DESIGN AND IMPLEMENTATION OF THE AUTONOMOUS TRAM DECISION-MAKING SYSTEM

• Main Author: Salsabila Suhaimi, Khansa
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/70819
• Institution: Institut Teknologi Bandung

Trams and trains differ in their operational environment. Trams can pass through open environments where there are road users. This causes tram drivers to always pay attention to the situation around them and predict possible situations that may arise. However, high workloads and fatigue can reduce tram drivers’ situational awareness of the situation around them. Therefore, the task of always predicting the behavior of road users and planning actions to be taken are the most common problems for tram drivers. This study aims to build a decision-making system based on an evaluation of the risk assessment of object collisions with trams, which can be implemented and tested on embedded systems. The stages of decision-making include assessing the risk of collision using the risk assessment method, making decisions using the finite state machine (FSM) method, and planning speed using the predictive control model (MPC) and fuzzy logic methods. The decision-making system is designed so that the tram can run from one station to another and overcome situations where there are other trams or road users on the rails or where road users cross the rails. The actions taken include adaptive cruise control, collision avoidance, and emergency braking. The designed decision-making system was simulated using the Carla simulator and implemented in NVIDIA Drive AGX Pegasus. The simulation results show that the tram can run autonomously with a safety percentage of 80%; that is, the system can predict the trajectory of at least 11 objects and avoid collisions in busy intersection situations. The decision-making of adaptive cruise control, collision avoidance, and emergency braking can maintain tram safety with percentages of 96.94%, 100%, and 100%, respectively. The system was tested on trams on the PT INKA rail in limited situations, namely, safe situations and situations where pedestrians were on the rails. The result is that the tram is able to run autonomously; that is, it can run according to the maximum speed of the track and avoid collisions.