Abstract
Minimizing human intervention in engines, such as traffic lights, through automatic applications and sensors has been the focus of many studies. Thus, Deep Learning (DL) algorithms have been studied for traffic signs and vehicle identification in an urban traffic context. However, there is a lack of priority vehicle classification algorithms with high accuracy, fast processing, and a lightweight solution. For filling those gaps, a vehicle detection system is proposed, which is integrated with an intelligent traffic light. Thus, this work proposes (1) a novel vehicle detection model named Priority Vehicle Image Detection Network (PVIDNet), based on , (2) a lightweight design strategy for the PVIDNet model using an activation function to decrease the execution time of the proposed model, (3) a traffic control algorithm based on the Brazilian Traffic Code, and (4) a database containing Brazilian vehicle images. The effectiveness of the proposed solutions were evaluated using the Simulation of Urban MObility (SUMO) tool. Results show that PVIDNet reached an accuracy higher than 0.95, and the waiting time of priority vehicles was reduced by up to 50%, demonstrating the effectiveness of the proposed solution.
Highlights
Traffic congestion is a worldwide problem because it affects a large part of the population, and the economy, through delays in the delivery of goods and fuel consumption, causing an inability to estimate travel time [1]
The Deep Learning algorithm in which the Lightweight Priority Vehicle Image Detection Network (PVIDNet) is introduced, and the performance validation metrics used in the tests are treated
The following validation metrics are used: accuracy, sensitivity or recall, and F-Measure. These metrics are composed of true positive (TP), false positive (FP), false negative (FN), and true negative (TN)
Summary
Traffic congestion is a worldwide problem because it affects a large part of the population, and the economy, through delays in the delivery of goods and fuel consumption, causing an inability to estimate travel time [1]. In a traffic lights context, where there are many images that need to be processed in real time, there is a necessity to improve the existing current models, reducing further the processing speed without a negative impact on accuracy In this context, an improved version of YOLOV3 called the Priority Vehicle Image Detection. Network (PVIDNet) is proposed in the present research To this end, a lightweight design strategy for the PVIDNet model is implemented through an improved Dense Connection model, based on [22], using feature concatenation to obtain a high accuracy and using the Soft-Root-Sign (SRS) [23] activation function for reducing the detection processing speed. A priority vehicle image detection network (PVIDNet) is proposed based on an improved YOLOv3 model using feature concatenation, and it presents a better detection accuracy than the original.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
