Pengaruh Jarak dan Kecerahan terhadap Waktu Respons Alat Simulasi Sistem Keamanan Kendaraan Berbasis Pengenalan Wajah

The rise in motor vehicle theft cases in various regions indicates the weakness of the security systems implemented by most users. Systems such as manual locks and alarms often fail to prevent crime, either because they are easily hacked conventionally or due to user negligence in their operation. In today's technological era, a system is needed that is not only secure, but also intelligent and practical. One promising solution is the implementation of a facial recognition-based security system. This study aims to design and test a vehicle security simulation system using facial recognition technology integrated with Arduino Uno and MATLAB. This system utilizes a laptop camera to capture the user's facial image, then performs a detection and verification process using the FaceNet algorithm. If the face is recognized and verified with data stored in the database, the Arduino will activate the actuator components in the form of a DC motor to simulate starting the engine, and a servo motor to simulate opening the vehicle door. This study uses a quantitative experimental approach to analyze the effect of variations in distance (30, 40, and 50 cm) and lighting brightness levels (10–20, 21–30, and 31–40 lux) on the system's response time. A total of 27 combinations of conditions were tested, and the data obtained were analyzed using Microsoft Excel and ANOVA tests in Minitab software. The results of the analysis showed that the optimal response time was obtained at a distance of 40 cm with a medium level of illumination (21–30 lux). In addition, both distance, brightness, and the interaction between the two factors were shown to have a significant effect on the system's response time (P-Value < 0.05). These findings indicate that the system is quite sensitive to environmental changes, so further testing is highly recommended, especially to measure the actual delay, the detection error rate, and the development of a more robust face detection algorithm so that the system can be used reliably in various lighting conditions and face capture angles in the real world.