Optimasi Perencanaan Lintasan Berbasis Multi Objective Genetic Algorithm Untuk Mobile Robot Di Lingkungan Dinamis

Rijalul, Haq (2023) Optimasi Perencanaan Lintasan Berbasis Multi Objective Genetic Algorithm Untuk Mobile Robot Di Lingkungan Dinamis. Masters thesis, Institut Teknologi Sepuluh November.

Text 07111850040001-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (4MB) | Request a copy

Abstract

Perkembangan teknologi dan semakin maraknya perdagangan online menimbulkan semakin banyaknya pengiriman barang dan semakin luas daerah pengiriman. Sebelum barang dikirim ke alamat pemesan barang terkumpul pada suatu wherehouse. Dalam wherehouse tersebut terjadi proses pemilahan barang. Untuk mencegah tertukarnya barang dan untuk mengoptimalkan proses pemilahan, dikemudian hari diperlukan suatu proses otomatisasi dalam pemilahan barang. Proses otomatisasi ini dapat dilakukan dengan menggunakan robot. Proses pemilahan barang dilakukan dengan robot dengan mengantarkan barang dari posisi awal menuju posisi tertentu. Untuk melaksanakan tugas tersebut robot memerlukan perencanaan lintasan gerak robot. Perencanaan lintasan dalam penelitian ini menggunakan metode multi objective genetic algorithm dengan dua objective yaitu mencari jarak terpendek dan meminimalisir jumlah tikungan. Pemilihan metode multi objective genetic algorithm dikarenakan genetic algorithm merupakan metode yang dapat beradaptasi bergantung kondisi. Selain itu genetic algorithm membutuhkan memori dan waktu komputasi yang lebih kecil dibanding metode lain.
Pengujian dalam penelitian ini dilakukan dalam dua tahap yaitu pengujian algoritma dan penerapan algoritma pada mobile robot. Pengujian algoritma diakukan dengan tiga fungsi fitness (F_1,F_2 dan F_3) dan empat population size yang berbeda yaitu 10, 20, 30 dan 40 untuk melihat respon solusi perencanaan lintasan dan waktu komputasi yang diperlukan. Tahap selanjuatnya adalah pengujian algoritma pada mobile robot yang disimulasikan menggunakan software CoppeliaSim Edu dengan lingkungan yang berisi obstacle yang letaknya dapat berubah – ubah (dinamis) untuk melihat rute yang diambil mobile robot dan waktu tempuh mobile robot pembawa barang.
Hasil pengujian menunjukan algoritma dengan fungsi fitness 3(F_3) memiliki hasil yang paling optimal dan konsisten dengan waktu komputasi 202 ms dan tingkat keberhasilan 20/20 untuk population size 40. Algoritma juga berhasil dterapkan pada mobile robot dan dapat menghindari obstacle.
===================================================================================================================================
Technological developments and the increasingly widespread online trade have resulted in more and more shipments of goods and wider shipping areas. Before the goods are sent to the customer's address, the goods are collected at a wherehouse. In the wherehouse there is a process of sorting goods. In order to prevent the exchange of goods and to optimize the sorting process, in the future an automation process is needed in sorting goods. This automation process can be done using a robot. The process of sorting goods is carried out by robots by delivering goods from the initial position to a certain position. To carry out these tasks the robot requires planning the trajectory of the robot's motion. Path planning in this study uses a multi-objective genetic algorithm method with two objectives, namely finding the shortest distance and minimizing the number of turns. The choice of the multi-objective genetic algorithm method is because a genetic algorithm is a method that can adapt depending on conditions. In addition, genetic algorithms require less memory and computation time than other methods.
Testing in this study was carried out in two stages, namely testing the algorithm and applying the algorithm to the mobile robot. Algorithm testing is carried out with three fitness functions (F_1,F_2 dan F_3) and four different population sizes, namely 10, 20, 30 and 40 to see the response of path planning solutions and the required computation time. The next stage is testing the algorithm on the mobile robot which is simulated using the CoppeliaSim Edu software with an environment that contains obstacles whose locations can change (dynamic) to see the route taken by the mobile robot and the travel time of the mobile robot carrying goods.
The test results show that the algorithm with the fitness function 3(F_3) has the most optimal and consistent results with a computation time of 202 ms and a success rate of 20/20 for a population size of 40. The algorithm is also successfully applied to mobile robots and can avoid obstacles.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Perencanaan Lintasan, Multi Objective Genetic algorithm, Mobile Robot, Lingkungan Dinamis, Path Planning, Multi Objective Genetic Algorithm, Mobile Robots, Dynamic Environment
Subjects:	Q Science > QA Mathematics > QA402.5 Genetic algorithms. T Technology > TJ Mechanical engineering and machinery > TJ211.415 Mobile robots
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Rijalul Haq
Date Deposited:	20 Jul 2023 03:26
Last Modified:	20 Jul 2023 03:26
URI:	http://repository.its.ac.id/id/eprint/98650

Actions (login required)

View Item