Correction: Samsung released their "Navibot" in Europe in 2010 which was one of the first "Vslam" robots. It beat Dyson by 4 years! Also someone in the comments here made an excellent point that Lidar bots have more moving parts and thus is less reliable over the years. I have heard of laser motors needing to be replaced after a couple years of heavy use, so it is something to consider as well.
Yes LDS motor failure indeed it is something to consider, but if i were to wory about LDS failure i'dd worry more about the failure of the laser emiter or capturing diode failure. LDS motors are VERY easy to replace and VERY cheap (for Xiaomi for example it costs around 8$ on aliexpress). The laser and diode cost around the same ammount but the problem is instaling them. Recalibrating them is a nightmare (without proffesional factory tools). I had the error 1 LDS sensor error on the Roborock S5 and fixed it (dissasamble the motor and all) in under 10 minutes, but if it were the laser or diode i would have had to buy the entire LDS alltoghether (around 80-100$). Luckily most of the time the motor will go first and i know people that use the Xiomi 1st gen since it was released with no problems.
@@vennic We need to have like a 'False wall' setting, then the roomba will actively try to drive at the barrier we designate, and then it maps out what it can.
Thats is not an issue of "lidar" robots but the programmer who decided to use the distance information of the lidar to not drive into objects. The same functionality is possible if one uses a camera with depth information. One could only use the lidar for navigation and drive bases on the bumper sensor👌
I’m finding this an issue with Ecovacs T8 cleaning under beds. The lining of the underside of bed sags. The “go go zone” would be handy, like in this zone it only uses bump stops. Or, a mode to make the reliance on the camera and bump zones and not the laser. It seems like the hardware is already there, it just needs a new line of code in the app etc.
You also forgot to mention the privacy aspect. LIDAR can only record distances whereas a camera is a camera. "Smart" devices are famously easy to hack into especially on the budget end, so having a camera that can go anywhere you tell it to can be a problem. Not saying that it has or hasn't happened yet, but it's definitely a possibility.
Lidar is getting more advanced by the day. Today's lidar system can make a very high resolution point cloud of the environment so privacy is no longer a good argument.
@@ZeroInDaHouse it is still 2D though. All your gonna get if you hack into a Neato robot or something similar is a floor plan, which you can probably get online anyway.
The truth of the matter is that Neato holds the US patent for the use of lidar on robotic vacuums. This is the reason Irobot and Shark "choose" not to use lidar.
That's silly. Apparently you can patent everything in the US as long as you are the first one to do so. iRobot holds thousands of patents, including the base idea of automatic cleaning robots, as well. So every manufacturer has to make sure to avoid the most simple things described in their patent.
@@brandon_nope Are they officially selling in the US? Because where I live they only sell on ebay and ali, send from china or some other foreign country.
Can you do a video on maintenance of Robot vacuums? Their long terms costs of ownership, replacement part availability for midrange vs high end, and costs for replacement “consumable” parts(Hepa, Brushes, Rollers). Example, My Roomba spends a lot of time on hard tile floors and it seems like the dual rollers might be experiencing more wear than a brush based system.
I have multiple original xiaomi robots. The only maintenance I do is vacuuming the HEPA filter with a regular vacuum cleaner. The brushes don't seem to wear much. But it should be noted, that it can be hard to buy original consumable parts.
He recently reviewed one with a forward camera to avoid obstacles, which would solve the feet problem, but yeah voice control to tell it to go clean other rooms would be neat... and not hard to implement.
@@CorporateShill Not so much with the neatos. Neither respond directly to the f bomb, although the d80 generally seemed more responsive to being kicked than newer models.
The most important thing you said during this entire video is LiDAR is better. I will second that notion and say there is no competition. LiDAR in my opinion, is 10 times better than cameras. My $225 DeeBot 930 from eBay can run circles around my $1200 irobot Which constantly get stuck in the no vacuum zones
Great comparison! Thanks for this nice video. I own a 2007 Roomba 581 and later I got a Neato XV-21 from 2012. I use both daily. The Roomba gives a lot of headbutt in the furniture but captures a lot of dust, precisely by using its chaotic navigation, passing several times in the same place. The rotating brush helps a lot too. Already Neato is very delicate, it does not hit the furniture or scratch the legs of the chairs. It doesn't capture so much dust because it lacks a rotating side brush. I bought these two old ones with years of use and minor defects, but with some love, screws and glue, they still work well to date. About LIDAR. An advantage of Roomba is its ability to "check" whether the obstacle is true or not, as in the case of the bed "skirt". My Roomba slow down and force the passage very slowly. Without noticing resistance, it can get in and out under the bed. Neato can't do that, LIDAR sees the fabric as a solid barrier. But maybe Neato could solve this problem modifing the "no go lines" feature by using it in reverse. If the user could draw "yes go lines" around the bed, the robot would know that there is a transposable "obstacle" even if the LIDAR sees a solid wall. What do you think of the idea?
Good overview of the pros and cons of each technology! I'd just like to point out the distinction between localizing the robot (knowing where it is) and obstacle avoidance. The robots with V-SLAM systems use it only for localizing, so they typically use a bumper and/or IR proximity sensors to detect obstacles and avoid them. How well they avoid "pet accidents" and still push on overhangs to get in under your couch, depends on how correctly those other sensors are tuned. A 2D lidar, on the other hand, can offer both localization and obstacle avoidance, but only at the specific height of the lidar sensor. That's why lidar robots bump into lots of stuff, or need bumper/IR proximity sensors to compensate for its blind spots. This is why we at Electrolux chose a third path for our Pure i9 robot, launched in 2017: It has a 3D lidar with no moving parts, so it can localize itself like a lidar robot, while also seeing obstacles at any height and with high resolution. This gives a lot of data to analyze what it sees, such as measuring the height of thresholds before climbing them, avoiding socks on the floor and even seeing the edge of a staircase or carpet and cleaning along it.
Exceptional video on the difference of vslam camera based robots VS LIDAR. I agree that camera can offer more data once the hardware allows it. Maybe in the future they will incorporate 3D Lidar systems? Anyways camera based systems tend to be more durable since they have no mechanical parts. Keep up the Amazing Work 👍🏻
We have Purei9 from Electrolux. It is a laser based system. But instead a spinning laser on top it is placed in front with two laser scanners do provide a 3d laser image. There you have the no cons which you have explaned. Poo -> no problem, no light -> no problem, more image information -> check.
Speaking about user experience and the global feeling you have, my 2015 Neato is way softer than my brand new Roomba. The roomba keeps bumping everywhere as if it had no cartography at all. It is very stressful to see it force its way on cables and fragile stuff... A LIDAR in that case seems to be way better than any VSLAM. The hybrid system of course seems to be the most intersting
I have a vacuum bot with LiDAR in it, which is perfect cleaning companion except when it comes to large mirrored wardrobes that I have, which wreck the sensors and it’s ability to accurately map the house. I was actually thinking of upgrading to Vslam tech bot, do they have a similar problem?
I had VSLAM based robot from MAMIBOT and returned it due to navigation problems in my house (f.e. I used it for two weeks and only once it found way back to charging base). It was not a bad robot but poor app and navigation problems were just horrible so decided to go with lasers. Xiaomi Viomi V2 PRO (with laser) that I bought later had no problems with mapping entire house right away, it always returnes to base and never gets confused during cleaning (even when I move it during cleaning). Viomi has much better app with "no-go" lines and all the good staff that is available via Xiaomi Home App (MAMIBOT VSLAM was unable to save any map in app after cleaning so did not create "no go" lines). For me laser is significantly more precise, reliable and many times offers better options in app.
You have precision floor mapping with lidar so what else do you need your vacuum to do? I argue against the concept of "future proof" because this really is the most i need out of a robot vacuum.
Vslam has another advantage. It lasts longer. In lidar you have a Camera, a Laser and a Motor to rotate that stuff at a higher speed. And most of the time, the motor fails in the first 2 years and has to be replaced. Cameras (vslam) on the other side, doesn’t have that problem
But, required processing power and cost of it, also in terms of batteries, is much worse with the camera. With LIDAR you barely need a computer. Raspberry Pi level hardware is more than enough. For camera you might need something like 15 Watts sucking Jetson nano. So it must be used at extremely low frame rate. I don't understand however why they cheap out on the motor/driver on LIDAR. Motor should almost never fail. The belt yes, but not the motor. However laser should see socks too. It's just that is mounted too high. Front facing camera OK, but that "fake wall" would be a "problem" too. Instead of camera stil, you can use cheap fixed lasers (range finders) or ultrasonic sensors. Yes, it would cost more than a bumper with switches. Yet none has now water or piss sensor...
On my Neato that broke I thought it was the little motor and bought a new motor. The lasers still would not turn. I tested both motors and with 5 volts dc they both turned fine. I discovered it was a broken navigation board. This is on a Neato D7 that cost me $900 and only ran 60 times before breaking down. For that reason I would never buy another Lidar robot. Too unreliable. This is why we have such a huge problem with waste in North America. Too many things are thrown out because you can’t get parts or the parts cost as much as buying a complete new one. The government needs to get involved with this if they want to really reduce pollution and landfill issues.
@@ds99, so a board not specific to lidar robots broke and you are never buying a lidar vacuum? Makes sense. Especially considering that camera robots have bigger and more complicated navigation boards.
@@ds99 Louis Rossmann fights here on YT for the right to repair just for this kind of reasons. In your case the culprit was probably (I guess) a bad solder job, a blown mosfet costing perhaps cents to replace. And then the consumer is left with absurdities as if you have to replace the house to change the dead light bulb. But the same could happen with the camera or anything else. Because the market is driven to replace not to repair. I have a dead temperature sensor (AC) and a small speaker (for blinker sound) in my VW SUV, less than few dollars total worth, repair is not available, replacement is in thousands. And then as you said.... I could not agree more.
Lidar have a little trouble with mirrors. It’s reflect laser beam and I have double size of my room it the app, or artificial map objects if it’s small mirror.
Being an IT professional, I find it odd how willing some people would be to having a motorized camera, connected to wifi, running around their house taking video/pictures of everything... and especially considering how some of the user agreements and ToS govern under the people's "Republic of China" and having network connections out of the US country (Ecovacs...) ... I'll pass, thanks. Probably nothing major but the implications are astounding to consider. I don't need the POSSIBILITY of someone seeing me (or my wife!) in my underwear - I'll keep my privacy as much as I can, thanks... Another reason all of my webcams are blocked off too.
The Lidar robots should do a mapping run where they use also use the bump sensors to see if something is actually a wall or just a light piece of fabric where they could go through so that the floor under the sofa gets cleaned as well. I also think that right now you should not buy a camera robot. It might be the better technique in the future but the processors that are installed in the current robots are not fast enough to handle artificial intelligence and flawless navigation. You would have to buy a new one with a better chip.
I've tried Shark IQ and Roborock S5 (VSLAM and lidar). Lidar is definitely better and VSLAM was completely useless on the Shark. The Shark never made a map of my house and never cleaned in consistent straight lines. The Roborock got it right the first time and consistently cleans straight and well
erm, the first robot vacuum was electrolux trilobite from 2000, which had random navigation, and first camera robot was Samsung NaviBot sr8855, which is today old like Roomba 500 series. so older than dyson and Roomba 980.
I went through a Roomba, a Dyson and a Neato, but now I found the Lidar based, Chinese Smart 360 S6 and it is so much better and way cheaper than e.g. a Roborock. App control works flawlessly too.
Coming from team Tesla I would say camera is much better. But putting it to the sealing is stupid - except if you clean sealing. Best would be black and white stereocameras in front of device. That's all. It's not that complicated to understand. Just 2 cameras.
I got my wife the S5 Max for Christmas...It is awesome. The mapping is incredible. I spent what seemed like hours deciding what to get. I think it was a coin flip between the Deebot 950 and the S5 Max. It seems like it has all the features/components of the S6 but with a bigger mop reservoir. Although i didn’t see many reviews specifically on the S5 Max, I went for it...we mainly have hard floors but even on carpet in the bedrooms, it’s great. The brush doesn’t jam and it kicks on its boosted suction. 👍
wonder if they can add an "ignore barrier" function to the app that ignores obstacles in certain areas unless they trigger bumpers. that would deal with issues like false bottoms on couches.
there is one more thing You should mention - privacy. Considering how bad those robots are secured - they can be quite easy hacked into spying (which some of them are already doing...). So not only Your house plan and any sound from home is send away, but also a pic/vid feed can be :( So till full maping/AI can be done only in vacuum(offline), not shared to cloud for AI checks, I would avoid them.
I don't believe that any of the VSLAM robots share anything with the cloud. If they did, why would they need such powerful processors to interpret the data from the camera if it was all done on a server somewhere else?
Hello thank you for your videos they are great and have helped us out a lot. I have a question looking for a robot vacuum that is able to go under some coaches that we have that have cloth skirts. We also like the ability to use a smart phone app. that will do a mapping of the house. I don't know if there is also one that systematically cleans instead of bouncing all over the place. I think I am looking for the impossible do you have any recommendations thanks.
So we are still working on our search tool - but try looking at some models and their features here: www.vacuumwars.com/types/robot-vacuum-cleaners?p46=2460,2351,2364,2338,132,133,134,135,136&p60=178&p37=104 In that link I have preselected all the smart ones that use Wifi and have no-go zone (which means their is an app involved). It's not perfect but it's a good start.
@@rebeccakalyani2486 yes I have two, they are both mapping and working great after a year, although there were a couple issues that were eventually resolved
hi. is this true... i read that lidar sensors break more then the others due to the movement they make? because they are moving parts. thanks in advance.
you are right and wrong about vslam. vslam is faster and gets a better maping but the trouble is you need a good camera programmed properly and aimed right. if thats dun its will get a better capture of whats around it and faster time since its now useing a lasor that has to spin a rotation each time. the down side tho is you have to get a verry expensive camera that has a super whide angel lence and high resalution to see father distance as the laser can do a larger distance no matter how great the quality. also you have to rember the laser can be interuped buy shiny objesct and buy things that will absorb the light.
It would be nice if the user could choose to turn the A.I. on or off. Some users have complained that it's too sensitive and wont' vacuum places where they would like.
The biggest reason for a company to choose cameras is cost. You can source a smartphone camera for under a dollar and then write vSLAM software that costs zero dollars to duplicate. Lidar is more expensive to manufacture, less reliable, and requires more resources for design / repair. Whether the company passes these benefits onto the consumer in the form of e.g. better cleaning power for the same price is up to them.
is lidar ok with floor length mirrors or will the cleaner try to enter a room that doesn"t exist? If so, would setting up a no go area in front of the mirror solve the problem? Or should I limit a purchase to vacuum and mop cleaners with no lidar?
I love my 50 dollar super clean robot. However it gets stuck easy. Also got a irobot braava 380 and I find that it doesnt run very long even after hours of charging. Possibly a bad battery Considering getting a nice robot vaccum and the braava 240 mop robot.
Thiiiiiis is the 2020 Roborock S5 Max and it is one of the best robot vacuums yet. Today I will take an in depth look and show you all of its quirks and features, then I'm gonna see how it cleans and then I'm gonna give it a DougScore. For more of my thoughts on the S5 Max, click the link below or visit vacuumtrader.com/robosteer
Why not have all three? The top camera could be mounted on top of the LIDAR housing. It could aid the LIDAR in mapping the room. The front camera could still be good for object recognition & avoidance. Personally, I don't mind if a floor robot stays away from skirts & doesn't get stuck under a chair or couch.
With regards to more and more advanced processors (the 32-bit cpu part of the video), the first robot vacuum, Electrolux Trilobite, used a MC68332 32-bit processor, so it is kind of back to basics :-)
The 32 bits only refers to the main CPU bus architecture. Todays vacuums use ARM-based System-on-Chip controllers that are more or less the same power of a smartphone, many megabytes of RAM, and up to 8 gigabytes of flash memory. The MC68332 is extremely primitive in comparison. The only similarity is the 32 bit bus used in the architecture. Nevertheless it's an important stepping stone in the development of household robots.
@@nrdesign1991 I beg to differ. The Trilobite is in my opinion more advanced than most Roomba's (all less than the 900-series) and more advanced that the Electrolux vacuum release a couple of years ago. Regarding the ARM chips used, there is a big difference in speed and memory capacity between Cortex-A chips used in smartphones and Cortex-M chips used for controlling robots - one are good at interfacing people and one is good at interfacing hardware :-) Also, Cortex-A chips tend (there are exceptions like iMX.5, BCM2835[ARM11] and Allwinner A10) to become obsolete and unavailable quite fast, while Cortex-M chips (and embedded SOC's in general) are available for very long periods.
Great video. One thing I feel it lacks is that you relate everything to the future of robot vacuums being AI - but you don't explain what this "AI" will bring to the table. Robot vacuums already clean homes very well. They already have scheduling through apps, they already have go back to recharge, and continue, they already have no go zones, they already have bump protection, and fall protection. What benefits could AI bring, with the additional cameras in the future?
I always wish my Neato Botvac 80 would go over the same place twice. Of the dirt that it does miss, it always just misses it and never goes near it again. I know my Botvac is old tech, but sometimes I wish it could switch between Lidar and Vslam, like do one first, then the other automatically. Precise is not always better than random, but the two together I bet would be very effective. In fact, they should make a test robot vac that allows switching of the two systems or using a combination of both and see which is used more by testers in their homes.It would be neato to see the results.
Let's break it down a different way. Quick show of hands. How many people would OK with having an Internet-connected camera roaming around their house taking video footage of everything and everybody? Ya, I don't see too many hands up! For me, that's a pretty big reason to go with Lidar.
@@marcusy7421 I won't argue w/ that and don't get me wrong, i am not "paranoid" about it , at first i am semi joking in replying to you :D (WTH, i even see myself buying next gen deebot 960, roborock S7, S8? :p in the future) But the more i think about it, the more i can imagine how it can be used by "ordinary" criminal. Just imagine a thief using your vacbot just to checking / make sure that the house is empty. btw, on LIDAR vs VSLAM,,, i am on LIDAR camp :D
I have several furnitures with 9cm height.and only vslam robot can do the job.i bought yeedi that luckily only 1 cm away from bottom part of the furnitures.
Maybe review a panasonic stick vacuum. I heard their it vacuum is really good. Also maybe try the new black and decker stick vacuum. My father has it and he likes it so maybe a good one to try.
Maybe a little late to the party but I thought I'd let you know that the lidar robots basically use the same algorithm which is called slam. In my opinion you shouldn't call it lidar Vs vslam but camera.
You should always include a night more for your vacuuming test. In the winter night time come early and my Roomba does not clean when I need it to. Will try lidar bot next.
As soon as a lifespan of a robot vacuum is 2-4 years, it makes sense to focus on the technology that works now. And it is lidar. When VSLAM will offer more, then just buy a new electronic "pet" :-) For now this "future-proofness" doesn't mean much for the end customer.
I think a key idea for the lifespan of a robo vacuum has to do with the availability of replacement parts and the savvyness of the end consumer as well. For example, I can buy a new motherboard for the roborock S5 that I have, and have the technical knowhow on how to replace it should I have to.
The new Xiaomi Mijia 1S also has both camera and lazer navigation. I'm surprised you didn't mention it. Are you planning on reviewing this model soon? Love the channel by the way!
Great video! Love the channel, you guys convinced me to go buy a used Xiaomi gen 1 a while back and now looking for my next one haha. Btw what’s the song at the end? Hehe
Correction: Samsung released their "Navibot" in Europe in 2010 which was one of the first "Vslam" robots. It beat Dyson by 4 years! Also someone in the comments here made an excellent point that Lidar bots have more moving parts and thus is less reliable over the years. I have heard of laser motors needing to be replaced after a couple years of heavy use, so it is something to consider as well.
Yes LDS motor failure indeed it is something to consider, but if i were to wory about LDS failure i'dd worry more about the failure of the laser emiter or capturing diode failure. LDS motors are VERY easy to replace and VERY cheap (for Xiaomi for example it costs around 8$ on aliexpress). The laser and diode cost around the same ammount but the problem is instaling them. Recalibrating them is a nightmare (without proffesional factory tools). I had the error 1 LDS sensor error on the Roborock S5 and fixed it (dissasamble the motor and all) in under 10 minutes, but if it were the laser or diode i would have had to buy the entire LDS alltoghether (around 80-100$). Luckily most of the time the motor will go first and i know people that use the Xiomi 1st gen since it was released with no problems.
Can you test the Navibot?
Isn’t it vSLAM, not Vslam? (Pedantic arguments over stylized capitalization intensifies)
I think that is why Roborock sells replacement laser navigation modules, although I haven’t heard of anyone having problems with it yet.
Lol thank god someone pointed it out. 1st vslam bots in 2014? South Koreans released their Arte in 2012
Lidar units only need a "you can try to clean under this area" function. So bed and furniture skirts would not be an issue.
Go go lines
@@vennic We need to have like a 'False wall' setting, then the roomba will actively try to drive at the barrier we designate, and then it maps out what it can.
@@Tonatsi that could make things very interesting
Thats is not an issue of "lidar" robots but the programmer who decided to use the distance information of the lidar to not drive into objects. The same functionality is possible if one uses a camera with depth information. One could only use the lidar for navigation and drive bases on the bumper sensor👌
I’m finding this an issue with Ecovacs T8 cleaning under beds. The lining of the underside of bed sags. The “go go zone” would be handy, like in this zone it only uses bump stops. Or, a mode to make the reliance on the camera and bump zones and not the laser. It seems like the hardware is already there, it just needs a new line of code in the app etc.
You also forgot to mention the privacy aspect. LIDAR can only record distances whereas a camera is a camera. "Smart" devices are famously easy to hack into especially on the budget end, so having a camera that can go anywhere you tell it to can be a problem. Not saying that it has or hasn't happened yet, but it's definitely a possibility.
Totally! I was looking for this comment!
Lidar is getting more advanced by the day. Today's lidar system can make a very high resolution point cloud of the environment so privacy is no longer a good argument.
@@ZeroInDaHouse it is still 2D though. All your gonna get if you hack into a Neato robot or something similar is a floor plan, which you can probably get online anyway.
@@ZeroInDaHouse Even if you get a high res map, you would know, that there is a TV, but the cam would "see" the brand, the exact size and so on.
Yes, I heard people hacked into someone's Roommate and was saying on them...that is why I chose Lidar. He didn't mention the privacy issue.😳
The truth of the matter is that Neato holds the US patent for the use of lidar on robotic vacuums.
This is the reason Irobot and Shark "choose" not to use lidar.
That's silly. Apparently you can patent everything in the US as long as you are the first one to do so. iRobot holds thousands of patents, including the base idea of automatic cleaning robots, as well. So every manufacturer has to make sure to avoid the most simple things described in their patent.
Ok but what about roborock?
@@brandon_nope Are they officially selling in the US? Because where I live they only sell on ebay and ali, send from china or some other foreign country.
@@brandon_nope Not made in the US.
tmicecave Amazon. The roborock s5 for $499
Can you do a video on maintenance of Robot vacuums? Their long terms costs of ownership, replacement part availability for midrange vs high end, and costs for replacement “consumable” parts(Hepa, Brushes, Rollers). Example, My Roomba spends a lot of time on hard tile floors and it seems like the dual rollers might be experiencing more wear than a brush based system.
I have multiple original xiaomi robots. The only maintenance I do is vacuuming the HEPA filter with a regular vacuum cleaner. The brushes don't seem to wear much. But it should be noted, that it can be hard to buy original consumable parts.
Honestly, whoever can design one to stop running over my feet like a toddler or will go clean another room when I tell it to f*** off wins my money
vennic well said.
I have mine set up in a way that it only starts when no one is home.
He recently reviewed one with a forward camera to avoid obstacles, which would solve the feet problem, but yeah voice control to tell it to go clean other rooms would be neat... and not hard to implement.
vennic
Roborock/ any lidar robot is perfect for you
@@CorporateShill Not so much with the neatos. Neither respond directly to the f bomb, although the d80 generally seemed more responsive to being kicked than newer models.
The most important thing you said during this entire video is LiDAR is better. I will second that notion and say there is no competition. LiDAR in my opinion, is 10 times better than cameras. My $225 DeeBot 930 from eBay can run circles around my $1200 irobot Which constantly get stuck in the no vacuum zones
I’ve been waiting this for so long
Cool, I have been waiting to see how Ecovacs did with the AIVI before I decided how I felt about it all.
@@VacuumWars fch
Great comparison! Thanks for this nice video.
I own a 2007 Roomba 581 and later I got a Neato XV-21 from 2012. I use both daily. The Roomba gives a lot of headbutt in the furniture but captures a lot of dust, precisely by using its chaotic navigation, passing several times in the same place. The rotating brush helps a lot too.
Already Neato is very delicate, it does not hit the furniture or scratch the legs of the chairs. It doesn't capture so much dust because it lacks a rotating side brush. I bought these two old ones with years of use and minor defects, but with some love, screws and glue, they still work well to date.
About LIDAR. An advantage of Roomba is its ability to "check" whether the obstacle is true or not, as in the case of the bed "skirt". My Roomba slow down and force the passage very slowly. Without noticing resistance, it can get in and out under the bed.
Neato can't do that, LIDAR sees the fabric as a solid barrier. But maybe Neato could solve this problem modifing the "no go lines" feature by using it in reverse. If the user could draw "yes go lines" around the bed, the robot would know that there is a transposable "obstacle" even if the LIDAR sees a solid wall.
What do you think of the idea?
Good overview of the pros and cons of each technology! I'd just like to point out the distinction between localizing the robot (knowing where it is) and obstacle avoidance. The robots with V-SLAM systems use it only for localizing, so they typically use a bumper and/or IR proximity sensors to detect obstacles and avoid them. How well they avoid "pet accidents" and still push on overhangs to get in under your couch, depends on how correctly those other sensors are tuned. A 2D lidar, on the other hand, can offer both localization and obstacle avoidance, but only at the specific height of the lidar sensor. That's why lidar robots bump into lots of stuff, or need bumper/IR proximity sensors to compensate for its blind spots.
This is why we at Electrolux chose a third path for our Pure i9 robot, launched in 2017: It has a 3D lidar with no moving parts, so it can localize itself like a lidar robot, while also seeing obstacles at any height and with high resolution. This gives a lot of data to analyze what it sees, such as measuring the height of thresholds before climbing them, avoiding socks on the floor and even seeing the edge of a staircase or carpet and cleaning along it.
i literally laughed at "pet accidents" lmaooo
This guys deserves more subscribers
Exceptional video on the difference of vslam camera based robots VS LIDAR. I agree that camera can offer more data once the hardware allows it. Maybe in the future they will incorporate 3D Lidar systems? Anyways camera based systems tend to be more durable since they have no mechanical parts. Keep up the Amazing Work 👍🏻
Good point!
Oh hi lol just scrolling through the comments
My mom is soooo happy for the S6 I gave her 2 weeks ago :)
We have Purei9 from Electrolux. It is a laser based system. But instead a spinning laser on top it is placed in front with two laser scanners do provide a 3d laser image. There you have the no cons which you have explaned. Poo -> no problem, no light -> no problem, more image information -> check.
Speaking about user experience and the global feeling you have, my 2015 Neato is way softer than my brand new Roomba. The roomba keeps bumping everywhere as if it had no cartography at all. It is very stressful to see it force its way on cables and fragile stuff... A LIDAR in that case seems to be way better than any VSLAM. The hybrid system of course seems to be the most intersting
I have a vacuum bot with LiDAR in it, which is perfect cleaning companion except when it comes to large mirrored wardrobes that I have, which wreck the sensors and it’s ability to accurately map the house. I was actually thinking of upgrading to Vslam tech bot, do they have a similar problem?
I love all your vacuum videos they are the best.
How can the video about robot vacuum be this interesting?! Good job 👍🏻
Just got the Roborock s5 max and love it.
I had VSLAM based robot from MAMIBOT and returned it due to navigation problems in my house (f.e. I used it for two weeks and only once it found way back to charging base). It was not a bad robot but poor app and navigation problems were just horrible so decided to go with lasers. Xiaomi Viomi V2 PRO (with laser) that I bought later had no problems with mapping entire house right away, it always returnes to base and never gets confused during cleaning (even when I move it during cleaning). Viomi has much better app with "no-go" lines and all the good staff that is available via Xiaomi Home App (MAMIBOT VSLAM was unable to save any map in app after cleaning so did not create "no go" lines). For me laser is significantly more precise, reliable and many times offers better options in app.
You have precision floor mapping with lidar so what else do you need your vacuum to do? I argue against the concept of "future proof" because this really is the most i need out of a robot vacuum.
Vslam has another advantage. It lasts longer. In lidar you have a Camera, a Laser and a Motor to rotate that stuff at a higher speed. And most of the time, the motor fails in the first 2 years and has to be replaced.
Cameras (vslam) on the other side, doesn’t have that problem
But, required processing power and cost of it, also in terms of batteries, is much worse with the camera. With LIDAR you barely need a computer. Raspberry Pi level hardware is more than enough. For camera you might need something like 15 Watts sucking Jetson nano. So it must be used at extremely low frame rate.
I don't understand however why they cheap out on the motor/driver on LIDAR. Motor should almost never fail. The belt yes, but not the motor. However laser should see socks too. It's just that is mounted too high. Front facing camera OK, but that "fake wall" would be a "problem" too. Instead of camera stil, you can use cheap fixed lasers (range finders) or ultrasonic sensors. Yes, it would cost more than a bumper with switches. Yet none has now water or piss sensor...
On my Neato that broke I thought it was the little motor and bought a new motor. The lasers still would not turn. I tested both motors and with 5 volts dc they both turned fine. I discovered it was a broken navigation board. This is on a Neato D7 that cost me $900 and only ran 60 times before breaking down. For that reason I would never buy another Lidar robot. Too unreliable. This is why we have such a huge problem with waste in North America. Too many things are thrown out because you can’t get parts or the parts cost as much as buying a complete new one. The government needs to get involved with this if they want to really reduce pollution and landfill issues.
@@ds99, so a board not specific to lidar robots broke and you are never buying a lidar vacuum? Makes sense. Especially considering that camera robots have bigger and more complicated navigation boards.
@@ds99 Louis Rossmann fights here on YT for the right to repair just for this kind of reasons. In your case the culprit was probably (I guess) a bad solder job, a blown mosfet costing perhaps cents to replace. And then the consumer is left with absurdities as if you have to replace the house to change the dead light bulb. But the same could happen with the camera or anything else. Because the market is driven to replace not to repair. I have a dead temperature sensor (AC) and a small speaker (for blinker sound) in my VW SUV, less than few dollars total worth, repair is not available, replacement is in thousands. And then as you said.... I could not agree more.
Lidar have a little trouble with mirrors. It’s reflect laser beam and I have double size of my room it the app, or artificial map objects if it’s small mirror.
Being an IT professional, I find it odd how willing some people would be to having a motorized camera, connected to wifi, running around their house taking video/pictures of everything... and especially considering how some of the user agreements and ToS govern under the people's "Republic of China" and having network connections out of the US country (Ecovacs...) ... I'll pass, thanks. Probably nothing major but the implications are astounding to consider. I don't need the POSSIBILITY of someone seeing me (or my wife!) in my underwear - I'll keep my privacy as much as I can, thanks... Another reason all of my webcams are blocked off too.
Why not just by a LIDAR robot if you're so concerned?
Another draw back with lidar it adds height that little bump dome can be enough to stop it from going under low objects
He sits in the middle like they are his pets :D
If possible in a future video... aren't cameras a privacy concern?
Agreed
But, you already have computers and mobile phone cameras. Why stop at vacuum cleaners and say its a privacy concern?
@@rickmartony9566, most computer cameras have lights that show if they are on, moreover, they are stationery.
@@MrPaukann true that. But still the same if you think about privacy.
@@rickmartony9566, not really, a robot vacuum sees the whole house, meanwhile a computer sees a wall, your head and your shoulders.
The Lidar robots should do a mapping run where they use also use the bump sensors to see if something is actually a wall or just a light piece of fabric where they could go through so that the floor under the sofa gets cleaned as well. I also think that right now you should not buy a camera robot. It might be the better technique in the future but the processors that are installed in the current robots are not fast enough to handle artificial intelligence and flawless navigation. You would have to buy a new one with a better chip.
And then they destroy your lego creation, spill your milk, knock over your trash can, and more disasters
I've tried Shark IQ and Roborock S5 (VSLAM and lidar). Lidar is definitely better and VSLAM was completely useless on the Shark. The Shark never made a map of my house and never cleaned in consistent straight lines. The Roborock got it right the first time and consistently cleans straight and well
erm, the first robot vacuum was electrolux trilobite from 2000, which had random navigation, and first camera robot was Samsung NaviBot sr8855, which is today old like Roomba 500 series. so older than dyson and Roomba 980.
What a opening scene!
I went through a Roomba, a Dyson and a Neato, but now I found the Lidar based, Chinese Smart 360 S6 and it is so much better and way cheaper than e.g. a Roborock. App control works flawlessly too.
Coming from team Tesla I would say camera is much better. But putting it to the sealing is stupid - except if you clean sealing. Best would be black and white stereocameras in front of device. That's all. It's not that complicated to understand. Just 2 cameras.
Great content! Looking forward for new robot vacuum models with both laser and camera combo!
Vslam is actually older than 2014. I had a Samsung one with a camera, but it got fully confused when it drove under a couch. 😂
You forgot the Dyson 360 eye
Still waiting for ROBOROCK S5 MAX review. When is it gonna happen?
I got my wife the S5 Max for Christmas...It is awesome. The mapping is incredible. I spent what seemed like hours deciding what to get. I think it was a coin flip between the Deebot 950 and the S5 Max. It seems like it has all the features/components of the S6 but with a bigger mop reservoir. Although i didn’t see many reviews specifically on the S5 Max, I went for it...we mainly have hard floors but even on carpet in the bedrooms, it’s great. The brush doesn’t jam and it kicks on its boosted suction. 👍
I made the mistake of ordering it for oversees before it was available. The shipping is taking way too long. It should be here Monday.
Random UA-cam Watcher used to be available before official launch this December from gearbest . Com
Random UA-cam Watcher It’s on there, I just searched S5 Max Roborock...prime delivery for the win!
Random UA-cam Watcher nah I bought it on the 20th. Roborock also had a $60 coupon as well.
Very nice comparaison
I have both Roborock S50 and IROBOT I7+ and by far the LIDAR in real situation the LIDAR are far better and more efficient
oh yes, a much needed discussion! Thanks
wonder if they can add an "ignore barrier" function to the app that ignores obstacles in certain areas unless they trigger bumpers. that would deal with issues like false bottoms on couches.
How do you track the movement of the vacuums? I have an Eufy and I'd like to map out what routes it takes.
I saw this tumblr joke once about “releasing” 24 roomba with an Alexa, and this guy could actually do that and that scares me
Could you please review one of Miele robot vacuums?
there is one more thing You should mention - privacy. Considering how bad those robots are secured - they can be quite easy hacked into spying (which some of them are already doing...).
So not only Your house plan and any sound from home is send away, but also a pic/vid feed can be :(
So till full maping/AI can be done only in vacuum(offline), not shared to cloud for AI checks, I would avoid them.
I don't believe that any of the VSLAM robots share anything with the cloud.
If they did, why would they need such powerful processors to interpret the data from the camera if it was all done on a server somewhere else?
Hello thank you for your videos they are great and have helped us out a lot. I have a question looking for a robot vacuum that is able to go under some coaches that we have that have cloth skirts. We also like the ability to use a smart phone app. that will do a mapping of the house. I don't know if there is also one that systematically cleans instead of bouncing all over the place. I think I am looking for the impossible do you have any recommendations thanks.
So we are still working on our search tool - but try looking at some models and their features here:
www.vacuumwars.com/types/robot-vacuum-cleaners?p46=2460,2351,2364,2338,132,133,134,135,136&p60=178&p37=104
In that link I have preselected all the smart ones that use Wifi and have no-go zone (which means their is an app involved). It's not perfect but it's a good start.
Best robot vacuum navigation method is having both Vslam and Lidar in one robot vacuum.
This is really vacuum historical wikipedia ive been looking for
aw thanks!
Thanks for this and the best robot vac 2019 vid, helped me decide on the deebot 901 with lidar, it's quite amazing
I just bought a 920 Eco vac. I’m having a lot of trouble mapping my large studio apartment. Does yours map well?
@@rebeccakalyani2486 yes I have two, they are both mapping and working great after a year, although there were a couple issues that were eventually resolved
hi. is this true... i read that lidar sensors break more then the others due to the movement they make? because they are moving parts. thanks in advance.
Shouldn't they use a Kinect style camera in the future
Three years have passed since this video was recorded. Is lidar still ahead? Do you think there's a chance for VSLAM to take the lead?
you are right and wrong about vslam. vslam is faster and gets a better maping but the trouble is you need a good camera programmed properly and aimed right. if thats dun its will get a better capture of whats around it and faster time since its now useing a lasor that has to spin a rotation each time. the down side tho is you have to get a verry expensive camera that has a super whide angel lence and high resalution to see father distance as the laser can do a larger distance no matter how great the quality. also you have to rember the laser can be interuped buy shiny objesct and buy things that will absorb the light.
It would be nice if the user could choose to turn the A.I. on or off. Some users have complained that it's too sensitive and wont' vacuum places where they would like.
The biggest reason for a company to choose cameras is cost. You can source a smartphone camera for under a dollar and then write vSLAM software that costs zero dollars to duplicate. Lidar is more expensive to manufacture, less reliable, and requires more resources for design / repair. Whether the company passes these benefits onto the consumer in the form of e.g. better cleaning power for the same price is up to them.
I have a Roman i7+ and it works perfectly at night and in the dark
is lidar ok with floor length mirrors or will the cleaner try to enter a room that doesn"t exist? If so, would setting up a no go area in front of the mirror solve the problem? Or should I limit a purchase to vacuum and mop cleaners with no lidar?
I'm pretty sure Samsung robot vacuums had a top mounted camera before Dyson or iRobot.
It would be a good idea in it can use other cameras like a smart tvs cam or a security cam .. ets
I love my 50 dollar super clean robot. However it gets stuck easy. Also got a irobot braava 380 and I find that it doesnt run very long even after hours of charging. Possibly a bad battery
Considering getting a nice robot vaccum and the braava 240 mop robot.
The Doug Demuro of vacuums!
Thiiiiiis is the 2020 Roborock S5 Max and it is one of the best robot vacuums yet. Today I will take an in depth look and show you all of its quirks and features, then I'm gonna see how it cleans and then I'm gonna give it a DougScore. For more of my thoughts on the S5 Max, click the link below or visit vacuumtrader.com/robosteer
Why not have all three? The top camera could be mounted on top of the LIDAR housing. It could aid the LIDAR in mapping the room. The front camera could still be good for object recognition & avoidance. Personally, I don't mind if a floor robot stays away from skirts & doesn't get stuck under a chair or couch.
Is so many hours of daily exposure to laser not dangerous?
With regards to more and more advanced processors (the 32-bit cpu part of the video), the first robot vacuum, Electrolux Trilobite, used a MC68332 32-bit processor, so it is kind of back to basics :-)
The 32 bits only refers to the main CPU bus architecture. Todays vacuums use ARM-based System-on-Chip controllers that are more or less the same power of a smartphone, many megabytes of RAM, and up to 8 gigabytes of flash memory. The MC68332 is extremely primitive in comparison. The only similarity is the 32 bit bus used in the architecture. Nevertheless it's an important stepping stone in the development of household robots.
@@nrdesign1991 I beg to differ. The Trilobite is in my opinion more advanced than most Roomba's (all less than the 900-series) and more advanced that the Electrolux vacuum release a couple of years ago. Regarding the ARM chips used, there is a big difference in speed and memory capacity between Cortex-A chips used in smartphones and Cortex-M chips used for controlling robots - one are good at interfacing people and one is good at interfacing hardware :-) Also, Cortex-A chips tend (there are exceptions like iMX.5, BCM2835[ARM11] and Allwinner A10) to become obsolete and unavailable quite fast, while Cortex-M chips (and embedded SOC's in general) are available for very long periods.
Great video. One thing I feel it lacks is that you relate everything to the future of robot vacuums being AI - but you don't explain what this "AI" will bring to the table. Robot vacuums already clean homes very well. They already have scheduling through apps, they already have go back to recharge, and continue, they already have no go zones, they already have bump protection, and fall protection. What benefits could AI bring, with the additional cameras in the future?
I always wish my Neato Botvac 80 would go over the same place twice. Of the dirt that it does miss, it always just misses it and never goes near it again. I know my Botvac is old tech, but sometimes I wish it could switch between Lidar and Vslam, like do one first, then the other automatically. Precise is not always better than random, but the two together I bet would be very effective. In fact, they should make a test robot vac that allows switching of the two systems or using a combination of both and see which is used more by testers in their homes.It would be neato to see the results.
Video is really informative!
Only cost comparison was missing.
Let's break it down a different way. Quick show of hands. How many people would OK with having an Internet-connected camera roaming around their house taking video footage of everything and everybody? Ya, I don't see too many hands up! For me, that's a pretty big reason to go with Lidar.
@@marcusy7421 but, "smart" phone don't have wheels and can go around autonomously :p .... hahaha...
@@marcusy7421 I won't argue w/ that and don't get me wrong, i am not "paranoid" about it , at first i am semi joking in replying to you :D
(WTH, i even see myself buying next gen deebot 960, roborock S7, S8? :p in the future)
But the more i think about it, the more i can imagine how it can be used by "ordinary" criminal.
Just imagine a thief using your vacbot just to checking / make sure that the house is empty.
btw, on LIDAR vs VSLAM,,, i am on LIDAR camp :D
@@marcusy7421, smartphones don't have alternatives, but vacuums have.
I have several furnitures with 9cm height.and only vslam robot can do the job.i bought yeedi that luckily only 1 cm away from bottom part of the furnitures.
There's a spanish brand claiming to have the more powerfull vacuum bot. I would love to see a review for Cecotec conga 5090 and 4090
What about replacing the laser with a camera but still have the spinning part?
thank you for the presentation. Very clear and helpful
How well do vslam vacuums work vaulted/cathedral ceilings if the angles of the ceiling don’t match the floor?
Will you be testing the zipper wand?
Can you please make a video on 1 brush v/s two brushes? The Roombas come with 2 and the 360 S5/S7 come with 1. Which would clean better/faster? Thanks
Maybe review a panasonic stick vacuum. I heard their it vacuum is really good. Also maybe try the new black and decker stick vacuum. My father has it and he likes it so maybe a good one to try.
Maybe a little late to the party but I thought I'd let you know that the lidar robots basically use the same algorithm which is called slam. In my opinion you shouldn't call it lidar Vs vslam but camera.
You should always include a night more for your vacuuming test. In the winter night time come early and my Roomba does not clean when I need it to. Will try lidar bot next.
Can you guys do a review video of the Eureka Mighty Mite and Eureka Quick Up?
Is Shark IQ lidar based?
Are those lasers fine whit pets eyes?
Thanks for the video!
Would you do a video on Indian company Milagrow robovacs sometime later?
Hey, can you review the Ecovacs Deebot M82 please?
Thank you great explanation
You are welcome!
As soon as a lifespan of a robot vacuum is 2-4 years, it makes sense to focus on the technology that works now. And it is lidar. When VSLAM will offer more, then just buy a new electronic "pet" :-) For now this "future-proofness" doesn't mean much for the end customer.
I think a key idea for the lifespan of a robo vacuum has to do with the availability of replacement parts and the savvyness of the end consumer as well.
For example, I can buy a new motherboard for the roborock S5 that I have, and have the technical knowhow on how to replace it should I have to.
Can you review LG’s ThinkQ robot vacuum.
The new Xiaomi Mijia 1S also has both camera and lazer navigation. I'm surprised you didn't mention it. Are you planning on reviewing this model soon? Love the channel by the way!
LG cordzero review please.. thanks for the quality reviews
Hi. When Roborock S5 Max on your channel?
Vslam is the system used by Cecotec that are the best vacuum robots.
Laser eye is new technology and much better than old camera eyes... No contest!
You kinda missed the most important distinction, which is that VSLAM robots are often significantly cheaper than LIDAR robots.
what’s the song at the end?
the most important point is that lidar is much more expensive for the companies to use, cameras are very cheap
Elon musk said the same thing about tesla's detection technology 🤔
What if they used lasers and cameras on the top? Superbot?
I saw the Roborock s5 bumping into my arcade1up cabinet. Cos it can’t see black object but can the camera robot see black objects?
Random is fine for apartments. at least it goes around multiple ways
I really like the Eco vac. I would like to see all three options on one vac though.👍🙂👍
And I would like to see a vacuum cleaner that has vacuum ducks on the corner that I can turn on and off by itself.
Forgot to mention laser don't have the problem of video leaks or hacks that have already happened.
I think LIDAR technology will also improve substantially in the near future
Great video! Love the channel, you guys convinced me to go buy a used Xiaomi gen 1 a while back and now looking for my next one haha.
Btw what’s the song at the end? Hehe