Sistem Deteksi Pothole Untuk Kendaraan Beroda Dua Demi Meningkatkan Keamanan Berkendara

Khusnanda, Mohammad Mikail Dwi (2022) Sistem Deteksi Pothole Untuk Kendaraan Beroda Dua Demi Meningkatkan Keamanan Berkendara. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05311840000028-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Kondisi jalan yang rusak sering kali disebabkan karena kondisi cuaca, hujan, dan banjir, hal itu menyebabkan perjalanan pengendara kendaraan bermotor menjadi tidak aman, terlebih kepada pengendara roda dua. Sistem IoT pendeteksi pothole atau jalan berlubang penting agar pengendara dapat mengantisipasi kondisi jalan yang buruk dan meningkatkan keamanan berkendara di jalan raya. Meninjau dari penelitian terdahulu telah menghasilkan sistem deteksi pothole menggunakan kamera, gyro ataupun proximity sensor, penulis membuat sistem deteksi pothole yang dapat mendeteksi kondisi jalan menggunakan mikrokontroller dan tersambung dengan sensor ultrasonik, sensor GPS, dan sensor proximity IR-ToF. Sensor memberikan warning tentang adannya obstacle dengan led dan buzzer yang menyala. Data diambil sebanyak 100 data jalan berlubang, 100 jalan pecah, dan 400 jalan normal yang dibersihkan dengan menghapus data tidak normal sebagai tahap preprocessing, lalu data diolah menggunakan algoritma machine learning SVM, LR, KRR, KNN, dan naive bayes. Tingkat akurasi tertinggi dari machine learning yang telah di uji adalah 99.08% dengan menggunakan algoritma SVM. Hasil ditampilkan pada middleware yaitu thingspeak.
====================================================================================================================================
Damaged road conditions are often caused by weather conditions, rain, and floods, which make the journey of motorized vehicle drivers unsafe, especially for two-wheelers. An IoT system that detects potholes or potholes is important so that motorists can anticipate bad road conditions and improve driving safety on the highway. Reviewing from previous research that has produced a pothole detection system using a camera, gyro or proximity sensor, the author makes a pothole detection system that can detect road conditions using a microcontroller and is connected to ultrasonic sensors, GPS sensors, and IR-ToF proximity sensors. The sensor gives a warning about the presence of an obstacle with a lit LED and buzzer. The data was taken as many as 100 data of potholes, 100 broken roads, and 400 normal roads which were cleaned by deleting abnormal data as a preprocessing stage, then the data was processed using machine learning algorithms SVM, LR, KRR, KNN, and naive bayes. The highest level of accuracy from machine learning that has been tested is 99.08% using the SVM algorithm. Testing of the tool has succeeded in detecting potholes, but the tool has shortcomings, namely, the tool detects the incline road as a pothole, this is due to the absence of reflections from objects in front of the proximity sensor. The data from the gps is displayed on the middleware i.e. thingspeak.

Item Type:	Thesis (Other)
Additional Information:	RSIf 004.678 Khu s-1 2022
Uncontrolled Keywords:	Pothole, SVM, LR, KNN, GPS, arduino, Pothole, SVM, LR, KNN, GPS, arduino
Subjects:	T Technology > T Technology (General) > T55.3.H3 Hazardous substances--Safety measures.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	22 Apr 2026 06:59
Last Modified:	22 Apr 2026 08:27
URI:	http://repository.its.ac.id/id/eprint/132868

Actions (login required)

View Item