Blurobi has introduced a new generation of its AI roadside perception system, built around an 8MP starlight sensor, motorised optics and an on-board inference unit. The system is designed for the conditions that make curbside monitoring hard in practice — low light, rain, occlusions and mixed traffic — rather than for ideal demonstration conditions.
Instead of streaming raw video to a data centre, the units decide locally: vehicles, occupancy and events are recognised at the edge, and only structured results are uploaded. That keeps bandwidth and storage predictable across a street, a district or an entire city, and it keeps sensitive footage inside the device.
The platform is delivered with the operating layer that curbside projects actually need: a live map of bays and poles, occupancy and turnover history, exception-based enforcement, and health monitoring that flags a unit before it fails. Projects of this shape are already running in more than 70 cities, including the large roadside-sensing programme in Beijing.
The new generation is available for pilots now, with deployment support for cities that want to start on a single corridor and extend along the network as results are proven.
