Perencanaan Lintasan dan Penghindaran Rintangan pada UAV dengan Metode Improved Grey Wolf Optimizer dan Partially Observable Markov Decision Process

Atalla, Baihaqi Nofal (2025) Perencanaan Lintasan dan Penghindaran Rintangan pada UAV dengan Metode Improved Grey Wolf Optimizer dan Partially Observable Markov Decision Process. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5022211116-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (4MB) | Request a copy

Abstract

Penelitian ini menerapkan metode perencanaan lintasan dan penghindaran rintangan untuk Unmanned Aerial Vehicle (UAV) menggunakan integrasi Improved Grey Wolf Optimizer (IGWO) dengan Partially Observable Markov Decision Process (POMDP). IGWO digunakan untuk menghasilkan lintasan optimal yang menghindari rintangan seperti radar, rudal, artileri dan zona larangan terbang, sedangkan POMDP diimplementasikan untuk penghindaran tabrakan dinamis dengan multiple intruder yang memiliki ketidakpastian arah gerak. Sistem menggunakan particle filter untuk mengestimasi heading intruder dan Monte Carlo Tree Search (MCTS) untuk pemilihan aksi. Pengujian dilakukan dalam simulasi lingkungan 3D dengan 5 intruder yang bergerak pada pola dan kecepatan berbeda. Hasil menunjukkan IGWO mampu menghasilkan lintasan optimal dengan penurunan fungsi biaya dari sekitar 5000 menjadi sekitar 2000 dalam 165 iterasi dan constraint violation bernilai nol, menunjukkan bahwa solusi yang dihasilkan memenuhi semua batasan yang ditentukan. Dengan menggunakan metode POMDP, UAV berhasil mempertahankan jarak aman minimum 0.8 meter dari intruder melalui pengaturan kecepatan adaptif antara 2-10 m/s. Path tracking menghasilkan rata-rata RMSE total sebesar 0.3027 m, yang menunjukkan UAV mampu mengikuti lintasan referensi dengan baik. Integrasi kedua metode ini terbukti efektif dalam menghasilkan navigasi UAV yang efisien pada lingkungan dengan multiple dynamic obstacles.
===============================================================================================================================
This research applies a path planning and collision avoidance methods for Unmanned Aerial Vehicles (UAV) by integrating Improved Grey Wolf Optimizer (IGWO) with Partially Observable Markov Decision Process (POMDP). IGWO is used to generate optimal paths avoiding obstacles such as radars, missiles, artillery, and No-Fly Zones, while POMDP is implemented for dynamic collision avoidance with multiple intruders having uncertain heading directions. The system employs particle filters to estimate intruder headings and Monte Carlo Tree Search (MCTS) for action selection. The experiments were carried out in a 3D simulation environment with 5 intruders moving in different patterns and speeds. Results show that IGWO successfully generates optimal paths with fitness cost reduction from around 5000 to approximately 2000 over 165 iterations and zero value constraint violations , indicating that the solution satisfies all specified constraints. By using the POMDP method, the UAV successfully maintained a minimum safe distance of 0.8 meters from the intruder through adaptive speed control between 2-10 m/s. Path tracking resulted in an average total RMSE of 0.3027 m, demonstrating the UAV's ability to follow reference paths effectively. The integration of these methods proves effective in producing efficient UAV navigation in the environments with multiple dynamic obstacles.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Improved Grey Wolf Optimizer, Partially Observable Markov Decision Process, Perencanaan Lintasan, Penghindaran Rintangan, Unmanned Aerial Vehicles. Improved Grey Wolf Optimizer, Partially Observable Markov Decision Process, Path Planning, Obstacle Avoidance, Unmanned Aerial Vehicles.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Baihaqi Nofal Atalla
Date Deposited:	23 Jan 2025 03:57
Last Modified:	23 Jan 2025 03:57
URI:	http://repository.its.ac.id/id/eprint/116714

Actions (login required)

View Item