Development of dynamic path replanning algorithm for autonomous mobile robots

• Main Author: Huang, Zhiwei
• Other Authors: Low Kin Huat
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Mechanical engineering::Robots; DRNTU::Engineering::Mechanical engineering::Mechatronics
• Online Access: http://hdl.handle.net/10356/78519
• Institution: Nanyang Technological University

For an autonomous mobile robot to navigate within its environment, it must be able to plan a path from its current location to a desired goal location. Path planning algorithms have been a key concept in developing autonomous robot motion. The existing algorithms with node-based or sampling-based search are capable of generating a collision-free path in a static environment given enough time. However, in real-world situations, the environment of the robot usually consists of one or more moving obstacles. These obstacles may at any point in time obstruct the path of the robot. Whenever the robot encounters any unpredictable moving obstacles, it must dynamically determine the next course of action in order to prevent a collision. This report will cover the development of a dynamic path replanning algorithm for mobile robot navigation in a 3-dimensional configuration space with both static and randomly moving obstacles. A study was done on the existing path planning and replanning algorithms to first develop a basic replanning algorithm in 2-dimensional configuration space. After simulations were done to verify and improve on its effectiveness, the technique will be augmented and implemented in 3-dimensional space. The proposed replanning technique incorporates the use of rapidly exploring random tree (RRT) and its variants. Trajectory estimation of the moving obstacle is a core concept of the replanning algorithm. It allows the robot to alter its existing motion path based on the future position and estimated trajectory of the obstacle. Various simulation results show the effectiveness of the proposed method.