Methods and apparatuses for calibrating sensors
View Patent ↗Methods and apparatuses to perform calibration of imprecise sensors for power monitoring in a data-processing system are described. The system includes a load coupled to one or more sensors. An electronic load changes a first input signal through one or more sensors by a predetermined amount. A difference in an output signal from the one or more sensors in response to the changing is obtained. The output signal is measured and sampled. A distribution of samples of the output signal is determined. The estimated parameters of the distribution that most likely to explain actual data are determined. Next, a transfer function of the one or more sensors is determined based on the estimated parameters. The input signal through the load is accurately predicted using the transfer function of the one or more sensors to monitor the power usage by the load.
1. A method to calibrate one or more sensors coupled to a load for power monitoring, comprising:
providing a time-varying first input current signal through the one or more sensors;
sampling a time-varying output voltage signal from the one or more sensors to yield samples of the time-varying output voltage signal;
determining a statistical distribution of the samples of the time-varying output voltage signal, wherein the distribution has at least two distribution components;
determining estimated parameters of each of the distribution components; and
using the estimated parameters to determine a transfer function of the one or more sensors.
2. The method of claim 1 , further comprising storing the samples as a vector of sampled voltage values.
3. The method of claim 1 , wherein the determining the estimated parameters includes maximizing the probability that an estimated parameter explains actual distribution of the samples.
4. The method of claim 1 , further including
(a) determining for each of the samples the probability that a sample comes from one of the distribution components; and
(b) determining a mean value and a variance for each of the distribution components.
5. The method of claim 4 , wherein (a) and (b) are continuously repeated until the estimated parameters sufficiently converge.
6. The method of claim 4 , wherein the mean value rejects noise in the samples.
7. The method of claim 1 , further comprising predicting a second input current signal through the load based on the transfer function.
8. The method of claim 1 , wherein the one or more sensors include an inductance.
9. An article of manufacture comprising:
a machine-accessible medium including data that, when accessed by a machine, cause the machine to perform operations to calibrate one or more sensors coupled to a load for power monitoring, comprising:
providing a time-varying first input current signal through the one or more sensors;
sampling a time-varying output voltage signal from the one or more sensors to yield samples of the time-varying output voltage signal;
determining a statistical distribution of the samples of the time-varying output voltage signal, wherein the distribution has at least two distribution components; determining estimated parameters of each of the distribution components; and
using the estimated parameters to determine a transfer function of the one or more sensors.
10. The article of manufacture of claim 9 , wherein the machine- accessible medium further includes data that cause the machine to perform operations, comprising storing the samples as a vector of sampled voltage values.
11. The article of manufacture of claim 10 , wherein the determining the estimated parameters includes maximizing the probability that an estimated parameter explains actual distribution of the samples.
12. The article of manufacture of claim 10 , wherein the machine- accessible medium further includes data that cause the machine to perform operations, comprising
(a) determining for each of the samples the probability that a sample comes from one of the distribution components; and
(b) determining a mean value and a variance for each of the distribution components.
13. The article of manufacture of claim 12 , wherein (a) and (b) are continuously repeated until the estimated parameters sufficiently converge.
14. The article of manufacture of claim 13 , wherein the mean value is to reject noise in samples.
15. The article of manufacture of claim 10 , wherein the machine-accessible medium further includes data that cause the machine to perform operations, comprising predicting a second input current signal through the load based on the transfer function.
16. The article of manufacture of claim 10 , wherein the one or more sensors include an inductance.
17. A system, comprising
means for providing a time-varying first input current signal through the one or more sensors coupled to a load;
means for sampling a time-varying output voltage signal from the one or more sensors to yield samples of the time-varying output voltage signal;
means for determining a statistical distribution of the samples of the time-varying output voltage signal, wherein the distribution has at least two distribution components;
means for determining estimated parameters of each of the distribution components; and
means for using the estimated parameters to determine a transfer function of the one or more sensors; and
means for predicting a second input current signal through the load based on the transfer function.
18. A method, comprising:
calibrating a sensor element within a data-processing system at least once,
wherein the calibrating includes switching in an additional load for at least a period of time; and measuring a first parameter of-associated with the sensor element when the additional load is switched in: measuring a second parameter associated with the sensor element when the additional load is switched out;
determining a transfer function of the sensor element based on the first parameter and the second parameter;
measuring an electrical parameter of the data-processing system based on the transfer function of the sensor element; and
in response to the measuring, adjust an operating parameter of the data-processing system.
19. The method of claim 18 , wherein the sensor element is calibrated one or more times during operation of the data-processing system.
20. The method of claim 19 , wherein the calibrating is performed repeatedly over time.
21. The method of claim 19 , wherein the calibrating is improved by taking a small number of measurements over a large number of samples, wherein each of the samples contains a small number of measurements.