Blogathon Part 4: Quest for Robotology

11 Sep / by: M&R / 13 comments /

We’re incredibly pressed for time today, so sadly this final daily post may not be the detailed and frankly hilarious post you’ve come to expect. If you’re just tuning into this blogstravaganza, here are some links you might want to check out.

 

Yesterday we provided some insight into the other things we like to spend our time doing, such as getting crafty, and subtly implied that we’d be returning to the dev-related today. If you guessed that today’s much anticipated topic is Robotology, you totally get 28 points.

We generally tend to vacillate between supreme confidence and terrifying anxiety about this game. Despite months and months of work, it’s still nowhere near complete, but at the same time we’ve managed to get most of the vital parts figured out. Many of these systems are currently up and running, and we’ve learned a lot in the process.

Our roadmap for the next few months is basically “crunch crunch crunch” (mmm! sounds so delicious.) trying to get something complete for the IGF deadline, and then crunch some more trying to get something finished for the end of the year (our hard deadline with the sincerely wonderful OMDC). We also need to wrap up the Office Yeti prototype.

In order to meet these deadlines, we’ve spent the past week getting organized so that we can properly apply some sort of triage to the remaining tasks (we are getting so good at triage). Previously we simply had lists of tasks in no particular order, broken up only by topic (i.e graphics, editor, bugs, etc) which we would trawl through randomly to find the next problem to work on; now we have a nice hierarchical series of tasks which has been sorted according to priority vis-a-vis our deadlines.

This was a somewhat depressing process, since it required cutting features entirely, and postponing others for the foreseeable future; still, it was ultimately worthwhile since we now know exactly what we have to do in the immediate future, and moreover we know we can get it all done.

 

So what specifically are we working on now? We’ve been stuck on control and animation for a while; basically we have a great simulator which lets us model motorized contraptions, but we now need to add AI/logic which drives the motors in order to accomplish things. From the high-level AI architecture to the low-level motor control, this is a problem which isn’t very well documented — most people seem to be content to have dynamics-based animation systems which amount to ragdolls following keyframe animation data, and leave it at that.

This is itself is far from trivial to get working, but even harder is figuring out some way to add feedback to the system, so that the ragdoll isn’t being dragged forward blindly by animation playback, but is instead controlled by a system which takes the current state of the ragdoll into account.

Of course, once we do find a working strategy, the real challenge will be to develop it into a flexible solution that can be applied to the various types of mechanisms we want to control — we don’t want to have to hand-code control systems from the ground up for each type of robot or moving object! Yikes. Ideally we’ll be able to plug together pre-existing control components to create each robot’s control system; barring that, we’ll have a set of common control functionality/utilities and then we’ll write each controller by hand using this “toolbox” rather than from scratch. Still better than nothing!

Aside from control, our other immediate task is getting a proper graphics system up and running. We have various prototypes which demonstrate that our basic ideas work, but we need to take these and create a proper rendering pipeline. This would be a relatively straightforward problem, except for our complete ignorance of how graphics systems are typically organized — we’re used to having this problem already taken care of by the Flash runtime, so we have no idea where to even begin.

 

Anyway, that’s about all we have time to get into right now, unfortunately! We really can’t wait to write up some tutorials on everything we’ve learned, but of course that will be a while since we need to get the actual game done first 😀

And so ends the legendary blogathon. We accomplished pretty much what we set out to, namely to get you readers up to date on where we’re at, and to give our abysmal blog posting statistics a much-needed kick in the ass. We’ll be back in the not-too-distant future with more news about Robotology, hopefully with some more video or screenshots so you can see how things are progressing along with the explanation.

That’s it for now! Thanks for reading, and for all your comments — we really appreciate it.

comments ( 13 )

  • Fantastic four day post-a-thon. You both have such an upbeat and quirky style to your writing and regardless of whether you’re talking about robotology or not I enjoy reading it. Great to hear the biggest mountain has since been climbed with Robotology, and I simply can’t wait.

    We’ve been spoiled the last 4 days, but would it be too much to ask for you guys to post something every like two weeks or so? That seems reasonable to me.

  • Mmmm tasty updates.

    I can’t wait for TNG and TNNG now. Please keep us posted (as in post blog posts frequently)!

  • Good thon, folks. DS9 is my favorite Star Trek series, so you can imagine my intense passion for Office Yeti. If only for conclusion as to what happens to the Ferengi after Nog is left Emporer.

  • Hi,

    I was wondering where are you based?. If for any chance you based in Vancouver I would like to get in touch with you to help with your projects?

    Keep it up!

    Rob

  • @Rob: sadly we’re in Toronto 🙁

    Slick, the developer for N+ XBLA is in Vancouver though! http://www.slickentertainment.com/

  • (seanG I think you mean five day!)
    It was an incredible shock to come by and check this website and find myself five days late for a supermassive blog-a-thon (light cannot escape its event horizon)! >:

    But it was enjoyable and it’s nice to be kept up-to-date by means of awesome blog posts 😀

  • yes yes yes, I’m sorry. I should know better leaving off the 0th post, since I’m a CS person.

  • Oh, the work you’re doing on behavior and animation for physics based critters sounds like all kinds of fun. Your mention of feedback systems reminds me of this, which is rather crude but has always been inspirational to me:
    http://www.karlsims.com/evolved-virtual-creatures.html

  • Yeah, that was one of the first things we read about this sort of thing, it was definitely inspirational! Although for now we’re trying to avoid getting into machine-learning stuff, since it seems pretty complex 🙂

    More recently, Van de Panne (who made Ski Stunt Simulator) has some pretty good papers: http://people.cs.ubc.ca/~van/papers/2009-TOG-taskControl/index.html

  • Hmm, the modular approach sounds like what introversion are working on for subversion, you should really check them out.

    (ps, as for once the Brits got there first lol)

  • We do read their blog.. but they’re almost as bad as we are with updating it 🙂

  • I wish you do it again… But you can’t!!!!!!!!1111!!!

    I’m SOOOOoooo 1337!!!!111!!!!!!! LOLZ!!!

    Sorry… just a mere mocking of n00bs.

    May the Robot be with Us all — otherworld99

    p.s.: you’re wondering why I have different sigs??

    Because they are 3 games!!!

    Also I <3 (love) but D:< (hate) n.

    (hopefully I’ll love the new 2.0 version [of n]!!!)

  • sorry forgot a can in my last post…

    May the Ninja/Robot/Yeti be with Us all — otherworld99

Leave a reply

Archives