This operator uses a kalman filter to estimate the distribution of one or more attribute values.
...
Kalman Filter to produce a statistically optimal estimate of the underlying system state.
Parameter
- Variables: The name of the variables
- Attributes: The attributes to perform feed the filter on
- Transition: The transition matrix 'A'
- ProcessNoise: The process noise matrix 'Q'
- Measurement: The measurement matrix 'H'
- MeasurementNoise: The measurement noise matrix 'R'
- InitialState: The initial state vector 'x' (optional)
- InitialError: The initial error matrix 'P' (optional)
- Control: The control matrix 'B' (optional)
Example
| Code Block | ||||
|---|---|---|---|---|
| ||||
out = KALMAN({MEASUREMENT
VARIABLES = ['x'[1.0]', ],
ATTRIBUTES = ['m'],
TRANSITION = '[1.0]', ProcessNoise
PROCESSNOISE = '[2.0]', ATTRIBUTES
MEASUREMENT = '['x'],1.0]',
MEASUREMENTNOISE = '[4.0]'},
in) |
| Code Block | ||||||
|---|---|---|---|---|---|---|
| ||||||
inout = KALMAN({ VARIABLES = ['x','y','dx','dy'], ATTRIBUTES = ['vx','vy'], INITIALSTATE = '[0.0, 0.0, 0.0, 0.0]', INITIALERROR = '[1.0,0.0,0.0,0.0;0.0,1.0,0.0,0.0;0.0,0.0,1.0,0.0;0.0,0.0,0.0,1.0]', MEASUREMENT = '[0.0,0.0,1.0,0.0;0.0,0.0,0.0,1.0]', TRANSITION = '[1.0,0.0,1.0,0.0;0.0,1.0,0.0,1.0;0.0,0.0,1.0,0.0;0.0,0.0,0.0,1.0]', PROCESSNOISE = '[1/4, 1/4, 1/2, 1/2;1/4, 1/4, 1/2, 1/2; 1/2, 1/2, 1, 1; 1/2, 1/2, 1, 1]', ATTRIBUTES MEASUREMENT = ['x1','x2'], '[0.0,0.0,1.0,0.0;0.0,0.0,0.0,1.0]', MEASUREMENTNOISE = '[10.0,0.0;0.0,10.0]'}, out in) |