This book presents analysis of various intelligent approaches, including applicable heuristic graph theoretic and bio inspired techniques for achieving optimal point-to-point navigation. Usage of LiDAR and RGB-D sensor data as source input has been preferred, ensuring total workspace coverage and minimization of action performed by the robot in carrying out realistic applications in certain congested environment. Sampling-based approaches which use arbitrary information gain formulation and Learning-based techniques, both are studied extensively. Uniform Sampling-based techniques are found feasible in exploring the state space without any complexity in geometrically modeling the configuration area ensuring embedded intelligence into the mobile robots in finding optimal execution. The navigation over both static as well as dynamic obstacles are analysed and the observations are presented in a comparative manner.
For dynamic environment, it is somewhat comparatively difficult for achieving proper path navigation. VSLAM (Visual Simultaneous Localization And Mapping) uses the data captured by externally perceived sensors for the purpose of self-locating and simultaneous map-building leading to understanding the unknown environment. This thesis also proposes a keen way to detect onroute obstacles using training of model through adversarial neural network along with 3D reconstruction of a concerned surrounding followed by memory tracing of already explored path by the mobile agent for ease in achievement of optimized path from start to desired goal position. In case of GPS-denied indoor environment primarily the robot works based on its first hand sensor data, for example, proximity analysis, distance measure etc. In various scientific works it is observed that indoor robots face not only constraint space challenge but also systematic maneuver, path planning and path finding in case of cluttered environment.
Primary contributions of the work include LiDAR data inference by 2D Hect SLAM, Construction of Fusion SLAM accumulating 2D and 3D depth features and Geometric Optimization of navigation planning algorithms. The thesis concludes with the graphical and numerical analysis of the accuracy achieved using mentioned algorithms and specific benchmarking of the performance of used techniques.
Inhaltsverzeichnis (Table of Contents)
- Significance of Reactive navigation with motion plan for Point-to-Point mobile robots
- Introduction
- Background of the work
- Perspective Localization of objects
- Understanding Point-to-point navigation
- Basic principle of Reactive Navigation
- Map based motion plan
- Need of research in Path planning strategies for mobile robots
- Research questions, objectives and identified research gap
- Research Framework and Hypothesis
- Experimental Platform Design
- Multi Sensor Fusion
- Visualizer Option
- Real-time map based planning
- Perception of indoor environment accompanied with obstacle localization
- Use of path planning approaches
- Thesis Contribution
- LiDAR data inference by 2D Hect SLAM
- Construction of Fusion SLAM accumulating 2D and 3D depth features
- Geometric Optimization of navigation planning algorithms
- Organization of the Thesis
- Literature Review
- Introduction
- Review Method
- Localization and mapping followed by path planning accompanied with Visualization
- SLAM mapping approaches followed by static and dynamic obstacle detection for proper path navigation
- Application of Obstacle Detection Algorithm for executing smooth navigation
- Introduction and application of path planning strategies for robot navigation
- Introduction of Reactive navigation and application
- Optimization in path finding
- Visual Representation of mobile robot trajectory and environment recreation
- Functional analysis of SLAM
- Features of 2D Hect SLAM
- Features of 3D depth SLAM
- Depth Estimation Approach
- Obstacle Estimation and Mapping Approach
- Pose Estimation
- 2D Hect Mapping
- Point cloud mapping
- 3D reconstruction of indoor environment with SLAM Data
- Analysis of mapping algorithms
- On path Object Detection and Recognition for collision free navigation
- General ideas of object Detection and Recognition
- Derivation of obstacles from objects
- Double stage Object Detection (DSOD)
- Single-stage Object Detection (SSOD)
- YOLO v4 in on route obstacle detection
- YOLO v5 in Microbial Object Detection (MOD)
- Fundamentals of Mobile robot Navigation
- Introduction
- Map based path plan
- Graph theoretic Approaches of path plan
- Bio inspired approaches for path navigation
- Geometric Optimization approach for path navigation
- Analysis of path plan Algorithms
- Experimental Methodology for executing path planning strategy
- Experimental setup
- Workstation setup
- Camera calibration
- Stepwise Methodology of experiment set up
- Integration of RPLidar with ROS
- Integration of Realsense with ROS
- Integration of Arduino with mobile robot for navigation
- Data Analysis and validate experiment
- SLAM results
- 2D Hect results
- 3D depth
- Results obtained from sensor fusion accompanied with SSOD
- Reactive navigation results
- Graph Theoretic approach results
- Bio inspired approach results
- Discussions and Conclusion
- Discussion
- Summary of Contribution
- Limitations of the Research and Scope of Future Work
- Suggestions for the Next Researcher
Zielsetzung und Themenschwerpunkte (Objectives and Key Themes)
This thesis aims to develop a robust and efficient path planning strategy for point-to-point navigation of a mobile robot in a GPS-denied indoor environment. The primary goal is to ensure collision-free movement by accurately detecting and avoiding obstacles, both static and dynamic, while achieving an optimized path.
- Sensor Fusion: Combining data from LiDAR and RGB-D sensors for comprehensive obstacle detection and environment perception.
- SLAM-based Mapping: Utilizing both 2D Hect SLAM and 3D Depth SLAM for accurate environment modeling and trajectory tracking.
- Obstacle Detection with Machine Learning: Implementing YOLO v4 for object detection and obstacle recognition to improve accuracy and efficiency.
- Path Planning Algorithms: Analyzing and comparing the performance of various graph theoretic and bio-inspired algorithms, including A*, Dijkstra, RRT, RRT*, Genetic Algorithm, and Particle Swarm Optimization.
- Geometric Optimization: Introducing a novel approach to optimize navigation by analyzing obstacle geometry and prioritizing dynamic obstacles.
Zusammenfassung der Kapitel (Chapter Summaries)
Chapter 1 introduces the research problem, objectives, and the proposed framework. The main contributions of the thesis are also detailed, focusing on LiDAR data inference, fusion SLAM, and geometric optimization techniques. Chapter 2 reviews relevant literature on SLAM, object detection, and path planning strategies, highlighting the significance of sensor fusion and the need for robust methods to handle dynamic obstacles. Chapter 3 explores the visual representation of the robot's trajectory and environment recreation, discussing the functionalities of SLAM algorithms, the characteristics of Hector SLAM and RGB-D SLAM, and the process of 3D reconstruction. Chapter 4 delves into object detection and recognition, examining different machine learning approaches and presenting the application of YOLO v4 for obstacle detection. Chapter 5 analyzes the fundamentals of mobile robot navigation, discussing various graph theoretic and bio-inspired path planning algorithms, including their advantages, disadvantages, and space-time complexities. Chapter 6 outlines the experimental methodology, describing the setup, hardware, software, and camera calibration procedures. Chapter 7 presents the results of the experiments, showcasing the performance of different SLAM techniques, obstacle detection using YOLO v4, and the comparative analysis of path planning algorithms. Chapter 8 concludes with a discussion of the research findings, summarizing the main contributions, highlighting the limitations of the study, and suggesting potential future directions for this research field.
Schlüsselwörter (Keywords)
This thesis explores the intersection of sensor fusion, SLAM, machine learning, and path planning algorithms to achieve robust point-to-point navigation for mobile robots in challenging indoor environments. Key terms and concepts include: point-to-point navigation, adversarial neural network, heuristic algorithms, GPS-denied environment, path planning, VSLAM, sampling-based technique, learning-based technique, geometrical optimization, object detection, obstacle avoidance, and zone prioritization.
- Arbeit zitieren
- Rapti Chaudhuri (Autor:in), 2022, Reactive Navigation with Sensor Fusion and Geometric Optimization for Map-Based Path Planning. A Systematic Point-to-Point Navigation for AGV, München, GRIN Verlag, https://www.grin.com/document/1247164
