+1 for moderation. Various moderation/voting/rating schemes work well for many sites. I think this is a critical mass issue, once you have way too many contributors, you might have the luxury of restricting access. Otoh winds change easily for these.

The robot makes use of the tiled raised floor of the data center and navigates tile-by-tile during initial layout discovery. It uses a frontier-based navigation, instead of the standard navigation from the device. It completes a fuil run once there is no visitable tiles left. You can find some of the details on the speed, which are not bad, from the ICRA and the ICAC papers.

There are a lot of great answers related to this post, and maybe I can share a bit of first-hand info on this as well--I am one of the folks behind this robotic data center monitoring project, and it is nice to see some of the key challenges and opportunities highlighted in the comments. We use static sensors as well for data centers, and there is trade off among the temporal/spatial density you get with static sensors, the cost of deployment, maintenance and measurement reliability. The robot can give you a full body scan of the data center by sampling at every tile (at several verticals), however, it obviously cannot be everywhere at once. Then, if you decide for finer spatial/temporal granularity, you can think of deploying static sensor based solutions. There are many cool technologies that also help with this approach. Some challenges with sensors are, the cost, maintenance and measurement reliability as the scales go up. One important point to make with the robot is, you can set it down and let it run, w/o any prior installation, as it discovers the data center layout and performs the measurements as it moves along, by navigating the tile structure.