So, I got a Roborock S6 MaxV, and I'm very happy with it. It works as intended, and does a good job. So on the hardware side, I don't have any real suggestions for improvement.
But I have a few on the software/app side:
- Whenever it sees an object, it should take a photo. Not only the ones that the AI feature recognizes. Right now it just shows a cone-icon and says "Avoided", but no picture of what it actually is that it avoided. It's great to show what the AI recognizes (or thinks that it recognizes) but it would be even better if I could actually see what it avoided but didn't recognize, so I can try and remove the obstacle before the next run.
- Request a live/instant image from the camera through the app. The best would be a live stream, but if not, then just a button to "Take a picture", so in case it gets stuck while i'm not home, at least I can see if it's a shoe, a carpet, or my cat's tail! ;) ... or simply if I'm just curious what's in front of it right now.
- When editing the map to add no-go zones and all these things, it would be nice if it was possible to change the individual corners of a square, so it's not just a square. Also, it would be nice if you could add circles or other shapes instead of just a square (e.g. I have a chair in the living room with a round foot, that it keeps getting stuck on. Would be nice if I could circle it out, instead of a square, that makes it go way too far away from the chair.
- An "Initialize map" feature. Would be nice if you could just let it drive around on it's own, to map the room, the first time you set it up, or put it in a new house/floor. It's not necessary to do a full clean of the house, with the very "narrow" vacuum-pattern, when the LiDAR could probably do it pretty quickly just by driving around, following the walls first, followed up by a bit more "rough" square pattern in the room, to catch other objects. This could probably do the mapping in less than half the time of a vaccuum. After that, you can add all the No-go zones, room divisions and stuff, and then do a clean to confirm everything is working properly.
- It would be nice if you could "erase walls" in the map editor. For some reason, my couch (which is more than 15cm above the ground, so plenty of space for it to get under it) is recognized as a wall, which makes the vacuum-pattern a bit more confusing. If I could go to the map-editor, and then with a "Wall eraser"-feature, simply erase walls (that are not really walls) manually by sliding my finger over it, then it would be perfect. Of course, the robot should remember in the future, that it shouldn't recognize walls in this area again.
- An option to show/hide all the map configurations. It's nice to see all the no-go zones and no-mop zones and walls, when you're in map editor mode. But when you're just looking at the map of the house, when it's running, then it can be a huge mess to look at. So a toggle-feature to show/hide all the zones would be nice, so it's just the room-colors (that is added when dividing rooms), all areas that are no-go zones are grayed out (made more transparent), and maybe also the settings (vacuum-strength and water level). The best would be to have a menu where you check off all the stuff you want to see on the map-overview when you look at the map, but a fixed toggle feature would be fine.
- Correct AI recognition? I'm actually a bit surprised that it's not there. A lot of the time, the AI is very wrong. Most of the stuff it sees on the floor is categorized as shoes, for some reason. And yesterday it recognized a window curtain as poop. Wouldn't it be nice if I could click a button and correct whatever the AI thought it was? And when it doesn't recognize things (like in my 1st suggestion) I would be able to set a category that matches? That would not only help you improve your AI, with help from you, but in the future, maybe the robot would simply know that the thing in my living room, that looks like a shoe, is actually a cat toy. Or that the curtain is actually not a poop, so it's okay to move to that corner of the room.
Tell me what you guys think? Any of these features you would like to see, and if yes, do you have any extra suggestions to any of them, to make them even better? :)
Hope that the Roborock Developers would consider some of them, or maybe say if some of them are already on your roadmap for upcoming features?