Autonomous Robot Navigation Research Articles

Striving for Less: Minimally-Supervised Pseudo-Label Generation for Monocular Road Segmentation

Identifying traversable space is one of the most important problems in autonomous robot navigation and is primarily tackled using learning-based methods. To alleviate the prohibitively high annotation-cost associated with labeling large and diverse datasets, research has recently shifted from traditional supervised methods to focus on unsupervised and semi-supervised approaches. This work focuses on monocular road segmentation and proposes a practical, generic, and minimally-supervised approach based on task-specific feature extraction and pseudo-labeling. Building on recent advances in monocular depth estimation models, we process approximate dense depth maps to estimate pixel-wise road-plane distance maps. These maps are then used in both unsupervised and semi-supervised road segmentation scenarios. In the unsupervised case, we propose a pseudo-labeling pipeline that reaches state-of-the-art Intersection-over-Union (IoU), while reducing complexity and computations compared to existing approaches. We also investigate a semi-supervised extension to our method and find that even minimal labeling efforts can greatly improve results. Our semi-supervised experiments using as little as 1% and 10% of ground truth data, yield models scoring 0.9063 and 0.9332 on the IoU metric respectively. These results correspond to a comparative performance of 95.9% and 98.7% of a fully-supervised model's IoU score, which motivates a pragmatic approach to labeling.

P‐16.12: Simultaneous Localization and mapping technology based on solid‐state lidar

The problems of autonomous navigation include positioning, mapping and path planning. SLAM has become one of the most popular and mature solutions for autonomous navigation of mobile robots. SLAM literally translates to locating and mapping simultaneously. In general, SLAM frameworks use cameras, IMU, wheel encoders, and a combination of light detection and ranging (lidar) sensors to provide superior accuracy for navigation systems. However, the price of mechanical lidar is exceptionally high compared to other sensors, which is the main reason for introducing a new generation of lasers called solid State Lidar (SSL)

OHM: GPU Based Occupancy Map Generation

Occupancy grid maps (OGMs) are fundamental to most systems for autonomous robotic navigation. However, CPU-based implementations struggle to keep up with data rates from modern 3D lidar sensors, and provide little capacity for modern extensions which maintain richer voxel representations. This article presents Occupancy Homogenous Mapping (OHM), our open source, GPU-based OGM framework. We show how the algorithms can be mapped to GPU resources, resolving difficulties with contention to obtain a successful implementation. The implementation supports many modern OGM algorithms including Normal Distributions Transform-Occupancy Maps (NDT-OM), Normal Distributions Transform-Traversability Maps (NDT-TM), decay-rate and Truncated Sign Distance Function (TSDF). A thorough performance evaluation is presented based on tracked and quadruped Uncrewed Ground Vehicle (UGV) platforms and UAVs, and data sets from both outdoor and subterranean environments. The results demonstrate excellent performance improvements both offline, and for online processing in embedded platforms. Finally, we describe how OHM was a key enabler for the UGV navigation solution for our entry in the Defense Advanced Research Projects Agency (DARPA) Subterranean Challenge, which placed second at the Final Event.

Spatial Expression of Multifaceted Soft Decoration Elements: Application of 3D Reconstruction Algorithm in Soft Decoration and Furnishing Design of Office Space

In China’s modern market economy under the rapid development of the general situation, we work more and more problems, and work pressure is also increasing. The so-called office space refers to the space layout, style, and the physical and psychological division of the space. Office space must take into account many factors, involving technology, technology, humanities, aesthetics, and other elements, while the office space is the space where people work and relax. In recent years, as people’s requirements for the work environment are increasingly high, therefore, the design of the office space is also more and more attention to people. The concept of soft furnishing design into the work space will help improve the overall corporate and office space design of cultural taste which is one of the main methods to show the quality and human connotation of the enterprise. The three-dimensional reconstruction refers to the creation of a mathematical model suitable for computer display and processing of three-dimensional space objects. It is an important basic tool for data processing, computing, and researching the performance of mathematical models in the computer environment, which can be applied in various fields such as autonomous navigation of mobile robots, aviation and remote sensing computing, industrial monitoring information system, medical imaging, and virtual reality. The 3D environment reconstruction technology has become one of the popular research areas in computer vision and increasingly attracts the attention of design practitioners. This paper takes the 3D environment reconstruction technology of office space soft decoration design as the basis and discusses the important elements and modeling ideas in soft decoration design, which adds to the interior design of office space, and uses Kinect to obtain the depth data in the 3D environment, so as to complete the realistic 3D reproduction of the interior environment based on computer vision technology.

Autonomous navigation of mobile robot in radiation environment with uneven terrain

Autonomous navigation of mobile robot in radiation environment with uneven terrain

Robot visual navigation estimation and target localization based on neural network

Abstract The high computational cost, complex external environment, and limited computing resources of embedded system are some major problems in traditional autonomous robot navigation methods. To overcome these problems, a mobile robot path planning navigation system based on panoramic vision was proposed. This method first describes the structure and functions of the navigation system. It explains how to use the environment to explore and map in order to create a panoramic vision sensor. Finally, it elaborates on the breadth-first search based on regression neural network (RNN) method, the Voronoi skeleton diagram method, the algorithm principle, and how to navigate by the planning path implementation of practical strategies. The simulation results illustrate that the breadth-first search method and the Voronoi skeleton graph method based on panoramic view have a high speed. The accessibility of RNN planning algorithm can effectively solve the difficult problems such as high computing overhead, complex navigation environment, and limited computing resources. In the actual robot navigation experiment, the difference in real-time performance and optimality performance that exists between the two algorithms is reflected in the length and duration of the course taken by the robot. When applied to a variety of site environments, the breadth-first search method requires between 23.2 and 45.3% more time to calculate the planned path than the Voronoi skeleton graph method, despite the fact that the planned path length is between 20.7 and 35.9% shorter using the breadth-first search method. It serves as a guide for choosing the appropriate algorithm to implement in practical applications.

A Brief Study on Object Detection and Tracking

The detection and tracking of objects in computer vision systems is a critical and challenging subject. Object identification and tracking currently has applications in a wide range of industries, including surveillance, autonomous robot navigation and vehicle navigation because to the increased accessibility of computing power and vast public datasets. When analyzing an image, the object tracking algorithm goal is to divide the region of interest, monitor its motion and location and identify any blobs. Every tracking necessitates an object detection mechanism either in each frame or whenever an object appears for the first time in the video sequence. The availability of high performance processors and low cost-high quality video cameras has enhanced the need of automated object tracking video sequences. Multiple object Identification and tracking in a dynamic environment has recently become focus of computer vision research. In this paper we discuss about the various object detection and tracking algorithms.

Mobile robot path planning based on fusion of improved A* algorithm and adaptive DWA algorithm

For mobile robot autonomous navigation, path planning algorithm matters. For the reason that traditional A* algorithm is unable to achieve the global path planning perfectly and avoid unknown obstacles at the same time, the mobile robot algorithm combining improved A* with adaptive DWA was proposed for path planning. Firstly, safe distance judgment is added in A* algorithm to avoid the problem that the path crosses the vertex of obstacle diagonally. Secondly, the dynamic tangential method is used to smooth the path of A* algorithm, which is more consistent with the actual motion curve of mobile robot. Then, the path key points explored by improved A* algorithm are extracted as intermediate target points of the velocity adaptive DWA algorithm. Unknown obstacles avoidance based on the optimal global path planning is realized. Finally, the proposed fusion algorithm’s feasibility is verified through mobile robot ROS simulation.

Real-Time Localization and Mapping Utilizing Multi-Sensor Fusion and Visual–IMU–Wheel Odometry for Agricultural Robots in Unstructured, Dynamic and GPS-Denied Greenhouse Environments

Autonomous navigation in greenhouses requires agricultural robots to localize and generate a globally consistent map of surroundings in real-time. However, accurate and robust localization and mapping are still challenging for agricultural robots due to the unstructured, dynamic and GPS-denied environmental conditions. In this study, a state-of-the-art real-time localization and mapping system was presented to achieve precise pose estimation and dense three-dimensional (3D) point cloud mapping in complex greenhouses by utilizing multi-sensor fusion and Visual–IMU–Wheel odometry. In this method, measurements from wheel odometry, an inertial measurement unit (IMU) and a tightly coupled visual–inertial odometry (VIO) are integrated into a loosely coupled framework based on the Extended Kalman Filter (EKF) to obtain a more accurate state estimation of the robot. In the multi-sensor fusion algorithm, the pose estimations from the wheel odometry and IMU are treated as predictions and the localization results from VIO are used as observations to update the state vector. Simultaneously, the dense 3D map of the greenhouse is reconstructed in real-time by employing the modified ORB-SLAM2. The performance of the proposed system was evaluated in modern standard solar greenhouses with harsh environmental conditions. Taking advantage of measurements from individual sensors, our method is robust enough to cope with various challenges, as shown by extensive experiments conducted in the greenhouses and outdoor campus environment. Additionally, the results show that our proposed framework can improve the localization accuracy of the visual–inertial odometry, demonstrating the satisfactory capability of the proposed approach and highlighting its promising applications in autonomous navigation of agricultural robots.

Effective Safety Strategy for Mobile Robots Based on Laser-Visual Fusion in Home Environments

The proven efficacy of safety strategies based on 2-D laser rangefinder (LRF) strongly stimulates their application to mobile robots operating in the home environment. However, it remains a challenge for the robot to avoid collisions with all obstacles in the environment. Since LRF can only scan a horizontal slice of the world, some objects cannot be fully observed, such as tables and chairs. In this article, an effective solution based on laser-visual fusion is presented to enhance the safety of the robot. First, a vision sensor is adopted to help detect obstacles that are not fully visible to LRF. Then we propose a method to convert the depth information of the visual image into 2-D <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">pseudo-laser data</i> representation. With this representation, a strategy for 2-D mapping is developed. On this basis, a novel map fusion algorithm is proposed to generate an improved grid map that amends the incorrect representation of obstacles on the traditional 2-D grid map. We further investigate a robot autonomous navigation strategy that considers LRF data and pseudo-laser data to avoid all obstacles. Experimental results show that the improved grid map together with the presented navigation strategy allows the robot not only to plan a “real” collision-free path, but also to navigate safely in both static and dynamic scenarios, and the proposed strategies can significantly enhance the performance of robot navigation in terms of safety, reliability and robustness.

A Semantic Classification Approach for Indoor Robot Navigation

Autonomous robot navigation has become a crucial concept in industrial development for minimizing manual tasks. Most of the existing robot navigation systems are based on the perceived geometrical features of the environment, with the employment of sensory devices including laser scanners, video cameras, and microwave radars to build the environment structure. However, scene understanding is a significant issue in the development of robots that can be controlled autonomously. The semantic model of the indoor environment offers the robot a representation closer to the human perception, and this enhances navigation tasks and human–robot interaction. In this paper, we propose a low-cost and low-memory framework that offers an improved representation of the environment using semantic information based on LiDAR sensory data. The output of the proposed work is a reliable classification system for indoor environments with an efficient classification accuracy of 97.21% using the collected dataset.

Autonomous live working robot navigation with real‐time detection and motion planning system on distribution line

In this study, an autonomous robot navigation system is designed for live working on distribution line. The developed system features a real-time detection and motion planning system, incorporating a manipulator capable of grasping power components. In order to accurately identify targets, the authors propose an object detection method based on the Larger Scale ‘You Only Look Once’ Version 4 (LS-YOLOv4) algorithm for detecting the insulators and drop fuses. The LS-YOLOv4 extracts features of power components by Convolutional Neural Network (CNN), and then performs feature fusion. Then the authors develop a motion planning method based on the Node Control Optimal Rapidly Exploring Random Trees (NC-RRT*), which can drive the robot to realise the autonomous robot motion planning and obstacle avoidance. On the grasping function, the authors present a reliable Lightweight-based Convolutional Neural Network (L-CNN) grasping point detection method. Finally, the authors evaluate fully autonomous robotic system in both simulated and real-world experiments. The experimental results demonstrate that the proposed system can effectively identify the target and complete the grasping task in an efficient way. Notably, the proposed motion planning method can take into account both planning efficiency and accuracy to manipulation tasks.

Distributed supervisory control for multiple robot autonomous navigation performing single-robot tasks

This paper proposes two new approaches based on the Supervisory Control Theory (SCT) of Discrete Event Systems (DES) for autonomous navigation of multiple robots with single-robot tasks being assigned by a centralized scheduler. The two planning strategies differ regarding their implementation, which can be centralized or distributed. Nevertheless, both approaches share the central objective of considering the scheduler and robots as DES, allowing the use of SCT to model and control the behavior of the whole multi-robotic system. Particularly, SCT is used to gather deliberative and reactive motion planning algorithms through a structured procedure, aiming to safely navigate in cluttered, dynamic, partially-known environments. The open-loop behavior of the autonomous navigation system is modeled for both approaches. Moreover, standard supervisors are obtained for the centralized planning approach. In addition, the supervisor localization method is considered to allow a distributed planning approach. Furthermore, as a case study, real experiments considering mobile robots are carried out to corroborate the proposed framework.

An Adaptive Threshold Line Segment Feature Extraction Algorithm for Laser Radar Scanning Environments

An accurate map is needed for the autonomous navigation of mobile robots in unknown environments. The application of laser radars has the advantages of high ranging accuracy and long ranging distances. Due to the small amount of data on laser radars and the influence of noise on the sensor itself, these amount to causing problems such as low accuracies of map construction and large positioning errors. Currently, the feature extraction of environmental line segments based on radar scanning data generally adopts the idea of recursion. However, the amount of calculations for applying recursion is large, and the threshold of extracted feature points needs to be set manually. Moreover, the fixed segmentation threshold will cause under-segmentation or over-segmentation. In this paper, an adaptive threshold-based feature extraction method for environmental line segments is proposed. The method denoises the original data first, and then an adaptive threshold of the nearest neighbor algorithm is provided to improve the accuracy of breakpoint judgment; next, the slope difference between adjacent line segments is evaluated according to the line segment fitting error in order to obtain the optimal corner feature. Finally, the point set is segmented to fit line-segment features. Based on actual environment tests, the environmental similarity of the line segment features extracted by the new algorithm in this paper increases by 8.3% compared with the IEPF (Iterative End Point Fit) algorithm. The algorithm avoids recursive operations, improves the efficiency by four times, and meets the real-time requirements of line segment fitting.

Applications of machine vision in agricultural robot navigation: A review

Many tasks in smart agriculture have further requirements for the autonomous navigation of agricultural robots. Due to irreplaceable visual information and low-cost hardware costs, machine vision (MV) is widely used in agricultural robot navigation. This paper reviewed applications of MV in agricultural robot navigation. First, the advantages, disadvantages and roles of different vision sensors and MV algorithms in agricultural robot navigation were introduced. And then, the development status of MV in agricultural robot navigation was reviewed. Finally, the challenges of MV in agricultural robot navigation were discussed. To address these challenges, this paper looks forward to the development directions of future research.

Eddy Current Magnetic Localization of Nonmagnetic Metal Targets Based on Metal Shell Model

Target detection technology based on magnetic anomalies, due to its wide detection range, high localization accuracy, all weather conditions and other characteristics, plays an important role in geological prospecting, medical monitoring, disaster rescue and other fields. Some nonmagnetic or weakly magnetic metal targets are difficult to observe with passive magnetic anomaly detection (MAD) methods. The active MAD method based on the eddy current magnetic effect can locate nonmagnetic metal targets. However, traditional detection methods based on magnetic dipoles require knowledge of the target magnetic moment in advance and are not suitable for near-field detection. To solve these problems, this paper establishes a detection model based on a nonmagnetic metal shell target and derives the eddy current magnetic field (ECMF) expression of the nonmagnetic metal shell target. An ECMF extraction method that separates the ECMF from the excitation magnetic field and background interference from the perspective of the signal phase is further proposed. Then, the particle swarm optimization (PSO) algorithm is used to locate the nonmagnetic metal target. When the metal target is located within the range of 1.2 m*1.2 m, the localization error is approximately 2.5%. This method is important and valuable in the fields of medical pill tracking, drug dose conversion monitoring, trapped miner locating, indoor autonomous robot navigation, unexploded object detection and other fields.

REAL-TIME MULTIPLE OBJECT DETECTION AND TRACKING

In many computer vision applications, such as surveillance, vehicle navigation, and autonomous robot navigation, object recognition and tracking are critical and difficult problems. One of the flow testing research subjects in PC vision is video observation in a powerful climate, particularly for people and cars. It is a critical innovation in the fight against illegal intimidation, crime, public safety, and effective traffic control. The endeavour entails developing an effective video surveillance system for use in complex contexts. Detecting moving things from a video is critical for object identification and target tracking in video surveillance. The detection of moving objects in video streams is the first significant phase of information, and background subtraction is a common method for foreground segmentation. Due to an increased demand for such systems in public spaces such as airports, subway stations, and mass events, intelligent and automated security surveillance systems have been an important study topic in recent years. In this context, one of the most significant needs for surveillance systems based on the tracking of abandoned, stolen, or parked vehicles is the tracking of fixed foreground regions. Because they work reasonably well when the camera is stationary and the change in ambient lighting is gradual, object tracking-based techniques are the most popular choice for detecting stationary foreground objects. They are also the most popular choice for separating foreground objects from the current frame.

RL-DOVS: Reinforcement Learning for Autonomous Robot Navigation in Dynamic Environments.

Autonomous navigation in dynamic environments where people move unpredictably is an essential task for service robots in real-world populated scenarios. Recent works in reinforcement learning (RL) have been applied to autonomous vehicle driving and to navigation around pedestrians. In this paper, we present a novel planner (reinforcement learning dynamic object velocity space, RL-DOVS) based on an RL technique for dynamic environments. The method explicitly considers the robot kinodynamic constraints for selecting the actions in every control period. The main contribution of our work is to use an environment model where the dynamism is represented in the robocentric velocity space as input to the learning system. The use of this dynamic information speeds the training process with respect to other techniques that learn directly either from raw sensors (vision, lidar) or from basic information about obstacle location and kinematics. We propose two approaches using RL and dynamic obstacle velocity (DOVS), RL-DOVS-A, which automatically learns the actions having the maximum utility, and RL-DOVS-D, in which the actions are selected by a human driver. Simulation results and evaluation are presented using different numbers of active agents and static and moving passive agents with random motion directions and velocities in many different scenarios. The performance of the technique is compared with other state-of-the-art techniques for solving navigation problems in environments such as ours.

Close Proximity Time-to-collision Prediction for Autonomous Robot Navigation: An Exponential GPR Approach

Fusion of X-band Doppler radar and infrared sensors can offer a great advantage for close proximity time-to-collision (TTC) prediction in the field of autonomous robot navigation due to precise obstacle speed detection and direction sensing. Nevertheless, poor ranging performance from the infrared sensors may result owing to fluctuating reflectivity of the moving obstacle. The TTC prediction accuracy may also be further degraded when the obstacle’s trajectory is non-parallel to the robot’s heading direction due to uncertainty in the radar’s radiation pattern which typically increases at the side lobes of the antenna. Thus, to enhance the performance, we propose an exponential Gaussian Process Regression (E-GPR)-based prediction model which is able to approximate an unknown function in a probabilistic manner. The proposed method is validated via a series of experiments with an obstacle approaching the robot from different viewing angles with a speed ranging between 30 and 63 cm/s. Results demonstrate that the average TTC error based on the sensor fusion is 0.31s; but with the E-GPR method, the error is successfully reduced to 0.0937s, which is the largest error reduction when compared against other competing machine-learning models such as multilayer perceptron neural network, support vector machine and boosted tree.

Biomimetic navigation system using a polarization sensor and a binocular camera.

With the vigorous development of vision techniques, simultaneous localization and mapping (SLAM) has shown the capability of navigating autonomous robots in global-navigation-satellite-system-denied environments. However, the long-term robust navigation of lightweight autonomous robots in outdoor environments with complex interferences, such as illumination change, dynamic objects, and electromagnetic interference, is still a great challenge. In this paper, a polarization sensor-aided SLAM (POL-SLAM) that can provide absolute heading constraints for pure SLAM is proposed. POL-SLAM is a lightweight, tightly coupled system consisting of a polarization sensor and binocular camera. By means of an initialization that uses a polarization sensor, an absolute heading angle for the entire map is designed. Additionally, an algorithm to eliminate mismatching points using the matching point vector is proposed. The objective function of bundle adjustment is then deduced according to the re-projection error and polarization sensor. The vehicle test shows that the yaw and trajectory accuracies of POL-SLAM are significantly improved compared to pure SLAM. The yaw and trajectory accuracies are increased by 43.1% and 36.6%, respectively. These results indicate that the proposed POL-SLAM can improve the reliability and robustness of pure SLAM and can be used in lightweight autonomous robots in outdoor environments.