Control a flexible link using feedback and an Arduino
How do you move a flexible system (think crane, slewing satellite, light robotic arm) from one position to another rapidly without having vibrations on arrival ? It’s not a easy problem.
I’m interested in trying out a control scheme called Wave Based Control. So I’m making a little rig to test it, to see if I can get it working.
The rig itself is simply a hobby servo, with a flexible arm attached (cable tie with a weight at the end) and an Arduino to do the processing.( EDIT changed from Cable tie to small metal ruler , should be easier to put strain gauges on.)
Goals
1) Physically construct the rig ( Done, used a CD case as a base)
2) Hack the potentiometer inside a hobby servo to get position feedback ( Done, wrecked a few servos doing it)
3 ) Measuring the force in the link (Not done but I’m thinking strain gauge)
4) Feeding the sensor data into a Arduino, creating a control signal for the servo (I’ve written some of the code, but its not complete)
I’ll try a get some photos up soon. (EDIT got one up. Its not wired up properly though. There’s two (set of three) wires coming from the servo, one is for controlling the servo, the other is connected to the servos internal pot pins, the rest is just batteries and a protoshield on top of an arduino)
Are you going to use input shaping techniques?
Thanks for the suggestion,
Really I’m looking for something more robust to changes in natural frequency.
I want to be able to change the weight (double/half/x10) at the end and for it to still work without changing the controller. That’s not really possible with input shaping.
I might try it out though, since its nice a simple to implement and will be a good test before I do feedback.
I’m going to use wave based control. It happens to be part of the work I do and has lots of nice features. Very robust to change in natural frequency and you can add more links (more cable ties with weights) and still keep the controller the same.
The only links I have are to academic papers, so there might be access problems if the particular journal need subscription, I’m not sure. Here’s a link to a paper in the IEEE Transactions on Robotics april 2007 anyway.
(complete instructions)