I'm actually genuinely surprised it did not have persistent maps yet.
- It is obviously useful to plan cleaning.
- Once calculated, storing them should be cheap.
- Detecting that the stored map is wrong if you manually move the Roomba to another location should be an issue; just ignore the stored map after too many inconsistencies with current reality.
Synchronizing a digital map with a physical device without a lot of expensive sensors and expensive specialized hardware is extremely hard. For example, consider that an iPhone X can only mostly keep track of a room's boundaries with its onboard sensors and camera... and it has a market price of $1000.
I think it’s $1,000 more because it’s a powerful, tiny, general-purpose computer and less because it can keep track of a room’s boundaries. That is to say, the Roomba doesn’t need to do everything a smartphone does and thus wouldn’t need $1,000 or smartphone technology.
well that's not a fair comparison. the iPhone is built to be capable of doing a lot more than a Roomba, thus the cost. Also, much of the specialized expensive equipment is required for the iPhone because of its form factor. For the Roomba, the only new pieces of data one would need is a feed of the current orientation relative to its starting point and rate of travel to mapping out a room's boundaries. The logic is more complex but not to a point where expensive hardware is required.
That is not exactly correct. You don't need expensive sensors to synchronize maps with reality. The coarsest estimate of where walls are can be used to infer a position in a map; it will just not be very precise. But errors of cms should be ok.
And Roomba only moves in a plane as a result of its (slow) actuators, whereas an iPhone X tries to do hyper-accurate 6dof real time position, from a very quick external source (your hand), for augmented reality. Thus the sensor and computation cost premium. But that does not apply to the Roomba.
Following nintendos lead, you only want small improvements in each product version so the user sees a benefit of buying the next version of the same thing they already have. When you are getting close to perfect, you start removing features and add them back in later versions again. (audio out on whatever handheld it was).
I am convinced that they could easily do this with the existing 900 line that has WiFi and "advanced vision" technology. They basically threw the hardware in, added some marketing buzzwords, and never really changed the software other than to show a coverage map. The software is the weakest part of roomba. It's navigation system is so naive despite the load of sensors inside. I wish they would unlock the damn thing and just let me script it myself.
I think it could make sense if they start with a hard requirement that it must be able to do a good job of cleaning a not-yet-mapped house. (They might feel that is important for making a good first impression or avoiding user confusion.)
If they manage to solve that case, then the benefit of retaining the map is less compelling. By definition, you can do an adequate job without it. So then it's a question of what you gain in cleaning performance/efficiency vs. the extra complexity and any potential problems it might bring.
Roombas came out in the early 2000s and still can't vacume a floor half as good in 2 hours as a human can in 10 minutes. If the floor has carpet it doesn't even begin to do 10% of the job.
Has there been any studies on how good the robot hovers realy are for capturing smaller particles? Does it have enough power to use a HEPA filter? Or does it just blow the dust out the backside like old hovers?
Anecdata: I'm allergic to pollen and cats(a little bit). This spring I got myself a Roomba 895 which is equipped with a HEPA filter and hadn't experienced any symptoms since. I have no real data on its effectiveness aside from "works for me", though.
I had a roomba, but the combination of constant brush maintenance, hair cutting and emptying made it a pain in the ass. Not to mention how it totally ate wires on the floor. The time it saved was taken up by different things. And it did a relatively bad job compared to a human
A dyson handheld does a better job cleaning under your bed, and you can clean the couches fairly easily with it.
Your $1000 is better spent on a miele canister and a dyson v7.
They make sense within a pretty narrow use case: You have a largish room with hard floor, which is usually kept uncluttered. You don't have pets.
We use it for a large kitchen and it's okay. Run it before leaving the house and the floor is clean when you get back. Who cares how long it takes or that it is noisy or whatever. It does like 75% as good as a human would and that's good enough most of the time.
But it's true that a part of the time you save is spent on maintenance instead.
>part of the time you save is spent on maintenance instead.
There's a quote I always remember while taking 30 seconds to clean the dishwasher:
"The world of A.D. 2014 will have few routine jobs that cannot be done better by some machine than by any human being. Mankind will therefore have become largely a race of machine tenders." -- Isaac Asimov, 1964
I've had a Roomba and Neato. They've always picked up more dust after using a different vacuum cleaner. There's some bit of necessity with a dog that sheds though. And there is something relaxing coming home to freshly vacuumed carpet.
Even calculated at only 10 minutes of time savings, that's 1040 minutes a year at 2x a week; giving me back an entire waking day of time a year (17/hours).
I don't know how you're measuring "half as good". But I have a high end Dyson. When I first got my Roomba, I cleaned my office floor with the Dyson as good as I could going over it several times. Then I set the Roomba loose on it to see if it pulled up any dirt. When it was finished the bin was completely full. I assume not all of it came from the edges of the carpet where the trim meets and that since it was moving much slower it actually pulled up dirt the Dyson had missed.
And don't misunderstand this at all, I absolutely love my Dyson.
I use a knockoff Roomba from Eufy. The benefit of it is that it can run while I make dinner. I don't care if it's imperfect or takes forever - it's an improvement and I don't have to monitor it.
Ours handles cat toys and random junk fairly well, usually doesn't eat cables and does surprisingly well on our hardwood floors. It also gets the lip under the cabinets that usually requires me to get the hand vac, which is nice.
My take is this is the biggest selling point here - it's not going to replace your vacuum cleaner entirely, but it surely makes it easier to keep a certain level of cleanness.
> Downside: Haven't met a robot able to detect cat poo/barf yet.
That's why I leave some dirt inside when emptying the container. By pure coincidence last time there was enough of this residue to mix with the cat barf and form a neat, dry, cigar-shaped... thing. I had to clean the brush area and the brushes themselves though.
I always think “It’d be nice to have a Roomba”. Then I remember I live in a three storey house with lots of rugs, a single step in the middle of the hallway, children who leave toys in the front room and rugs everywhere.
I own one 8XX model (~3 years old, I think) and can attest that "Roomba proofing" the spaces prior to operation and having furniture with high clearings and basic shape legs are kind of needed to prevent it getting stuck.
This, minus the kids. As much as I hate vaccuuming I'd love to have a roomba, but the hassle of roomba proofing the house is way too great. I live in a rather farmlike environment so the dirt in my house can range from the tiniest dust particles to anything the cats will drag inside. I would not like to have the roomba smearing around parts of a dead bird all over, I also have rather high doorsteps and I don't feel like having those roomba ramps all over my house. So at the moment, the obstacles of having a roomba are much greater than just wipping out the good old vaccuum cleaner and use 45 minutes of that day to clean on my own.
It sounds like you need a different grade of Roomba, one that avoids really nasty things and can gracefully fall all the way down stairs to go from the top to the bottom of the house.
A fair amount of machine learning is needed if a Roomba is able to do things like identify 'nuggets' left by pets, 'cat sick', mystery yellow liquids, slugs by patio doors and beheaded shrews. We as humans have no problems identifying such horrors so long as the lights are on, but for a Roomba things are not so easy.
A version of a Roomba needs to positively look for such unsavouries and report back via the app that the house is clear. This is not that far off what the other iRobot robots do - the military ones - so hang in there, one day there will be a Roomba for you.
Toys and rugs are no match for our Neato, but it eats cables and unfortunately scratches the wooden floor with the metal plated contacts. So it is not possible to leave it on automated schedule. Every time the room must be cleared of cables.
3 story house, 2 kids and several rags here - our Roomba changed our lives. It forces you (and the kids) to clean all toys in the evening, it will clean the living room every night and then I carry it to other stories as I see fit. I have also adjusted all our furniture to either sit on the floor or be "roomba tall" - and now I can not imagine how people live with that dust under a sofa or a wardrobe :)
Yes, the toys thing is great - we have the cleaning set at 10am every morning and you've never seen a kid scramble to tidy up so quickly as when she hears the robot start up!
I also have a couple of kids and have thought this would be a great side effect of getting a Roomba, but then I always think, after I clean up all the toys running an upright would only add 5% more effort to my total clean up time. Then again maybe that's a nice forcing function.
Been using Neato XV for some time; the way it navigate the rooms is a lot better than my previous Roomba. But apparently Roomba caught up navigation-wise. Dirt disposal definitely looks interesting as cleaning is an issue if you have pets at home.
The robot vacuums cleaner technology seems to finally be useful enough, although it still has improvements to make.
I've just recently acquired a $400 RoboRock s50. Comparing its features with the $900 Roomba, I can say it's worth it. The s50 tries to mop the floor albeit just with water. It is very good with carpets though. I ordered it from China and I was anxious if I would have to use translation software to understand the manual but thankfully the s50 has the voice commands, manual and the mobile app translated into English. The scheduled cleaning is really worth it.
Make sure to keep away cables and curtains that reach the floor. Once, the wind got the curtain around the cleaner and the sensor detected obstacles everywhere and stopped. The app notified me that I need to remove obstacles which I did promptly and continued to clean until it finished. I am not really sure what's stopping Roborock to make the maps permanent like the Roomba.
Also owner of a Roborock S50, can confirm, bang for the buck is much better than with Roombas. My only gripe is indeed that the maps are not persistent. In fact it's even more absurd, they are persistent in the sense that you can see previous maps, they are stored, but it doesn't take them into account to optimize the next cleaning session and, more importantly, you cannot use them for the ever useful zoned cleaning function. Why? It seems that all the components needed are already there, and looks like it should be a trivial software update to enable this.
Anyway, an amazing vacuum and absolutely great for the price.
I also recently bought one and noticed the same, but then I think I understood the reason for not persisting the maps, when the robot found a closed door: after I opened the door, the robot just wouldn't cross it because of the (temporarily) stored temporary obstacle.
Storing a map also means you need to have a very good way to update it.
Also own a s50, most annoying thing for me is that you can't do reverse clean up zones, ie. Tell it where CAN'T clean.
Also, Saving zoned cleanups would be a time saver.
And rotating of the map for when it misalign them, which it keep doing sometimes even after following their instructions.
This is their answer when I emailed about rotatable map;
"Dear
The map takes the environment near the starting point as a reference. If there are some things near the starting point, it will cause the Robot to draw a deformed map. In this case, please move the dock charger to an open area to solve this problem. Please ensure to retain 0.5 m or more in two sides and 1.5 m or more in the front of the dock charger. "
> Super cleaning like a tornado. Nothing is left after a tornado, you can walk on the floor with bare feet
> Easy to climb 2cm of Obstacle. Nothing can stop the robot's steps
> With a strong power, the robot absorbs all the debris, steel beads, soybeans, dust, millet or any other wastes all in a sudden (is that's what typically lying around?)
GearBest suggests that while buying the S50 robot, I should also adds two accessories and save: an Espresso machine and another portable Espresso machine.
These are underrated. I have the original Xiaomi and it works great. I sold my $700 Neato botvac connected and paid half for this one. It isn't even the best and can do our whole house with no intervention consistently.
This was one reason I got a stupid robovac (Eufy) with no network capabilities, just a remote control. It doesn't even try to map my floor, but rather just zooms around randomly. Frankly, the cheap price didn't make the decision any harder either.
Works surprisingly well, you just have to give it a bit of time and let it do its thing in peace.
Like most other "privacy" policies, this was written up by lawyers.
Note that they make no clarifications as to what is defined as "customer's" data, only that "Clean Map Reports" are not shared with 3rd parties. i.e. once the map data is stored on their servers, does it still belong to the customer or to Roomba? Technically if it becomes Roomba's data they are not violating their privacy policy. Additionally, my reading of it suggests the latter, because the policy says they will honor requests to delete it. A better policy would be to automatically delete it after 90 days, for example.
A good argument could be made to use that dataset for improving the product, but IMO the risks of de-anonymizing it and exposing that data far outweigh the benefits.
Stay away from Neato cleaners, they sell them with a type of battery that only lasts for 6 months (that is what the guarantee claims), so prepare to spend $45x2 every year.
I had mine for 2 years without battery replacement and it still cleans my house in roughly the same time as it did before and using the same number of total charges (one to start with and one extra after some time).
That was the case in my older neato, 5+ years away. But my botvac had performed flawlessly for the last two years (and the app app works amazingly, including saving older cleanup maps, sharing the dirt in status, etc)
Weekend project: reuse the persistent map data for VR by creating a virtual environment containing in-world obstacles at the same place as physical obstacles.
Tangentially related idea[0]: A programmable robovac simulation environment. Can be gamified or an actual research environment, think "Kerbal Vacuum Program".
Roombagotchi: You have to buy special Febreze scented color coded confetti, and compulsively sprinkle it all around the floor to feed, exercise, entertain, play fetch, perform tricks, and constantly keep your robotic pet alive and happy.
The most impressive thing about Roomba is its physical durability. My Roomba must have traveled many many miles over the past 5 years and it looks battered like WALL-E. I have replaced all the peripherals over the time but the gear goes strong and my Roomba still does a great job.
While my Nilfisk hoover requires me to operate it manually, it is nearly 30 years old and still works perfectly. Only thing I've had to replace is the power cord which started to come apart.
Having to replace a bunch of Roomba peripherals after under 5 years and calling it 'durable' is laughable.
Sounds great, but unless they offer a model with no internet connectivity it'll be a no from me. I've got an older Roomba that only has 802.15.4 and IR connectivity, with no cameras. That feels a lot safer.
I had the same concerns, particularly since I bought the 980 with the camera on it. I ended up creating an access point with my phone to set it up, and now I have an old beaten phone as controller for it that has the same access point but no internet. Getting to that point was a pain, and the reason why I'm currently convinced that this will be my last iRobot.
My concerns aren't about what the Roomba would do to other devices on my network, but what kind of data it's sending out and how that data could be used. That can't be mitigated by just having the thing connect to a separate network.
I'd rather not have a 3rd party have floor-plans of my house and continuously images of every inch of my home. I'd rather all of that data stay local to a device that I own and control.
I believe the grandparent comment said he cut off internet access for it and the old phone he uses to control it. It sounds like it works on a strictly LAN connection unhindered.
It does sound like a lot of trouble for a vacuum cleaner, though.
I mean, how clean do people need their floors to be? I vacuum like once every couple weeks and it’s not like I’m getting sick or my floors look gross.
We have ours run three times a week due to the amount of dust and dirt our two dogs track into the house. And even then the floor doesn't look pristine.
Carpets I've had could go months without noticing anything, but the amount of dust bunnies and random crap of unknown origin that got stuck to my feet shocked me the first place I got with a hardwood floor.
My take on their comment was that they do have a separate wifi connection that is just their phone, vacuum device and an access point with no internet on any of them that is physically separate from the rest of their network.
I'm with you though. No way would I EVER get anything like that that connected to the internet. I can't think of any benefits that would remotely come close to justifying that behavior.
Even if it meant insurance would drop to $0, one "leak" and anyone with a bit of resources can get a nice up to date map of your house, when you're around, what lights are on, when their on, etc. It would be much worse than that site that scarfed social media and documented when people were on vacation, left the house or whatever.
> I'm with you though. No way would I EVER get anything like that that connected to the internet. I can't think of any benefits that would remotely come close to justifying that behavior.
Next up: it'll just connect to 4G. I hope that doesn't happen, but it just seems all too likely.
Im hoping not, since it's more expensive in manufacturing, and people who go through the trouble of not having their Roomba on the internet are more edge cases than normal customers. Plus consumer watchdogs and privacy groups will start to get on your nerves.
The phone opening the access point for the router doesn't have a SIM card, and is just there to provide a WiFi connection for the Roomba to be able to control it. That way, the vacuum can't access the internet but I can still control it.
> That way, the vacuum can't access the internet but I can still control it.
Interesting! The camera functionality still works without an internet connection? I'd always assumed that some of the processing was getting offloaded to the "cloud," since I couldn't see another reasonable need for the wifi connection.
What if when you died you were shown a list of all the things you worried about in life for no reason?
I suspect being concerned that a hacker will obtain and sell or use robot vacuum floor plan data facilitate a physical robbery would be high on that list.
Probably the same risks as with any IoT device. Any device is dangerous to some extent if it:
(1) has internet connectivity,
(2) has poor security (due to bad design/implementation and/or lack of updates), and
(3) is on your internal network.
It may not be a valuable target itself, but it can serve as a beachhead to launch further attacks on other devices on the same network. Assuming you rely on some firewall-style network security (a notion of inside and outside), which most people do.
So basically, it's dangerous in the same way that laptops, cell phones, tablets, video game consoles, security cameras, ebook readers, and printers are all dangerous?
It's the same type of risk, but the degree of risk varies.
Laptops run operating systems written by and supported by software companies. Because software is one of their core competencies, they are more likely to understand vulnerabilities and release updates to patch them in a timely manner. And they are more likely to continue releasing updates for years.
IoT devices (security cameras, etc.) are usually made by companies whose core competency is hardware. They tend to think in terms of making money by selling hardware, and the software may be viewed as overhead and of secondary importance. You are lucky to get any updates at all, and updates tend to cease long before the hardware stops working.
I also think there's a higher chance of smaller IoT device makers selling your data to third parties. On the flip side you could probably add in any device from Alphabet companies to collecting, analyzing and selling that data.
Oddly, the Roomba has been great so far getting is to tidy the house. Some errant cables and poor laundry habits are the only thing that breaks the daily run of these things.
I have an older random walk Roomba and a newer Neato BotVac. The Botvac is nearly usable, but is also battling and eventually getting stuck on a dustpan in a corner more often than not.
The Botvac get caught on a lot of random junk and is beeping miserable most days, and would be much better if it backed down and marked the area to be avoided.
I made some little ramps from a piece of wood I bought at a hardware store. They were curved pieces designed for an oak table or something. One side low to the ground, one side 3cm high.
I bought a couple sets, sawed them to the right length for each frame, and varnished them so they look like the floor. Now the robot can go anywhere.
I didn't have to attach them to the floor or anything, they mostly stay in place.
This will be a big step up from what I have now where my Roomba immediately heads to one never-travelled corner of my apartment and does its best to get stuck there.
It also has an uncanny knack for doing stuff like getting into the bathroom when I'm taking a shower, then banging the door shut behind it and proceeding to thrash around until it dies. Or heading out a just-opened front door and trying hurtle itself off my landing.
Once after a conversation with a QA coworker, I tried cornering a Roomba with the walls and my feet, hoping it would stop operating. I watched it continuously bumping into the walls and my feet, for 5 minutes.
I have to say: I love my Roomba; this device meant I had to use an upright once in the past three years - when I moved out of my previous apartment.
And one of the things that I love about it is that it's simple. It uses a random walk for its cleaning algorithm, which means that it gets its job done on the average if you let it run long enough.
Random walks are easy to understand. I can re-position furniture in my apartment to make it a bit more Roomba-friendly, or I can shift an office chair before a run to make sure Roomba gets under the table when I'm going away for work.
The nice thing about it is that there's nothing that can break in this scheme. It's dead simple. If Roomba doesn't get to a corner, I know exactly why, and what to do about it.
With persistent maps, that aspect is out of the window. If a Roomba doesn't map your apartment right.. good luck figuring out how to deal with it.
KISS, I'll keep repairing my Series 600 until the spares run out.
Seems like you've invented hypothetical problems to complain about. The existing Roomba has none of the issues you speculate it has, and as long as there is a way to force-delete the persistent map it is unlikely the new one will either.
>Seems like you've invented hypothetical problems to complain about. The existing Roomba has none of the issues you speculate it has
Current one doesn't feature persistent maps, either.
The new one does, and is app-controlled. 10 years into the future, you will still be able to find spare parts for the now-10-year-old-Roomba, but good luck with the app.
I'll bet it still can't avoid getting caught on rug tassels, or wedged under furniture that's exactly the wrong height, or have long hair completely bind up the roller.
122 comments
[ 3.1 ms ] story [ 93.9 ms ] thread- It is obviously useful to plan cleaning.
- Once calculated, storing them should be cheap.
- Detecting that the stored map is wrong if you manually move the Roomba to another location should be an issue; just ignore the stored map after too many inconsistencies with current reality.
And Roomba only moves in a plane as a result of its (slow) actuators, whereas an iPhone X tries to do hyper-accurate 6dof real time position, from a very quick external source (your hand), for augmented reality. Thus the sensor and computation cost premium. But that does not apply to the Roomba.
Apple's strategy too?
Third parties implement the feature and then you pay $9-$49 for the software. Sometimes you can add $4-$19 for a software update for every OS update.
It's a great ecosystem for developers.
If they manage to solve that case, then the benefit of retaining the map is less compelling. By definition, you can do an adequate job without it. So then it's a question of what you gain in cleaning performance/efficiency vs. the extra complexity and any potential problems it might bring.
Can't say how they work on carpet, but for wood flooring, they are most excellent!
A dyson handheld does a better job cleaning under your bed, and you can clean the couches fairly easily with it.
Your $1000 is better spent on a miele canister and a dyson v7.
We use it for a large kitchen and it's okay. Run it before leaving the house and the floor is clean when you get back. Who cares how long it takes or that it is noisy or whatever. It does like 75% as good as a human would and that's good enough most of the time.
But it's true that a part of the time you save is spent on maintenance instead.
There's a quote I always remember while taking 30 seconds to clean the dishwasher:
"The world of A.D. 2014 will have few routine jobs that cannot be done better by some machine than by any human being. Mankind will therefore have become largely a race of machine tenders." -- Isaac Asimov, 1964
Even calculated at only 10 minutes of time savings, that's 1040 minutes a year at 2x a week; giving me back an entire waking day of time a year (17/hours).
And don't misunderstand this at all, I absolutely love my Dyson.
Ours handles cat toys and random junk fairly well, usually doesn't eat cables and does surprisingly well on our hardwood floors. It also gets the lip under the cabinets that usually requires me to get the hand vac, which is nice.
They're great at keeping pet fur and "everyday dust" at bay, though. It's easier to clean your house when the baseline is already taken care of.
Downside: Haven't met a robot able to detect cat poo/barf yet. Gotta be careful...
> Downside: Haven't met a robot able to detect cat poo/barf yet.
That's why I leave some dirt inside when emptying the container. By pure coincidence last time there was enough of this residue to mix with the cat barf and form a neat, dry, cigar-shaped... thing. I had to clean the brush area and the brushes themselves though.
A fair amount of machine learning is needed if a Roomba is able to do things like identify 'nuggets' left by pets, 'cat sick', mystery yellow liquids, slugs by patio doors and beheaded shrews. We as humans have no problems identifying such horrors so long as the lights are on, but for a Roomba things are not so easy.
A version of a Roomba needs to positively look for such unsavouries and report back via the app that the house is clear. This is not that far off what the other iRobot robots do - the military ones - so hang in there, one day there will be a Roomba for you.
Anyway, an amazing vacuum and absolutely great for the price.
Storing a map also means you need to have a very good way to update it.
This is their answer when I emailed about rotatable map;
"Dear The map takes the environment near the starting point as a reference. If there are some things near the starting point, it will cause the Robot to draw a deformed map. In this case, please move the dock charger to an open area to solve this problem. Please ensure to retain 0.5 m or more in two sides and 1.5 m or more in the front of the dock charger. "
I'm struggling to think of any open space that large within my apartment. 3 m² is about 6.5% of my apartment's total area.
> Super cleaning like a tornado. Nothing is left after a tornado, you can walk on the floor with bare feet
> Easy to climb 2cm of Obstacle. Nothing can stop the robot's steps
> With a strong power, the robot absorbs all the debris, steel beads, soybeans, dust, millet or any other wastes all in a sudden (is that's what typically lying around?)
GearBest suggests that while buying the S50 robot, I should also adds two accessories and save: an Espresso machine and another portable Espresso machine.
Works surprisingly well, you just have to give it a bit of time and let it do its thing in peace.
Like most other "privacy" policies, this was written up by lawyers.
Note that they make no clarifications as to what is defined as "customer's" data, only that "Clean Map Reports" are not shared with 3rd parties. i.e. once the map data is stored on their servers, does it still belong to the customer or to Roomba? Technically if it becomes Roomba's data they are not violating their privacy policy. Additionally, my reading of it suggests the latter, because the policy says they will honor requests to delete it. A better policy would be to automatically delete it after 90 days, for example.
A good argument could be made to use that dataset for improving the product, but IMO the risks of de-anonymizing it and exposing that data far outweigh the benefits.
0 - https://github.com/cretz/software-ideas/issues/89
Having to replace a bunch of Roomba peripherals after under 5 years and calling it 'durable' is laughable.
I'd rather not have a 3rd party have floor-plans of my house and continuously images of every inch of my home. I'd rather all of that data stay local to a device that I own and control.
It does sound like a lot of trouble for a vacuum cleaner, though.
I mean, how clean do people need their floors to be? I vacuum like once every couple weeks and it’s not like I’m getting sick or my floors look gross.
Carpets I've had could go months without noticing anything, but the amount of dust bunnies and random crap of unknown origin that got stuck to my feet shocked me the first place I got with a hardwood floor.
I'm with you though. No way would I EVER get anything like that that connected to the internet. I can't think of any benefits that would remotely come close to justifying that behavior.
Even if it meant insurance would drop to $0, one "leak" and anyone with a bit of resources can get a nice up to date map of your house, when you're around, what lights are on, when their on, etc. It would be much worse than that site that scarfed social media and documented when people were on vacation, left the house or whatever.
Next up: it'll just connect to 4G. I hope that doesn't happen, but it just seems all too likely.
Interesting! The camera functionality still works without an internet connection? I'd always assumed that some of the processing was getting offloaded to the "cloud," since I couldn't see another reasonable need for the wifi connection.
I suspect being concerned that a hacker will obtain and sell or use robot vacuum floor plan data facilitate a physical robbery would be high on that list.
In what way?
(1) has internet connectivity,
(2) has poor security (due to bad design/implementation and/or lack of updates), and
(3) is on your internal network.
It may not be a valuable target itself, but it can serve as a beachhead to launch further attacks on other devices on the same network. Assuming you rely on some firewall-style network security (a notion of inside and outside), which most people do.
Laptops run operating systems written by and supported by software companies. Because software is one of their core competencies, they are more likely to understand vulnerabilities and release updates to patch them in a timely manner. And they are more likely to continue releasing updates for years.
IoT devices (security cameras, etc.) are usually made by companies whose core competency is hardware. They tend to think in terms of making money by selling hardware, and the software may be viewed as overhead and of secondary importance. You are lucky to get any updates at all, and updates tend to cease long before the hardware stops working.
Right now I have to spend thirty minutes cleaning before the roomba can even run.
(Consider this Prior art so no one tries to patent such a simple idea)
The Botvac get caught on a lot of random junk and is beeping miserable most days, and would be much better if it backed down and marked the area to be avoided.
I bought a couple sets, sawed them to the right length for each frame, and varnished them so they look like the floor. Now the robot can go anywhere.
I didn't have to attach them to the floor or anything, they mostly stay in place.
It also has an uncanny knack for doing stuff like getting into the bathroom when I'm taking a shower, then banging the door shut behind it and proceeding to thrash around until it dies. Or heading out a just-opened front door and trying hurtle itself off my landing.
I think my Roomba may be trying to off itself.
Getting stuck in unimaginable places.
Unusual desire to frequently visit particular places for unknown reasons.
Going into the bathroom, only while you’re showering.
Banging on the door because you’re behind it.
Heading out just-opened doors, only to quickly return.
I look forward to the version 2.0’s 95% accurate dust disposal system.
I let go of it eventually.
And one of the things that I love about it is that it's simple. It uses a random walk for its cleaning algorithm, which means that it gets its job done on the average if you let it run long enough.
Random walks are easy to understand. I can re-position furniture in my apartment to make it a bit more Roomba-friendly, or I can shift an office chair before a run to make sure Roomba gets under the table when I'm going away for work.
The nice thing about it is that there's nothing that can break in this scheme. It's dead simple. If Roomba doesn't get to a corner, I know exactly why, and what to do about it.
With persistent maps, that aspect is out of the window. If a Roomba doesn't map your apartment right.. good luck figuring out how to deal with it.
KISS, I'll keep repairing my Series 600 until the spares run out.
Current one doesn't feature persistent maps, either.
The new one does, and is app-controlled. 10 years into the future, you will still be able to find spare parts for the now-10-year-old-Roomba, but good luck with the app.