Future member, new projects: exoskeletons, DIY laptops and NFC. Anyone interested ?

[Alexandre Coninx]

Hi,

I'm a new potential hackspace member. I came tonight and I intend to
join to work on a few projects I have. I'm just dropping a line here so
anyone interested can contact me or react.

Here is what I am up to:
- Some not well defined yet but cool stuff around assistive robotics and
exoskeletons, especially using air muscles:
http://www.instructables.com/id/How-to-make-air-muscles!/
Air muscles are powerful, they are easy very cheap and to make, they
have a nonlinear behaviour close to biological muscles, and they are
much safer than electrical actuators when it comes to strapping
something to your body. Plus, they have a nice steampunk vibe :)
Japanese researchers have made a nice upper body exoskeleton:
http://www.youtube.com/watch?v=QpZ2_a0SMmA
And there are some people in the US who made a cool soft suit (sounds
abusive to call it an exoskeleton since there is no "skeleton"):
http://cwwang.com/2008/04/08/soft-pneumatic-exoskeleton/
I would probably start by trying to make some air muscle, power them,
learn to control them efficiently (typically using motor valves and
arduino) and go on working on something like upper body enhancement.

- A long term project would be building my own laptop, including the
hull and structural parts. I have built my share of towers and
opened/fixed/hacked a few laptops but apart from that I am in completely
unknown waters here; in fact it is as much a pretext to learn how to do
CAD properly and use plastic/metal working tools than anything else. I
suppose I would design the parts with CAD software, make the hull and
bigger structural elements using the lab's metalworking tools (I like my
laptops tough and sturdy, not brittle and shiny), and maybe use 3D
printers to design smaller and oddly shaped parts. The various
electronics and "computer" parts would be salvaged, bought or built (if
possible).

- NFC security explorations, especially around the contactless payment
systems (paypass/paywave). I've read a lot of material on the subject
and have some knowledge of NFC and RFID in general, but I couldn't go
further, first and foremost because I don't personally own such an NFC card.

If you want to discuss anything of this, I'm in the hackspace tonight;
I'm currently in the sofa space, with a red shirt, in the chair behind
that awesome presence bot they are building. If you're not here or
stumble on this mail later, just drop me a mail.

--
endy

[Alexander Baxevanis]

Welcome to the Hackspace :) 

Your projects sound interesting, I'm definitely curious to see if you can communicate with contactless payment cards (and other similar things, such as passports).

If you want to experiment with payment cards, there's 2 prepaid contactless cards that you can easily get for free:

O2 Money card: http://www.o2.co.uk/money/moneyaccountcard (Visa)

Paypal Access Card: https://www.paypal.com/uk/webapps/mpp/accesscard-apply-premium (Mastercard)

[Michael James]

Regards,

Michael James

> --
> You received this message because you are subscribed to the Google Groups "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to london-hack-sp...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.

[Michael James]

Hi

Your project look awesome, im a relatively new member myself (having been a sponsor for a few months and joining fully about a month ago) tho I've only been to the space a few times.

I'd love to get involved with the Air Muscles project. My mind is already overrun with possibilities. My field is primarily electronic engineering and I'm pretty sure a given a decent micro and RTOS some clever stuff can be made.

Keep me posted on how u want to proceed, I'm definitely interested in getting involved. I may even buy some if the components this week.. I'm very excitable.

:-)

MikeOJ

On 17 Sep 2013, at 22:04, Alexandre Coninx <en...@viletfourbe.org> wrote:

[Eugene Nadyrshin]

I'm interested in the air muscles for robotics applications. I know ShadowRobot seels/uses them and I'd be interested in making and controlling them. So do give me a shout when you are about, I can often be spotted working on the robot arms

[Mark Steward]

On Tue, Sep 17, 2013 at 10:04 PM, Alexandre Coninx <en...@viletfourbe.org> wrote:

- NFC security explorations, especially around the contactless payment systems (paypass/paywave). I've read a lot of material on the subject and have some knowledge of NFC and RFID in general, but I couldn't go further, first and foremost because I don't personally own such an NFC card.

This is relevant to my interests - I've just written some code to speak EMV and found that it's trivial to pull out the equivalent of the magnetic stripe from a Paywave card. However, I'd like to investigate ways of fingerprinting individual cards without storing the PAN.

Mark

[Alexandre Coninx]

On 18/09/13 08:23, Alexander Baxevanis wrote:
> Welcome to the Hackspace :)
>
> Your projects sound interesting, I'm definitely curious to see if you
> can communicate with contactless payment cards

It seems that Mark has already done much of what I intended (see his
reply). The investigations he suggests are interesting, though. And
since it is trivial to pull the magstripe data from the NFC, I would be
interested to see if you can somehow clone the card for payment purpose
using that. You could not clone the EMV secure element of course, but
maybe you can tweak things so the payment terminals would fall back on
magstripe data. It would be a huge security breach.

There should be some way to do that since if I understand correctly this
is (part of the way) SimplyTapp works [1] (or worked, I don't really
understand if they are still in business).

> If you want to experiment with payment cards, there's 2 prepaid
> contactless cards that you can easily get for free:
>
> O2 Money card: http://www.o2.co.uk/money/moneyaccountcard (Visa)
> Paypal Access
> Card: https://www.paypal.com/uk/webapps/mpp/accesscard-apply-premium
> (Mastercard)

Thanks for the info, man.



[1] http://www.simplytapp.com/

--
endy
There is an old Chinese saying that goes, "When you have nothing to say,
quote an old Chinese saying."

[Mark Steward]

On Wed, Sep 18, 2013 at 12:19 PM, Alexandre Coninx <en...@viletfourbe.org> wrote:

On 18/09/13 08:23, Alexander Baxevanis wrote:

Welcome to the Hackspace :)

Your projects sound interesting, I'm definitely curious to see if you
can communicate with contactless payment cards

It seems that Mark has already done much of what I intended (see his reply). The investigations he suggests are interesting, though. And since it is trivial to pull the magstripe data from the NFC, I would be interested to see if you can somehow clone the card for payment purpose using that. You could not clone the EMV secure element of course, but maybe you can tweak things so the payment terminals would fall back on magstripe data. It would be a huge security breach.

There should be some way to do that since if I understand correctly this is (part of the way) SimplyTapp works [1] (or worked, I don't really understand if they are still in business).

If I remember correctly, as long as you assert that the customer is present, you only _need_ the PAN and expiry date to make a valid transaction on a payment gateway. It'll probably be far more likely to be flagged as dubious than one with more data, though.

Mark

[Alexandre Coninx]

On 18/09/13 09:24, Michael James wrote:
> I'd love to get involved with the Air Muscles project. My mind is already overrun with possibilities. My field is primarily electronic engineering and I'm pretty sure a given a decent micro and RTOS some clever stuff can be made.

Well, I went on digging on the topic, the main problem of air muscles
seems to be reactivity. If you look at the video of the guy with the
soft suit, you can hear the noise of the muscles inflating/deflating and
the timing seems to be a bit off, I wonder if his suits actually works
(ie enhances his strength, endurance, or something). The japanese design
clearly works (increases the load you can carry), but the tester is not
doing much more than statically bearing 50kg of rice bags, that's an
achievement but I would be more impressed if he could actually pickup
and unload heavy things (beer fermentation vats, anyone ?). Clearly
there is still room for progress.

Maybe it has more to do with the control of pressure than with the air
muscle technology itself. Then with higher pressure, good quality
solenoid valves, clever sensors and as you say something realtime to
process the data we could achieve a good result.

Another way I ultimately would like to explore is using other fluids
such as water. It would probably solve the supposed reactivity problem
since water are hardly compressible, but the power source would be more
problematic and it could become expensive to find tubes, bladders and
connectors that can take that stress. (The whole system would be much
heavier, too.)

To get back from speculation: what I am up to right now is making a few
muscles, make them lift stuff and maybe operate a dummy robotic arm with
two opposite muscles. Then move on to designing a limited system (for an
arm, maybe).

About the previous work, I found a better video of the japanese design
[1]. Visibly, all the air muscles are attached to a rigid frame in the
back, and operate the arm joints through what looks like Bowden cables
(same system as bicycle brakes). This is not as elegant as directly
putting the muscles on the arms and legs I think but may make more sense
to control an exoskeleton.

> Keep me posted on how u want to proceed, I'm definitely interested in getting involved. I may even buy some if the components this week.. I'm very excitable.

Well, the problem is that right now I'm completely broke -_- so I can't
buy any stuff or even a hackspace membership. Things will get better in
about one week. Also I won't be in London this weekend. I would probably
be OK to start experimenting next week if we can get some materials on time.

In order to start there are a number of tutorials on the subject. There
are two instructables. One [2] is really nice and has an excellent
description of the hardware used. Another [3] was written by the guy who
made the US soft suit but is much rougher. There are also plenty of
tutorials on robotics forums. I think we should start by gathering some
hardware and trying to follow one of them.

I don't know how this usually works with the space, but maybe we should
stop polluting the general ML and create a wiki project page or
something ? I am also on IRC (EndyFourbe), I'm mostly lurking but query
me if you want to chat :)


[1] http://www.youtube.com/watch?v=_cw1jFhRWU8
[2] http://www.instructables.com/id/How-to-make-air-muscles!/
[3] http://www.instructables.com/id/Pneumatic-Muscles/

--
Alexandre Coninx

[Alexandre Coninx]

On 18/09/13 10:52, Eugene Nadyrshin wrote:
> I'm interested in the air muscles for robotics applications. I know
> ShadowRobot seels/uses them and I'd be interested in making and
> controlling them.

I love ShadowRobot, their wirk on the shadow hand is really impressive.
But the muscles they sell are really expensive (�150 + VAT for one 30cm
muscle !). It may be worth it to read their documentation, though, since
they obviously got the design right and we could learn from it.

[Filthy hipster scum]

Hey, i helped design the shadow hand as used to work there and know their system inside and out an how to build the air muscles. Would love to get involved with that project.

[Alexandre Coninx]

On 18/09/13 14:30, Filthy hipster scum wrote:
> Hey, i helped design the shadow hand as used to work there and know their system inside and out an how to build the air muscles. Would love to get involved with that project.

That is extremely cool :)

Would everyone interested be interested by meeting somewhere during the
next week to discuss all this ?

[Alexandre Coninx]

To anyone interested: I created a project page
https://wiki.london.hackspace.org.uk/view/Project:Air_Muscles

It's just a starting point. Put your name, hack, improve, complement,
change and discuss as much as you want :)

[chrisbob12]

Hi Alexandre,

You may find this thread interesting.

What sort of applications do you have in mind for an exoskeleton?

Regards, Chris.

[Adrian Godwin]

I built the air muscle controllers for Robothespian.

They're awkward things to get working properly as the air is compressible, the load on the muscle tends to change as a function of the geometry of the limb, and there's nonlinearity in the movement / pressure response. We got them working pretty well (see videos on youtube) but a guy at Festo has done some further work that gets way more out of them.

However, you need some pretty fancy pressure control to get it to work - we started with rather expensive Festo pressure control valves but ended up with a custom made device based around an SMC valve. I still have a Festo muscle around that I can demonstrate but not sure I can put together a position controller. The improved controller models the valve and muscle a lot better but needs much heavier processing.

--

[Alexandre Coninx]

On 19/09/13 00:43, Adrian Godwin wrote:
> I built the air muscle controllers for Robothespian.

Nice work :)

> They're awkward things to get working properly as the air is
> compressible,

That was my guess. Did anyone ever try running them with liquids ? I
find frequent comments that it "should work" but nobody seems to have
actually experimented and given feedback.

> the load on the muscle tends to change as a function of
> the geometry of the limb, and there's nonlinearity in the movement /
> pressure response.

Well, that kind of nonlinear behavior is also the whole point of having
them instead of, say, electric actuators or even regular pistons, if I
understand correctly. I only read a few articles and the documentation
of the Shadow Robot Company air muscles and I understand these things
are a hell to model properly (just like biological stuff).

> We got them working pretty well (see videos on
> youtube) but a guy at Festo has done some further work that gets way
> more out of them.
>
> However, you need some pretty fancy pressure control to get it to work -
> we started with rather expensive Festo pressure control valves but ended
> up with a custom made device based around an SMC valve.

It is actually rather encouraging; it means that with fancy pressure
control you *can* make them work the way you want.

I am quite interested in how exactly the control system operate, from a
hardware but also software point of view. Is finding/building the right
hardware the tricky part and did you just use good ol control theory
stuff once your sensors and actuators were good enough? Or did you have
to use alternate approaches? My own training would suggest me that the
problem is well suited to machine learning approaches (simple inputs,
quite simple outputs, and a big nonlinear black box we are afraid to
open in the middle).

> I still have a
> Festo muscle around that I can demonstrate but not sure I can put
> together a position controller. The improved controller models the valve
> and muscle a lot better but needs much heavier processing.

I would be really interested in seeing your muscle and discussing these
issues with you.

What is nice with processing power is that it goes cheaper and cheaper
with time :) If this is really the bottleneck, we can consider many
approaches (CPUs, GPUs, even FPGAs if needed) (and Michael James was
speaking of using a system running some realtime OS to control the whole
stuff).

--
endy

[Ian Lewis]

this may be of interest for air muscles http://www.flickr.com/photos/montani87/8591179971/in/set-72157633095088408/  https://www.youmagine.com/designs/soft-robotica-bending-mechanism

looks like you 3D print the mould and then use rubber(?) for the 'muscle'

Ian 

On Tuesday, September 17, 2013 10:04:06 PM UTC+1, Alexandre Coninx wrote:

[Andrés Muñiz Piniella]

Hi,

about air muscles. I for sure did not read enough documentation but the exoskeleton does not seem to go all the way to the foot. this probably puts strain on your ankles if you are loading your upper body?

[Antaeus]

Hey Alex,

Welcome to the hackspace! I would like to meet up and chat about the air muscle project. My mech eng skills should come in handy. I'm looking to get into robotics as a career and keen to meet people with similar interests.

Regards,

Antaeus

On Tuesday, 17 September 2013 22:04:06 UTC+1, Alexandre Coninx wrote:

[Andres]

>>
>> - A long term project would be building my own laptop, including the
>> hull and structural parts. I have built my share of towers and
>> opened/fixed/hacked a few laptops but apart from that I am in
>completely
>> unknown waters here; in fact it is as much a pretext to learn how to
>do
>> CAD properly and use plastic/metal working tools than anything else.
>I
>> suppose I would design the parts with CAD software, make the hull and
>
>> bigger structural elements using the lab's metalworking tools (I like
>my
>> laptops tough and sturdy, not brittle and shiny), and maybe use 3D
>> printers to design smaller and oddly shaped parts. The various
>> electronics and "computer" parts would be salvaged, bought or built
>(if
>> possible).
>>

I also have the long term goal of doing a franken laptop like those described in Cory Doctrow books. Hakerspace is really far away but I would colaborate on a wiki. I am interested openhardware and rather steer away from closed hardware such nvidia. For that I got as far as finding http://h-node.org/.

--
Enviado desde mi teléfono con K-9 Mail.

[Alexandre Coninx]

On 21/09/13 20:59, Ian Lewis wrote:
> this may be of interest for air muscles
> http://www.flickr.com/photos/montani87/8591179971/in/set-72157633095088408/
> https://www.youmagine.com/designs/soft-robotica-bending-mechanism
>
> looks like you 3D print the mould and then use rubber(?) for the 'muscle'

According to that guy's flickr stream, it is actually silicon:
http://www.flickr.com/photos/montani87/with/8592280510/

I did see some 3D-printed designed but nothing as interesting at that
one, thanks for the reference. I wonder what he uses it for and the
power of that thing.


--
Alexandre Coninx

[Alexandre Coninx]

On 22/09/13 18:14, Andres wrote:
> I also have the long term goal of doing a franken laptop like
> those described in Cory Doctrow books.

I didn't read any of Cory's hardcopy books, but I would be interested in
knowing what he says about that.

> Hakerspace is really far away but I would colaborate on a wiki. I am
> interested openhardware and rather steer away from closed hardware
> such nvidia. For that I got as far as finding http://h-node.org/.

I'll try to avoid anything obviously non-free too, but I'm actually more
interested in designing the whole laptop as a tool than by the exact
electronics inside. I think nowadays the difference between a decent
laptop and a great laptop is much more related to the form factor,
materials used, build quality and ergonomics than to the details of the
hardware inside. That's (in part) why stuff like T-series (and above)
Thinkpads and Apple laptops sell despite being priced way higher than
the competition.

For now the only thing I could find is that:
http://www.bunniestudios.com/blog/?p=3265
http://www.kosagi.com/w/index.php?title=Novena_Main_Page

They don't have any hull/structure yet but they did design their own
motherboard -_-

(BTW I created a project page:
http://wiki.london.hackspace.org.uk/view/Project:DIY_Laptop)

[Andres]

Alexandre Coninx <en...@viletfourbe.org> escribió:

>On 22/09/13 18:14, Andres wrote:
>> I also have the long term goal of doing a franken laptop like
> > those described in Cory Doctrow books.
>
>I didn't read any of Cory's hardcopy books, but I would be interested
>in
>knowing what he says about that.
>

Nothing much, he mentions you can buy a shell of the laptop and fill it up. Books are 'little brother' and 'pirate cinema'. You can get them in ebook format as well as pdf from hus website.

>
>> Hakerspace is really far away but I would colaborate on a wiki. I am
> > interested openhardware and rather steer away from closed hardware
> > such nvidia. For that I got as far as finding http://h-node.org/.
>
>I'll try to avoid anything obviously non-free too, but I'm actually
>more
>interested in designing the whole laptop as a tool than by the exact
>electronics inside. I think nowadays the difference between a decent
>laptop and a great laptop is much more related to the form factor,
>materials used, build quality and ergonomics than to the details of the
>
>hardware inside. That's (in part) why stuff like T-series (and above)
>Thinkpads and Apple laptops sell despite being priced way higher than
>the competition.
>
>For now the only thing I could find is that:
>http://www.bunniestudios.com/blog/?p=3265
>http://www.kosagi.com/w/index.php?title=Novena_Main_Page
>
>They don't have any hull/structure yet but they did design their own
>motherboard -_-
>
>(BTW I created a project page:
>http://wiki.london.hackspace.org.uk/view/Project:DIY_Laptop)

Great thanks! I see what you mean by the design, right now I have £200 netbook from 6 years ago so anything with a better ergonomics than that is not difficult.