Work

Hybrid Simulation-Driven Digital Twin Approach for Optimizing Multiscale Product Design and Prototype Development Processes

Year: 2026

Type: conference-paper

Abstract: Traffic signal recognition is extremely important for improving safety on roadways and for auxiliary purposes in autonomous driving systemsThe existing methods usually have a problem with the accuracy and interpretability when applied in complex and naturalistic settings. In this paper, we offer a novel hybrid deep learning framework to combine Convolutional Neural Network (CNN) models with Support Vector Machines (SVM)-based feature extraction to classify feature based on it. The whole system was tested and trained on the GTSRB which contains numerous traffic sign images in varying weather conditions. This model is done on the Python platform and is able to perform classification at 99.60% which is a sign that the model is strong and stable in generalization to different circumstances. The framework provides developers of autonomous vehicles, developers of traffic safety systems, and city planners a reliable, interpretable, and computationally efficient approach to real-time traffic signal recognition.

Source: International Conference on Intelligent and Sustainable Electronics & Computing Technologies (INSECT)

DOI: 10.1109/insect68872.2026.11663867

Authors: Syed Hassan Imam Gardezi, Chamandeep Kaur, Mohammed Saleh Al Ansari, Aruna Gude, D. Jyothi Preshiya, Alamelu alias Rajasree. S

Institutions: Jazan University, University of Bahrain, Pace University, Saveetha University, Sri Ramakrishna Engineering College

Language: English

Topic: Advanced Neural Network Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

SDGs: Industry, innovation and infrastructure, Sustainable cities and communities

Open Access status: hybrid

OpenAlex ID: W7204850758

FWCI:

Cites: 15

Cited by:

Related to: 10