Improved Dynamic Window Approach Using Sugeno Fuzzy Logic For Mobile Robot Collision Avoidance

Rachman, Falah Ibnu (2025) Improved Dynamic Window Approach Using Sugeno Fuzzy Logic For Mobile Robot Collision Avoidance. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5022211238-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB)

Abstract

Penelitian ini menyajikan Dynamic Window Approach (DWA) yang telah ditingkatkan dengan menggunakan Sugeno Fuzzy Logic untuk meningkatkan penghindaran tabrakan robot bergerak di lingkungan yang dinamis. Algoritma DWA menggunakan parameter bobot tetap yang membatasi kemampuan beradaptasi mereka terhadap perubahan kondisi lingkungan, terutama dalam skenario dengan rintangan yang bergerak. Untuk mengatasi keterbatasan ini, penelitian ini mengembangkan Fuzzy Dynamic Window Approach (FDWA) yang mengintegrasikan logika fuzzy Sugeno dengan DWA untuk memungkinkan penyesuaian bobot adaptif berdasarkan real-time. Metodologi penelitian ini melibatkan implementasi dan perbandingan tiga algoritma navigasi: DWA standar, Improved DWA, dan FDWA dengan penyesuaian bobot fuzzy yang kontinu. Semua algoritma diuji dalam lingkungan simulasi yang mewakili laboratorium B104 di Institut Teknologi Sepuluh Nopember, yang menampilkan skenario rintangan statis dan rintangan dinamis. Hasil simulasi menunjukkan bahwa FDWA secara signifikan mengungguli DWA standar dan DWA yang telah ditingkatkan dalam lingkungan dinamis. Dalam skenario rintangan dinamis, FDWA mencapai efisiensi navigasi 99,63%, menyelesaikan navigasi dalam 81 detik tanpa tabrakan, dan mempertahankan path following error hanya 0,079 meter. Sebaliknya, DWA standar menunjukkan kinerja yang buruk dengan efisiensi 80,37%, kesalahan jalur 5,173 meter, dan tingkat tabrakan 0,6%. Improved DWA mencapai kinerja menengah dengan efisiensi 98,84% dan kesalahan jalur 0,245 meter.
========================================================================================================================================
This research presents an improved Dynamic Window Approach (DWA) using Sugeno Fuzzy Logic for enhanced mobile robot collision avoidance in dynamic environments. Traditional DWA algorithms utilize fixed weight parameters that limit their adaptability to changing environmental conditions, particularly in scenarios with moving obstacles. To address these limitations, this study develops a Fuzzy Dynamic Window Approach (FDWA) that integrates Sugeno fuzzy logic with DWA to enable adaptive weight adjustment based on real-time environmental feedback. The research methodology involves implementing and comparing three navigation algorithms: standard DWA, improved DWA, and FDWA with continuous fuzzy weight adjustment. All algorithms were tested in simulated environments representing the B104 laboratory at Institut Teknologi Sepuluh Nopember, featuring both static multi-obstacle and dynamic obstacle scenarios. Simulation results demonstrate that FDWA significantly outperforms both standard DWA and improved DWA in dynamic environments. In dynamic obstacle scenarios, FDWA achieved 99.63% navigation efficiency, completed navigation in 81 seconds with zero collisions, and maintained a path following error of only 0.079 meters. In contrast, standard DWA showed poor performance with 80.37% efficiency, 5.173 meters path error, and 0.6% collision rate. The improved DWA achieved intermediate performance with 98.84% efficiency and 0.245 meters path error.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Mobile Robot, Dynamic Window Approach, Penghindaran Rintangan, Sugeno Fuzzy Logic. Mobile Robot, Dynamic Window Approach, Collision Avoidance, Sugeno Fuzzy Logic.
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:	Falah Ibnu Rachman
Date Deposited:	22 Jul 2025 08:39
Last Modified:	22 Jul 2025 08:39
URI:	http://repository.its.ac.id/id/eprint/120589

Actions (login required)

View Item