Problem
In Lab 6 we go head to head and challenge ourselves to see if we can manually control the feedback network better than a PI controller can. In our Man vs Machine challenge we are controlling the speed of a DC motor. Our system is built inside MatLab’s Simulink toolbox interfacing with an Arduino.
In the previous lab assignment we introduced PID controllers which minimizing the error of the system. P is the proportional gain, I is the integral gain, and D is the derivative gain. A summation of all PID gains is fed back into the motor model. In our manual control side of the lab we will have to adjust a slider adjusting the gain being fed into the motor model. The problem with using a human as a controller is the reaction time we have in adjusting the gain. In a simulation of only five seconds the accuracy of the manual control will depend on the individual. We also introduce the need to use a filter on our output since the noise from position encoder is amplified when we take the discrete derivative.
Procedure
The first step of the procedure was to download the Simulink block from blackboard.
Once the block was set up we had to find the top speed for our motor by adjusting the slider gain and setting the normalization gain to 1. Once we identified the top speed of our motor, we 1 normalized it by setting the normalization gain, b, to 1/(top speed). Then we configured the low pass filter block as follows.
Then we started the simulation and attempted to get the top speed to settle at .5 by adjusting the slider gain. Then we attempted to do the same thing by using a PI controller designed in parallel.
Once the block was configured, we started testing by setting the P gain to 1 and the I gain to 0. We adjusted the P and I gains in attempting to get the best possible output. The results are shown in the Results section of the report.