Autonomous Systems in the Internet of Vehicles
by Balamurugan Balusamy, Sandeep Kumar Mathivanan, Prabhu Jayagopal, S. K. B. Sangeetha, Ali Kashif Bashir
14Spatiotemporal Attention-Based CNN-BiLSTM Model for Robust Lane and Obstacle Detection in IoV-Enabled Autonomous Driving
Suresh Kulandaivelu1*, Syied Mazar2, Sangeetha N.3, Sathiyapriya Rajavelu1 and Anita Garhwal4
1School of Computing Science and Engineering, Galgotias University, Greater Noida, India
2Department of Electronics and Communication Engineering, Sreenidhi University, Hyderabad, India
3Department of AI&ML, Amruta Institute of Engineering and Management Sciences, Bidadi, Karnataka, India
4Department of Computer Science, Asia Metropolitan University, Jalan Lembah, Bandar Baru Seri Alam, Masai, Johor, Malaysia
Abstract
For safe operation in resource-constrained and changing environments, self-driving cars in Internet of Vehicles (IoV) networks require accurate and effective perception systems. However, under poor environmental conditions such as” different light, interference by weather, and occlusion of scenes, traditional vision-based models often fall short performing below par. Here, spatiotemporal attention-enhanced convolutional neural network (CNN)–bidirectional long short-term memory (BiLSTM) is introduced that combines attention-based refinement, bidirectional temporal context modeling, and spatial feature extraction. For improved motion understanding, the BiLSTM extracts forward and backward temporal relations, and the CNN module pools dense visual features of adjacent frames. These are followed by a multihead spatial attention mechanism that identifies ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access