Memory-efficient and rapidly-converging path planner

• Main Author: Immanuel, Ivan
• Other Authors: Low Kin Huat
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Aeronautical engineering; Engineering::Mechanical engineering
• Online Access: https://hdl.handle.net/10356/136879
• Institution: Nanyang Technological University

UAS (Unmanned Aerial Systems) Traffic Management, or UTM in short, aims to provide services to multiple UAVs (Unmanned Aerial Vehicles) occupying an airspace and manage them to ensure safe and efficient operation. For that purpose, a robust path planning algorithm is critical in determining the route which needs to be traveled by each UAV. The planner employed by UTM must be able to return optimal, or at least near optimal path, to avoid convolutions and reduce energy consumption of each UAV. Furthermore, the planner must be fast enough to handle multiple path generations of multiple UAVs and should require minimal memory space for efficient use. In addition to being used by UTM, the planner can also be used on individual UAV, providing it has the capability to autonomously plan their path, to generate their own path or re-plan quickly should the need arise due to unforeseen on-flight circumstances. In this research, a modified sampling-based path planner has been developed. This planner borrows the idea of optimizing with the use of intelligent sampling from RRT*-Smart and limiting maximum nodes stored from RRT*FN. The algorithm was written and compared with other existing planning algorithms in MATLAB and converted into C++ for ROS simulation. The newly developed algorithm has been proven to be feasible in both 2-dimensional and 3-dimensional environments and furthermore, possess the ability to produce path with high degree of optimality within a reduced execution time and requiring less memory capacity.