Page 2 of 2

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: March 25th, 2013, 3:37 pm
by glnc222
I have transitions and molding larger than 1/4" and there does not seem to be a problem.

Moldings with sufficient ramp width -- most apparently -- don't cause a problem; it's not the height alone which matters. There are some which are too square edged, and narrow metal ones causing a problem.

Some moldings could be treated as obstacles to back off if only Neato would detect blockage of forward motion when the wheels are turning, a very strange gap in the programming given sophistication in other parts. The laser was used to guide motion, but not to detect its lack -- an even simpler use. Reminds of how scholars found parts of the bible were written by different authors.

Other moldings and similar furniture parts are capable of trapping Neato because it does climb up on them, only then getting stuck on top; no software fixes would help. These require boundary markers as long as the bumper remains as designed, and is not modified to detect these lower obstacles.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: April 16th, 2013, 8:58 pm
by glnc222
another tight space navigation failure

Under cabinets against a wall usually handled, Neato backed up against the outside chair leg facing in, and wedged the rear up tight against the cabinet bottom with the wheels under constant rear torque and just sat there, laser spinning. It apparently could not complete an expected reverse motion and kept trying. It is pure chance what position it will get into in this space. Walls blocking two sides are clearly an extra constraint. Pressing the button to pause and pushing the bot into the middle of the space, restarting had it continue as normal.

It seems pretty clear there is no software use of laser information showing no movement under wheel torque in some cases; yet I have seen it back off and retry some obstacles. All it had to do here was move forward and it would have released, so maybe they skipped backing up cases. Its like they gave up coding halfway through a list, if they ever had one in the first place. I am concerned it kept powering the motors. I may have to block these cabinets off with mag strips or wood bars hidden underneath to be sure it isn't damaged when running unattended.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: April 20th, 2013, 2:34 pm
by glnc222
It turns out only 3/4" can make a difference in ability to navigate tight spots. The cabinet bottom causing a trap had a wood strip securing wires against the baseboard against the wall underneath, lacking on a similar cabinet never causing problems. Don't really need that strip so removing it probably will prevent that occasional trap. A bit more room to back up when it needs to, after bumping the back wall. Rare so hard to be certain, can't just reproduce the events.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: April 24th, 2013, 1:13 pm
by chris-s
glnc222 wrote:1. A tight cul de sac
Hey !!! You speak french ??? :mrgreen: :lol:

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: April 24th, 2013, 3:08 pm
by glnc222
English is famous for adopting terms from other languages -- british empire, Norman conquest, etc. Architectural, artistic terms in street layouts, like latin stuff. Read French in school but lack conversational training. Been there a couple times, Paris and Alsace, driving in from Germany and Zurich. C'est trop d'annees apres l'ecole. Allons enfants de les Neato.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: June 4th, 2013, 2:16 pm
by glnc222
example of stuck backing up
stuck.jpg
Neato stuck backed up onto lamp base

stucknot.jpg
Neato avoids getting stuck when front facing lamp


Shown is an example of Neato stuck backing up, where the rear got lifted on top of a 2 inch thick circular lamp base against a wall about 18" from a corner. Most of the time Neato approaches the side wall with the lamp along the intersecting wall, behind the lamp, turning the corner towards the lamp and bumping off cleanly from the front. But if Neato does it's rotating thing and bounces off the wall opposite the lamp in the corner, it will back up onto the lamp and get stuck. Traction is insufficient to pull off.

It's annoying to have to use boundary strips for such a case.

I may investigate limiting how far the rear can lift over obstacles. This occurs only when the rear hits something when backing up, stopping motion and directing the wheel torque instead onto the wheel suspension arm pivot, rotating the body. The wheel arm is pivoting the wheels down, keeping the front down and lifting the back instead, from the direction of wheel rotation in reverse gear. A block might be placed to limit rotation of the wheel suspension arm. Effect on navigation to be determined. Since lifting the back slightly reduces the footprint, new ways of getting stuck could emerge. Push one place and it pops up elsewhere...

[edit] a mod has been made to reduce the tilt range: http://www.robotreviews.com/chat/viewtopic.php?f=20&t=17095

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: July 2nd, 2013, 2:17 pm
by Montyleo
My neato is less than a year old. It runs for a minute or less and then either says replace my dirt bin(it's in properly) or it gets stuck in an Unobstructed corner like a naughty child in punishment and says it can't navigate. Is there a fix, or do I need to return it?

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: July 2nd, 2013, 3:09 pm
by vic7767
Since your unit is still under warranty, contact NeatoRobotics CS for a replacement.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: July 10th, 2013, 11:49 pm
by glnc222
backing up to a wall

In this case Neato backs into a wall at the corner of a mag strip running into the wall. Because the wheels are slipping on carpet, it never detects the obstruction and never stops running in place, eventually damaging the rug. Software shows no reliance on the lidar plainly detecting the wall as well as having scanned that wall in numerous passages around it. In contrast, Neato seems very aware of walls approached from the front, turning before contact with the bumper. The behavior here might be associated with the mag strip invoking a different section of software compared to other cases of backing into a wall; would require testing, the factory's job, not mine.

This is a very specific, reproducible gap in navigation software. Why this sort of thing which prevents cleaning is not fixed instead of tinkering with other minor details of efficiency is a mystery. Perhaps their lab has no carpet of the right kind. The solution at the moment seems to be not relying on mag strips. Perhaps a simple barrier the lidar sees will work better.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: December 26th, 2018, 8:37 pm
by Scuba_steve
By raising furniture and placing no go strips I have remedied all problematic areas except one: a La-Z-Boy rocker. The poor vacuum tries so hard to wedge it self underneath the rocker and the rocker gives way to its insistent nudges until eventually the vacuum is toofar under the rocker and the rocker is sitting too heavy on the vacuum and it can no longer go anywhere. I have tried to prevent this with no go strips placed just under the rocker but it’s not working. Any ideas?

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: December 27th, 2018, 2:25 am
by JEfromCanada
Scuba_steve wrote:By raising furniture and placing no go strips I have remedied all problematic areas except one: a La-Z-Boy rocker. The poor vacuum tries so hard to wedge it self underneath the rocker and the rocker gives way to its insistent nudges until eventually the vacuum is toofar under the rocker and the rocker is sitting too heavy on the vacuum and it can no longer go anywhere. I have tried to prevent this with no go strips placed just under the rocker but it’s not working. Any ideas?


Neato will penetrate about 2" BEYOND the mag strip. So, I would place the strip across the area where the Lazyboy's legs extend behind the chair. My Lazyboy has skirts on it, so when I'm not actually sitting on it, the skirt is detected as a "wall". But if I'm sitting on the chair and leaning forward, I see the same symptoms you are describing.

Re: Where Neato's Get Stuck Preventing Cleaning

PostPosted: December 27th, 2018, 4:39 pm
by glnc222
This is the oldest problem I addressed with mods. From the Mods & Repairs wiki thread top of forum,
Thread "Overhead pinch trap"
http://www.robotreviews.com/chat/viewtopic.php?f=20&t=15853&hilit=pinch
youtube video http://youtu.be/Vq3RXH6kzxc
No permanent alteration of the vacuum is necessary, taping parts to the front as shown.
It is interesting the Vorwerk VR200 model has an overhead sensor on the turret which Neato models did not include, though even that might not work in this case, photos would be needed.

Note I preferred to modify the vacuum instead of placing barrier strips everywhere. A related problem is the bumper bottom being too high, which affects flat chair runner legs, scandinavian styles -- as well as some circular base bar stools (one amusing video shows the mod with a whole forest of bar stools).

The Samsung I use, with camera guidance and IR proximity sensors, detects the overhead pinch from the cushion bottom of my Norwegian Eurosleeper convertible sofa and does not go under it, as did the Neato (and Roomba), getting trapped. Some stick on rubber cones were once sold to put atop the Roomba bumper for this.