MEASURING VEHICLE SAFE DISTANCE USING DEEP LEARNING METHODS ON SMARTPHONE DEVICES

Hidayat, Rachmat (Pusat Riset Kecerdasan Artifisial dan Keaman Siber, Badan Riset dan Inovasi Nasional)
Insani, Asep (Pusat Riset Teknologi Pengujian dan Standar, Badan Riset dan Inovasi Nasional)
Khusni, Uus (Pusat Riset Kecerdasan Artifisial dan Keaman Siber, Badan Riset dan Inovasi Nasional)
Hidayat, Asep Rahmat (Pusat Riset Teknologi Pengujian dan Standar, Badan Riset dan Inovasi Nasional)

Jurnal Instrumentasi
Vol 48, No 2 (2024)

Maintaining a safe driving distance is an essential part of road safety. Innovations in technology, such as cameras and sensors, assist in calculating vehicle distances but come with limitations like inaccurate results and high computational demands. This research aims to develop a deep learning algorithm that helps drivers maintain a safe distance and enhance safety. Our study found that the Inception V3+DBND method allows for efficient real-time vehicle distance estimation using a smartphone camera, offering an optimal distance prediction accuracy of 97.20%. The research findings provide a practical and effective solution for measuring safe vehicle distances.