Arduino controlled wheelchair

[Burgerman]

JUST a mathematical problem?

The Mathematical Universe
NASA/JPL-Caltech/R. Kennicutt/SINGS

The history of mathematics is a history of humanity seeking to understand the universe. Therefore, many consider the holy grail of mathematics to be the same as that of physics: a theory of everything, a unified theory that explains all physical reality.

Math generally plays a vital role in any theory of everything, but contemporary cosmologist Max Tegmark even goes so far as to theorize that the universe itself is made of math. In his mathematical universe hypothesis, he proposes that math is indeed a human discovery and that the universe is essentially one gigantic mathematical object. In other words, mathematics no more describes the universe than atoms describe the objects they compose; rather math is the universe. Tegmark even goes so far as to predict that a mathematical proof for a theory of everything could eventually fit on a T-shirt.

More than 60 years earlier, however, Austrian mathematician Kurt Gödel put forth a theory that argues quite the opposite. Gödel's first incompleteness theorem concerns axioms, logical mathematical statements that we assume to be true but can't be proven with a mathematical proof. A simple example of this would be the axiom of equality (X = X). We assume this to be a true statement, but we can't actually back it up with a mathematical proof. Gödel's theorem states that any adequate axiomatizable theory is incomplete or inconsistent.

The implication, according to theoretical physicist and mathematician Freeman Dyson, is that mathematics is inexhaustible. No matter how many problems we solve, we'll inevitably encounter more unsolvable problems within the existing rules [source: Feferman]. This would also seem to rule out the potential for a theory of everything, but it still doesn't relegate the world of numbers to either human invention or human discovery.

Regardless, mathematics could stand as humanity's greatest invention. It composes a vital part of our neural architecture and continues to empower us beyond the mental limits we were born with, even as we struggle to fathom its limits.

[Irving]

@max71 While I applaud your thinking and determination this isn't something easily done.

...

But don't let that stop you... maybe you'll be the one to crack it!

Cant be that insurmountable, its a maths / parabolic problem. Instead of the chair driving to the gap directly, drive to it so that it is going through square...or square ish.

The math is relatively easy.. deciding what and where the obstacles/gap are accurately, reliably and quickly enough whilst moving through it turns out to be quite tricky

[Max71]

@max71 While I applaud your thinking and determination this isn't something easily done. A colleague of mine runs a research project involving teams from UK and France looking at 'shared control' systems for wheelchairs - think of it like riding a horse; you tell it which way, it does the job of (mostly) avoiding obstacles etc. The problems and challenges are non-trivial - even getting obstacle detection to work reliably in a home environment with varying hard and soft surfaces/material etc is a challenge; what works well to avoid walls and steer down a 1.5m+ wide corridor is all but useless to steer through a 1m gap between a sofa and a coffee table for example. And they've tried everything from simple ultrasonics, laser scanners, lidar, millimetric radar, machine vision, etc, etc. No one sensor technology works reliably in all environments.

But don't let that stop you... maybe you'll be the one to crack it!

Hi Irving, first of all I thank you for the warning.
I agree with you that this project is simple in thinking but difficult to make it. What I think I do is, try. Various levels of redundant control, different types of visual and sensorial recognition, locking system for each electrical fault, many tests and I supervise myself with that nice red stop button on the robot. I don't know if I will succeed, but I try.
In the worst case scenario I will try to steal the google car project!

[Max71]

@max71 While I applaud your thinking and determination this isn't something easily done. A colleague of mine runs a research project involving teams from UK and France looking at 'shared control' systems for wheelchairs - think of it like riding a horse; you tell it which way, it does the job of (mostly) avoiding obstacles etc. The problems and challenges are non-trivial - even getting obstacle detection to work reliably in a home environment with varying hard and soft surfaces/material etc is a challenge; what works well to avoid walls and steer down a 1.5m+ wide corridor is all but useless to steer through a 1m gap between a sofa and a coffee table for example. And they've tried everything from simple ultrasonics, laser scanners, lidar, millimetric radar, machine vision, etc, etc. No one sensor technology works reliably in all environments.

But don't let that stop you... maybe you'll be the one to crack it!

Cant be that insurmountable, its a maths / parabolic problem. Instead of the chair driving to the gap directly, drive to it so that it is going through square...or square ish.

When learning to drive we are taught to look in the distance and our brains will figure out the rest, its triangulation. The closer we focus the sharper the angles become and the more sever our driving/steering correction's are.

If distance sensors were on each side of a wheelchair, a narrow 'field of view' and mounted on servo's, directed by a lidar.
....maybe two sets on each side, left side measuring the distance the left side of the obstacle and the left side measuring the right side to the obstacle and visa versa.
Creating double triangles.

Surely its just a maths problem to get the 'base' of the quad triangles to equal each other. To fit squarely though a gap.... ?

At the moment my idea is to implement various layers of control. The first is only sensorial, to teach the robot to avoid obstacles, through a small ANN, there are many examples on the web that seem to work well. Having established this aspect, I will pass to the second level, the best navigation in tight environments. Yesterday I was discussing this with an engineer who was in charge of drone navigation for the army, he told me that it is very difficult but not impossible, and I believed it!

[Max71]

JUST a mathematical problem?

The Mathematical Universe
NASA/JPL-Caltech/R. Kennicutt/SINGS

The history of mathematics is a history of humanity seeking to understand the universe. Therefore, many consider the holy grail of mathematics to be the same as that of physics: a theory of everything, a unified theory that explains all physical reality.

.......

Regardless, mathematics could stand as humanity's greatest invention. It composes a vital part of our neural architecture and continues to empower us beyond the mental limits we were born with, even as we struggle to fathom its limits.

I do not exclude asking God for the solution.

[hatam1996]

hi
can you give me diagram and code
More explanation plz

[gcebiker]

Hi if you search my posts there are a few hundred posts, i am sure among them (more to the end of the threads) you will find what you seek.

[hatam1996]

gcebiker

thank you for replying
You did a great job
Just another question

Can I use Joystick Module

[woodygb]

[hatam1996]

woodygb thanks
you are so cooooooooooool guys

[woodygb]

Can I use Joystick Module

Yes ...but you will need to rewrite part of the Arduino sketch.

[hatam1996]

Where to edit the code
I don't know ?

[hatam1996]

CAN YOU GIVE ME sketch Modified code

[woodygb]

Can you make the control board?

This is NOT just an Arduino.

Not sure ...but there might be a sketch by gcebiker for a Blue tooth Arduino Nunchuck if you look thru the threads.

[hatam1996]

Yes, I made it
But it didn't work

[woodygb]

Concentrate on getting that working as it is needed WITH any Arduino Sketch.

[gcebiker]

hi

I made control board
And used Joystick Module
But it didn't work
Please give me the code I want it to work with the joystick model
CAN YOU GIVE ME sketch Modified code PLZ

Hi Hatam, if you go to the thread posting.php?mode=reply&f=2&t=6503
You will find new versions of code that work better, if you post there you will have the most up to date information.
I have started working on it again, its summer here In Australia and I am getting ready for Squid season early in the new year.

I am now working on using the emulator to drive a Storm Arrow power chair with Brushless or 'GB' motors as Invacare calls them (Gearless Brushless)

The Emulator works with the POWER MODULE, it replaces the JOYSTICK MODULE.

[hatam1996]

this my control board

[woodygb]

Do you have the original joystick module?

[hatam1996]

yes i have
I have already asked for help with this problem on the site
viewtopic.php?f=2&t=9046

[hatam1996]

I've got a problem with the joystick OF my wheelchair
I have two controllers burnt in the joystick type controller atmega8a



https://www.dropbox.com/s/u6zxj5lpl6bne ... 191629.jpg
https://www.dropbox.com/s/3cow0fl0e0n98 ... 191707.jpg
https://www.dropbox.com/s/fkvj7l3z95m4l ... 191737.jpg
https://www.dropbox.com/s/7gqhux7ntx7rt ... .LARGE.jpg
https://www.dropbox.com/s/7u2lapig3pqia ... XFXXXN.jpg
https://www.dropbox.com/s/i2hy8cb0uxatr ... be.com.png
https://www.dropbox.com/s/i2hy8cb0ux...e.com.png

[gcebiker]

yes i have
I have already asked for help with this problem on the site
viewtopic.php?f=2&t=9046

Hi Hatam, i will reply on that thread.

Rebuilding a joystick is incredibly complicated.

Lets talk on the other thread about the faults you have with your current factory joystick.
We can try to help with that problem first.

Tony

[kaban]

Hello folks,

Does anyone know the baudrate of the CAN in a DX Controller?

I am working on getting a Invacare G50 controlled by a Arduino Nano.

Already got some telegrams out of it (100kBaud):
42 3 B0 81 21
42 3 1F FC 1C
42 6 B0 81 21 92 FF BF
42 3 FC 1C 1
42 3 B0 81 21
42 3 C2 50 B0
42 6 FF BF 1C B B0 81
42 3 2D 50 B0
42 3 48 10 1C
42 3 48 10 1C
42 6 A5 DE 1C B B0 81
42 6 C2 B0 81 27 20 89 //beginning to steer somewhere here
42 6 20 45 FF D8 50 B0
42 6 41 F2 1C B B0 81
42 6 45 F9 1C B B0 81
42 6 41 F2 1C B B0 81
42 6 20 45 F5 1C B B0
42 6 41 EA 1C B B0 81
42 6 B0 81 23 20 45 F3
42 6 41 DB 1C B B0 81

But when sending one or many telegrams back to the Controller, nothing happens. I've tried it with many different tools but not every tool is working with their own code. (Tried tools with Lawicel (CAN232) protocoll and ASCII console output from Arduino. The console output from the arduino appears to be the most reliable way of reading but it's definetely not the most comfortable way of doing so.
With watching the Behavior of these telegrams for an hour and trying to understand which telegram does what I neither had luck.
So I went here to ask if someone has got information about the DX Bus..

Working with this:
-Arduino Nano
-MCP2515 + TJA1050 CAN-Shield (https://www.ebay.de/itm/MCP2515-CAN-Bus-Modul-TJA1050-Transceiver-5V-Arduino-Raspberry-Pi/253369172743?hash=item3afdfac307:g:Z4wAAOSw9tlb4ifz)

[kaban]

Telegram scheme:
ID Length Data

[woodygb]

RS485 is the coms protocol.
WheelchairDriver • View topic - Dynamic Controls Power chair, RC Boat trailer project
viewtopic.php?f=2&t=6503&start=300#p117806

[LROBBINS]

I think that the Shark is RS|485, but I'm pretty sure that the DX (& DX2) is a proprietary version of CAN.

[gcebiker]

Hi,

I am not sure it if will help with the DX but I've attached the protocol for the Shark (RS485 38400 baud)

Shark Bus Communications version5.pdf

(71.85 KiB) Downloaded 310 times

[stevelawiw]

@gcebiker

I am now working on using the emulator to drive a Storm Arrow power chair with Brushless or 'GB' motors as Invacare calls them (Gearless Brushless)

Sorry I know nothing about your emulator, will you use the existing controller or perhaps RoboteQ? I've been trying to get RoboteQ and the GB Motors to play nicely, I was hoping there might be some common ground!

[gcebiker]

@gcebiker

I am now working on using the emulator to drive a Storm Arrow power chair with Brushless or 'GB' motors as Invacare calls them (Gearless Brushless)

Sorry I know nothing about your emulator, will you use the existing controller or perhaps RoboteQ? I've been trying to get RoboteQ and the GB Motors to play nicely, I was hoping there might be some common ground!

The present emulator i have built up for the Shark Controller will turn on the GB motors on the Storm Arrow Chair but not do any thing else.
I have yet to hook it up to the Logic Analyzer to see the data stream.

If it turns out to be RS485 then i will mod the code i have for the Shark and use that to control the factory Power Module.

I have also tried swapping the Original Joystick for the Shark (as manual says it can use either) but same issue.
I have the programming SD card for the Original Joy and OEM Wizard for the Shark, just been to busy to get to tinkering...

Enter COVID19 , now lots more time as house bound again....

[stevelawiw]

The present emulator i have built up for the Shark Controller will turn on the GB motors on the Storm Arrow Chair but not do any thing else.
I have yet to hook it up to the Logic Analyzer to see the data stream.

Oh ok,well good luck with your project, and stay safe!