The Korean AI traffic information company Nota has been included in the global technical company NVIDIA’s Success Story which showcases examples of solutions using deep learning and AI.
Nota’s ITS solutions, including real-time traffic signal control, have been implemented as part of the Pyeongtaek Smart City project in South Korea. The project proved the actual improvement of traffic flow enough to be announced as the first Korean startup in NVIDIA Success Story.
Nota’s ITS solutions are based on the image recognition technology to analyse traffic volumes and queues, share data to analyse saturation rate and optimise signal timing at intersections.
The deep learning-based video detection model developed by Nota extracts road information using CCTV images on the road. All processes from the execution of the AI model to the calculation of the traffic signal time operate on the edge-device installed on the site, which does not go through the server. This means any issues around privacy or network disconnection can be prevented in advance.
Nota explains that the overall procedure from running the AI model to calculating the time for the traffic signal can be operated on the edge-devices. This architecture can prevent privacy infringement and network disconnection in advance.
Through this case, Nota was able to solve critical challenges around heavy traffic congestion during rush hour, real-time traffic signal control for emergency vehicles, the need for optimised on-device AI without the chronic latency, high cost, and privacy issues of traditional cloud-based solutions with high accuracy, uptime, and the robustness to environmental variables.
Nota adds it was able to reduce the model’s size to a tenth of its original size while maintaining high performance with mAP of 94.34. Furthermore, Nota reduced its hardware cost by 85% compared to the server-based solution that’s offered by most incumbents in the Intelligent Transportation System industry.
Nota is participating in ITS World Congress (Hamburg, Germany) and GITEX Technology Week (Dubai, UAE) in October, which it says are the most significant events focused on smart mobility and the digitalisation of transport. Nota is bringing one of the main solutions, Nota ITS Solutions, along with NetsPresso, the automatic model compression platform.
Nota ITS Solutions optimizes intersections enabling more efficiency and safety with AI using a Smart Camera with real-time traffic monitoring with edge AI, Smart Traffic Signal Control giving signal control optimisation with Reinforcement Learning (RL) Smart Crossing, and real-time incident prediction on edge devices.
Nota’s technology optimises intersections enabling more efficiency and safety with AI. Image recognition analyses the congestion level by identifying the traffic volume, vehicle queue, and occupancy at the intersection and effectively controls the traffic signal accordingly.
Highways News is a media partner of the ITS World Congress and is providing the official podcasts. We will endeavour to find out more about the Nota solution during the week in Hamburg and bring you the interview on Highways Voices.
(Picture – Nota)