Algoritma Pembagian Tugas Terdistribusi Multi Kamikaze UAV Dengan Penghindaran Halangan Dinamis Menggunakan Dynamic Artificial Potential Field

Ramadhan, Muhammad Faiz (2025) Algoritma Pembagian Tugas Terdistribusi Multi Kamikaze UAV Dengan Penghindaran Halangan Dinamis Menggunakan Dynamic Artificial Potential Field. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5022211013-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (12MB) | Request a copy

Abstract

Operasi multi-UAV di ruang tiga dimensi selalu berhadapan dengan trade off antara efisiensi lintasan dan waktu tempuh di satu sisi serta margin keselamatan terhadap tabrakan antarpesawat maupun rintangan di sisi lain. Penelitian ini menawarkan kerangka terpadu yang menggabungkan Dynamic Artificial Potential Field dengan Distributed Assignment Switch sehingga penghindaran rintangan dan pembagian tugas berlangsung serempak tanpa koordinator pusat, lalu dievaluasi terhadap Modified Artificial Potential Field sebagai pembanding. Hasil pengujian menunjukkan DAPF menjaga pemisahan antar agen dan mempertahankan jarak ke rintangan sekitar 2.5 kali. Pada lingkungan dengan rintangan statis total lintasan berkurang sekitar sepuluh persen meskipun waktu tempuh sedikit lebih panjang dibanding MAPF. Sebaliknya pada rintangan dinamis MAPF menyelesaikan misi sekitar 26% lebih cepat dan lintasannya hanya 3.6% lebih pendek namun margin keselamatan turun tajam. Integrasi DAPF dan DAS dengan demikian menghasilkan sistem multi UAV yang konsisten mencapai target sekaligus mempertahankan jarak aman dengan kebutuhan komputasi yang tetap realistis untuk operasi daring.
====================================================================================================================================
Multi UAV operation in three dimensional space faces a persistent trade off between path and time efficiency on one hand and safety margins against inter vehicle and obstacle collisions on the other. This study proposes an integrated framework that combines Dynamic Artificial Potential Field with Distributed Assignment Switch so obstacle avoidance and task allocation proceed simultaneously without a central coordinator, and evaluates it against Modified Artificial Potential Field as a baseline. Experiments show that DAPF maintains inter agent separation and obstacle clearance about two and a half times larger. In static obstacle environments the total path length is reduced by roughly 10% although mission time becomes slightly longer than with MAPF. In dynamic obstacle settings MAPF completes the mission about 26% faster and its paths are only 3.6% shorter but the safety margin drops sharply. The integration of DAPF and DAS therefore yields a multi UAV system that consistently reaches its targets while preserving safe separation with computational demands that remain practical for online operation.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Dynamic Artificial Potential Field, Distributed Assignment Switch, MultiAgent, Collision Avoidance
Subjects:	Q Science > Q Science (General) > Q337.3 Swarm intelligence U Military Science > UG1242 Drone aircraft--Control systems. (unmanned vehicle)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Faiz Ramadhan
Date Deposited:	25 Jul 2025 01:25
Last Modified:	25 Jul 2025 01:26
URI:	http://repository.its.ac.id/id/eprint/121215

Actions (login required)

View Item