RoseRolls ballin’

Yes, dump puns are my specialty. Yes, even in English. Especially in English, actually – because dumb puns are all I’m capable of, and you better go see Nathan Arthur for actually clever puns in English. That’s probably the ultimate criterion about your knowing English, or any language for that matter – it’s the ability to make clever puns. That’s like level C3. And the day I make clever puns in Chinese (or any pun for that matter) – well, as the song sings, “that’ll be the day” – the day pigs have wings. However (and notice the smooth back-to-topic transition) I think after reading so many ambiguous objdump lines, I’m damn close to making witty puns in ARM assembly language. Teachers, your line here is : “Amazing !”, and friends, yours is “Charles, we’re lost you !”. If that wasn’t all a joke, both’d be right in their way.

So yes, before the urge for blather about puns ROSE in me, I meant to talk about where RoseRolls was at. You people from this year’s run of ROSE have already read that in French (oh boy, I have to translate), but for you other followers of this sect adventure here is what we’re looking at for our project’s newer Sphero-sprinkled taste :

Objectives :

  • A spherical robot that can move around and emit LED lights
  • Android cell phone (Bluetooth) remote control/drive app
  • Preset trajectories (sequences of instructions), sent by phones
  • (Bluetooth) communication with a Sphero (basic aim here : mimetism. We’ll see about more complex stuff later.)

I guess I’ll skip the newer PSSC as they’re only interesting for us and the teachers (and we’ve already posted them on the mailing list) (so ask if you reaaally want them that bad).

Now new questions and needs arise, and for each of them, we tried to think of a practical way to answer them.

  • Kalman : as Loik explained, we need to see where we can go without, how hard it is in its concept, in its implementation, and in its live calculation. So we need to define exactly what version of the Kalman filter we’ll use (define the state it predicts, define the captors it relies on – so we need to find a compromise about them too – and their calculated noise …) and infer from that the necessary calculus power (along with the need for a side or integrated FPU). That’s the part we don’t really know how to do : to estimate, to quantify the physical hardware needs of an algorithm – that’s probably benchmarking and the like.
  • The spherical shell : what material ? We aim for transparent for fun, light for ease of movement, and not too  fragile for obvious reasons. Also – important – it has to be easily openable – we’ll do that often. So we need to find where we can buy that kind of stuff.
  • Sphero tests, to see what this little ball can actually do and with what precision in terms of speed, position … and see what we can hope to do from there. We’ve managed to connect to a Sphero from Minicom (it turns light blue) but before sending any commands we need to go to User Hack mode, and that is done by sending a string of hex-coded ASCII chars which don’t exist in the regular ASCII table. So cool, we didn’t get far there.
  • Another thing is the question of charging, which Loïk tackled : so we have this great (this amazing, this wonderful) induction-driven charging setup, but it appears to do more than charging the batteries with cool (with blazing awesome) cordless induction : it also seems to communicate data, since the base detects when Sphero’s on it and lights up a blue LED, and Sphero detects when it’s on the base and goes to sleep. There’s even a reset button on the base that actually wirelessly resets Sphero ! So we’ll have to see wether all that is really necessary or if we can still charge without implementing any of this and just handle the flow of energy. We’ll have to perform tests and read the documentation (or ask Ian, whose answer we still impatiently await).
  • Components and parts : Last but not least, we’ve talked about CPU needs regarding Kalman filtering, but then there’s the rest, and especially captors. See Loïk’s post about that, but not of all Sphero’s captors can be used. Some we haven’t identified, and some are too small to be soldered by the means available at school. So we’ll have to see what else we still need, see what’s available and make a relevant choice. In particular, we know they had a problem with their quartz clock, because plastic rolling on wool carpets create great deals of static electricity and the thing was often jammed or behaving abnormally because of it – so they had to solder an extra wire from the metal body of the quartz to the ground of the board pretty much by hand ! It’s not good looking – but we, learning from their misadventures, will get a quartz that already has an extra pin connected just for that to its body and which we’ll connect to the ground.

So, the more we think about it, the more questions arise – as half of what we have to do we don’t know how to do. Hopefully (probably, in fact) tomorrow’s design lab will teach us some of the stuff we need there, and for the rest we’ll carry on working – since work, if we listen to Sarkozy who was gone but is now back (or soon to be), is the solution to everything.

PS : Sorry for the long post. Some people, and I must be one of those, just have to write it out, to write it all out, and so they just go on and on, with a profusion of words and a copious quantity of ideas, and very long complex-compound sentences that only seem to be a never-ending sequence of independent clauses, separated by commas and linked by coordinating conjunctions, which can be easily repeated over and over, and that really just annoys the burning hell out of you when it comes to its fourth line (unless you’re Proust). I’ll do something no one ever does on 9gag any more – and it’s a shame – I’ll give you a potato, because I’m sorry for the long post. Actually, I’ll do better, since that PS was long : I’ll give you a double potato AND the history of the “Sorry for the long post” potato at the same time. Sh*t yeah.


Charles THIN

ROSE 2013

Commentaires fermés.