Notifications
Clear all

My new robot test base

21 Posts
3 Users
1 Likes
6,369 Views
robotBuilder
(@robotbuilder)
Member
Joined: 5 years ago
Posts: 2037
Topic starter  

Decided to build another test robot base to replace the large box I used earlier. The electronics are exactly the same. I have tested it with the dead reckoning moves I was coding on the larger base. Next I want to mount eight sonar sensors (four shown in the image, I have to get four more) along with eight infrared sensors arranged the same way. Later I will add physical bumper switches to deal with hitting the leg of a chair or table. I have some ideas on how to use them to navigate and even map the house.

newBase

 


   
Quote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 

Looking good! ? 

I'm building my larger robots out of wood too.   In fact, I've even been thinking about building the robot arms out of wood.  Wood is nice to work with and can actually look pretty nice when stained and finished. 

I need to draw up my robot in Fusion 360.  Right now the only drawing I have is in my imagination.  It's not doing much good there. ?

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
robotBuilder
(@robotbuilder)
Member
Joined: 5 years ago
Posts: 2037
Topic starter  

It is of course only a test robot base upon which more actuators and structures could be built only limited by the weight the puny robot vacuum motors can sustain. It is only a test base for testing code for a "brain" which can be transferred literally as a black box to a larger or stronger base with bigger motors. Ultimately it is all about code to process the sensor data and act in accordance with some goal using the given hardware.

I only have limited funds for the hobby and spending on a more powerful robot base would require making it able to do something practical such as a helper robot. I would have made my own version of db1 but as Bill stated, "This is a HUGE project, and as such it will end up being a fairly expensive one."

Also I can't decide if a rectangle or circular base is best as they both have pros and cons.

I have a round base I may work on later. It is a gutted cheap robot vacuum cleaner coming with battery, charger and power supply circuit and of course the two geared motors with encoders. This is much cheaper than buying the parts separately. It also has a plastic bumper with infrared sensors (drop and side wall). The bumper moves two rods through two interrupt sensors which I guess is better than a physical switch in terms of wear and tear.

newBase2

Another robot that reflects the kind I would like to build is:


   
Duce robot reacted
ReplyQuote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 
Posted by: @casey

I would have made my own version of db1 but as Bill stated, "This is a HUGE project, and as such it will end up being a fairly expensive one."

Exactly my thoughts as well.   This is a project that Bill is admittedly designing as he goes.    And while he has laid out his basic design as three "layers", there isn't a lot of technical details in precisely what kind of hardware he's going to end up employing.    I heard in another thread, in fact I'm pretty sure it was you who said, that Bill is going to move up from using the $55  RPi4 to instead using a $239 HiKey960. 

If that's true, and he keeps making those kinds of changes DB1 can easily end up being several thousand dollars to build.   I believe he's also going to incorporate the Jetson Nano into DB1 as well?   I don't know, but I just don't see how DB1 is going to remain a low-cost project with all these expensive parts being added.

I just hope everything works out very well for Bill and anyone who's building DB1.   I wish them all the very best.   Without knowing what the final cost is going to be it's not a project that I'll be starting anytime soon. I'll be glad to contribute to the project via patreon, and I'll enjoy all the lectures and learn a lot from them.  But as for me building a replica of DB1 that's just not in the stars.   I'd already be dropping out if I had to buy a HiKey960. ? 

I'm just too financially distraught to try to keep up.

Posted by: @casey

I have a round base I may work on later. It is a gutted cheap robot vacuum cleaner coming with battery, charger and power supply circuit and of course the two geared motors with encoders. This is much cheaper than buying the parts separately. It also has a plastic bumper with infrared sensors (drop and side wall). The bumper moves two rods through two interrupt sensors which I guess is better than a physical switch in terms of wear and tear.

Yes, it's really nice when you can make use of an already existing base.

I'm currently starting with these little robot cars because I'm cheap.  But you can do quite a bit of programming and prototyping with these and all of that work can then  be easily transferred over to a larger physical robot later.

Alysha Baby

I'm only just starting on the base for my larger robot.  I don't even have a drive system or motors on it yet.  But I do have a full machine shop so I can fabricate most of what I need.

Robot Base (2)
Robot Base (4)

I know the wheels look like overkill and probably are.  But this thing might get heavy so I figure it won't hurt to have some heavy duty wheels.

And yes, Arlo is a really nice robot.  That's about the size I'm planning on building Alysha to be.  Including two arms like Arlo has.

Bill hasn't given us any drawings of what he expects DB1 to look like in the end.  I haven't heard any mention of any arms, or head.   So people who are building DB1 at this stage have no clue what they're actually going to end up with when they are finished.   Hopefully it will be a really cool robot!

There was a fellow who posted when the forums first began offering to design and 3D-print a body for DB1.  But I haven't seen him post since.   I'll have to see if I can find that thread and post to it to see if any progress has been made toward that body design.

It's all fascinating to me.  I'm sure Bill will end up with a really cool robot whatever he does.  It's just not crystal clear to me at this point exactly where he's headed.  If it were me I'd build the cheapest DB1 I possibly could and keep it simple.   Then perhaps move on to a more expensive DB2 after that.  

While I don't know where Bill is going with DB1, I still LOVE his YouTube channel in general. ? 

Keep up the great work Bill, and best wishes on DB1 wherever she takes you. ? 

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
robotBuilder
(@robotbuilder)
Member
Joined: 5 years ago
Posts: 2037
Topic starter  
Posted by: @robo-pi

I'm currently starting with these little robot cars because I'm cheap.  But you can do quite a bit of programming and prototyping with these and all of that work can then  be easily transferred over to a larger physical robot later.

My suggestion is you can keep all the electronics and motors but use a light weight but larger base.  I am using a  base 13 inches in diameter. There is no reason you can't use a weak robot base to develop navigation using light weight sensors including vision. I did buy one of those small yellow motor bases but when I got hold of the robot vacuum motors it was a no brainer to use those instead.

The issue with four wheels is skidding during a turn which tears up the ground (or carpet) assuming you are driving them like a tank wheels. Also pneumatic wheels go flat that is why I used solid rubber wheels in my first robot base. I also used a welded heavy duty steel frame, strong bearings and two 24v window wiper motors. This was powerful enough to carry a man or move a large couch or pull itself through soft grass. I was going to use it as a motorized wheel barrow until one of the motors failed!

The other type of base I thought of building was a tricycle with a powered front wheel. Easy to drive in a straight line or turn by some degree unlike coordinating two independent motorized wheels. The front wheel pivot would become the point of reference. That is what Frank DaCosta used with his robot pet.

http://cyberneticzoo.com/cyberneticanimals/1979-robot-pet-frank-dacosta-american/

And an attempt to reproduce it,
https://ezrasrobots.wordpress.com/2016/12/03/the-robot-pet/


   
ReplyQuote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 
Posted by: @casey

I did buy one of those small yellow motor bases but when I got hold of the robot vacuum motors it was a no brainer to use those instead.

I absolutely agree.   I didn't have those available.  I was going to build my own base like you did, but I'm not just cheap.  I'm also lazy. ?   So I went with these robot cars chassis.

Posted by: @casey

The issue with four wheels is skidding during a turn which tears up the ground (or carpet) assuming you are driving them like a tank wheels

I don't see this as being an issue precisely because they aren't tank treads.   They should turn on a dime with very little skidding required.   There will be some skidding but I don't expect it to be a serious problem.  I also don't  have floors that I'm worried about scuffing up.   I live in a forest and half the forest floor gets tracked into my house on a daily basis anyway.   So my floors are perpetually "dirty" although I prefer to refer to them as "natural environment substrate" this way it sounds like I'm a naturalist instead of a bum. ? 

Posted by: @casey

Also pneumatic wheels go flat that is why I used solid rubber wheels in my first robot base.

These wheels are so cheap I could just replace the entire wheel if it goes flat.  They are only $3.99 a piece including the the metal wheel, the tire, and even already inflated up to pressure.   I also wanted wheels large that are sturdy enough to deal with some rough terrain should I happen to take the robot outside in the yard.   There are also some uneven thresholds between the rooms in my house.  Small hard wheels can have sometimes have difficulty with those.

Posted by: @casey

The other type of base I thought of building was a tricycle with a powered front wheel. Easy to drive in a straight line or turn by some degree unlike coordinating two independent motorized wheels. The front wheel pivot would become the point of reference. That is what Frank DaCosta used with his robot pet.

Yeah the Robot Pet looks like a good design too.  So many options to choose from.   I don't know how well my base is going to work.  I did spin it around on a dime on a piece of plywood to see what it felt like and I didn't feel any skidding resistance.   So only time will tell how well it ends  up working.   But yeah, I took the potential skidding into consideration and it just didn't look like it was going to be a major issue.   One possible reason is because on my design the front and rear axles are very close together.   Basically as close at they can be without the tires hitting into each other.  Skidding might be a bigger problem with a longer wheelbase.  These wheels are almost on top of each other already.

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
robotBuilder
(@robotbuilder)
Member
Joined: 5 years ago
Posts: 2037
Topic starter  

One practical thing your robot could do is keep the floor nice and clean? However if you spot it surreptitiously sweeping the dirt under the carpet your AI may be too advanced 🙂 

Although I personally have no problem with dust and dirt my wife does so I have to take off my boots before entering the house and sweep, mop and vacuum it on a regular basis.  I tell her that one day if I ever get my robot working it will keep the house nice and clean after I am gone 🙂

Yes wheels are cheap, the cost is in the geared motor and the hard part is coupling it to the wheel. Have you sourced any motors? Are you going to drive the pair of wheels on each side via a chain or belt? 

I wasn't suggesting small wheels just solid rubber wheels. A powered wheel will climb over a 90 degree obstacle if it is less than the radius of the wheel. Tyres going flat when you are not there could be an issue perhaps needing sensors to detect such an occurrence?

Yes, the closer the wheels are together and the wider the base the easier it will rotate. I can't tell how wide your base is but it seems to me you could double it. I would make it at least as wide as it is long.


   
ReplyQuote
Duce robot
(@duce-robot)
Member
Joined: 5 years ago
Posts: 672
 

@casey

I love it can't wait to see the system good job


   
ReplyQuote
Duce robot
(@duce-robot)
Member
Joined: 5 years ago
Posts: 672
 

@robo-pi

No set plan ..........that's how I built duce'ee ??


   
ReplyQuote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 
Posted by: @casey

Have you sourced any motors? Are you going to drive the pair of wheels on each side via a chain or belt? 

My current plan is to drive them with bicycle chains.   I am also planning on using only two motors and coupling the front and rear wheels together with a single chain drive.   So it will be 4WD with one motor driving both wheels on each side of the chassis.    I'd love to get this thing up and running, but unfortunately I haven't had time to work on it.   I also don't yet have the sprockets.  Darn things are quite expensive to buy.  I wanted to go around to yard sales looking for junk bicycles cheap to get some cheap sprockets and chains.  But I haven't even had time to go yard sale searching this summer. 

I have several different motors laying around.  Plus I also have a car that I'll be disassembling soon.  It has four power windows and I've been thinking about maybe using those motors.   But I have lots of motors laying around to choose from already.

Posted by: @casey

I can't tell how wide your base is but it seems to me you could double it. I would make it at least as wide as it is long.

It's already 18" wide.  Much wider and it won't fit through my doorways.   It's probably a lot larger than it appears to be in the photos.   I'm feeling good about the dimensions.  I made it as large as I dare.  I could have squeaked out a couple more inches, but then navigation would through doorways would need to be super precise.  I only have a couple inches of clearance on each side of the doorway as it is now.  Plus my entire cottage is quite small.  It's only 25 feet square.  Or 625 square feet.   And that's divided up into 5 different rooms.  So it's a tight ship.  The robot is going to be about as big as I dare make it.

 

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
robotBuilder
(@robotbuilder)
Member
Joined: 5 years ago
Posts: 2037
Topic starter  
Posted by: @robo-pi
It's already 18" wide.  Much wider and it won't fit through my doorways.   It's probably a lot larger than it appears to be in the photos.  

 What is its length? What is the radius of the wheels and their width? From the photo I imagined it was twice a long as it is wide.

Is that the quaint little house shown in the image you posted of the fallen tree?

 


   
ReplyQuote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 
Posted by: @casey

What is its length? What is the radius of the wheels and their width? From the photo I imagined it was twice a long as it is wide.

As far as the actual wheelbase is concerned it's basically square.     The plywood base give the optical illusion that it's twice as long as it is wide.   This is occurs for two reasons.  First the plywood base fits down inside the tires so it's a lot skinnier than than the actual wheelbase width.  And secondly it extends quite a it further toward the front and back than the location of the axles giving the illusion of being far longer than the actual wheelbase.

I don't have it here to measure.  It's out in the shop.  But I think the tires are about 3" wide, maybe a bit wider. The metal wheels are 4" in diameter.  And I think the outside diameter of the tires is about 8".   The tires are separated by about 2 inches.   So this gives 18" from the front of the front tired to the back of the rear tire.  

The width as measured from the outside of the tires is about 18" as well.   So given that the tires are about 3" wide.  Half of that give 1.5 inches to the center of the tires.   Then double it again for the other side and subtract 3" from 18".  This yields 15" for the center of the tires wheelbase in width.

From front to back we need to subtract 8" total from the 18" to find the center wheelbase there and we get a 10" wheelbase.   So strangely it's actually 5" wider than it is long in terms of where the tires actually meet the carpet. ? 

Just the opposite of what the photos appear to be suggesting due to the optical illusion created by the rectangular plywood base.

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 
Posted by: @casey

Is that the quaint little house shown in the image you posted of the fallen tree?

Yep that's it.  I actually cut a bunch of lumber to install a new roof.  I was  hoping to do that this fall, but it's not going to happen.  I'm far behind schedule on my wood boiler project so I'll be lucky to get that project done in time for winter.  The new roof will have to wait for next year.   That's ok, that will give the lumber more time to cure anyway.   I just cut the lumber for the roof this spring so it won't hurt to let it sit over the winter.  I just hope I live long enough to finish it. ? 

By the way that tree is still there.   It's actually cut down further.  I've been chipping away at it.   But it's not completely down  yet.   It's a monster tree.  Solid oak and was quite alive so it's really heavy.   If it were a dead tree it would be a lot lighter, but this baby was very much alive when it broke off.  So it's heavy with moisture and sap.

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
robotBuilder
(@robotbuilder)
Member
Joined: 5 years ago
Posts: 2037
Topic starter  
Posted by: @robo-pi

I'd love to get this thing up and running, but unfortunately I haven't had time to work on it.   I also don't yet have the sprockets.  Darn things are quite expensive to buy.  I wanted to go around to yard sales looking for junk bicycles cheap to get some cheap sprockets and chains.  But I haven't even had time to go yard sale searching this summer. 

I have several different motors laying around.  Plus I also have a car that I'll be disassembling soon.  It has four power windows and I've been thinking about maybe using those motors.   But I have lots of motors laying around to choose from already.

The reason I wouldn't use chains is because they stretch. Even with a tension wheel you eventually have to replace them or reduce their length. We had one to drive a pump on the spray tank and the chain was forever jumping off its cogs as it stretched.  To me chains are just ugly compared with the nice worm gear used in a window wiper motor to drive the single gear attached to the shaft. The power window motors would probably be ok for a small robot if you have the machinery to attach wheels and encoders to them. As I wrote elsewhere finding suitable geared wheeled motors at a price I could afford for a medium to large robot base has been a big reason I gave up trying to build one.  So instead it is a matter of just using a light weight robot base and concentrate on the code and sensors required and their layout. In truth it is the coding that will take up most of the time to get a working robot up and running.  A robot base can be made very quickly if you can source the parts at a price you can justify.

Keeping motivated has also been an obstacle to completing the project. It has gone in spits and spurts of activity with months and many years in between when I just do other things instead.

If I was in your situation I think I would want to build a remote controlled Mars Rover type robot with visual feedback to move about outside during the winter months.  Put a bucket on it and it can carry stuff for you like maybe fire wood or tools.  Another toy I would like is one of those flying drones where you just attach your smart phone for video feedback. You would be free to fly around the forest and check things.

 


   
ReplyQuote
Robo Pi
(@robo-pi)
Robotics Engineer
Joined: 5 years ago
Posts: 1669
 
Posted by: @casey

The reason I wouldn't use chains is because they stretch. Even with a tension wheel you eventually have to replace them or reduce their length. We had one to drive a pump on the spray tank and the chain was forever jumping off its cogs as it stretched. 

A pump on a spray tank is hardly a fair comparison with a home robot that will be moving at extremely slow speeds.  I can't even imagine bicycle chains wearing out or stretching much in this application.   They won't even see nearly as much wear or stress as they would in their original intended use on a pedal driven bicycle.  And a pedal driven bicycle typically lasts for years unless it's being used daily by a serious bicycle enthusiast. 

I've always had good luck with chain drives.  I've owned several motorcycles throughout my life.  All but one were chain drive.   I don't recall ever having to replace a chain on any of them.   So my experience with chains is apparently quite different from your experience.   However, I would  say that a chain driven pump sprayer does seem like a design looking for trouble.  I would have used a belt in that situation.

DroneBot Workshop Robotics Engineer
James


   
ReplyQuote
Page 1 / 2