Active SLAM Research Articles

LiDAR-Based End-to-End Active SLAM Using Deep Reinforcement Learning in Large-Scale Environments

LiDAR-Based End-to-End Active SLAM Using Deep Reinforcement Learning in Large-Scale Environments

Exploration- and Exploitation-Driven Deep Deterministic Policy Gradient for Active SLAM in Unknown Indoor Environments

This study proposes a solution for Active Simultaneous Localization and Mapping (Active SLAM) of robots in unknown indoor environments using a combination of Deep Deterministic Policy Gradient (DDPG) path planning and the Cartographer algorithm. To enhance the convergence speed of the DDPG network and minimize collisions with obstacles, we devised a unique reward function that integrates exploration and exploitation strategies. The exploration strategy allows the robot to achieve the shortest running time and movement trajectory, enabling efficient traversal of unmapped environments. Moreover, the exploitation strategy introduces active closed loops to enhance map accuracy. We conducted experiments using the simulation platform Gazebo to validate our proposed model. The experimental results demonstrate that our model surpasses other Active SLAM methods in exploring and mapping unknown environments, achieving significant grid completeness of 98.7%.

NAVS: A Neural Attention-Based Visual SLAM for Autonomous Navigation in Unknown 3D Environments

Navigation in unknown 3D environments aims to progressively find an efficient path to a given target goal in unseen scenarios. A challenge is how to explore the navigation quickly and effectively. An end-to-end learning approach has been proposed to extract geometric shapes from RGB images, but it is not suitable for large environments due to its exhaustive exploration with exponential search space. Active Neural SLAM (ANS) presents a Neural SLAM module to maximize the exploration coverage to tackle the active SLAM task. However, ANS still frequently visits the explored areas due to the inappropriate local target selection. In this paper, we propose a Neural Attention-based Visual SLAM (NAVS) model to explore unknown 3D environments. Spatial attention is provided to quickly identify obstacles (such as similarly colored tea table or floor). We also leverage the priority of unknown regions in the short-term goal decision to avoid frequent exploration with a channel attention. The experimental results show that our model can build a more accurate map than ANS and other baseline methods with less running time. In terms of relative coverage, NAVS achieves a 0.5%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} improvement over ANS in overall and a 1.1%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} improvement over ANS in large environments.

Collaborative Autonomous Navigation of Quadrotors in Unknown Outdoor Environments: An Active Visual SLAM Approach

Collaborative Autonomous Navigation of Quadrotors in Unknown Outdoor Environments: An Active Visual SLAM Approach

Topological belief space planning for active SLAM with pairwise Gaussian potentials and performance guarantees

Determining a globally optimal solution of belief space planning (BSP) in high-dimensional state spaces directly is computationally expensive, as it involves belief propagation and objective function evaluation for each candidate action. However, many problems of interest, such as active SLAM, exhibit structure that can be harnessed to expedite planning. Also, in order to choose an optimal action, an exact value of the objective function is not required as long as the actions can be sorted in the same way. In this paper we leverage these two key aspects and present the topological belief space planning (t-bsp) concept that uses topological signatures to perform this ranking for information-theoretic cost functions, considering only topologies of factor graphs that correspond to future posterior beliefs. In particular, we propose a highly efficient topological signature based on the von Neumann graph entropy that is a function of graph node degrees and supports an incremental update. We analyze it in the context of active pose SLAM and derive error bounds between the proposed topological signature and the original information-theoretic cost function. These bounds are then used to provide performance guarantees for t-bsp with respect to a given solver of the original information-theoretic BSP problem. Realistic and synthetic simulations demonstrate drastic speed-up of the proposed method with respect to the state-of-the-art methods while retaining the ability to select a near-optimal solution. A proof of concept of t-bsp is given in a small-scale real-world active SLAM experiment.

Symmetry-aware Neural Architecture for Embodied Visual Navigation

AbstractThe existing methods for addressing visual navigation employ deep reinforcement learning as the standard tool for the task. However, they tend to be vulnerable to statistical shifts between the training and test data, resulting in poor generalization over novel environments that are out-of-distribution from the training data. In this study, we attempt to improve the generalization ability by utilizing the inductive biases available for the task. Employing the active neural SLAM that learns policies with the advantage actor-critic method as the base framework, we first point out that the mappings represented by the actor and the critic should satisfy specific symmetries. We then propose a network design for the actor and the critic to inherently attain these symmetries. Specifically, we use G-convolution instead of the standard convolution and insert the semi-global polar pooling layer, which we newly design in this study, in the last section of the critic network. Our method can be integrated into existing methods that utilize intermediate goals and 2D occupancy maps. Experimental results show that our method improves generalization ability by a good margin over visual exploration and object goal navigation, which are two main embodied visual navigation tasks.

Active SLAM: A Review on Last Decade.

This article presents a comprehensive review of the Active Simultaneous Localization and Mapping (A-SLAM) research conducted over the past decade. It explores the formulation, applications, and methodologies employed in A-SLAM, particularly in trajectory generation and control-action selection, drawing on concepts from Information Theory (IT) and the Theory of Optimal Experimental Design (TOED). This review includes both qualitative and quantitative analyses of various approaches, deployment scenarios, configurations, path-planning methods, and utility functions within A-SLAM research. Furthermore, this article introduces a novel analysis of Active Collaborative SLAM (AC-SLAM), focusing on collaborative aspects within SLAM systems. It includes a thorough examination of collaborative parameters and approaches, supported by both qualitative and statistical assessments. This study also identifies limitations in the existing literature and suggests potential avenues for future research. This survey serves as a valuable resource for researchers seeking insights into A-SLAM methods and techniques, offering a current overview of A-SLAM formulation.

Actively-exploring thermography-enabled autonomous robotic system for detecting and registering HVAC thermal leaks

The inefficiency, labor-intensity, and possibility of automation of the current practice of HVAC system inspections, which heavily relies on human inspectors, are being questioned. This paper describes a lightweight and reproducible robotic system dubbed AcTEA-bot, that automatically detects and locates thermal leaks in ceiling environments while self-navigating. AcTEA-bot is designed and built upon three modules, including an active visual SLAM module, a thermal leak detection module, and a thermal leak registration module. To evaluate the effectiveness of AcTEA-bot, a true-to-scale ceiling environment containing five thermal leaks is built, and 30 experiments are conducted where AcTEA-bot starts with different poses and conducts the inspection task. Results demonstrate that thermal leaks of multiple types in ceiling environments can be accurately detected and located.

Gaze Control for Active Visual SLAM via Panoramic Cost Map

This paper focuses on improving the positioning accuracy of the visual simultaneous localization and mapping (VSLAM) through actively controlling the gaze of the positioning camera mounted on an autonomous mobile robot. We present a gaze control method for the VSLAM with a novel panoramic cost map. In the method, we add a panoramic camera besides the positioning camera to aid the gaze control of the positioning camera. The panoramic camera is modeled as a unit sphere model and the positioning camera is modeled as a pinhole model that is concentric with the unit sphere model. The panoramic camera captures a panoramic image for the evaluation of an environment. The effective factors, such as feature points and moving objectives, of the VSLAM are detected in the panoramic image. When the pinhole model pointing at a specific orientation, a cost value can be calculated according to the effective factors. After calculating the above-mentioned cost in all available orientations of the pinhole model, a panoramic cost map that is determined by the individual costs can be established. Next, a target orientation is selected according to the panoramic cost map and the positioning camera is rotated to the target orientation by a pan-tilt. The test results in different scenery show that the proposed method can improve the positioning accuracy of the VSLAM compared to stationary orientation.

Perception-Aware Planning for Active SLAM in Dynamic Environments

This paper presents a perception-aware path planner for active SLAM in dynamic environments using micro-aerial vehicles (MAV). The “Next-Best-View” planner (NBVP planner) is combined with an active loop closing, which is called the Active Loop Closing Planner (ALCP planner). The planner is proposed to avoid both static and dynamic obstacles in unknown environments while reducing the uncertainty of the SLAM system and further improving the accuracy of localization. First, the receding horizon strategy is adopted to find the next waypoint. The cost function that combines the exploration gain and the loop closing gain is designed. The former can reduce the mapping uncertainty, while the latter takes the loop closing possibility into consideration. Second, a key waypoint selection strategy is designed. The selected key waypoints, instead of all waypoints, are treated as potential loop-closing points to make the algorithm more efficient. Moreover, a fuzzy RRT-based dynamic obstacle avoidance algorithm is adopted to realize obstacle avoidance in dynamic environments. Simulations in different challenging scenarios are conducted to verify the effectiveness of the proposed algorithm.

IRotate: Active Visual SLAM for Omnidirectional Robots

In this paper, we present an active visual SLAM approach for omnidirectional robots. The goal is to generate control commands that allow such a robot to simultaneously localize itself and map an unknown environment while maximizing the amount of information gained and consuming as low energy as possible. Leveraging the robot’s independent translation and rotation control, we introduce a multi-layered approach for active V-SLAM. The top layer decides on informative goal locations and generates highly informative paths to them. The second and third layers actively re-plan and execute the path, exploiting the continuously updated map and local features information. Moreover, we introduce two utility formulations to account for the presence of obstacles in the field of view and the robot’s location. Through rigorous simulations, real robot experiments, and comparisons with state-of-the-art methods, we demonstrate that our approach achieves similar coverage results with lesser overall map entropy. This is obtained while keeping the traversed distance up to 39% shorter than the other methods and without increasing the wheels’ total rotation amount. Code and implementation details are provided as open-source, and all the generated data is available online for consultation.

The New Method of Active SLAM for Mapping Using LiDAR

In this article, we propose a new approach to addressing the issue of active SLAM. In this design, we used the already functional SLAM algorithm, which we modified for our case. Matlab was used as the main software tool. All proposed methods were experimentally verified on a mobile robotic system. We used LiDAR as the primary sensor. After mapping the environment, we created a grid map. The grid map may be used for the navigation of the mobile robotic system, but the navigation and control of the mobile robotic system are not involved in this article. The result of the whole process is an autonomous mapping of the environment.

Active SLAM: Place Classification and Viewpoint Planning from Domain-Invariant Scene Graphs

Active SLAM: Place Classification and Viewpoint Planning from Domain-Invariant Scene Graphs

Map Making in Social Indoor Environment Through Robot Navigation Using Active SLAM

Robotics has come a long way from industrial robotic arms and is all set to enter our homes. The capability of a robot to navigate in an unknown human populated environment with obstacles and making map simultaneously is one of the significant characteristics in the domain of autonomous robotics. Further, the problem of robot navigating in a social environment while ensuring human safety and comfort through social norms needs to be addressed. This article presents a solution for mapping of unknown terrains with dynamic obstacles using simultaneous localization in social environments through Adaptive Squashing Function based artificial neural network training, which is able to track the target orientation angles more efficiently as compared to conventional fixed slope squashing function based backpropagation training algorithm. The performance of different state of the art techniques have been compared with proposed work through simulation models. Simulation results demonstrated the effectiveness of the proposed algorithm in complex environment where the proposed algorithm converged in less than 50% of the iterations taken by the exhaustive search algorithms and approximately 33% of the iterations taken by random search algorithm. Further, the proposed approach was tested in the real world settings, wherein the robot was deployed to create map for the Kalpana Chawla Center for Research in Space Science and Technology, Chandigarh University with mobile humans.

A Deep Reinforcement Learning Approach for Active SLAM

In this paper, we formulate the active SLAM paradigm in terms of model-free Deep Reinforcement Learning, embedding the traditional utility functions based on the Theory of Optimal Experimental Design in rewards, and therefore relaxing the intensive computations of classical approaches. We validate such formulation in a complex simulation environment, using a state-of-the-art deep Q-learning architecture with laser measurements as network inputs. Trained agents become capable not only to learn a policy to navigate and explore in the absence of an environment model but also to transfer their knowledge to previously unseen maps, which is a key requirement in robotic exploration.

Active Asteroid-SLAM

In this paper, we propose an active real-time capable 3D graph based simultaneous localization and mapping (Graph SLAM) approach, which actively estimates the state of an autonomous spacecraft relative to a simultaneously established map estimate. The graph is constructed in a tightly-coupled fashion, where an Extended Kalman Filter estimates the relative offset between two of its vertices. An additional relative measurement is derived by matching point clouds obtained by a light detection and ranging (LiDAR) system. In order to yield a significant speed-up, scan matching is implemented on the GPU. To reduce the uncertainty of either the state or the map estimate, we present an approach to actively control the system resting on an extended representation of uncertainty in the map. Furthermore, it adapts its behavior depending on the current uncertainty distribution in order to find a dynamic trade-off between exploitation (improve localization performance) and exploration (improve knowledge about the environment). Finally, we present a post-processing approach to discover landing sites in the map estimate without prior knowledge. The evaluation is conducted in a numerical simulation, where the spacecraft explores the real 3D model of Itokawa in its actual dynamic environment. Within that simulation, we use a shader-based GPU implementation for simulating LiDAR measurements. We evaluate the performance of the active SLAM approach and demonstrate that the use of the adaptive approach improves navigation and exploration performance at the same time.

NAO robot obstacle avoidance based on fuzzy Q-learning

PurposeThis paper aims to propose a novel active SLAM framework to realize avoid obstacles and finish the autonomous navigation in indoor environment.Design/methodology/approachThe improved fuzzy optimized Q-Learning (FOQL) algorithm is used to solve the avoidance obstacles problem of the robot in the environment. To reduce the motion deviation of the robot, fractional controller is designed. The localization of the robot is based on FastSLAM algorithm.FindingsSimulation results of avoiding obstacles using traditional Q-learning algorithm, optimized Q-learning algorithm and FOQL algorithm are compared. The simulation results show that the improved FOQL algorithm has a faster learning speed than other two algorithms. To verify the simulation result, the FOQL algorithm is implemented on a NAO robot and the experimental results demonstrate that the improved fuzzy optimized Q-Learning obstacle avoidance algorithm is feasible and effective.Originality/valueThe improved fuzzy optimized Q-Learning (FOQL) algorithm is used to solve the avoidance obstacles problem of the robot in the environment. To reduce the motion deviation of the robot, fractional controller is designed. To verify the simulation result, the FOQL algorithm is implemented on a NAO robot and the experimental results demonstrate that the improved fuzzy optimized Q-Learning obstacle avoidance algorithm is feasible and effective.

Real-Time Active SLAM and Obstacle Avoidance for an Autonomous Robot Based on Stereo Vision

In this article, the problem of real-time robot exploration and map building (active SLAM) is considered. A single stereo vision camera is exploited by a fully autonomous robot to navigate, localize itself, define its surroundings, and avoid any possible obstacle in the aim of maximizing the mapped region following the optimal route. A modified version of the so-called cognitive-based adaptive optimization algorithm is introduced for the robot to successfully complete its tasks in real time and avoid any local minima entrapment. The method’s effectiveness and performance were tested under various simulation environments as well as real unknown areas with the use of properly equipped robots.

On the Importance of Uncertainty Representation in Active SLAM

The purpose of this work is to highlight the paramount importance of representing and quantifying uncertainty to correctly report the associated confidence of the robot's location estimate at each time step along its trajectory and therefore decide the correct course of action in an active SLAM mission. We analyze the monotonicity property of different decision-making criteria, both in 2-D and 3-D, with respect to the representation of uncertainty and of the orientation of the robot's pose. Monotonicity, the property that uncertainty increases as the robot moves, is essential for adequate decision making. We analytically show that, by using differential representations to propagate spatial uncertainties, monotonicity is preserved for all optimality criteria, A-opt , D-opt , and E-opt , and for Shannon's entropy. We also show that monotonicity does not hold for any criteria in absolute representations using Roll-Pitch-Yaw and Euler angles. Finally, using unit quaternions in absolute representations, the only criteria that preserve monotonicity are D-opt and Shannon's entropy.

Autonomous planetary rover navigation via active SLAM

PurposeThis paper aims to introduce an efficient active-simultaneous localization and mapping (SLAM) approach for rover navigation, future planetary rover exploration mission requires the rover to automatically localize itself with high accuracy.Design/methodology/approachA three-dimensional (3D) feature detection method is first proposed to extract salient features from the observed point cloud, after that, the salient features are employed as the candidate destinations for re-visiting under SLAM structure, followed by a path planning algorithm integrated with SLAM, wherein the path length and map utility are leveraged to reduce the growth rate of state estimation uncertainty.FindingsThe proposed approach is able to extract distinguishable 3D landmarks for feature re-visiting, and can be naturally integrated with any SLAM algorithms in an efficient manner to improve the navigation accuracy.Originality/valueThis paper proposes a novel active-SLAM structure for planetary rover exploration mission, the salient feature extraction method and active revisit patch planning method are validated to improve the accuracy of pose estimation.