Traffic Road Sign Recognition Using Multi-CNN Model for Autonomous Ground Vehicles

Abstract

Detecting and recognizing the traffic road sign plays a significant role in autonomous vehicles while processing the details regarding each road signs. This research proposed a novel Multi Convolutional Neural Network (M-CNN) for recognizing different traffic signs. In this model, four different CNN structures are applied for processing various road signs. So, the high-level and low-level features are extracted to classify the signs into a proper class in driver-assisted autonomous vehicles. Here, the developed model has produced outstanding results while classifying different real-time traffic signs. The efficiency of the proposed model has been assessed by a well-known database named German Traffic Sign Recognition Benchmark (GTSRB). Moreover, the simulation outcomes demonstrate that the M-CNN model has certain advantages while detecting various traffic signs. Simulation results indicate that the proposed M-CNN model has a higher recognition rate and reduced processing time when compared to Machine Learning (ML) and Deep Learning (DL) models.