Innovative AI-Based Traffic Management System Developed by Researchers at NIT Rourkela
In a significant stride towards advancing traffic management systems, researchers at the National Institute of Technology Rourkela (NIT Rourkela) have successfully developed an Artificial Intelligence-based multi-class vehicle detection (MCVD) model. Accompanying this development is a light fusion bi-directional feature pyramid network (LFBFPN) tool, both of which are poised to enhance traffic management immensely.
The innovative research project is spearheaded by Santos Kumar Das, an Associate Professor in the Department of Electronics & Communication Engineering, along with his team. They have harnessed the power of an intelligent vehicle detection (IVD) system that utilizes computer vision technologies to detect vehicles in various images and videos, showcasing significant improvements in traffic analysis.
Central to this system is its ability to collect real-time traffic data, which plays a crucial role in optimizing traffic flow, reducing congestion, and aiding in future road planning and infrastructure development. This advancement not only addresses the current traffic woes but also offers a prospective solution for future urban planning strategies.
The research has been recognized for its significant impact and published in the prestigious Institute of Electrical and Electronics Engineers (IEEE) Transactions on Intelligent Transportation Systems. The paper highlighting these findings includes contributions from Santos Kumar Das and research scholars Prashant Deshmukh, Krishna Chaitanya Rayasam, alongside Upendra Kumar Sahoo from the ECE department of NIT Rourkela, and Sudhan Majhi from the Indian Institute of Science, Bangalore.
Revolutionizing Vehicle Detection with the MCVD Model
The newly developed MCVD model stands out due to its incorporation of a video de-interlacing network (VDnet). This network is essential for efficiently extracting key features from traffic images, regardless of the varying sizes and shapes of vehicles. Moreover, the model’s capability is further enhanced by the LFBFPN tool, which finetunes the extraction of these vital image data.
Professor Das elaborated on the unique aspect of the LFBFPN, stating, “What sets LFBFPN apart is its simplified methodology that does not compromise on accuracy. It processes input data through a vehicle detection head (MVDH), enabling precise detection and classification of vehicles across diverse traffic scenarios.”
Performance and Application of the MCVD Model
In comparison to existing methods, the MCVD model significantly improves accuracy. It has undergone rigorous real-time performance evaluations on advanced technology platforms like the Nvidia Jetson TX2, highlighting its robustness and adaptability. These state-of-the-art evaluations indicate that the system could be effectively utilized in traffic-heavy and mixed vehicular environments commonly spotted in developing countries.
“By tackling the limitations associated with older models and addressing the unique challenges posed by mixed traffic environments, the MCVD model provides a scalable solution for real-time vehicle detection, which is particularly crucial for developing nations,” concluded Professor Das.
The innovative research from NIT Rourkela underscores a leap forward in traffic management systems, leveraging AI’s potential to streamline and make urban commuting more efficient. As cities worldwide grapple with increasing traffic challenges, such cutting-edge solutions offer a glimmer of hope for sustainable traffic management and smoother transit experiences.