisation and mapping. Experiments show that compared with the current popular SLAM technology, the semantic-based SLAM technology proposed in this paper can help the robot to obtain more accurate localisation and mapping, in addition, using this semantic information, the robot can also better identify the surrounding objects, which lays the foundation for performing more complex tasks.

Efficient Point-Line Visual-Inertial SLAM Combined with EDLines and Line Flow Optimization

Abstract Visual-inertial Simultaneous Localization and Mapping (SLAM) technology has received extensive attention in the field of autonomous driving due to its characteristic of not requiring prior information. However, point feature-based SLAM systems are limited in performance in low-texture environments, struggling to extract sufficient reliable feature points, which leads to a decrease in positioning accuracy and even system failure. Although line features have an advantage in low-texture environments, existing line feature matching methods have the following limitations: low matching efficiency, low matching accuracy, and insufficient system robustness. This paper proposes an efficient point-line fusion visual-inertial SLAM system (EPL-VINS) aimed at addressing the aforementioned issues. When applying line features, the following technical challenges need to be overcome: improving the efficiency and accuracy of line feature detection; reducing over-segmentation to decrease mismatching; and enabling line feature tracking to adapt to dynamic scenes. To address these challenges, this paper presents the following innovations: an optimized line feature detector based on short line filtering and similar line fusion techniques, utilizing optical flow features for line feature tracking to enhance tracking speed and accuracy; and the introduction of an adaptive line feature residual to mitigate the negative impact of the sparse number of line features on state estimation. Experiments demonstrate a remarkable 70% improvement in the efficiency of line feature detection and tracking tasks compared to the combined use of the LBD and LSD algorithms on the EuRoc dataset. Furthermore, without loop closure constraints, EPL-VINS outperforms several leading open-source algorithms, showing reduction in RMSE of 25.85% over PL-VINS and 23.82% over PL-VIO on the same dataset. This approach offers a robust solution to the real-time challenges of online feature extraction and tracking, while significantly enhancing the quality and reliability of line features.

Enhancing SLAM algorithm with Top-K optimization and semantic descriptors

To address the computational challenges faced by edge devices using deep learning to process LiDAR point cloud data, this paper proposes a SLAM algorithm incorporating Top-K optimization to generate semantic descriptors and global semantic map for laser data efficiently. This approach aims to reduce computational complexity while enhancing processing speed. The algorithm extracts semantic information from LiDAR data, constructs two-dimensional semantic descriptors, and improves the robot’s semantic understanding of its surrounding environment. In the loop closure detection phase, the algorithm identifies loop candidates by calculating the geometric and semantic similarities of the descriptors. It utilizes front-end odometry to stitch together subgraphs from these loop candidates, thereby detecting true loop closures. Finally, true loop closures add constraints in the factor graph, facilitating pose optimization. Experimental results show that this descriptor can match more loop closures without affecting accuracy. The algorithm enhances the pose estimation accuracy of the robot and generates global point cloud maps rich in semantic information. Under the influence of the Top-K strategy, the average inference time is reduced by 10.7%, and the memory usage decreases by 19.5% compared with before in the Network Inference module. This Top-K strategy significantly conserves computational resources for optimizing edge-device deep learning algorithms, particularly when processing LiDAR point cloud data. Additionally, it effectively reduces the computational load in practical applications while maintaining inference accuracy and efficiency.

VS-SLAM: Robust SLAM Based on LiDAR Loop Closure Detection with Virtual Descriptors and Selective Memory Storage in Challenging Environments

LiDAR loop closure detection is a key technology to mitigate localization drift in LiDAR SLAM, but it remains challenging in structurally similar environments and memory-constrained platforms. This paper proposes VS-SLAM, a novel and robust SLAM system that leverages virtual descriptors and selective memory storage to enhance LiDAR loop closure detection in challenging environments. Firstly, to mitigate the sensitivity of existing descriptors to translational changes, we propose a novel virtual descriptor technique that enhances translational invariance and improves loop closure detection accuracy. Then, to further improve the accuracy of loop closure detection in structurally similar environments, we propose an efficient and reliable selective memory storage technique based on scene recognition and key descriptor evaluation, which also reduces the memory consumption of the loop closure database. Next, based on the two proposed techniques, we develop a LiDAR SLAM system with loop closure detection capability, which maintains high accuracy and robustness even in challenging environments with structural similarity. Finally, extensive experiments in self-built simulation, real-world environments, and public datasets demonstrate that VS-SLAM outperforms state-of-the-art methods in terms of memory efficiency, accuracy, and robustness. Specifically, the memory consumption of the loop closure database is reduced by an average of 92.86% compared with SC-LVI-SAM and VS-SLAM-w/o-st, and the localization accuracy in structurally similar challenging environments is improved by an average of 66.41% compared with LVI-SAM.

A visual SLAM loop closure detection method based on lightweight siamese capsule network

Loop closure detection is a key module in visual SLAM. During the robot’s movement, the cumulative error of the robot is reduced by the loop closure detection method, which can provide constraints for the back-end pose optimization, and the SLAM system can build an accurate map. Traditional loop closure detection algorithms rely on the bag-of-words model, which involves a complex process, has slow loading speeds, and is sensitive to changes in illumination or viewing angles. Therefore, aiming at the problems of traditional methods, this paper proposes an algorithm based on the Siamese capsule neural network by using the deep learning method. We have designed a new feature extractor for capsule networks, and in order to further reduce the parameter count, we have performed pruning based on the characteristics of the capsule layer. The algorithm was tested on the CityCentre dataset and the New College dataset. Our experimental results show that the proposed algorithm in this paper has higher accuracy and robustness compared to traditional methods and other deep learning methods. Our algorithm demonstrates good robustness under changes in illumination and viewing angles. Finally, we evaluated the performance of the complete SLAM system on the KITTI dataset.

GCENet: A geometric correspondence estimation network for tracking and loop detection in visual-inertial SLAM

GCENet: A geometric correspondence estimation network for tracking and loop detection in visual-inertial SLAM

Tunnel crack assessment using simultaneous localization and mapping (SLAM) and deep learning segmentation

Tunnel crack assessment using simultaneous localization and mapping (SLAM) and deep learning segmentation

Advancing autonomous SLAM systems: Integrating YOLO object detection and enhanced loop closure techniques for robust environment mapping

Advancing autonomous SLAM systems: Integrating YOLO object detection and enhanced loop closure techniques for robust environment mapping

A Survey of Robotic Monocular Pose Estimation.

Robotic monocular pose estimation is an important part of neural monocular pose estimation-driven methods, which includes monocular simultaneous localization and mapping (SLAM) and single-view object pose estimation (OPE) driven by neural methods. The mapping thread leeches onto robotic monocular pose estimation. Robotic monocular pose estimation consists of the localization part of monocular SLAM and the object pose solving part of single-view OPE. Depth prediction neural networks, semantics, neural implicit representations, and large language models (LLMs) are neural methods that have been important components of neural monocular pose estimation-driven methods. Complete robotic monocular pose estimation is a potential module in real robots. Possible future research directions and applications are discussed.

Privacy-Preserved Visual Simultaneous Localization and Mapping Based on a Dual-Component Approach

Edge-assisted visual simultaneous localization and mapping (SLAM) is widely used in autonomous driving, robot navigation, and augmented reality for environmental perception, map construction, and real-time positioning. However, it poses significant privacy risks, as input images may contain sensitive information, and generated 3D point clouds can reconstruct original scenes. To address these concerns, this paper proposes a dual-component privacy-preserving approach for visual SLAM. First, a privacy protection method for images is proposed, which combines object detection and image inpainting to protect privacy-sensitive information in images. Second, an encryption algorithm is introduced to convert 3D point cloud data into a 3D line cloud through dimensionality enhancement. Integrated with ORB-SLAM3, the proposed method is evaluated on the Oxford Robotcar and KITTI datasets. Results demonstrate that it effectively safeguards privacy-sensitive information while ORB-SLAM3 maintains accurate pose estimation in dynamic outdoor scenes. Furthermore, the encrypted line cloud prevents unauthorized attacks on recovering the original point cloud. This approach enhances privacy protection in visual SLAM and is expected to expand its potential applications.

CYCLICCAE: A CONFORMATIONAL AUTOENCODER FOR EFFICIENT HETEROCHIRAL MACROCYCLIC BACKBONE SAMPLING.

Macrocycles are a promising therapeutic class. The incorporation of heterochiral and non-natural chemical building-blocks presents challenges for rational design, however. With no existing machine learning methods tailored for heterochiral macrocycle design, we developed a novel convolutional autoencoder model to rapidly generate energetically favorable macrocycle backbones for heterochiral design and structure prediction. Our approach surpasses the current state-of-the-art method, Generalized Kinematic loop closure (GenKIC) in the Rosetta software suite. Given the absence of large, available macrocycle datasets, we created a custom dataset in-house and in silico. Our model, CyclicCAE, produces energetically stable backbones and designable structures more rapidly than GenKIC. It enables users to perform energy minimization, generate structurally similar or diverse inputs via MCMC, and conduct inpainting with fixed anchors or motifs. We propose that this novel method will accelerate the development of stable macrocycles, speeding up macrocycle drug design pipelines.

Loosely Coupled PPP/Inertial/LiDAR Simultaneous Localization and Mapping (SLAM) Based on Graph Optimization

Navigation services and high-precision positioning play a significant role in emerging fields such as self-driving and mobile robots. The performance of precise point positioning (PPP) may be seriously affected by signal interference and struggles to achieve continuous and accurate positioning in complex environments. LiDAR/inertial navigation can use spatial structure information to realize pose estimation but cannot solve the problem of cumulative error. This study proposes a PPP/inertial/LiDAR combined localization algorithm based on factor graph optimization. Firstly, the algorithm performed the spatial alignment by adding the initial yaw factor. Then, the PPP factor and anchor factor were constructed using PPP information. Finally, the global localization is estimated accurately and robustly based on the factor graph. The vehicle experiment shows that the proposed algorithm in this study can achieve meter-level accuracy in complex environments and can greatly enhance the accuracy, continuity, and reliability of attitude estimation.

IVU-AutoNav: Integrated Visual and UWB Framework for Autonomous Navigation

To address the inherent scale ambiguity and positioning drift in monocular visual Simultaneous Localization and Mapping (SLAM), this paper proposes a novel localization method that integrates monocular visual SLAM with Ultra-Wideband (UWB) ranging information. This method enables high-precision localization for unmanned aerial vehicles (UAVs) in complex environments without global navigation information. The proposed framework, IVU-AutoNav, relies solely on distance measurements between a fixed UWB anchor and the UAV’s UWB device. Initially, it jointly solves for the position of the UWB anchor and the scale factor of the SLAM system using the scale-ambiguous SLAM data and ranging information. Subsequently, a pose optimization equation is formulated, which integrates visual reprojection errors and ranging errors, to achieve precise localization with a metric scale. Furthermore, a global optimization process is applied to enhance the global consistency of the localization map and optimize the positions of the UWB anchors and scale factor. The proposed approach is validated through both simulation and experimental studies, demonstrating its effectiveness. Experimental results show a scale error of less than 1.8% and a root mean square error of 0.23 m, outperforming existing state-of-the-art visual SLAM systems. These findings underscore the potential and efficacy of the monocular visual-UWB coupled SLAM method in advancing UAV navigation and localization capabilities.