I had to make some rough calculations for my system response. First, I need to ensure that the sinusoid peaks were even spaced in the x-direction. This was easy to do. I chose 16 seconds as my period of oscillation with a 1 to 1 conversion rate from seconds to inches. The harder part was to make it so that the decay rate of the sinusoid peaks was exponential. I started with a peak that was 18 inches above the reference input. Then each peak after that had to be smaller and on top of that, the delta between each peak had to be smaller each time too. I ended up with a descent oscillating response that reached the reference input with no steady state error.