Autonomous Navigation of Mobile Robots: Integrating Lidar, RRT*, and Frontier-Based Algorithms for Path Planning, Exploration, and Motion Control

Huy Nguyen1,2       Million Angesom1,2       Mohsin Kabir1,2       Narcı́s Palomeras1      
1University of Girona   2EMJM in Intelligent Field Robotic Systems  
Video Code arXiv

Abstract

Autonomous mobile robots have emerged as a significant area of research and development, with potential applications ranging from warehouse automation to planetary exploration. One critical aspect of autonomous mobile robots is their ability to navigate in complex and dynamic environments while efficiently exploring unknown areas. This project presents a comprehensive approach that combines Lidar sensing, Rapidly-exploring Random Trees*(RRT*), and Frontier-Based algorithms to achieve enhanced path planning, exploration, and motion control for mobile robots. The integration of these techniques addresses the challenges of obstacle avoidance, real-time decision-making, and efficient exploration, leading to improved autonomy and reliability in robotic systems. The experimental evaluation demonstrates the effectiveness and performance of the proposed approach, showcasing its potential for a wide range of applications.

Video

Simulation

Methodology

functional_block_diagram

Initially, lidar data is fed into the grid mapper, generating a grid map representation. Concurrently, localization algorithms continuously provide position transformations in the world frame. The grid map is connected to the path validity checker, ensuring that the planned paths are collision-free. The exploration algorithms, employing a frontier-based strategy, provide the next best view point to explore the environment as a goal point to the planner. The planner, implemented as RRT*, plans a path and verifies the feasibility of the generated path through the path validity checker. Valid paths between the start and goal positions are then passed to the motion controller. The motion controller, utilizing the Dynamic Window Approach (DWA), calculates velocities for the actuators, enabling the robot’s movement. The subsequent sections will illustrate each component and method of the system.



Results

results1
Frontier centroids detected at the start of exploration
results2
Detected frontier centroids during the exploration process

One of the system’s key strengths lies in its integration of lidar sensing and grid mapping. Lidar sensors provide detailed and real-time perception data, allowing for accurate representations of the environment. The grid mapping module effectively converts the lidar data into a grid map, providing a reliable basis for path planning and motion control. Lidar data enhances the system’s perception capabilities and enables the robot to navigate complex and dynamic environments. Employing the RRT* algorithm as the core planner offers several advantages. RRT* efficiently explores the high- dimensional configuration space and generates optimal paths between the start and goal positions. Furthermore, including a path validity checker ensures the feasibility and safety of generated paths. By considering static and dynamic obstacles, the path validity checker verifies the collision-free nature of the path the planner produces. This feature significantly reduces the risk of collisions and improves the system’s overall reliability. The motion control component based on the Dynamic Window Approach (DWA) provides agile and reactive motion control for the mobile robot.



Acknowledgements

This project is part of the IFRoS program course. If you are interested in the project or intend to use it for any purpose, please contact the author first.

We borrow the page template from  Nerfies project page. Special thanks to them!
This website is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.