Simulation of the scheme

Can it work?

• Method
  1. Start with equally spaced phase angles, ie equal intervals between a stimulus in one train and the stimulus in the next train.
  2. Generate n stimulus trains, each at p stimuli per second.
  3. Assuming a standard twitch shape, generate a train of twitches for each input.
  4. Add these twitch trains to get total tension
     • Pass the result through a saturating non-linearity if high stimulation rates are to be simulated.
  5. Find the ac component of tension.
     • There are several acceptable methods, but the best is to subtract from tension a "smoothed tension" formed by box car averaging tension over one period of the stimulation.
  6. Fit a sine wave of frequency p Hz to the ac component of tension.
     • The amplitude of this is the fundamental ripple.
  7. Shift the phase of each train by an amount proportional to the amplitude of the fitted sine wave at the time of its stimulus pulse.
  8. repeat from step 2.
• Results
  • It works!
  • The fundamental ripple decreases quickly and without instability towards zero.

  

  • Other measures of ripple also decrease
  • Order of twitches doesn't affect the ability to zero fundamental component.

  

  (p=5pps. Twitch parameters to match Gastrocnemius)

  • The optimum phase angles found remain optimum when the basic rate p is changed.

  

   

Return to overview