Reflective surfaces and lamp/chair bases

[SimonB]

Hi, firstly I'd like thank the forum, it's been a helpful resource when deciding which robot vacuum to buy.

I recently bought an XV-25 which has the 3.2 firmware, and translucent wheels. I'm not sure about the upward firing exhaust, but it does seem to fire more up than horizontal.

I've discovered a few snags that I'd be interested to see if others have also experienced and if there is a way to avoid them. Unfortunately I'm finding more and more issues with the bot's navigation that is rendering it not fit for purpose, which is a real shame, as I want it to work. We suffer from large numbers of dead flies and insects on the floor during the night, so I wanted something that would vacuum them up automatically before we get up in the morning. The combination of the bot getting stuck as well as not having sufficient battery to complete the job overnight is making me consider returning it.

Reflective surfaces - the kickboard in our kitchen is shiny silver and on straight runs the bot seems to cope with this, however there is one corner where it persistently and repeatedly tries to move forward, bounces, turns a little, bounces in another spot and never concludes that there is an obstruction. Eventually it reports the clear my pathway error. What I don't understand is why the bumper being triggered is not causing the bot to map out an obstruction, much like it would at a wall corner. Is it because the laser is not detecting an obstruction because of the shiny surface?

Climbing - a few times now the bot has climbed over lamp bases, low chair bases (eg breakfast bar chairs with a large circular base) and a glass side table that has a rising slope. From reading this doesn't seem to be an uncommon problem. What I feel the bot lacks is a means to detect the angle of approach is adopting and pulling out when it must realise that it's climbing an obstacle and not a transition, or is it designed to be able to move across some quite substantial transitions.

Right now I'm trying with the mag strip, but the bot is moving the short pieces around the lamp base and glass table and got stuck this morning. I'm also excluding the kitchen area (open plan) since this is not where the majority of the flies die and drop and it will reduce the run time to hopefully allow it to complete it < 7 to 8 hours. However it's a pain to have to do this every evening and I don't want to make the strips permanent, or at least semi-permanent so the bot won't move them. I'm tempted to try the bumper extension from another thread to lower the bumper, but this might well prevent it moving over rugs we have.

Does anyone have any knowledge of improvements that Neato are working on that might address either of these two problems?

Thanks.

[glnc222]

It's still somewhat an experimental technology. If you want it to work in problem areas some modifications to the unit and/or the premises are necessary in some circumstances. It is also common to avoid certain rooms entirely. These products are so specialized and the market so small, there is little investment in major design changes.
Some modifications to the bumper and tilt are referenced in http://www.robotreviews.com/chat/viewto ... 20&t=16475

[snav]

You are referring to a cleaning-time under 7-8 hours, it sounds strange that the Neator would clean beyond 3-4 hours.
I think the Neato will try to recharge two times if it doesn't complete the house in the first charge.
Lets says you have got really lucky and got o good battery that lasts an hour.. then it will only try to clean your house for three hours before it gives up.
So when you wake up in the morning it would probably have tried to clean a couple of hours tops... or am i misunderstanding the time aspect?

[glnc222]

kickboard in our kitchen is shiny silver and on straight runs the bot seems to cope with this, however there is one corner where it persistently and repeatedly tries to move forward, bounces, turns a little, bounces in another spot and never concludes that there is an obstruction. Eventually it reports the clear my pathway error. What I don't understand is why the bumper being triggered is not causing the bot to map out an obstruction, much like it would at a wall corner. Is it because the laser is not detecting an obstruction because of the shiny surface?

A polished, mirror-like surface deflects the laser instead of returning reflection to the sensor, thereby appearing as an open space it tries to enter. Lidar cannot work with mirrors. Neato's software does not appear to integrate bumper hits into a map of the wall, but it may do this with a mag strip along the wall (it may need to be out a bit in order to fit under the sensors). A few vertical strips of masking tape at intervals close to Neato's size reflecting the laser will identify the wall and also assists with some very dark furniture, as on the inside, unseen, side of the legs. When I added that in one case it won't bump the legs anymore.

[uncle_bob]

then it will only try to clean your house for three hours before it gives up.

I really, really wish this was user selectable. My newly purchased XV-21 is going back for refund unless I can get it to clean my whole (smallest) house within the three attempts.

I have shag pile carpet and the machine only give 30 minute run times and can only clean 3/4 of the house before it gives up :/

[glnc222]

shag pile carpet

If it is moving very slowly, bogged down on the carpet, a slider addition may improve results. See the end of thread at http://www.robotreviews.com/chat/viewto ... 20&t=16830 for the simplest version, some cardboard and tape.
It moves slower when first circumnavigating the perimeter, but should then move what looks like to me at least one foot per second. Slowing down wastes power and reduces area covered by battery charge.

[SimonB]

Thanks for all the replies, apologies I didn't get notification although I thought I had that enabled.

Odd that bumper and laser are not combined, that would seem an obvious and, dare I say it, easy thing to do to cope with situations like ours.

Not being able to judge its own height also seems poor, but then I don't know how the laser mapping works.

The 7-8 hour comment was simply that it was scheduled at 00:00 and was still running at 7.30am. I happened to hear its first run last night and it seemed to do 65 minutes, which seems pretty good from what I've read on here.

I now exclude the kitchen and dining room with strips to avoid the kickboard problem, as well as it getting stuck under the dining chairs. Those are not the problem areas where we get all the insects, so I am somewhat happy to exclude them and it means the unit has finished by the time we get up.

I think I've about resolved all issues where it does clean with mag strips and lifting one piece of furniture. Thanks for the suggestion of masking tape, but that's a concession I'm not willing to make, when the appearance of the house starts to suffer for the sake of the cleaning.

I've also used it for some spot cleaning, which seems to work quite well. Whether it's quicker or more convenient than grabbing the dusbuster I'm not sure, but it's always fun to watch it work.

[glnc222]

When all the walls are mirrored Neato can't fix position with the Lidar, so can't map anything at all. Another case reported here, however, of a polished stone wall reveals more possibilities than Neato yet employs.

With enough walls to fix position, the Wall Sensor can detect mirrored walls same as any wall, and map accordingly. It is already quite good at following walls with this sensor. But apparently, it relies on the lidar to map walls -- sort of union rules. The software just isn't thorough, more of a demo for an abstract floor on a graphics screen, or trade show booth.

News just reported major delays in shipping out of the Port of New York because of a faulty upgrade to trucking management software, and recently an entire airline was shut down a day for failure of their operations management system. So Neato's not the only one with the "standard sub-standard software", to quote a famous Disney movie. Where's IBM's pixie dust -- soy based -- when you need it?