It's been a short week this week, since I've been off two days providing childcare during the summer holidays, but there's been a lot of progress. The week kicked off with a meeting with Brian Henson and visiting researcher Miyong Lee to look into some experiments on weight perception using FATKAT. This is exciting - having actually made the system, it's good to be getting as much use out of it as we can! I also had a meeting with Will Shaw to discuss the priorities for his upcoming studies, particularly in terms of the new version of FATKAT we're building. I have a couple of priorities for rebuilding the electronics next week, so Will's study can get underway, whereas the new casings we've designed can wait. All good stuff.
I also had a productive meeting with our Research and Innovation Services chaps and external consultants about commercialising some of Ian Flatters' work. Again, very exciting - but also hush hush, since we don't want to give details away.
But most of the week has been spent writing and debugging LabVIEW code for fitting Gaussian peaks to velocity profiles for reach-to-grasp actions. As to why, I'll explain in due course - but suffice to say there's been an awful lot of building and rebuilding code to make it more modular, teasing out why particular bugs crop up (missing markers return a value of 37, 000, 000, 000, 000, 000, 000, 000, 000, 000. Put that as a step change through a butterworth filter, and you get some very bizarre results. Then there's breaking the actual reach-to-grasp movement down into its component parts - how do we differentiate where feedforward ends and feedback kicks in? Anyway, it's mostly done now - the technical bits are in place: next week I can focus on the actually analysing data. Hurrah!
No comments:
Post a Comment