This is going to be a fairly short update most likely on the Roborock S7 MaxV Ultra. OK evidently not. Corporate test you can read about in the links at the bottom. Had three nights where it’s not completed the run this week but two of them are not its fault. One, however, is. I caught and corrected that one tonight.
Every robot vacuum is prepping for tomorrow when I start up round two of the head to head (assuming our ILIFE replacement finishes its mapping tonight, which it looks like it will.)
Night one of not completing the run involves people being in the building at odd hours and me calling it off. Night two involved someone putting a trash can blocking the unit … well, it didn’t block the unit really but I got an error message that the unit could not reach the dock, took remote control, rammed a plastic trash can repeatedly until it moved far enough out of the way, docked it, thought I restarted it but evidently did not.
The Reactive AI Obstacle Avoidance fail
What’s interesting here is night three (tonight.) I noticed the Roborock S7 MaxV was in the robot trapping chairs area and had been there for a minute. I initially thought that it was locked up, hung up or caught so I jumped in on video and watched. I couldn’t see what it was stuck on but it kept moving forward a bit and acting like it was hung, spinning around like a dog, backing stopping, spinning, etc. This happened over and over again.
The unit appeared to be right next to the chairs from what I could tell on the map and what I was watching on video I couldn’t see that it was actually even touching anything.
I would see the picture above, it would rotate, and then move forward a bit away from the robot trap, and stop. It appeared as though it was stuck on a string or something. Nothing I could see from the S7’s camera or from the atrium cameras made any sense as to why it was stuck.
It finally threw an error after multiple attempts to drive itself away and I took over on remote control. I drove it right out without any issues. As I recall I pressed the up/forward button and just drove straight out. I turned around and looked to see if there was any evidence as to what has trapped it there twice (it got stuck once last week in that area.) Nope.
No strings, wires, anything I could think of that would result in this having an issue. It’s like the chairs, knowing they could not hold this robot captive, used mind control. Whatever it was it appears to have figured out how to mess with their obstacle Reactive AI obstacle avoidance. As such I’ve zoned off that area for now and consider this the second repeat RAI OA AI fail I’ve discovered.
This would be at 45+ hours or so of cleaning. 5-6 hours a night, 10 nights at least… me playing with having it super scrub sometimes, yeah… that’s damned impressive.
Oh yeah, I’m going to keep nitpicking but let me tell you this thing is amazing. I hope they get this chair robot trap recognized in the AI but even if they don’t, it’s amazing.
Roborock S7 MaxV going the distance… er, too far
Note the map above and the long hallway that’s two colors. I can’t tell you how long that hallway is but I can say it’s over 80 feet pretty confidently. This messed with the S7 MaxV in interesting ways. When it was all one color (or was listed as Hallway, one room,) the unit would continually have issues thinking it was at the end of the hall and be several feet off. New worlds were discovered. Walls were driven through. The blue room bottom right grew a whole area that just doesn’t exist and can’t be reached.
I ended up slicing mt 80+ foot hallway in two and that fixed the distance and some of the discovery of lands unknown coming back to the charger, but the blue room is still messed up. The map at that point has not updated in a few runs and the room still doesn’t exist in Einsteinian space as far as I can see. There’s several feet beyond the physical wall that can’t exist there due to another office being there. And a wall.
Only probably something to consider if you’re cleaning 4000+ square feet of space and have an obscenely long hallway. Yeah, I’ve got screenshots at 3,548 square feet of cleaning one night during a clean that’s missing several rooms at that point, why do you ask?
I highly suspect that if the robot goes too far in a single direction down a narrow hall it loses where it’s at and attempts to reposition. That hallway looks about the same all the way down. Suddenly you’re about the distance off between two large support beams.
Also I’ll note that unless I’m mistaken that hallway is straight. It appears that at that distance the map indicates it’s slightly curved. I don’t claim to be an expert on these things, but it appears to show enough distance off at the end that I can see it (or I’m hallucinating.)
Once again, 80+ foot hall I do not expect a home vacuum to handle gracefully, however it did up until recently when it mapped out some more areas. I start to think I may have found the hidden upper limit of the map (somewhere around 5000 square feet of mapped area) – could be mistaken but I bet the company I would find it.
Full disclosure, I get these things to test and am under no obligation to be nice. Fuller disclosure, the Roborock series is the only series I’ve tested that I’ve 1) had named by my kids and wife, 2) purchased items for, 3) not had a give away plan (I never sell items from Pocketables).
All sorts of weirdness in notifications tonight
I started getting push notifications at 9:42 tonight. It claimed the downstairs vacuum was finished cleaning, and at 9:45 it said the MaxV was returning to dock because it had a low battery. That’s initially why I noticed that it was trapped, because 9:42 was too early for it to have a low battery unless something was wrong.
I didn’t realize at that point that the first notification had indicated the downstairs vacuum. That’s not at that building. That’s at my house and her name is Rhonda. She was not running at 9:42.
At 9:50 I got notification that my downstairs vacuum was finished cleaning and returning to dock. As Rhonda (known to Google Home and Roborock as “downstairs vacuum” was on the dock and had not been cleaning all day that I knew of, this got me wondering what was up. That’s why I managed to jump in and notice that the S7 MaxV Ultra was stuck. My first screenshots were about 9:54. The MaxV was not low on battery, but it was trapped.
No idea why I got insanely delayed notifications.
It’s late, I’m tired of writing about vacuums at the moment, I bid y’all good night.
Previous / future Roborock S7 MaxV Ultra coverage
Day 1- The Roborock S7 MaxV Ultra I’m initially impressed
Day 3 – I’m more impressed
Day 6 – find some issues and wants
End of week 1 – some strange oddness
Roborock S7 MaxV Ultra ends week 2 like a lion
The end of week 3 / start of week 4
Week 4 – end of most testing