OBJECT DETECTION WITH CAMERA FOR SEMI-AUTOMATIC BATTERY TRAM STEERING ON TRAFFIC IN INDONESIA

• Main Author: Roberts Tantono, Evan
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/61804
• Institution: Institut Teknologi Bandung

Where at this time the development of Autonomous Vehicles in Indonesia began to develop starting with the presence of automatic cars, Autopilot on planes, and so on. So at this time PT Indonesia Kereta Api (PT INKA) is developing an Autonomous Tram. As is known, the basic component for an Autonomous Tram to work is to have an object detection camera to view the results in 2D and combine it with LiDAR. The dataset used is also a local dataset from PI Inka which has been annotated. So as to produce a 3D bounding box with class and distance to the source with very good accuracy. After knowing the surroundings, it is included in the decision making. After going through the decision making process, predictions for the tram are generated. The results of this Autonomous Tram can produce a tram that can help the driver to drive. The existence of an Autonomous Tram can reduce congestion, accidents due to driver errors and air emissions, because it uses electricity as a power source. In this section the author only performs 1 subsystem, namely Object Detection using YoloV5