Navigation In Dynamic Environments Research Articles

Long-term navigation for autonomous robots based on spatio-temporal map prediction

The robotics community has witnessed a growing demand for long-term navigation of autonomous robots in diverse environments, including factories, homes, offices, and public places. The core challenge in long-term navigation for autonomous robots lies in effectively adapting to varying degrees of dynamism in the environment. In this paper, we propose a long-term navigation method for autonomous robots based on spatio-temporal map prediction. The time series model is introduced to learn the changing patterns of different environmental structures or objects on multiple time scales based on the historical maps and forecast the future maps for long-term navigation. Then, an improved global path planning algorithm is performed based on the time-variant predicted cost maps. During navigation, the current observations are fused with the predicted map through a modified Bayesian filter to reduce the impact of prediction errors, and the updated map is stored for future predictions. We run simulation and conduct several weeks of experiments in multiple scenarios. The results show that our algorithm is effective and robust for long-term navigation in dynamic environments.

Local Path Planner for Mobile Robot Considering Future Positions of Obstacles

Local path planning is a necessary ability for mobile robot navigation, but existing planners are not sufficiently effective at dynamic obstacle avoidance. In this article, an improved timed elastic band (TEB) planner based on the requirements of mobile robot navigation in dynamic environments is proposed. The dynamic obstacle velocities and TEB poses are fully integrated through two-dimensional (2D) lidar and multi-obstacle tracking. First, background point filtering and clustering are performed on the lidar points to obtain obstacle clusters. Then, we calculate the data association matrix of the obstacle clusters of the current and previous frame so that the clusters can be matched. Thirdly, a Kalman filter is adopted to track clusters and obtain the optimal estimates of their velocities. Finally, the TEB poses and obstacle velocities are associated: we predict the obstacle position corresponding to the TEB pose through the detected obstacle velocity and add this constraint to the corresponding TEB pose vertex. Then, a pose sequence considering the future positions of obstacles is obtained through a graph optimization algorithm. Compared with the original TEB, our method reduces the total running time by 22.87%, reduces the running distance by 19.23%, and increases the success rate by 21.05%. Simulations and experiments indicate that the improved TEB enables robots to efficiently avoid dynamic obstacles and reach the goal as quickly as possible.

Shortest Path Planning and Efficient Fuzzy Logic Control of Mobile Robots in Indoor Static and Dynamic Environments

Efficient navigation in dynamic environments is a critical skill for of mobile robots, where obstacles can stochastically appear. This paper presents a complete navigation and control system that integrates effective path optimization and motion control capabilities for mobile robots evolving in indoor static and dynamic environments. This system consists primarily of two layers. In the Optimization Layer (Global planner), a Deterministic Constructive Algorithm (DCA) quickly generates the shortest path, as a sequence of nodes, to get to the goal position while avoiding the static obstacles. The Control layer (Local planner) employs an Efficient Fuzzy Logic Controller (EFLC) to continuously guide the robot around the detected dynamic obstacles and drive safely the robot along the intended path. Simulations conducted on various maps with different complexities demonstrate the effectiveness of the DCA planner. Finally, validations using V-REP software show the strength of the proposed EFLC that mimics human reasoning for mobile robots navigating in dynamic environments.

Enhancing Robotic Navigation in Dynamic Environments

Robotic navigation in dynamic environments represents a significant challenge and opportunity in the field of robotics, with profound implications for applications ranging from autonomous vehicles to service robots in humancentric spaces. This paper addresses the critical need for advanced navigation strategies that enable robots to operate efficiently and safely in environments characterized by continuous change and unpredictability. Traditional navigation methods, which rely heavily on static maps and predefined paths, often fall short in dynamic settings where obstacles and pathways can shift rapidly. Consequently, there is a pressing need for innovative approaches that combine robust sensing, real-time data processing, and adaptive decision-making to enhance robotic navigation capabilities. To tackle these challenges, we propose a comprehensive framework that integrates several cutting-edge technologies and methodologies. At the core of our approach is the use of advanced sensor fusion techniques, which combine data from multiple sensors, including LiDAR, cameras, and ultrasonic sensors, to create a detailed and dynamic representation of the robot's surroundings. This multi-modal sensing approach ensures that the robot can detect and respond to changes in the environment with high accuracy and reliabiality. Complementing the sensor fusion process is the implementation of realtime data processing algorithms, powered by machine learning and artificial intelligence. These algorithms enable the robot to analyze vast amounts of sensory data on-the-fly, recognizing patterns, predicting potential obstacles, and making informed navigation decisions in real-time. Machine learning models are trained using extensive datasets that include a variety of dynamic scenarios, ensuring that the robot can generalize from past experiences and adapt to new situations effectively. A key component of our framework is the development of adaptive path planning algorithms that can dynamically adjust the robot's trajectory in response to environmental changes. These algorithms leverage techniques from both classical robotics and modern AI, including probabilistic roadmaps, rapidly-exploring random trees (RRT), and deep reinforcement learning. By continuously updating the robot's path based on real-time sensory input, these adaptive planning methods ensure that the robot can navigate efficiently and avoid collisions even in highly dynamic environments. In addition to the technical advancements, our framework also emphasizes the importance of human-robot interaction (HRI) in dynamic environments. Effective HRI mechanisms are crucial for scenarios where robots operate alongside humans, such as in healthcare, hospitality, and collaborative manufacturing. We incorporate natural language processing (NLP) and gesture recognition systems to facilitate intuitive communication between humans and robots, allowing for seamless cooperation and enhanced safety. To validate our proposed framework, we conducted extensive experiments in a variety of dynamic settings, including urban environments, industrial warehouses, and public spaces. The results demonstrate significant improvements in navigation efficiency, collision avoidance, and overall operational safety compared to traditional methods. Our framework not only enhances the robot's ability to navigate dynamic

Enhancing Robotic Navigation in Dynamic Environments

Robotic navigation in dynamic environments represents a significant challenge and opportunity in the field of robotics, with profound implications for applications ranging from autonomous vehicles to service robots in humancentric spaces. This paper addresses the critical need for advanced navigation strategies that enable robots to operate efficiently and safely in environments characterized by continuous change and unpredictability. Traditional navigation methods, which rely heavily on static maps and predefined paths, often fall short in dynamic settings where obstacles and pathways can shift rapidly. Consequently, there is a pressing need for innovative approaches that combine robust sensing, real-time data processing, and adaptive decision-making to enhance robotic navigation capabilities. To tackle these challenges, we propose a comprehensive framework that integrates several cutting-edge technologies and methodologies. At the core of our approach is the use of advanced sensor fusion techniques, which combine data from multiple sensors, including LiDAR, cameras, and ultrasonic sensors, to create a detailed and dynamic representation of the robot's surroundings. This multi-modal sensing approach ensures that the robot can detect and respond to changes in the environment with high accuracy and reliabiality. Complementing the sensor fusion process is the implementation of realtime data processing algorithms, powered by machine learning and artificial intelligence. These algorithms enable the robot to analyze vast amounts of sensory data on-the-fly, recognizing patterns, predicting potential obstacles, and making informed navigation decisions in real-time. Machine learning models are trained using extensive datasets that include a variety of dynamic scenarios, ensuring that the robot can generalize from past experiences and adapt to new situations effectively. A key component of our framework is the development of adaptive path planning algorithms that can dynamically adjust the robot's trajectory in response to environmental changes. These algorithms leverage techniques from both classical robotics and modern AI, including probabilistic roadmaps, rapidly-exploring random trees (RRT), and deep reinforcement learning. By continuously updating the robot's path based on real-time sensory input, these adaptive planning methods ensure that the robot can navigate efficiently and avoid collisions even in highly dynamic environments. In addition to the technical advancements, our framework also emphasizes the importance of human-robot interaction (HRI) in dynamic environments. Effective HRI mechanisms are crucial for scenarios where robots operate alongside humans, such as in healthcare, hospitality, and collaborative manufacturing. We incorporate natural language processing (NLP) and gesture recognition systems to facilitate intuitive communication between humans and robots, allowing for seamless cooperation and enhanced safety. To validate our proposed framework, we conducted extensive experiments in a variety of dynamic settings, including urban environments, industrial warehouses, and public spaces. The results demonstrate significant improvements in navigation efficiency, collision avoidance, and overall operational safety compared to traditional m ethods. Our framework not only enhances the robot's ability to navigate dynamic environments but also provides a scalable solution that can be tailored to different applications and operational requirements.

Autonomous Robot Navigation in Dynamic Environments: A Temporal-Difference Learning Approach

Autonomous Robot Navigation in Dynamic Environments: A Temporal-Difference Learning Approach

A LiDAR Mapping System for Robot Navigation in Dynamic Environments

A LiDAR Mapping System for Robot Navigation in Dynamic Environments

3D LiDAR-based obstacle detection and tracking for autonomous navigation in dynamic environments

3D LiDAR-based obstacle detection and tracking for autonomous navigation in dynamic environments

Hierarchical Relational Graph Learning for Autonomous Multirobot Cooperative Navigation in Dynamic Environments

As a specific kind of Cyber-Physical Systems (CPSs), autonomous robot clusters play an important role in various intelligent manufacturing fields. However, due to the increasing design complexity of robot clusters, it is becoming more and more challenging to guarantee the safety and efficiency for multi-robot cooperative navigation in dynamic and complex environments. Although Deep Reinforcement Learning (DRL) shows great potential in learning multi-robot cooperative navigation policies, existing DRL-based approaches suffer from scalability issues and rarely consider the transferability of trained policies to new tasks. To address these problems, this paper presents a novel DRL-based multi-robot cooperative navigation approach named HRMR-Navi that equips each robot with both a two-layered hierarchical graph network model and an attention-based communication model. In our approach, the hierarchical graph network model can efficiently figure out hierarchical relations among all agents that either cooperate for efficiency or avoid obstacles for safety to derive more advanced strategies, and the communication model can accurately form a global view of the environment for a specific robot, thus the multi-robot cooperation efficiency can be further strengthened. Meanwhile, we propose an improved Proximal Policy Optimization (PPO) algorithm based on the Maximum Entropy reinforcement learning, named MEPPO, to enhance the robot exploration ability. Comprehensive experimental results demonstrate that, compared with state-of-the-art approaches, HRMR-Navi can achieve more efficient cooperative navigation with less time cost, lower collision rate, higher scalability, and better knowledge transferability.

End-to-End Autonomous Navigation Based on Deep Reinforcement Learning with a Survival Penalty Function.

An end-to-end approach to autonomous navigation that is based on deep reinforcement learning (DRL) with a survival penalty function is proposed in this paper. Two actor-critic (AC) frameworks, namely, deep deterministic policy gradient (DDPG) and twin-delayed DDPG (TD3), are employed to enable a nonholonomic wheeled mobile robot (WMR) to perform navigation in dynamic environments containing obstacles and for which no maps are available. A comprehensive reward based on the survival penalty function is introduced; this approach effectively solves the sparse reward problem and enables the WMR to move toward its target. Consecutive episodes are connected to increase the cumulative penalty for scenarios involving obstacles; this method prevents training failure and enables the WMR to plan a collision-free path. Simulations are conducted for four scenarios-movement in an obstacle-free space, in a parking lot, at an intersection without and with a central obstacle, and in a multiple obstacle space-to demonstrate the efficiency and operational safety of our method. For the same navigation environment, compared with the DDPG algorithm, the TD3 algorithm exhibits faster numerical convergence and higher stability in the training phase, as well as a higher task execution success rate in the evaluation phase.

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.

New players in intelligent transportation: Autonomous Segway in a dynamic environment

This paper heralds a mathematical treatment of Segways as autonomous robots for personal transportation and deliveries and courier services in constrained dynamic environments from a bird’s-eye view. New velocity-based stabilizing controllers of an autonomous nonholonomic two-wheeled self-balancing personalized Segway robot are extracted from a total potential developed by employing the Lyapunov-based Control Scheme (LbCS) for navigation in a partially known environment. Velocity controllers’ cost and time effectiveness and efficiency result from the interaction of the three prominent pillars of LbCS: smoothest, shortest, and safest path for motion planning. Furthermore, the autonomous personal transporter has an obstacle avoidance sensor with a limited detection range ideal for fast navigation in dynamic environments with narrow corridors, tracks, and pathways. This also successfully facilitates navigation in a partially known environment where the sensors only receive and avoid static and dynamic obstacles in a limited range. The results are numerically validated, and the efficacy of the new controllers is exemplified via computer simulations, which illustrate the forward, backward, and zero-turn radius maneuvers of the Segway robot. Introducing the particular autonomous personal transporter would contribute to transportation systems of smart cities.

Adaptive Feature Attention Module for Robust Visual–LiDAR Fusion-Based Object Detection in Adverse Weather Conditions

Object detection is one of the vital components used for autonomous navigation in dynamic environments. Camera and lidar sensors have been widely used for efficient object detection by mobile robots. However, they suffer from adverse weather conditions in operating environments such as sun, fog, snow, and extreme illumination changes from day to night. The sensor fusion of camera and lidar data helps to enhance the overall performance of an object detection network. However, the diverse distribution of training data makes the efficient learning of the network a challenging task. To address this challenge, we systematically study the existing visual and lidar features based on object detection methods and propose an adaptive feature attention module (AFAM) for robust multisensory data fusion-based object detection in outdoor dynamic environments. Given the camera and lidar features extracted from the intermediate layers of EfficientNet backbones, the AFAM computes the uncertainty among the two modalities and adaptively refines visual and lidar features via attention along the channel and the spatial axis. The AFAM integrated with the EfficientDet performs the adaptive recalibration and fusion of visual lidar features by filtering noise and extracting discriminative features for an object detection network under specific environmental conditions. We evaluate the AFAM on a benchmark dataset exhibiting weather and light variations. The experimental results demonstrate that the AFAM significantly enhances the overall detection accuracy of an object detection network.

Long-Range Navigation in Complex and Dynamic Environments with Full-Stack S-DOVS

Robotic autonomous navigation in dynamic environments is a complex problem, as traditional planners may fail to take dynamic obstacles and their variables into account. The Strategy-based Dynamic Object Velocity Space (S-DOVS) planner has been proposed as a solution to navigate in such scenarios. However, it has a number of limitations, such as inability to reach a goal in a large known map, avoid convex objects, or handle trap situations. In this article, we present a modified version of the S-DOVS planner that is integrated into a full navigation stack, which includes a localization system, obstacle tracker, and novel waypoint generator. The complete system takes into account robot kinodynamic constraints and is capable of navigating through large scenarios with known map information in the presence of dynamic obstacles. Extensive simulation and ground robot experiments demonstrate the effectiveness of our system even in environments with dynamic obstacles and replanning requirements, and show that our waypoint generator outperforms other approaches in terms of success rate and time to reach the goal when combined with the S-DOVS planner. Overall, our work represents a step forward in the development of robust and reliable autonomous navigation systems for real-world scenarios.

Research on Inertial Navigation Technology of Unmanned Aerial Vehicles with Integrated Reinforcement Learning Algorithm

With the continuous expansion of unmanned aerial vehicle (UAV) applications, traditional inertial navigation technology exhibits significant limitations in complex environments. In this study, we integrate improved reinforcement learning (RL) algorithms to enhance existing unmanned aerial vehicle inertial navigation technology and introduce a modulated mechanism (MM) for adjusting the state of the intelligent agent in an innovative manner [1,2]. Through interaction with the environment, the intelligent machine can learn more effective navigation strategies [3]. The ultimate goal is to provide a foundation for autonomous navigation of unmanned aerial vehicles during flight and improve navigation accuracy and robustness. We first define appropriate state representation and action space, and then design an adjustment mechanism based on the actions selected by the intelligent agent. The adjustment mechanism outputs the next state and reward value of the agent. Additionally, the adjustment mechanism calculates the error between the adjusted state and the unadjusted state. Furthermore, the intelligent agent stores the acquired experience samples containing states and reward values in a buffer and replays the experiences during each iteration to learn the dynamic characteristics of the environment. We name the improved algorithm as the DQM algorithm. Experimental results demonstrate that the intelligent agent using our proposed algorithm effectively reduces the accumulated errors of inertial navigation in dynamic environments. Although our research provides a basis for achieving autonomous navigation of unmanned aerial vehicles, there is still room for significant optimization. Further research can include testing unmanned aerial vehicles in simulated environments, testing unmanned aerial vehicles in realworld environments, optimizing the design of reward functions, improving the algorithm workflow to enhance convergence speed and performance, and enhancing the algorithm's generalization ability. It has been proven that by integrating reinforcement learning algorithms, unmanned aerial vehicles can achieve autonomous navigation, thereby improving navigation accuracy and robustness in dynamic and changing environments [4]. Therefore, this research plays an important role in promoting the development and application of unmanned aerial vehicle technology.

Chase and Track: Toward Safe and Smooth Trajectory Planning for Robotic Navigation in Dynamic Environments

In this article presents a trajectory planning approach toward safe and smooth robot motion in dynamic environments. We develop a hierarchical planning framework with a global planner generating the shortest path between the robot and the navigation target. Specially, a virtual target (VT) is set to run on the global path with a designed velocity. At the local level, the robot chases the VT and tracks the global path when traveling through the dynamic environment. We employ the model predictive control (MPC) framework for the local path generation. In particular, the prediction horizon of the MPC is adaptively changed concerning the distance between the robot and the VT. It implicitly reflects the crowdedness of the environment, which helps reduce the environmental uncertainty. Besides, we develop an event-triggered mechanism that executes the local plan aperiodically to release the computational burden. Based on the local chasing and tracking performance, we develop a global path replanning scheme in response to the untraversable area emerging in the dense environment. The developed framework is validated through extensive experiments in dynamic environments, demonstrating that the robot can reach the target faster and showcase a safer and smoother trajectory in the navigation.

Feature Descriptor based on Differential Evolution for Visual Navigation in Dynamic Environments

This work proposes to use a teach-and-repeat method to navigate in an outdoor environment. Instead of using state-of-the-art feature point descriptors to match the current images to the database, we rely on the Generated Binary Robust Independent Elementary Features (GRIEF), a model robust to changes in illumination and weather. These descriptors are obtained via a training process employing evolutionary methods and binary comparison tests. The objective is to obtain visual feature descriptors designed for a given task. For the teach-and-repeat task, they are trained to optimize the estimate of the robot heading error with respect to a learned path. In this work, we propose to investigate how Differential Evolution (DE) can improve the performance of the GRIEF in terms of the convergence rate during the training phase and of fitness score. The new approach, named GRIEF-DE, is trained with the Michigan data set, made of pictures of an outdoor environment with significant changes in appearance caused by seasonal weather variations and illumination, and compared with the original GRIEF. The proposed model was applied to 4 different data sets and the experimental results suggest that the use of Differential Evolution improves the training performance as well as the estimation of the robot orientation with respect to the path of reference.

Mobile Robot Navigation in Dynamic Environments by Offline Reinforcement Learning using Flow-Based Generative Model

Mobile Robot Navigation in Dynamic Environments by Offline Reinforcement Learning using Flow-Based Generative Model

COLREGS-based collision avoidance algorithm for unmanned surface vehicles using modified artificial potential fields

In this paper, to improve the safety of navigation in dynamic environments, a new collision avoidance method is presented for Unmanned Surface Vehicles (USVs). It is mainly focused on dynamic obstacles and safety requirements. Convention on the International Regulations for Preventing Collisions at Sea (COLREGS) is introduced and the original artificial potential field (APF) algorithm is modified to ensure safe collision avoidance of USVs. The new attractive and repulsive potential field functions are designed and a new process of dynamic collision avoidance is constructed to solve the issue. At last, theoretical simulations and field tests are launched to validate our method. The simulation results show that the proposed method can perform safe navigation well. Through our USV platform, the effectiveness of the improved APF method for collision avoidance is also verified. Meanwhile, it is fast, effective, and deterministic.

Deep learning of monocular depth, optical flow and ego-motion with geometric guidance for UAV navigation in dynamic environments

Computer vision-based depth estimation and visual odometry provide perceptual information useful for robot navigation tasks like obstacle avoidance. However, despite the proliferation of state-of-the-art convolutional neural network (CNN) models for monocular depth, ego-motion and optical flow estimation, a relatively low volume of work has been reported on their practical applications in unmanned aerial vehicle (UAV) navigation. This is due to well-known challenges — embedded hardware constraints, viewpoint variations, scarcity of aerial image datasets, and intricacies of dynamic environments. We address these limitations to facilitate real-world deployment of CNN in UAV navigation. First, we devise efficient confidence weighted adaptive network (Cowan) training framework that iteratively leverages intermediate prediction confidences to enforce cross-task consistency over corresponding image regions. This achieves competitive accuracy with a lightweight CNN capable of real-time execution on resource-constrained embedded systems. Second, we devise a test-time refinement method that adapts the network to dynamic environments while simultaneously improving accuracy. To accomplish this, we first update ego-motion using pose information from on-board inertial measurement unit (IMU). Then, we decompose the UAV’s motion into constituent vectors, and for each axis, we formulate geometric relationships between depth and translation. Based on this information, we triangulate corresponding points acquired through optical flow. Finally, we enforce geometric consistency between the initially updated pose and triangulated depth. Cowan with geometric guided refinement (Cowan-GGR) achieves significant accuracy and robustness. Field tests show the proposed model is capable of accurate depth and object-level motion perception in real-world dynamic environments, thus proving its efficacy in facilitating UAV navigation.