Madimack i60 Time Lapse Pool Cleaner - Raw Footage TEST 3
Вставка
- Опубліковано 19 січ 2024
- Madimack i60 Time Lapse Pool Cleaner - Raw Footage TEST 3
Full battery, two hours, same start condition as Test #2
Conclusion: The dumbest random mapping so far, and almost a complete failure to get up the walls.
If you find my videos useful you may consider supporting the EEVblog on Patreon: / eevblog
Web Site: www.eevblog.com
Main Channel: / eevblog
EEVdiscover: / eevdiscover
AliExpress Affiliate: s.click.aliexpress.com/e/c2LRpe8g
Buy anything through that link and Dave gets a commission at no cost to you.
T-Shirts: teespring.com/stores/eevblog
#ElectronicsCreators #Madimack - Наука та технологія
Must be hard to map such a regular shape....
Yup, the official Madimack GT i60 demo video shows it much more precise in navigating back and forth.....beginning to suspect a gyro/IR issue.....or buggy firmware....or over-extended sales video. Worth investigating all the same.
I'm suspecting the curvatire of the pool corners cauing a traction issue and hence it resets the mapping.
I still think that when it's mapped your pool entirely, it reports its findings to North Korea.
For added excitement, how about a Roomba cleaning the paving, suspense every time the cliff detector stops it taking a swim.
It'd be fun if some skilled math & video person could extract its route from the video and make a density map of its coverage. It seems like it hits some areas a lot of times. Also would be interesting to see thoughts on what an "optimal" path calculation would look like, even thought I'm sure Madimack put a lot of thought into it already.
I've made one from first timelapse 😅
All the thought in the world doesn't help if the physical issues in practice cause the mapping to reset itself, as seems to be the case here.
This would drive me up the wall.
It seems liek it is really loving your pool. Like espacially that little corner of the stair. Maybe you should leave them alone? Let them have some time together without your prying?
Again, an intriguing soundtrack. Made me watch in 1x speed just to listen to it! Reminds me of Tears for Fears :)
Beadcrete may be a bit too slippery for it, and it probably hasn't' been tested for it.
Would be interesting to get a Demo(they do them for free) from Maytronics, Dolphin or liberty(battery model) and see if they can do any better with the beadcrete. Maytronics are the robotic industry standard for pool cleaners, like a Fluke!
It looks like your pool cleaning robot is drunk lol
I recently saw a video about a robot vacuum that was hacked to work locally without "phoning home". That vacuum would display the map it had generated on a web page. Does this pool vacuum not have some way of doing the same, or at least extracting the data?
It also doesn't seem to be either covering the whole pool, not remembering previous sessions. - looks broken, or sales pitch is bs!
Nope, no ability to do that.
@eevblog2 The local software for those vaccums is called valuetuedo. And one good video about is called: "Sucking dust and cutting grass: reversing robots and bypassing security"
Could it be that it relies on its magnetometer so much that the rebar/steel pipes are messsing with it's mapping.
Interesting theory !
Interesting, but I doubt it.
My guess is that the Pool Robot tries to do some sort of pattern, but that as good as it may be, that *_The Butterfly Effect_* gets in the way - so it does "something", but is affected by error!
"Intelligent Navigation: Advanced infrared guidance and dual-drive navigation ensure your pool is cleaned with double the efficiency."
I didn't think that infrared worked underwater, that could explain why it looks like it drives drunk.
Looks like it doesnt have enough thrust to hold itself on the walls , hmmm
Yup. I suspect the curvature of the walls is causing a problem.
BOGO sort isn't an efficient algorithm.
It's just never going to be satisfied with the cleanliness of the corner of that step. I think it needs more RAM.
Since you are now more familiar with this thing, you don't have to call it by its formal name: "Randomized Willie-Nilly, III". Just "Randy" is good. Revealing that none of that "AI" "mapping" and "planning" rubbish is, well just "rubbish". G-day...
Hopefully someone smarter than me could use YOLO to track the bot from this video. That way you can determine its coverage.
snik2pl did it.
A Bump ‘n’ Go toy from the 1980s has better coverage.
I really dislike (aka hate) random movement pattern. That is so lazy...
It's not designed to be random, it's just having issues that cause the mapping to reset.
what a load of junk
Absolutely no logic in it's movements... Looks like it uses a random direction generator
It's not designed to be random, it's just having issues that cause the mapping to reset. This is the worst example, the other videos are better.