I have a system arm that free-ranges within a certain area. The arm swings, translates and rotates about a vertical axis which sits atop a mobile base. The mobile base can move any direction in the X-Y plane. It's movements are strictly programmed but I would like to start a new version that is semi-autonomous. This new version would need to receive feedback on 3-axis positioning, speed and rotation relative to a fixed point at the edge of the enclosed system (or relative to the position at which it is initialized). Sensor feedback is completely new to me but I am willing to dive in. Could someone recommend a great starting point for me? Thanks!
What kind of range and accuracy are you looking for?
From a repeated position accuracy standpoint, using a accelerometers and gyroscopes would accumulate errors and drift over time.
Rotary and linear position encoders would probably be the way to go.
If you've opened up a more recent mechanical mouse, you've probably seen the slotted plastic wheel running through the gap of a caliper-like component like this one. You count pulses to keep track of where you are. These are pretty cheap, as long as you have a way of easily making/obtaining the necessary rotary or linear code strip.
Accelerometers and Gyroscopes are going to be good, but not perfect. Depending on what your goal of the feedback is, you might want to consider using something like stepper motors. They will give you more of an absolute position of the current rotation of each motor.
If you want to use accelerometers and gyroscopes you should go take a look at this question: How to determine position from gyroscope and accelerometer input?