#robotics Logs

Feb 22 2015

01:18 Triffid_Hunter you'd have to either keep it stepping, or provide some counterweight for it to move around
01:18 caveman usually humanoid robots have flat feet so that they can stand still without the need of complicated balancing
01:18 Triffid_Hunter grab videos of stilt-walkers, see what they do
01:18 caveman i am thinking to give it some sort of a small backbone, so that it tilts accordingly to keep it stand still
01:21 caveman my concern right now is: what micro controller to use for such task?
01:25 ace4016 a 32-bit one most likely :P
01:26 ace4016 do you need floating point numbers for the balance computation? and are you going to be doing a lot of them (most likely)?
01:28 Triffid_Hunter caveman: something cortex-m4f
01:29 Triffid_Hunter the m4f runs up to 208MHz or so and has a hardware floating point unit for your EKF and IK math ;)
02:13 rue_bed yea, the tech changed on me again, looks like I'll be using a tablet for the realtime ik of my hexapod
12:15 delinquentme in full step holding, a stepper motor wont be throwing off fluctuating magnetic fields right?
12:16 delinquentme however in microstepping, depending on where within the travel of the resolution , the magnetic fields will be constantly switching right?
12:28 caveman ace4016 / Triffid_Hunter: sorry i was away. are you still there? it is about a robot balancing itself on two stick legs by moving its backbone around to keep itself standing. you said a 32-bit or m4f or something. Q) do i need doing floating point stuff? i don't know, i guess preferably. Q) do i need to perform lots of calculation? i honestly don't know what delay is toleratable, this is something i need to
12:28 caveman work on but any help is appreciated
13:09 ace4016 caveman, that depends on a few things; with fuzzy logic, you might be able to get away with integers or if you find a clever way to work with integers or fixed-point numbers'
13:10 ace4016 also, how often you have to compute your control law will depend on many things, including the inherent stability/instability of the system
13:10 ace4016 you can draw out some math if you'd like before implementing it to see how often and how much effort the control law will need
13:10 ace4016 depends on control scheme too
13:11 ace4016 Kalman, for example, is really good, but really expensive
13:16 caveman ace4016: how about the motors themselves? is servo my only option?
13:18 ace4016 no
13:18 ace4016 you can use any actuator that facilitates control of the CoG/CoM
13:19 ace4016 linear actuators, stepper motors, brushed/brushless motors, etc.
13:19 ace4016 most hobby people use servos because they're easy
13:19 ace4016 and cheap
13:19 ace4016 or a combination of
13:21 caveman cog, as in similar to those in bikes?
13:22 caveman any links?
13:23 caveman CoG?
13:52 ace4016 caveman, Center of Gravity (CoG) and Center of Mass (CoM)
13:53 ace4016 usually they're the same point
13:53 ace4016 i'm not sure of any examples where they're not, but i believe there are situations when they're not the same
13:54 ace4016 don't think they change much until you get to space level physics and such
15:57 dipt jpeg me please
15:57 dipt ))