DESIGN & IMPLEMENTATION OF YOLOV4-BASED OBJECT RECOGNITION ON THE EMBEDDED SYSTEM AND GRAPHICAL USER INTERFACE IN THE TRAFFIC SIGN RECOGNITION SYSTEM FOR AUTONOMOUS TRAM

• Main Author: Rayhan, Ilham
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/67895
• Institution: Institut Teknologi Bandung

One alternative to electric-powered public transportation that can be used by residents and implemented by the government as a safe and environmentally friendly mode of transportation in the city is the tram. Trams in developed countries have proven to be effective in reducing congestion problems and increasing the effectiveness of mobility within cities. PT. INKA is collaborating with Research.ai and the Bandung Institute of Technology to develop autonomous trams which will later be operated in major cities in Indonesia. In this collaboration, PT. INKA oversees developing the tram that will be used and Research.ai and ITB oversee developing an autonomous system from the tram. One of the obstacles faced in the operation of autonomous trams that do not have drivers in the future is how to operate autonomous trams in accordance with existing regulations and in accordance with good operating conditions. One way to make sure the tram can work properly is to follow the traffic signs, especially for trains. Based on the description above, a traffic sign recognition system based on artificial intelligence is proposed in the form of a Traffic Sign Recognition System. The introduction of this sign will be implemented on an autonomous system of trams using an embedded NVIDIA based system, namely NVIDIA Drive AGX Pegasus. So, it is necessary to process the artificial intelligence model to be able to work well on the embedded system. In addition, the recognition results from the Traffic Sign Recognition system will also be displayed on a part of the entire autonomous tram interface (GUI) that will be used by the tram driver or operator.