Manuver Autonomous Car ITS Di Bundaran Atau U-Turn Menggunakan Deep Reinforcement Learning

Meliaz, Muhammad Roychan (2021) Manuver Autonomous Car ITS Di Bundaran Atau U-Turn Menggunakan Deep Reinforcement Learning. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

Text 07211740000012-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2023. Download (9MB) | Request a copy

Abstract

Autonomus Car atau kendaraan otonom merupakan kendaraan yang memiliki kemampuan untuk berkendara secara mandiri layaknya dikendalikan manusia dengan mengunakan rangkaian kecerdasan buatan. Pada penelitian ini kami mengajukan riset pengembangan kendaraan otonom iCar ITS (Intelligent Car Institut Teknologi Sepuluh Nopember) dengan mengembangkan sistem manuver kendaraan otonom di bundaran atau u-turn dalam lingkungan yang disimulasikan. Dalan lingkungan simulasi, model yang digunakan adalah model kendaraan yang disesuaikan dengan iCar. Pengembangan sistem navigasi dan manuver kendaraan otonom dilakukan menggunakan metode Deep Reinforcement Learning, salah satu cabang dari Machine Learning. Pada penelitian ini, didapatkan hasil model reinforcement learning yang mampu melakukan manuver bundaran simpang empat dan bundaran tanpa simpang dengan nilai rerata deviasi sudut dari jalurnya masing-masing senilai 27.011° dan 30.068°, mampu bermanuver tanpa collision selama rerata 13.3 detik dan 7.9 detik, serta dengan kecepatan rerata 27.0 kmpj dan 28.5 kmpj.
=====================================================================================================
An autonomous car or autonomous vehicle is a vehicle which has the ability to drive independently as if controlled by humans using a series of artificial intelligence. In this study, we propose research on the development of an autonomous vehicle iCar ITS (Intelligent Car Institut Teknologi Sepuluh Nopember) by developing a maneuvering system for autonomous vehicles at roundabouts or u-turns in a simulated environment. In the simulation environment, the model used is a vehicle model adapted to iCar. The development of autonomous vehicle navigation and maneuvering systems is carried out using the Deep Reinforcement Learning method, a branch of Machine Learning. From this research, the results obtained are reinforcement learning models that are able to maneuver at roundabouts with intersections and roundabouts without intersections with a mean deviation value of the angle of the path of 27,011° and 30,068°, respectively, able to maneuver without collision for an average of 13.3 seconds and 7.9 seconds, and with an average speed of 27.0 kmph and 28.5 kmph.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	kendaraan otonom, reinforcement learning, deep learning, simulasi, autonomous vehicle, simulation
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > T Technology (General) > T57.62 Simulation T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL152.8 Vehicles, Remotely piloted. Autonomous vehicles.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	Muhammad Roychan Meliaz
Date Deposited:	01 Sep 2021 04:16
Last Modified:	01 Sep 2021 04:16
URI:	http://repository.its.ac.id/id/eprint/91160

Actions (login required)

View Item