Abstract: We propose an online path-planning algorithm using a variation of Monte Carlo tree search (MCTS) for navigating intelligently unmanned aerial vehicles (UAVs) to track mobile ground targets in complex environments, called the P-UAV algorithm. The proposed algorithm employs a non-myopic method applied to a partially observable Markov decision process (POMDP) model, accounting for long-term decision making. Our algorithm integrates a heuristic technique to efficiently generate paths.
The algorithm yields to parallel processing methods to significantly enhance its computational performance, making it suitable for real-time implementation. Simulation experiments show that our path-planning algorithm is efficient and achieves good exploration-exploitation tradeoff in finding a near-optimal solution despite the very large search space.

Adviser: Prof. Edwin K.P. Chong
Co-Adviser: -
Non-ECE Member: Prof. Olivier Pinaud
Member 3: Prof. Mahmood Azimi-Sadjadi
Addional Members: Prof. Ali Pezeshki

Publications:
1. A. Vasutapituks and E. K. P. Chong, ”Design of Autonomous UAV Guidance
System Using Monte Carlo Tree Search,” 2022 7th International Conference
on Business and Industrial Research (ICBIR), Bangkok, Thailand, 2022, pp.
677-682.

2. A. Vasutapituks and E. K. P. Chong, ” An Autonomous UAV Path-Planning Algorithm for Mobile Target
Tracking in Complex Environments,” The 27th Annual Meeting in Mathematics 2023 (AMM 2023) and
International Conference in Number Theory and Applications 2023 (ICNA 2023), Bangkok, Thailand, 2022

Program of Study:
ECE-520 Optimization Methods-Control and Communication
ECE-513 Digital Image Processing
ECE-656 Neural Networks and Adaptive Systems
ECE-514 Applications of Random Processes
ECE-681A1 Algebraic Coding Theory
ECE-516 Information Theory
-
-