9th International Conference on ICT for Sustainable Development

Authors - Ishan Rai Agarwal, Sahil Dhanani, Shikha Mundra, Ankit Mundra
Abstract - Vehicle-to-vehicle (V2V) communication technology advancements have changed traffic safety and efficiency recently. It has improved traffic flow and significantly reduced the risk of accidents by exchanging real-time information about their surroundings. This work offers a distinctive study which uses computer vision and natural language processing to give a ground-breaking analysis of Visual Question Answering (VQA), which improves inter-vehicle communication security. For experimentation and analysis we have created a small dataset and conducted rigorous testing to demonstrate the effectiveness of VQA in improving V2V communication and authentication by adopting multimodal features such as text and images. In order to learn multimodal features, recent deep learning networks like Visual Transformer and BERT is adopted. Overall, it is observed that even in contexts with limited resources, our findings highlight the potential of VQA to strengthen networked vehicles against new security threats.