Simulasi Autonomous Vehicle Pada Nvidia Jetbot Terhadap Rintangan Statis Dalam Situasi Berkendara Sehari-Hari

Dwayasutha, Bonaventura Bhama Prayudha (2021) Simulasi Autonomous Vehicle Pada Nvidia Jetbot Terhadap Rintangan Statis Dalam Situasi Berkendara Sehari-Hari. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 02111740000085-Undergraduate_Thesis.pdf] Text
02111740000085-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2023.

Download (2MB) | Request a copy
[thumbnail of 02111740000085-Undergraduate_Thesis.pdf] Text
02111740000085-Undergraduate_Thesis.pdf
Restricted to Repository staff only

Download (2MB) | Request a copy

Abstract

Ketergantungan masyarakat akan kendaraan bermotor berbanding lurus dengan angka kecelakaan lalu lintas di Indonesia. Pengembangan teknologi, khususnya pada otomotif, merupakan jawaban dari permasalahan ini. Salah satunya adalah penanaman kecerdasan buatan kendaraan untuk menggantikan peran manusia sebagai pengemudi atau yang biasa disebut dengan autonomous vehicle (AV). Pada penelitian ini akan dibahas lebih lanjut tentang implementasi AV dalam kondisi berkendara sehari-hari, terutama terhadap rintangan (collision) statis yang biasa ditemui di jalanan.
Penelitian ini memanfaatkan NVIDIA JetBot sebagai media simulasi model kecerdasan buatan yang telah dikembangkan. Simulasi dilakukan dengan memasukkan data awal berupa gambar kondisi sekitar dengan rintangan-rintangan berbagai ukuran berjumlah total 200 gambar. Data ini akan dipelajari oleh NVIDIA JetBot dengan metode deep learning berarsitektur CNN. Tahapan akhir dari simulasi ini adalah menjalankan NVIDIA Jetbot secara autonomous dengan level kecepatan 0.32 pada bidang uji dimana terdapat beberapa rintangan yang tersebar secara acak. Tingkat keberhasilan JetBot dicatat dan dibandingkan dengan tingkat keberhasilan minimum sebesar 58%. Apabila JetBot belum mencapai tingkat keberhasilan minimum, kecepatan dan jumlah data input akan divariasikan hingga diperoleh tingkat keberhasilan minimum pada performa JetBot. Pengujian akan dilakukan pada dua kondisi yaitu kondisi gelap dan kondisi terang.
Terdapat beberapa kendala yang ditemukan setelah melakukan penelitian ini, seperti susunan sistem penggerak JetBot yang kurang baik sehingga tidak mampu mentolerir kondisi permukaan bidang uji, ataupun kondisi daya baterai yang mempengaruhi performa JetBot. Setelah dilakukan beberapa penyesuaian, diperoleh kecepatan maksimum bagi NVIDIA JetBot untuk melakukan fungsi collision avoidance dengan baik adalah 0.857 m/s (level kecepatan 0.5). Jumlah data input minimum bagi NVIDIA JetBot untuk melakukan fungsi collision avoidance dengan baik adalah 110 gambar. Performa NVIDIA JetBot dalam melakukan fungsi collision avoidance dengan baik pada kondisi gelap mengalami penurunan secara signifikan jika dibandingkan dengan performa pada kondisi terang. Tingkat keberhasilan tertinggi yang diperoleh pada kondisi gelap sebesar 83% (200 data input), tingkat keberhasilan terendah sebesar 50% (110 data input).
=====================================================================================================
People's dependence on motor vehicles is directly proportional to the number of traffic accidents in Indonesia. The development of technology, especially in automotive, is the answer to this problem. One of them is the planting of artificial intelligence in vehicles to replace the role of humans as drivers, commonly known as autonomous vehicles (AV). This study will further discuss the implementation of AV in daily driving conditions, especially against static collisions that are commonly encountered on the road.
This research utilizes NVIDIA JetBot as a simulation medium for the artificial intelligence model that has been developed. Simulations are carried out by entering initial data in the form of pictures of ambient conditions with obstacles of various sizes with total amount of 200 images. This data will be studied by NVIDIA JetBot using a deep learning method with the CNN architecture. The final stage of this simulation is to run NVIDIA Jetbot autonomously with a speed level of 0.32 on a test field where there are several obstacles scattered randomly. JetBot's success rate is recorded and compared to a minimum success rate of 58%. If the JetBot has not reached the minimum success rate, the speed and amount of input data will be varied until the minimum success rate for JetBot performance is obtained. The test will be carried out in two conditions (dark conditions and bright conditions).
There were several problems that were found after conducting this research, such as the arrangement of the JetBot drivetrain system which is not good enough so that it could not tolerate the surface conditions of the test field; the battery power that affected JetBot’s performance. After some adjustments, the maximum speed for NVIDIA JetBot to perform the collision avoidance function properly is 0.857 m/s (speed level 0.5). The minimum amount of input data for NVIDIA JetBot to properly perform the collision avoidance function is 110 images. NVIDIA JetBot's performance in dark conditions has decreased significantly compared to performance in bright conditions. The highest success rate obtained in dark conditions was 83% (200 input data), the lowest success rate was 50% (110 input data).

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Traffic Accidents, Autonomous Vehicle, NVIDIA JetBot, Collision Avoidance, Success Rate; Kasus Kecelakaan, Autonomous Vehicle, NVIDIA JetBot, Collision Avoidance, Tingkat Keberhasilan
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
T Technology > TJ Mechanical engineering and machinery > TJ211 Robotics.
T Technology > TJ Mechanical engineering and machinery > TJ211.4 Robot motion
T Technology > TJ Mechanical engineering and machinery > TJ211.415 Mobile robots
T Technology > TJ Mechanical engineering and machinery > TJ223.P76 Programmable controllers
T Technology > TL Motor vehicles. Aeronautics. Astronautics
T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL152.8 Vehicles, Remotely piloted. Autonomous vehicles.
Divisions: Faculty of Industrial Technology and Systems Engineering (INDSYS) > Mechanical Engineering > 21201-(S1) Undergraduate Thesis
Depositing User: Bonaventura Bhama P D
Date Deposited: 30 Aug 2021 02:40
Last Modified: 30 Aug 2021 02:40
URI: http://repository.its.ac.id/id/eprint/90356

Actions (login required)

View Item View Item