PurerAir sensors are now riding along on China's Baize Zhi patrol robots

Most air quality monitoring is stationary. A sensor sits on a wall, a lamppost, or a rooftop, and it tells you about the air in that one spot. That works well for a room or a building. It works less well for a park, a business district, or a transport hub, where conditions change block by block and foot traffic moves faster than fixed infrastructure can track.

The Shenzhen Automotive Research Institute of Beijing Institute of Technology (BIT) is testing a different approach. PurerAir has integrated its air quality sensors into "Baize Zhi" (白泽智), an autonomous patrol robot developed by the Institute. Named after a mythical guardian beast from the Classic of Mountains and Seas (山海经), Baize Zhi is built on scaled-down autonomous driving technology and already handles outdoor security patrol, emergency response, and public communication in places where hiring and retaining human patrol staff has become difficult — exposed, repetitive, all-weather work that the industry is short-staffed for.

The robot's existing sensor suite is substantial: over 30 recognition types, a 360-degree sensing range out to 100 metres, lidar and vision-based perception, and one-touch emergency call and remote video support. It's already deployed or being evaluated across shopping districts, tourist sites, university and public parks, industrial parks, residential communities, ports, airports, exhibition centres, and government complexes.

What it hasn't measured, until now, is the air itself.

With PurerAir's sensors added to that suite, Baize Zhi becomes a mobile source of air quality data, generated as it moves through its patrol route in Shenzhen rather than sitting fixed at one point. That data feeds into the same real-time dashboard PurerAir uses across its other deployments, the education campus in Guangzhou and the commercial property monitoring at 67 Pall Mall, giving facility and city management teams a live, location-tagged read on conditions across a site rather than a single reading extrapolated outward.

This is also a useful test of an idea we've written about before: that air quality data is only as good as its resolution. A single fixed sensor can tell you the air was fine at the entrance. It can't tell you what it was like on the far side of the car park, or near the loading dock, or by the ventilation intake. A robot that's already moving through a site on a fixed patrol route solves that problem without adding new fixed infrastructure, it just adds a new sensor to a vehicle that was already going to be there.

It's a small case study in a broader pattern worth naming: as more physical infrastructure gets automated — patrol robots, delivery robots, cleaning robots, each one is a candidate for carrying environmental sensors it wasn't originally built for. The marginal cost of adding air quality monitoring to a vehicle that already has power, connectivity, and a reason to be moving through the space is much lower than installing a dedicated sensor network from scratch.

This is currently a pilot at a single site in Shenzhen, a first test of whether the approach holds up before any wider rollout.

We'll share real air quality readings from Baize Zhi's patrols once the pilot is complete.