Prediction of Traffic Density Using YOLO Object Detection and Implemented in Raspberry Pi 3b + and Intel NCS 2
Rosa Andrie Asmara, Bimo Syahputro, Dodit Supriyanto, Anik Nur Handayani
Abstract
Increasing the number of vehicles as the population increases causes traffic congestion, thus indirectly hampering community activities. One way to unravel traffic jams is to do activities outside of peak hours. This study discusses how to determine vehicle traffic density using the Double Exponential Smoothing forecasting model, which is to estimate in quantity, what will happen in the future based on relevant data in the past. While testing the forecasting model using the MAPE (Mean Absolute Percentage Error) method, it aims to measure how the error rate obtained from the results of forecasting. Vehicle identification as the object causing the traffic jam uses the Convolutional Neural Network (CNN) object detection method, especially the You Only Look Once (YOLO) method. The system design involves a Raspberry Pi 3 device, an Intel NCS 2 device, and a website application. Based on the test results obtained, accurate forecasting accuracy of more than 86%.