chip updates: back to plan A [updates]
So I did some thinking and some research. Getting this thing to walk like cheetah is going to be really hard because of the lack of compliance on the actuators.
So I think what we will do is get the robot to walk using a creep gait first. Where the robot knows where it’s CoG is. It moves its legs into a position where it can take weight off a leg. And it moves that leg. And then repeats.
The difference between this and what we tried before is we now know where the CoG is at any time and are dynamically adjusting the legs to achieve a CoG within the support triangle.
The rest, errors in the model and etc, will be accounted for by the IMU.
https://makezine.com/2016/11/22/robot-quadruped-arduino-program/ we're really going to follow something as simple as this for right now. Just to get it walking, turning, and strafing very simply. When we get this done, we'll start looking at state space and LQR control to get much more dynamic stability.
SIMPLE SIMPLE SIMPLE AT LEAST AT FIRST. We'll probably also turn off IMU stabilization for this.