Kalman filter demonstration using TrackSim program

-------------------- The basics --- use 's' to run, so can pause

SysNoise->Gaussian
Display scale:  128:1 (default)
MeasNoise scale:  1.0 (default)
SysNoise scale:  0.015625 (press 'k' 6 times)

Start off showing actual state only
Show 1D (no horizontal time axis), then time axis
Show measurements on top of actual
Show measurements only -- can you see the state?
Finally, show filter output + actual + measurements

-------------------- Explain ratio of dynamic noise : measurement noise
	note:  MeasNoise/SysNoise = actual MEAS_NOISE/DYN_NOISE = filter

Plot->Display->matrix values (so can see ratio of Q to R)
System->Constant(1D)
SysNoise->Gaussian
MeasNoise->Gaussian
Variables->KF variables
	DYN_NOISE = 0.1	notice state is smoother than meas but jumpy
	DYN_NOISE = 0.01	smoother state (notice Q:R change)
	DYN_NOISE = 0.001	smoother state (notice Q:R change)
	DYN_NOISE = 0.0001	smoother state (notice Q:R change)
	DYN_NOISE = 0.1  MEAS_NOISE = 100		emphasize ratio

-------------------- Inc/dec SysNoise (i/k) and MeasNoise (u/j),
	tuning Q:R to change filter based upon noises

system->reset (and at any time to center display)
actual system MeasNoise=1.0 SysNoise=0.015625
MEAS_NOISE=1.0 DYN_NOISE=0.0001 (looks pretty good)
change SysNoise=0.125 (up three times) (notice gaps in tracking)
change DYN_NOISE=0.01 (tracks better)  show DYN_NOISE=0.001, 0.01, 0.1

-------------------- Show predicted state, how filter works step by step

system->reset (and at any time to center display)
Plot->display->predicted state
MeasNoise=1.0 SysNoise=0.015625 MEAS_NOISE=1.0 DYN_NOISE=0.0001
green line (prediction) + measurement (black)= filter output (red)

-------------------- Explain filter initiation

system->reset (and at any time to center display)
Plot->display->predicted state
MeasNoise=1.0 SysNoise=0.015625 MEAS_NOISE=1.0 DYN_NOISE=0.0001
filter starts at zero; takes time to converge to stable tracking