Bengaluru: Bengaluru-based startup Radiantech Systems has developed a ruggedised AI edge-computing device designed to reduce the bandwidth, storage and server demands of AI-powered traffic surveillance. The hardware, named RTE-300-R Rugged High Performance Edge AI Compute, was showcased at the Bengaluru Tech Summit 2026.The device processes traffic camera footage locally at road junctions instead of continuously transmitting massive volumes of raw video to centralised servers. Once a violation, such as jumping a red light, is detected, only a brief alert detailing the incident is sent to the city command centre.Explaining the need for the technology, Radiantech Systems CEO Sumanth S said a network of 10,000 cameras, each transmitting 3–4 MB of data per second, could generate around 40,000 Mbps of continuous backhaul traffic. Such volumes increase bandwidth, storage and computing requirements, pushing authorities towards expensive GPU-powered data centres.The RTE-300-R aims to address this challenge through local processing. Designed for harsh outdoor conditions, the ruggedised, air-cooled unit eliminates the need for expensive, water-intensive cooling infrastructure. It can also store footage locally for 60 days, allowing authorities to retrieve specific clips when required, before automatically deleting the files.Radiantech has submitted a formal proposal to the Bengaluru Traffic Police (BTP), which has expressed interest in the architecture. The company is awaiting a Letter of Intent (LoI).Positioning itself as a hardware systems manufacturer, Radiantech plans to collaborate with third-party software developers to deploy AI-based violation-detection models on its devices. The approach could help cities reduce infrastructure costs while making AI-driven traffic enforcement more efficient and scalable.
