I'm trying to design and build a small direct-drive robot arm, inspired by this cool post: https://news.ycombinator.com/item?id=12536041 https://news.ycombinat

I'm trying to design and build a small direct-drive robot arm, inspired by this cool post: https://news.ycombinator.com/item?id=12536041 https://news.ycombinator.com/item?id=12536041 I haven't worked with electronics since college, so I was trying to find a suitable project. I unconscioussly bought some components and a couple of CHIP computers and I wanted to do something that interacts with the real world (sensors and actuators), but servo-based actuators seemed a bit lame. Initially, I was going to use two motors and a spring per DOF in order to simulate a muscle ("variable stiffness joint", according to the previous post - I was going by intuition). But then I read that direct drive was possible (the motor's magnetic field acts as the "spring") and that once that you have a suitable motor controller built all the rest is software defined, and I found the sweetspot for a project. Such a motor controller is sufficient to keep me researching and prototyping for some time! If I succeed with this, I'm planning on researching how to do some sort of artificial skin for robots. For the moment the idea revolves around connecting a USB camera to a bunch of fiber optic threads from a cheap lamp such as http://prosites-lottofun9.homestead.com/files/fiberlamp2.jpg http://prosites-lottofun9.homestead.com/files/fiberlamp2.jpg and try to detect pressure on loops of the fiber by changes in the light source. That way you can cheaply and efficiently have a huge number of data points, with off-the-shelf components... but if someone does it beforehand I'll also be happy!