Rancang Bangun Dan Analisis Prototype Automated Guided Vehicle (AGV) Pengangkut Toolkit Bengkel Perkakas Berbasis Kontrol Pid

Manual toolkit distribution in machining workshops causes production inefficiencies due to high operator downtime. Conventional navigation robots using infrared sensors often fail due to light reflections from oil spills and metal dust. This study aims to design and analyze a 0.5 kg Automated Guided Vehicle (AGV) prototype with a 1 kg payload capacity. Navigation is developed using two TCS34725 color sensors and a TCA9548A I2C multiplexer connected to an Arduino Mega 2560 to detect RGB and Clear values, rendering it oil-resistant. Movement stability is controlled by a PD algorithm driving an L298N driver and dual 1:48 DC gearbox motors, integrated with NRF24L01 wireless communication. Tuning results using the Ziegler-Nichols method determined optimal parameters at Kp=4.80 and Kd=1.58. This combination provides responsive steering, dampens oscillation, and eliminates integral windup during turns. Navigation tests demonstrated a 100% success rate in executing turns and station braking. At a 1 kg payload, the AGV achieved an average delivery time of 15.21 seconds compared to 11.32 seconds for manual delivery over an average distance of 12.5 meters. This is because increasing the payload mass raises the kinetic friction between the rubber wheels and the workshop floor, thereby reducing the AGV's linear speed at a constant PWM ratio.