Methods, devices and systems for determining the zero rate output of a sensor
A bias value associated with a sensor, e.g., a time-varying, non-zero value which is output from a sensor when it is motionless, is estimated using a ZRO-tracking filter which is a combination of a moving-average filter and a Kalman filter having at least one constraint enforced against at least one operating parameter of the Kalman filter. It achieves faster convergence on an estimated bias value and produces less estimate error after convergence. A resultant bias estimate may then be used to compensate the biased output of the sensor in, e.g., a 3D pointing device.
1. A method for filtering an output of a sensor to compensate for a bias error comprising:
sensing rotation of a device about at least one axis to generate at least one output associated therewith, said at least one output comprising at least one of an angular velocity, an angular rate, and an angular position; and
utilizing a zero-rate output (ZRO) tracking filter configured to receive said at least one output to compensate said at least one output for a zero-rate offset error,
wherein said ZRO filter is implemented as a combination of a Kalman filter having at least one constraint enforced on at least one parameter associated therewith and a moving average filter.
2. The method of claim 1 , wherein said bias is associated with a non-zero value output from said at least one sensor when said device is stationary.
3. The method of claim 1 , wherein said at least one constraint is a lower bound enforced on an error variance (P) of an estimate of said bias.
4. The method of claim 1 , wherein said at least one constraint is an upper bound enforced on an error variance (P) of an estimate of said bias.
5. The method of claim 1 , wherein said at least one constraint is a lower bound enforced on a gain of said Kalman filter.
6. The method of claim 5 , wherein said lower bound is enforced using at least one moving average.
7. The method of claim 6 , wherein said at least one moving average is a cumulative moving average during a first number of samples of said at least one output, and is an exponential moving average after said first number of samples.
8. The method of claim 1 , wherein said at least one constraint is an upper bound enforced on a gain of said Kalman filter.
9. The method of claim 1 , wherein said at least one constraint includes both an upper bound and a lower bound on an error variance (P) of an estimate of said bias and an upper bound and a lower bound on a gain of said Kalman filter.
10. The method of claim 1 , wherein said at least one constraint is selected based on whether said ZRO filter has converged on an estimate of said bias, such that a first constraint is applied before said ZRO filter has converged on said estimate of said bias and a second constraint, different from said first constraint, is applied after said ZRO filter has converged on said estimate of said bias.
11. The method of claim 10 , further comprising:
determining when said ZRO filter has converged on said estimate based on a convergence confidence factor of said ZRO filter.
12. A device comprising:
at least one sensor configured to sense rotation of said device about at least one axis and to generate at least one output associated therewith, said at least one output comprising at least one of an angular velocity, an angular rate, and an angular position; and
a zero rate offset (ZRO) filter configured to receive said at least one output and to compensate said at least one output for zero-rate offset associated with said at least one sensor,
wherein said ZRO filter is implemented as a combination of a Kalman filter and a moving-average filter.
13. The device of claim 12 wherein said Kalman filter has at least one constraint enforced on at least one parameter associated therewith.
14. The device of claim 13 , wherein said Kalman filter has at least one constraint enforced on at least one parameter associated therewith.
15. The device of claim 12 , wherein said ZRO filter switches its computation method for a coefficient from a first technique prior to convergence on a bias estimate to a second technique after convergence on said bias estimate.
16. The device of claim 12 , wherein said ZRO filter evaluates whether a current estimate of said bias has converged to a true value of said bias.
17. The device of claim 12 , wherein said ZRO filter computes measurement variance for a phase “before convergence” of an estimate of said bias using a technique which does not directly count a DC portion of said at least one output.
18. The device of claim 12 , further comprising:
a plurality of instances of said ZRO filters running together with at least one of different conditions and initializations to generate both a quick convergence to said bias and less likelihood of false convergence to said bias.
19. The device of claim 12 , wherein said ZRO filter enforces different bounds on prediction variation before convergence to a bias estimate and after convergence to a bias estimate to reduce backlash pickup.