Automating Tollbooths in India with Nvidia’s Accelerated Computing
India, with its massive road network and numerous tollbooths, faces significant challenges in managing traffic congestion and long commute times. To address this issue, Calsoft, an Indian-American technology company, has collaborated with Nvidia to implement automated tollbooth systems integrated with India’s unified payments interface (UPI) for a client. This partnership aims to streamline toll collection processes and reduce the reliance on manual labor.
One of the key hurdles in automating toll systems in India is the non-standardized license plates. These plates vary in color, sizing, font styles, and placement on vehicles, making automatic number plate recognition (ANPR) systems less accurate. However, Calsoft’s solution has been successful in overcoming this challenge by developing an ANPR pipeline that can detect and classify plates with about 95% accuracy. This technology reads passing vehicle plates and charges the associated driver’s UPI account, minimizing the need for manual toll collection.
Nvidia’s accelerated computing technology has played a vital role in the success of this project. Vipin Shankar, senior vice president of technology at Calsoft, highlights the challenges faced during the implementation, particularly in night-time detection and improving model accuracy under various environmental conditions like fog, heavy rains, and reflections. Nvidia’s Metropolis, an application framework that combines visual data and AI, has been instrumental in tracking and detecting vehicles throughout the toll collection process.
To deploy and manage their AI models, Calsoft engineers utilized Nvidia Triton, while the real-time streaming platform was built using the Nvidia DeepStream software development kit. These tools enable efficient processing and analysis of data streams, including real-time object detection and classification. Calsoft also leverages Nvidia hardware, such as Jetson edge AI modules and A100 Tensor Core GPUs, to ensure scalability, sustained performance, and adaptability as traffic conditions evolve.
The implementation of automated tollbooth systems in several leading metropolitan cities as part of a pilot program has showcased the potential impact of this solution. By reducing manual labor and improving accuracy in toll collection, India can effectively address traffic congestion and enhance overall road efficiency. With the support of Nvidia’s accelerated computing technology, Calsoft’s tollbooth solution sets a precedent for future growth and expansion, making India’s road network more efficient and adaptable to the evolving traffic conditions.