A cable is a different problem from a crumb
A good cleaning robot needs judgment before it needs another cleaning pass. The useful question is whether it can distinguish something to collect from something to avoid. A charging cable, a chair leg and a spilled drink call for different responses.
DJI’s September 3 ROMO 2 announcement focuses on that distinction. The company says the robot recognizes thin obstacles down to 2 mm, avoids transparent objects, and changes its cleaning approach for debris versus liquids. These are manufacturer claims from controlled testing.
The arm matters at the edges
A LiDAR-guided arm swings through 123 degrees to extend cleaning beneath furniture. The official product image also shows the translucent station and robot: a view of the hardware, rather than evidence of how well the obstacle detection works.
For a useful real-home comparison, watch one complete run through the same room. Count cable tangles, missed patches and interventions. Those observations tell you more about whether it saves time than a single maximum specification.
Check the feature you are buying
DJI lists September regional ordering, while Local Data Mode is promised later this year. Check local availability and which software features have actually shipped before treating the announcement as a complete feature checklist.
Source published 2026-09-03. Coverage is based on the maker’s announcement and demonstration.