I have a couple pieces of furniture where I've added custom parts to make them roomba-compatible so it wouldn't get stuck on them. I've also gotten rid of more than one piece of furniture because it didn't play nice with the robot. Sorry chair, if the robot doesn't like you, I just can't keep you.
Roads, are designed for the main users - people, driving. So self-driving cars have to cope with that. But we'll soon get to the point where some road markings / signs are changed around to make them more obvious to self-driving cars, in order to avert rare but costly accidents. The point to watch out for is when the self-driving cars are the primary users, and the needs of human drivers start to fall away as not important any more.
Domestic furniture will continue to be designed for both accepting the human posterior and for interacting well with domestic automation.
> The point to watch out for is when the self-driving cars are the primary users, and the needs of human drivers start to fall away as not important any more.
Yeah. I'm really not looking forward to streets being all fenced off to prevent us pesky humans from jaywalking and causing trouble to our robotic overlords.
This post is not really about robot vacuums and laundry racks, but about how simple ideas can elude human beings for long stretches of time, and then in retrospect look obvious.
I experienced this together with a friend while we were CS students. I went with him to a store to get several pieces of carpet. We needed each to be sized roughly about 1 meter by 3 or 4 meters. For some reason, both of us were quite worried about the long pieces not fitting in the car easily. Even when the person in the store had cut the pieces and they were on the floor in front of us we were looking puzzled at the lengths not fitting in the car. Then we were amazed to see our friendly salesperson stacking the pieces on top of eachother and rolling them up on the short end, creating a thick roll of carpet of about 1 meter long that would fit in the car without any issue. We were both laughing and feeling a bit ashamed that neither of us had thought of this simple solution and that we were really worried for nothing all this time.
What's also really cool is that people are willing to share these small discoveries in a format (the web) that means they can help others. How many people's pain will be alleviated by this tip, and others like it, thanks to the simple impetus to share freely? (And, more darkly, when should you choose to keep it to yourself?)
For a long time I felt like keeping on top of the laundry was a constant struggle because we would generate a new porch-load of wets almost as fast as it took the load on the porch lines to dry. If we had one or two days of rain, or were too busy to swap the loads, we'd get behind and it would take a few days to catch back up.
Finally I bought more clothesline and doubled the amount of hanging space on the porch. Now we have plenty of margin - but why didn't I try this years ago? The extra hooks were out there the whole time.
At least it only took me a few months with our tiny washing machine to realize that I could collect several wet loads in one laundry basket and hang them all up on the porch in one big batch, instead of waiting for the first load to be hung up before starting the next one in the washer.
So refreshing to see that others have this simple problem, too. Roombas really are like users. They will eventually find flaws and errors in your software, or, in this case, shoelaces and pieces of cloth lying on the ground.
Other issues with the roomba:
- plant raisers with thin wooden feet.
- wine rack with low base/feet
- ikea chair with low U-shaped feet. Especially when the roomba tries to return to station
I also noticed that it has a hard time detecting dark/black surfaces. Probably it has to do with absorption of whatever kind of sensor wavelengths it uses.
Difference being that tools/tech/developers should ideally, and very often do, make users' lives easier. Whereas here it's partly the other way around: users having to come up with workarounds. Instead of the tool being fixed.
I recently purchased a Roomba Evo i3 and I have this exact same problem, only with my bar stools [1]. The bar stools have horizontal supports that stick up about 25 mm from the ground—not enough to trigger the bump sensor but enough for the robot to be unable to drive over them.
I'm not super excited about this particular solution because it might affect the stability of the barstools when people are sitting in them, but I'm considering making some 3D-printed pieces to increase the height of the horizontal support enough to trigger the bump sensor.
(If you're reading this and work at iRobot, please file an internal bug report about this!)
The front of the robot would ride up until the wheels were against the bar and lost traction, so it could still back off easily. The (Wyze) firmware was smart enough to do so and record the spot as obstructed after a few seconds of wheelspin, but this always amounted to several minutes of loud tyre slippage as it got around the entire chair.
That firmware also trusts its LIDAR completely (thus bedskirts need a lifted up port or it won't try to go through), so tape strung between the chair legs at turret height worked at first. The tape kept getting stepped on, though, so spare adhesive bumpers beneath the existing feet turned out to be our final solution.
Oh and don't start me on the pedestal [1] bar stools (also bar table) I have and the robot keeps climbing the base discs until it gets stuck. You can't mark them as no-go areas as they're too small (you end up with dirt around them) and they also move around, so whenever cleaning that area I have to stick around to solve the unavoidable alarm notifications. Or I carry the stools someplace else and must unstuck the robot only from the table pedestal.
It's kind of interesting that this happens given how relatively old robot vacuums are nowdays. The robot getting wedged into high or low obstacles is a pretty common issue.
The expensive robots added lidar sensors for better navigation, but the basic obstacle sensing mechanisms hasn't really improved: front/side sensor, bumper, + cliff sensors for falling (and these last ones aren't 100% foolproof either).
A front camera doesn't really help these robots IMHO (roomba, cough). Nice gimmick, but unless you reach 99% reliability, you'll still see your robot squishing dog poop, falling, wedging itself into the sofa..
I guess it's a nasty tradeoff between robustness, service lifetime, and cost.
I bought a cleaning robot a year ago (totally recommended btw), and I also went the route to adapt the furniture and placement of things to help the robot.
Roombas could definitely be better. Mine gets stuck all the time. The reason I keep using it, however, is that it cleans a lot of stuff prior to getting stuck, so it still feels worth it to me.
We recently were gifted a DEEBOT OZMO T8 AIVI ... and while the hardware side works pretty flawless, the software side is just a major shitstorm. Like every time I open the control app the map looks different. So far I had the following problems:
1. The "split this room" feature does not work on two rooms that have been detected as one. I believe that that is due to the bot not recognising a central chimney as "not room".
2. While cleaning it detects the closed bathroom door as a wall and then does not path into there anymore. As far as I can tell this does not trigger any warning, the bot just drives back into the charging station after some time.
3. The position of the charging station was lost, which made the bot update the map in a weird way, superimposing where it thought it was skewed over the map.
I feel like some of those problems could be fixed fairly easily if users had more control over the SLAM process. I.e. the ability to permanently fix walls in the map.
If you’re lucky enough to have a model that supports Valetudo, reflashing the firmware with an open source alternative can alleviate some software problems:
I recently saw an announcement of a "brain replacement" project for (a specific) robotic mowers ... such an "after market" project for robotic vacuums would be great. But replacing the electronics is definitely no easy feat.
It’s a little sad to see 3D printing used for such trivial components, where you get all the weaknesses of a 3D printed part but none of the strengths.
Here’s a free startup idea: a structural search engine for off-the-shelf parts. E.g. “show me all parts that fits around round tubing of this diameter and has a flat bit at the other end”.
What? Small one-off fixings is one of the perfect uses of 3D printers. Your startup idea won't work precisely because of the huge variety of shapes that you'd need. Your searches would almost always return 0 results.
It's a free startup idea, if you take it and it doesn't pan out I'll happily refund you the $0 for it.
> precisely because of the huge variety of shapes that you'd need. Your searches would almost always return 0 results.
Sure, the risk is either zero results or large amounts of irrelevant results. But it's for seemingly impossibly large search spaces that search engines are the most valuable. It's no coincidence that Google is a trillion dollar company and pressing control or command-f on a web page isn't.
Edit: also, while we're poking holes in the idea, the main problem isn't the feasibility of the technology, the main problem is that people who need drying-rack-holder-uppers probably do not constitute a rapidly growing market, which is the other thing Google had going for it.
I would have gone with a piece of foam pipe insulation, available at any hardware store for about $3 (or McMaster-Carr [1], to stay at the topic of a one-stop shop).
I think it's very much a case of knowing what exists and where to search for it, while 3d printing allows you to just think about what you want to exist.
Yes, something like McMaster-Carr, but you search not by the name but the shape of the thing.
You'd probably have to take all their CAD files and put into some kind of searchable index, where the query itself is some kind of a CAD description of what you want, like a drying-rack-holder-upper.
- 15 minutes would be gathering the necessary tools and gathering at the workbench.
- Some other 10 would be measuring and thinking how to do it
- Then 15 minutes would be measuring, drawing, cutting and drilling. Oh, and I forgot I would need this tool, let's get it downstairs/from garage/house...
- Finding out it doesn't fit perfectly you already lost 5 mins. And now if hole is too big, you create new one, +15 mins. If hole is is too small, you search for your sandpaper. Sand it + fit, sand it + fit...
- Ahh, it broke, because wood was too thin. Retry.
- Putting tools back, cleaning up your place +10 min
Total: Little more than hour realistically :)
/from a perspective of DIYer, how it usually happens with me.
I feel like there's been a kind of meta-discussion happening around the HN peripherals about laundry racks. I got a laundry rack for the first time after seeing a post about them here on the front page. Later, mentioning the laundry rack resulted in a thread of laundry rack recommendations of sorts. And now we've got a brand new post on the front page related to managing said laundry racks. I wonder how many people, like me, are reading this as new laundry rack owners who are learning to integrate their new low-tech laundry appliance into the rest of their high-tech home.
Glad to see i'm not the only one solving problems that way :)
My Roomba would get stuck under the radiator in the living room. It fits under one end, but aparently it's not completely level, so as it was doing it's edge routine it would gradually wedge itself stuck under the radiator.
The solution was "simple" as 3D printing a couple of "stands" to place under the radiator where the Roomba fit and didn't fit.
As for my Robot lawnmower getting stuck on the trampoline, i simply dug down the trampoline legs far enough to make the robot drive safely over them :)
I used to be the same, but then I got the vacuum and it's amazing how good they really are at keeping the house clean, so I changed my habits and finally got around to doing all the necessary cable management. Also it's not just the floor, the whole house just has less dust in general when used regularly.
It really depends on the vacuum. Mine doesn't have the sensors to know where it has been etc and gets stuck often. It doesn't save us any time but it does still get places that we wouldn't typically get to.
Sometimes I wish I had gone with a newer vacuum. I got the original Xaiomi vacuum and while it has the LIDAR it's lacking on the software side of things and "forgets" the map with every clean. Luckily I have root and installed valetudo on it so I can bypass most of the limitations using HomeAssistant
My house is neat enough because I have a robot vacuum.
The vacuum runs every weekday at the same time. It's mentally easier for me to spend the 1-5 minutes running around the house and get stuff off the floor for the vacuum to do its job than it is to chase after it to get it to dock again.
The biggest motivator is the amount of dust it gathers even when it runs every day. I'd be breathing in all that crap if it didn't pick it up.
This is what works for me too - feeling relief that robot managed to collect all this dust I didn't even knew was there. And that's every single day! Totally worth those ~5 minutes collecting stuff all over the place.
Same here. I'm waiting for the next generation that comes with a claw to pick up stuff and put it in a "miscellaneous item" box on the first round, then go about the vacuum/mop process.
Having to force myself to keep the house neat is the reason why I got a robot vacuum in the first place. You make a deal with it: you keep the floors tidy, then it manages the dust bunny population.
Based on the replies to your comment I might try another model, but I had one and it was a huge hassle cleaning up before it, cleaning up after it and cleaning it multiple times per job because the brushes would get “tangled” on nothing. I stopped using it because it’s less work to just get the mop and broom and do it myself. It feels like it’s main selling point is keeping a clean floor clean, but if you have kids, hair, pets or a party you’re gonna have a bad time.
I have an iRobot Roomba J7 which has a camera and object avoidance. I had a cheap Eufy previously which tore itself apart after eating one too many phone chargers. The Roomba just works around it without a second thought. I'm very impressed, and it's made the change between 'running the vacuum' being a task I need to do 5-10 minutes of prep work for to a completely automated job that I don't even have to think about any more.
When I ordered my home 3 years ago, I chose furnitures that have bottom high enough to let a Dyson 360 pass, and carefully designed their layout so the robot can easily go from one room to the other without circling around or running into wires.
The robot still got stuck from time to time, but the efforts were quite worth it. :)
My dining chairs are perfect roomba traps. It can effortlessly navigate into the space between the four legs, but somehow seems unable to just drive out in a straight line. If I forget to lift the chairs to the table before I turn the roomba on, it inevitably ends up bonking against the four legs of the chair.
Simple but controversial answer: don't use a vacuum cleaner.
Depends on how many dogs, children and visitors you have.
No need to 3d print parts to fix the first world problems.
I do change my shoes at the door. It can be weeks before the dustpan and brush comes out. Or months for the mop. Otherwise, all is good. Cleanliness is maintained.
My solution was to zip-tie lengths of black poster board to the legs. Worked very well for a long time. Eventually the vacuum did bend it a bit, but it's not *too* ugly.
Preventing it from getting into the wires is the only option I know of.
Either cable-management to put all the cables off of the ground or in gutters, or block off the area using magnetic tape, virtual walls, or redzones depending on your vacuum.
You let it run when there's wires around? I went through and attached all wires to the wall or behind immovable things before I let the vacuum run. I just assumed it would rip the crap out of my wires and start a fire.
Still gets stuck with the rubber tube cable management, also makes changes to my setup a huge hassle. Considering buying trucking and sticking it to the floor.
I don't own a robotic vacuum, but two racks. They seldom end up in the exact same position every time, are frequently moved between rooms depending on state of drying, weather, and so on. Not to mention the kids.
I can't imagine spending effort exactly mapping the location of a rack, a passive/permanent solution like the one in the article is just right.
It doesn’t matter where I put them in the Roomba app, it still ignores them. Then it goes out into the deck and makes a racket while it looks for things to get stuck in/on/over/under.
Some do, but not all of them. We have the cheapest Eufy one and it doesn't have no-go areas. I just lift wires and have screwed a piece of wood underneath the shoe rack where it sometimes gets stuck because the floor isn't level. The piece of wood stops it heading under there.
The energy bill difference is probably a lot less than you might think, at least for a lot of people. Anyone running a dehumidifier or air conditioner is going to be paying a large chunk of the cost anyways in recondensing the moisture back out of the air.
