Autonomous Mobile Robot Navigation Research Articles

L-PCM: Localization and Point Cloud Registration-Based Method for Pose Calibration of Mobile Robots

The autonomous navigation of mobile robots contains three parts: map building, global localization, and path planning. Precise pose data directly affect the accuracy of global localization. However, the cumulative error problems of sensors and various estimation strategies cause the pose to have a large gap in data accuracy. To address these problems, this paper proposes a pose calibration method based on localization and point cloud registration, which is called L-PCM. Firstly, the method obtains the odometer and IMU (inertial measurement unit) data through the sensors mounted on the mobile robot and uses the UKF (unscented Kalman filter) algorithm to filter and fuse the odometer data and IMU data to obtain the estimated pose of the mobile robot. Secondly, the AMCL (adaptive Monte Carlo localization) is improved by combining the UKF fusion model of the IMU and odometer to obtain the modified global initial pose of the mobile robot. Finally, PL-ICP (point to line-iterative closest point) point cloud registration is used to calibrate the modified global initial pose to obtain the global pose of the mobile robot. Through simulation experiments, it is verified that the UKF fusion algorithm can reduce the influence of cumulative errors and the improved AMCL algorithm can optimize the pose trajectory. The average value of the position error is about 0.0447 m, and the average value of the angle error is stabilized at about 0.0049 degrees. Meanwhile, it has been verified that the L-PCM is significantly better than the existing AMCL algorithm, with a position error of about 0.01726 m and an average angle error of about 0.00302 degrees, effectively improving the accuracy of the pose.

The integration of GPS and visual navigation for autonomous navigation of an Ackerman steering mobile robot in cotton fields.

Autonomous navigation in agricultural fields presents a unique challenge due to the unpredictable outdoor environment. Various approaches have been explored to tackle this task, each with its own set of challenges. These include GPS guidance, which faces availability issues and struggles to avoid obstacles, and vision guidance techniques, which are sensitive to changes in light, weeds, and crop growth. This study proposes a novel idea that combining GPS and visual navigation offers an optimal solution for autonomous navigation in agricultural fields. Three solutions for autonomous navigation in cotton fields were developed and evaluated. The first solution utilized a path tracking algorithm, Pure Pursuit, to follow GPS coordinates and guide a mobile robot. It achieved an average lateral deviation of 8.3 cm from the pre-recorded path. The second solution employed a deep learning model, specifically a fully convolutional neural network for semantic segmentation, to detect paths between cotton rows. The mobile rover then navigated using the Dynamic Window Approach (DWA) path planning algorithm, achieving an average lateral deviation of 4.8 cm from the desired path. Finally, the two solutions were integrated for a more practical approach. GPS served as a global planner to map the field, while the deep learning model and DWA acted as a local planner for navigation and real-time decision-making. This integrated solution enabled the robot to navigate between cotton rows with an average lateral distance error of 9.5 cm, offering a more practical method for autonomous navigation in cotton fields.

Research on a Visual/Ultra-Wideband Tightly Coupled Fusion Localization Algorithm.

In the autonomous navigation of mobile robots, precise positioning is crucial. In forest environments with weak satellite signals or in sites disturbed by complex environments, satellite positioning accuracy has difficulty in meeting the requirements of autonomous navigation positioning accuracy for robots. This article proposes a vision SLAM/UWB tightly coupled localization method and designs a UWB non-line-of-sight error identification method using the displacement increment of the visual odometer. It utilizes the displacement increment of visual output and UWB ranging information as measurement values and applies the extended Kalman filtering algorithm for data fusion. This study utilized the constructed experimental platform to collect images and ultra-wideband ranging data in outdoor environments and experimentally validated the combined positioning method. The experimental results show that the algorithm outperforms individual UWB or loosely coupled combination positioning methods in terms of positioning accuracy. It effectively eliminates non-line-of-sight errors in UWB, improving the accuracy and stability of the combined positioning system.

A Hardware Realization Framework for Fuzzy Inference System Optimization

Fuzzy inference systems (FISs) are a key focus for decision-making in embedded systems due to their effectiveness in managing uncertainty and non-linearity. This study demonstrates that optimizing FIS hardware enhances performance, efficiency, and capabilities, improving user experience, heightened productivity, and cost savings. We propose an ultra-low power FIS hardware framework to address power constraints in embedded systems. This framework supports optimizations for conventional arithmetic and Most Significant Digit First (MSDF) computing, ensuring compatibility with MSDF-based sensors. Within the MSDF-computing FIS, fuzzification, inference, and defuzzification processes occur on serially incoming data bits. To illustrate the framework’s efficiency, we implemented it using MATLAB, Chisel3, and Vivado, starting from high-level FIS descriptions and progressing to hardware synthesis. A Scala library in Chisel3 was developed to connect these tools seamlessly, facilitating design space exploration at the arithmetic level. We applied the framework by realizing an FIS for autonomous mobile robot navigation in unknown environments. The synthesis results highlight the superiority of our designs over the MATLAB HDL code generator, achieving a 43% higher clock frequency, and 46% and 67% lower resource and power consumption, respectively.

Deep‐Q‐Network hybridization with extended Kalman filter for accelerate learning in autonomous navigation with auxiliary security module

AbstractThis article proposes an algorithm for autonomous navigation of mobile robots that mixes reinforcement learning with extended Kalman filter (EKF) as a localization technique, namely EKF‐DQN, aiming to accelerate the maximization of the learning curve and improve the reward values obtained in the learning process. More specifically, Deep‐Q‐Networks (DQN) are used to control the trajectory of an autonomous robot in an environment with many obstacles. To improve navigation capability in this environment, we also propose a fusion of visual and nonvisual sensors. Due to the ability of EKF to predict states, this algorithm is used as a learning accelerator for the DQN network, predicting future states and inserting this information into the memory replay. Aiming to increase the safety of the navigation process, a visual safety system is also proposed to avoid collisions between the mobile robot and people circulating in the environment. The efficiency of the proposed control system is verified through computational simulations using the CoppeliaSIM simulator with code insertion in Python. The simulation results show that the EKF‐DQN algorithm accelerates the maximization of rewards obtained and provides a higher success rate in fulfilling the mission assigned to the robot when compared to other value‐based and policy‐based algorithms. A demo video of the navigation system can be seen at: https://bit.ly/3reEZrU.

Exploring unknown environments: motivated developmental learning for autonomous navigation of mobile robots

Exploring unknown environments: motivated developmental learning for autonomous navigation of mobile robots

Development of rapidly exploring random tree based autonomous mobile robot navigation and velocity predictions using K-nearest neighbors with fuzzy logic analysis

Development of rapidly exploring random tree based autonomous mobile robot navigation and velocity predictions using K-nearest neighbors with fuzzy logic analysis

Fuzzy inference system-assisted human-aware navigation framework based on enhanced potential field

With the advent of Autonomous Mobile Robots (AMRs) in public areas such as malls and airports, their harmonious coexistence with humans is crucial. AMRs must operate in a manner that ensures human safety, comfort, and acceptability to reduce stress. This is called Human Aware Navigation. This study introduces a framework for AMR navigation that prioritizes safety and human comfort in such environments, utilizing an enhanced Potential Field approach augmented by Fuzzy Inference Systems. To achieve a smooth AMR trajectory, the framework employs these systems based on AMR, human, and obstacle information. The proposed approach is tested across various scenarios, including complex, cluttered environments that mimic practical situations. Simulation results demonstrate that AMRs using the proposed method navigate human-rich environments safely and comfortably while mitigating common issues associated with Potential Field-based approaches, such as local minima and obstacles near the goal.

Autonomous navigation of mobile robots in unknown environments using off-policy reinforcement learning with curriculum learning

Reinforcement learning (RL) is effective for autonomous navigation tasks without prior knowledge of the environment. However, traditional mobile robot navigation algorithms, based on off-policy RL, often face challenges such as low sample efficiency during training and lack of adequate safety mechanisms. In this paper, we present an off-policy RL navigation model named Soft Actor-Critic with Curriculum Prioritization and Fuzzy Logic (SCF). The model uses energy as a prioritized evaluation metric for experience replay. And through task-level curriculum, the agent’s learning sequence is formulated, thereby enhancing sampling efficiency and safety. We propose a Curriculum-based Energy Prioritization (CEP) approach. It selects a replay trajectory that matches the current agent’s capability based on trajectory energy. Our results show that robots using off-policy RL often have limitations in dynamic obstacle avoidance. To rectify this, our model uses a fuzzy logic controller to enhance real-time obstacle avoidance. The SCF approach enables mobile robots to navigate adeptly in unpredictable and dynamic environments, ensuring optimal planning control while being safe and robust. Experiments in Gazebo simulation environment and real world confirm the effectiveness of our proposed method. The comparison results show the superior performance of this method, especially in unknown and dynamic environments.

A comparative study of pose estimation algorithms for visual navigation in autonomous robots

Autonomous navigation is a research field that gives mobile robots the capacity to perform various tasks without human assistance. Autonomous navigation based on visual sensors can be used in GPS denied environments. Vision-based navigation performs feature detection, matching, and pose estimation using camera images. This paper presents a new approach to autonomous navigation for mobile robots using a Color-based Image Segmentation and Centroid Detection algorithm instead of traditional feature detection algorithms. The algorithm intentionally matches features across images using known feature points and uses conventional techniques such as Epipolar Geometry and Perspective-N-Points algorithms for camera pose estimation. The study includes camera calibration to estimate intrinsic parameters and their physical unit conversion ensuring accurate measurements. Experimental datasets are used to analyze the performance of the Epipolar geometry, Perspective-3-Point, and Efficient-Perspective-n-Point based pose estimation algorithms. To enhance accuracy, the P3P algorithm is modified to consider combinations of image points for pose estimation. The paper concludes with a comparative analysis between the original P3P algorithm and the modified version, providing valuable insights into their respective performances. Overall, the paper aims to provide a comparison of different pose estimation algorithms used for visual navigation. Keywords: pose estimation, visual navigation, camera calibration, perspective-n-point, efficient pnp, epipolar geometry Abbreviations PnP, perspective-n-point; P3P, perspective-3-point; EPnP, efficient perspective-n-point; SLAM, simultaneous localization and mapping; VO, visual odometry; GPS, global positioning system; SURF, speeded up robust features; SIFT, scale-invariant feature transform; EKF, extended kalman filter; RGB, red-green-blue; HSV, hue- saturation-value; CNN, convolutional neural network

A Comparison Study between Traditional and Deep-Reinforcement-Learning-Based Algorithms for Indoor Autonomous Navigation in Dynamic Scenarios.

At the beginning of a project or research that involves the issue of autonomous navigation of mobile robots, a decision must be made about working with traditional control algorithms or algorithms based on artificial intelligence. This decision is not usually easy, as the computational capacity of the robot, the availability of information through its sensory systems and the characteristics of the environment must be taken into consideration. For this reason, this work focuses on a review of different autonomous-navigation algorithms applied to mobile robots, from which the most suitable ones have been identified for the cases in which the robot must navigate in dynamic environments. Based on the identified algorithms, a comparison of these traditional and DRL-based algorithms was made, using a robotic platform to evaluate their performance, identify their advantages and disadvantages and provide a recommendation for their use, according to the development requirements of the robot. The algorithms selected were DWA, TEB, CADRL and SAC, and the results show that-according to the application and the robot's characteristics-it is recommended to use each of them, based on different conditions.

Autonomous Robotic Navigation Approach Using Deep Q-Network Late Fusion and People Detection-Based Collision Avoidance

In this work, we propose an approach for the autonomous navigation of mobile robots using fusion the of sensor data by a Double Deep Q-Network with collision avoidance by detecting moving people via computer vision techniques. We evaluate two data fusion methods for the proposed autonomous navigation approach: Interactive and Late Fusion strategy. Both are used to integrate mobile robot sensors through the following sensors: GPS, IMU, and an RGB-D camera. The proposed collision avoidance module is implemented along with the sensor fusion architecture in order to prevent the autonomous mobile robot from colliding with moving people. The simulation results indicate a significant impact on the success of completing the proposed mission by the mobile robot with the fusion of sensors, indicating a performance increase (success rate) of ≈27% in relation to navigation without sensor fusion. With the addition of moving people in the environment, deploying the people detection and collision avoidance security module has improved about the success rate by 14% when compared to that of the autonomous navigation approach without the security module.

Recent progress in trajectory smoothing of mobile robots

The mobile robotics is growing rapidly and is being used in a variety of applications, greatly improving productivity. Path planning is a central part of autonomous navigation for mobile robots. However, most paths generated by these methods have sharp turns, which are unacceptable for mobile robots due to their own motion limitations or specific task scenarios. Smoother paths are therefore needed to improve the reliability and efficiency of the robot while adapting to realistic scenarios. This paper summarizes the classical curve smoothing algorithms and relate new research results presented in recent years, and divide them into six categories: polynomial interpolation, non-uniformly interpolated curves, Beizer curves, B spline curves, Reeds-Shepp curves, geometric algorithms, etc. Besides, a brief mathematical description is given along with the algorithms, and some of the algorithms are also given as examples for reference. In the future, better curve smoothing algorithms to satisfy the kinematic laws of different vehicles (robots/ships/UAVs) in multiple environments and non-simple interpolation smoothing algorithms will be proposed.

STP4: spatio temporal path planning based on pedestrian trajectory prediction in dense crowds

This article proposes a means of autonomous mobile robot navigation in dense crowds based on predicting pedestrians’ future trajectories. The method includes a pedestrian trajectory prediction for a running mobile robot and spatiotemporal path planning for when the path crosses with pedestrians. The predicted trajectories are converted into a time series of cost maps, and the robot achieves smooth navigation without dodging to the right or left in crowds; the path planner does not require a long-term prediction. The results of an evaluation implementing this method in a real robot in a science museum show that the trajectory prediction works. Moreover, the proposed planning’s arrival times is 26.4% faster than conventional 2D path planning’s arrival time in a simulation of navigation in a crowd of 50 people.

A safe reinforcement learning approach for autonomous navigation of mobile robots in dynamic environments

AbstractWhen deploying mobile robots in real‐world scenarios, such as airports, train stations, hospitals, and schools, collisions with pedestrians are intolerable and catastrophic. Motion safety becomes one of the most fundamental requirements for mobile robots. However, until now, efficient and safe robot navigation in such dynamic environments is still an open problem. The critical reason is that the inconsistency between navigation efficiency and motion safety is greatly intensified by the high dynamics and uncertainties of pedestrians. To face the challenge, this paper proposes a safe deep reinforcement learning algorithm named Conflict‐Averse Safe Reinforcement Learning (CASRL) for autonomous robot navigation in dynamic environments. Specifically, it first separates the collision avoidance sub‐task from the overall navigation task and maintains a safety critic to evaluate the safety/risk of actions. Later, it constructs two task‐specific but model‐agnostic policy gradients for goal‐reaching and collision avoidance sub‐tasks to eliminate their mutual interference. Then, it further performs a conflict‐averse gradient manipulation to address the inconsistency between two sub‐tasks. Finally, extensive experiments are performed to evaluate the superiority of CASRL. Simulation results show an average 8.2% performance improvement over the vanilla baseline in eight groups of dynamic environments, which is further extended to 13.4% in the most challenging group. Besides, forty real‐world experiments fully illustrated that the CASRL could be successfully deployed on a real robot.

Map Construction and Path Planning Method for Mobile Robots Based on Collision Probability Model

A map construction method based on a collision probability model and an improved A* algorithm is proposed to address the issues of insufficient security in mobile robot map construction and path planning in complex environments. The method is based on modeling the asymmetry of paths, which complicates problem solving. Firstly, this article constructs a collision probability function model, and based on this model it is fused with the obstacle grid map, which is based on the grid method, to draw a collision probability grid map (CPGM) containing collision probability information. Secondly, incorporating the collision probability values from the CPGM into the actual cost function of the traditional A* algorithm improves the security of path planning in complex environments. The experimental results show that the improved A* algorithm decreases the percentage of dangerous nodes in complex environments by 69.23%, shortens the path planning length by 19.52%, reduces the search time by 16.8%, and reduces the number of turns by 46.67%. Therefore, the method in this paper solves the problem of traditional grid maps lacking security information and can plan a path with higher security and which is smoother, improving the security and robustness of mobile robot autonomous navigation in complex environments.

Enhancing navigation of autonomous mobile robots through permanent federated learning

Deep learning is crucial for advancing artificial intelligence in robotics. Deep learning enables autonomous robots to perceive and control their environment. Federated learning allows distributed robots to train models while maintaining privacy. This paper focuses on using FL to overcome vision-based obstacles in mobile robotic navigation. We evaluate FL’s performance in both simulated and real-world environments. Our research compares FL’s multiple image classifiers to cloud-based central learning using existing data. We also implement a continuous learning system on mobile bots with autonomous data generation. Training models in simulation and reality improves accuracy and enables continuous model updates. Глубокое обучение имеет решающее значение для развития искусственного интеллекта в робототехнике. Оно позволяет автономным роботам воспринимать окружающую среду и управлять ею. Федеративное обучение позволяет распределенным роботам обучать модели, сохраняя конфиденциальность. В этой статье основное внимание уделяется использованию федеративного обучения для преодоления препятствий, основанных на зрении, в мобильной роботизированной навигации. Производительность такого обучения оценивается как в смоделированных, так и в реальных условиях. Настоящее исследование сравнивает несколько классификаторов изображений с облачным централизованным обучением с использованием существующих данных. Также система непрерывного обучения внедряется на мобильных ботах с автономным формированием данных. Обучение моделей в симуляции и реальности повышает точность и обеспечивает непрерывное обновление модели.

Autonomous navigation and indoor mapping for a service robot

Abstract
 Objective: Simultaneous Localization and Mapping (SLAM) is a quite common and interesting problem in mobile robotics. It is the basis of safe autonomous navigation of mobile robots and the entrance to new combined applications with a manipulator for instance. Method: In order to find a solution to the SLAM problem, the ROS middleware and the MRPT were selected. Autonomous navigation was tested using two methods, the MRPT navigation ROS package, which is a reactive navigation method based on Trajectory Parameter Space (TP-Space) transformations, and the ROS navigation stack, a standard for differential drive and holonomic wheeled robots. Results: To validate the advantages and disadvantages of both approaches, a mobile robot with strong kinematic constraints (Ackermann-steering-type) known as Summit was used. As an additional work, an application using the mobile robot Summit and a robotic manipulator (Powerball) was carried out, with the intention of picking and placing objects of the mobile robot, a widely spread application among service robotics, especially, in the area of industrial logistics. Conclusions: Finally, it is concluded that with the tests carried out with the robot, it was possible to demonstrate autonomous navigation, using the two mentioned methods.

Robot navigation based on multi-sensor fusion

Over the past years, with the rapid development of artificial intelligence technology, mobile robots have appeared in more and more fields. Different types of robots play different roles. Due to the complex indoor environment, there are still many technical problems to be solved for robots. Aiming at the shortcomings of single sensor, low mileage accuracy, and poor security in traditional mobile robot navigation systems, we propose a mobile robot autonomous navigation system based on multi-sensor perception. This system integrates mainstream synchronous positioning and mapping algorithms, path planning algorithms, and positioning algorithms, and uses multiple sensors to compensate each other to achieve the automatic navigation function of the mobile robot. In order to verify the reliability of the method, We have carried out multi angle analysis in the simulation environment, and the experimental results show that the proposed navigation system is reliable in operation, high in mileage accuracy, and good in robustness, which broadens the application scenario of mobile robots.

Deep reinforcement learning-aided autonomous navigation with landmark generators.

Mobile robots are playing an increasingly significant role in social life and industrial production, such as searching and rescuing robots, autonomous exploration of sweeping robots, and so on. Improving the accuracy of autonomous navigation of mobile robots is a hot issue to be solved. However, traditional navigation methods are unable to realize crash-free navigation in an environment with dynamic obstacles, more and more scholars are gradually using autonomous navigation based on deep reinforcement learning (DRL) to replace overly conservative traditional methods. But on the other hand, DRL's training time is too long, and the lack of long-term memory easily leads the robot to a dead end, which makes its application in the actual scene more difficult. To shorten training time and prevent mobile robots from getting stuck and spinning around, we design a new robot autonomous navigation framework which combines the traditional global planning and the local planning based on DRL. Therefore, the entire navigation process can be transformed into first using traditional navigation algorithms to find the global path, then searching for several high-value landmarks on the global path, and then using the DRL algorithm to move the mobile robot toward the designated landmarks to complete the final navigation, which makes the robot training difficulty greatly reduced. Furthermore, in order to improve the lack of long-term memory in deep reinforcement learning, we design a feature extraction network containing memory modules to preserve the long-term dependence of input features. Through comparing our methods with traditional navigation methods and reinforcement learning based on end-to-end depth navigation methods, it shows that while the number of dynamic obstacles is large and obstacles are rapidly moving, our proposed method is, on average, 20% better than the second ranked method in navigation efficiency (navigation time and navigation paths' length), 34% better than the second ranked method in safety (collision times), 26.6% higher than the second ranked method in success rate, and shows strong robustness.