LARA The Robot Vacuum Cleaner M1 (Pivoted to LARA-M0)

[salihkanber]

I have been interested in robot vacuum cleaners for a long time, and I have now finally decided to create one. Its name will be LARA M1, and currently, I only have these motors;



I think this will be a good start, any comments will be appreciated, thanks.

[Ed.Kloonk]

I hope you don't mind me saying but I hope this project sucks.

[salihkanber]

Ok ok not just the motor I also do have a designwork I started alongside.

[Infraviolet]

Not fully sure what that motor's specs are, but when it comes to the wheels you'll probably do better with 37mm 12V motors with integrated gear heads, lots of gear ratios available and lots of motor models from many manufacturers all fitting the same face and mounting size.

[salihkanber]

Yes you are right, those geared DC's work pretty well with wheels. Actually for the wheels I am planning to do smt different. I will use geared small step motors for the wheels and will check the movements with an IMU onboard. Planning to use a Rasp Pico W as the main controller.



(For the previous photo, ignore the yellow its-everywhere-in-every-project wheel on the background, I am not sure who left it there for what ) The motor in the picture is for the suction fan and has these specs; 14.4V, 25.000rpm (seller says that, not tested yet , 70W max, DC.

Already watched many 3d printed fans scattered in many projects but yes I decided to work on another suction fan and scatter it with that 25.000rpm



I too hope the project sucks

[tom66]

Might be less fun, but you can buy whole robot vacuum fan assemblies.  There is also the art in a vacuum cleaner of designing the filter and dust catcher, too dense of a filter and the filter clogs before the dust box fills, too little filter and the fan and bearings are gummed up from all the dust going through, and the room dust just gets spread around everywhere.

IMO, the fun part of a robot vacuum is the room solving algorithm (avoiding furniture and pets) which my cheap robot vacuum at home doesn't quite get right.  Would be interesting to write new firmware for it sometime if we replace it.

[salihkanber]

IMO, the fun part of a robot vacuum is the room solving algorithm (avoiding furniture and pets) which my cheap robot vacuum at home doesn't quite get right.  Would be interesting to write new firmware for it sometime if we replace it.

The real value is in the software & algorithms side, absolutely agreed on this part. My aim is to create and continuously develop a DIY + Professional-like vacuum robot which will be versatile, upgradeable and modular (Will not compete with the sells-millions-type giant brands, some tens will be ok for us). Efforts on creating such a platform will also teach us a lot about the way ahead of us in robotics.

What path we are taking in sunday robotics now is;

ProBUDDY Kits ------------> Equipment & Small Sized Robotics -------------> Advanced Systems & AI

We've commercialised the ProBUDDY Kits, about a year after our Kickstarter Campaign (Damn a hard road but fun, we shipped the rewards, I'm also considering creating a tabletop game that teaches about the whole Kickstarter journey in a fun way).

So, now its time to move on with the robots and thought that a vacuum cleaner robot platform might be the first step.

If we can build such a robot (DIY + Professional-like vacuum robot which will be versatile, upgradeable and modular) then we will be able to develop algorithms on it (software is valuable but it need a proper hardware & platform) and also for anyone that is interested.

You can buy whole robot vacuum fan assemblies.  There is also the art in a vacuum cleaner of designing the filter and dust catcher, too dense of a filter and the filter clogs before the dust box fills, too little filter and the fan and bearings are gummed up from all the dust going through, and the room dust just gets spread around everywhere.

Definitely true. I experienced that 8-9 years ago while trying to design & build a ducted fan myself. Propeller and fan design (at a spesific speed) is not as easy as it looks. Same will be happening here I know. I will still try them out the home made parts first and then seek for commercial parts for the LARA Mark 1.

The platform will be a combination of the home-made parts + commercial parts. We will (together with you guys) determine which parts will be used as new versions of our product are released (Mark 1, Mark2,3,4 --- ?, on which version did Tony Stark stopped updating the Ironman suit (respect) ? Mark 69?

I think we wont go that far

[NiHaoMike]

Might be less fun, but you can buy whole robot vacuum fan assemblies.  There is also the art in a vacuum cleaner of designing the filter and dust catcher, too dense of a filter and the filter clogs before the dust box fills, too little filter and the fan and bearings are gummed up from all the dust going through, and the room dust just gets spread around everywhere.

Would be interesting to design a robot vacuum that uses a length of hose connected to a blower mounted in the window, no filters to replace and basically unlimited capacity except for an occasional cleaning of the screen used to protect the blower from larger objects. It would be good for places like wood shops that generate a lot of biodegradable dust.

[Ed.Kloonk]

[salihkanber]

Would be interesting to design a robot vacuum that uses a length of hose connected to a blower mounted in the window, no filters to replace and basically unlimited capacity except for an occasional cleaning of the screen used to protect the blower from larger objects. It would be good for places like wood shops that generate a lot of biodegradable dust.

A robot for industrial spaces, hungry for the big bites got it 

[salihkanber]

@Ed.Kloonk those people are nuts ) But I like the idea of putting a game mode on the robot  noted it  I really did here;

[salihkanber]

Ok that yellow thing rotates, no vibrations, what now?

[Ed.Kloonk]

Needs moar AI.

[salihkanber]

Hahaha.. ) More AI into LARA, noted that.. Then get ready for an amotional bro-like or (even better) wife-like cleaning machine being ) Yayyy

By the way, since the vacuum system will take time to get robust, besides the development of the LARA M1, I decided to put a LARA M0 on the line with no cleaning functions what-so-ever.

So that a platform just for the algorithm developments will be ready sooner.

[Infraviolet]

Post #11, now connect some tubing and measure how much suction it can actually produce when a fully housing covers the fan blades. Be sure to optimise to get enough suction at this point before trying to integrate with other parts of the robot.

[salihkanber]

Post #11, now connect some tubing and measure how much suction it can actually produce when a fully housing covers the fan blades. Be sure to optimise to get enough suction at this point before trying to integrate with other parts of the robot.

Got it. Here is the assambled vacuum motor v1.



Watch the motor in action;



*Yes it sucks
*No vibration, seems stable in full power
*Too much noise

Definitely it needs proper testing, optimisation & upgrades. So I ll stop working on the vacuum motor at this point now, and as I mentioned in the previous post, now its better to pivot the effords towards LARA M0 with no cleaning functions but the electronics hardware inside, to be created as an algorithm development platform.

I couldn't decide on what the description of it should be.;

*LARA M0: The vacuum robot that does not suck
*LARA M0: Non-Vacuumer Vacuum Robot
*LARA M0: DVRAADP (Dummy Vacuum Robot As A Development Platform)

Do they all suck and/or do you have a better suggestion??

[salihkanber]

I call out to the best software developers all over the world, develop algorithms for the hardware we have been developing and let the customers purchase your algorithms as well.

The question is: Can you write better algorithms than me? or us? (Chatgpt suggests: Dare to outshine and redefine excellence!

[salihkanber]

Can anyone suggest a proper cliff sensor for the LARA M0?

An example image;

[Ed.Kloonk]

[Infraviolet]

A cliff sensor would be any downwad looking proximity sensor which is able to operate mostly independently of floor surface finish. Time of flight IR may be overly precise, long ranegd and costly for this use, a reflection sensing (measuring level of returned IR light) could work but might be overly affected by black vs shiny surfaces. A SHARP type IR which internally focuses a returned spot to a different place on an array and uses thisposition to utput a voltage indicating range might be the best choice, you'll want a ring of them positioned in multiple places both to compare readings across all and avoid reversing over cliffs.
Cheapr versions (because if I were building this I'd want a whole ring of them) of this sort of thing:
https://uk.farnell.com/sharp/gp2y0a41sk0f/sensor-distance-analogue-output/dp/1618431

[Georgy.Moshkin]

one big disadvantage of current vacuum robots is their size. They can't go under the fridge, piano and other congested places. I think that sweeping cloth robots are underestimated: not chewing your socks and expensive MacBook cable, can be super thin and small, silent. Noise levels of my xiaomi vacuum are horrific, and it tries to climb up to xiaomi fan stand with other kind of clicking noise. My dream cleaner robot is a swarm of small sweeping-cloth bots. Or a combination, a team of vacuum and mini cloth bot.

[salihkanber]

I think that sweeping cloth robots are underestimated. My dream cleaner robot is a swarm of small sweeping-cloth bots. Or a combination, a team of vacuum and mini cloth bot.

Could you give an example to those sweeping-cloth bots please?

[max_torque]

Why not just buy an existing robot cleaner and make your own control board for it?  No point in having to painfully re-engineer all the driving stuff imo when it already exists cheaply.

If you choose a popular existing model you might even be able to sell a few "open source" control boards for those as well

[salihkanber]

Why not just buy an existing robot cleaner and make your own control board for it?  No point in having to painfully re-engineer all the driving stuff imo when it already exists cheaply.

You know.. Thats an amazing idea! Lets do that! (Discussing these issues with the brightest minds all over the world opens new doors.. So valuable..Thank you) Let's think about a name..

If you choose a popular existing model you might even be able to sell a few "open source" control boards for those as well

Can you suggest one?? Or anyone please suggest one ??

[Georgy.Moshkin]

Could you give an example to those sweeping-cloth bots please?

something like this (image attached), and I've seen triangular ones.
What if make it thinner? Adding metallic plate for weight will create enough friction.