SPEED CONTROL DESIGN OF THE AUTOMATED PEOPLE MOVER SYSTEM TRAIN USING PID PARTICLE SWARM OPTIMIZATION PARALLELED WITH FUZZY LOGIC SUGENO FOR SPEED AND POSITION TARGET TO NUMBER OF PASSENGERS VARIATIONS

• Main Author: Dwi Yulianto, Adi
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/52584
• Institution: Institut Teknologi Bandung

Technology development has made an impact to the transportation system and Automated People Movers (APMS) at Soekarno-Hatta Airport is one of it. APMS is a transportation system designed to operated automatically. However, at the moment, the operation is still carried out manually because the speed controller on Automatic Train Operation (ATO) subsystem in signal of Communication Based Train Control (CBTC) does not work optimally which results in low performances in the time schedule and the final position of the train. Because of that, this research's goal is to make a speed control design of the APMS that controsl the operation of the train optimally and automatically. The research begins with determination the dynamic system model of the train and the speed profile of the train. After that, the optimal speed control is designed to follow the reference of speed control using Proportional Integral Derivative Particle Swarm Optimization (PID PSO) with minimun errors as the objective function. Next, the fuzzy logic Sugeno is designed parallel with the PID PSO to compensate the error by the change of the number of passengers, so the train controller is able to overcome the uncertainty of the number of passengers. In this research, there were three parts of control simulation, PID auto tuning as a comparison, PID PSO, and PID PSO paralleled with Fuzzy Logic Sugeno. Simulation was carried out by five variations of the train mass based on the number of passengers to compare the performance of each controller in handling the variation of condition caused by the uncertainty of number of passengers. The simulation result showed that the speed controller design can follow the train profile determined by means of speed Root Mean Square Error (RMSE) is 7,940x10- 4 for the PID PSO and 7,620x10-4 for the PID PSO with Fuzzy Logic Sugeno and also standard deviation of speed RMSE is 8,182x10-5 for the PID PSO and 4,228x10-5 for the PID PSO with Fuzzy Logic Sugeno.