Optimizing bandwidth and power in wireless networks of smart sensors
View Patent ↗A system that uses statistical techniques to selectively transmit data from a sensor. During operation, the system receives a sequence of quantized values from the sensor. The system then determines whether a distribution for the sequence of quantized values indicates that the sensor is observing a real event. If so, the system transmits sensor data for the real event to a receiver.
1. A method for using statistical techniques to selectively transmit data from a sensor, comprising:
receiving a sequence of quantized values from the sensor; and
for each quantized value in the sequence of quantized values,
updating a bin of a histogram that is associated with the quantized value to indicate that the quantized value was observed;
calculating a frequency for the bin;
determining whether the calculated frequency for the bin has deviated by a specified amount from a mean frequency for the bin by using a Sequential Probability Ratio Test (SPRT), wherein the SPRT is applied to each bin in the histogram; and
if so,
determining that the sensor is observing a real event; and
transmitting sensor data for the real event to a receiver.
2. The method of claim 1 , wherein the SPRT includes at least one of:
a positive-mean test, which tests whether the mean of a time-series for the inferential variable is above a reference level;
a negative-mean test, which tests whether the mean of the time-series for the inferential variable is below a reference level;
a nominal-variance test, which tests whether the variance of the time-series for the inferential variable is proportional to a scale factor;
a inverse-variance test, which tests whether the variance of the time-series for the inferential variable is proportional to the inverse of the scale factor;
a positive first-difference test, which tests whether the time-series for the inferential variable is increasing;
a negative first-difference test, which tests whether the time-series for the inferential variable is decreasing;
a positive first-difference variance test, which tests whether the first difference of the variance of the time-series is increasing; or
a negative first-difference variance test, which tests whether the first difference of the variance of the time-series is decreasing.
3. The method of claim 1 , wherein prior to determining whether the calculated frequency has deviated by the specified amount from the mean frequency, the method further comprises determining the mean frequency and a variance for the quantized value by:
observing a sequence of quantized values during a training phase, wherein the sensor observes background variations during the training phase;
for each quantized value in the sequence of quantized values,
updating a bin of the histogram that is associated with the quantized value to indicate that the quantized value was observed; and
calculating a mean frequency and a variance for the bin.
4. The method of claim 1 , wherein the sensor is located on a smart mote, which measures and preprocesses physical variables, and transmits the resulting data wirelessly to a receiver.
5. A computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for using statistical techniques to selectively transmit data from a sensor, wherein the method comprises:
receiving a sequence of quantized values from the sensor; and
for each quantized value in the sequence of quantized values,
updating a bin of a histogram that is associated with the quantized value to indicate that the quantized value was observed;
calculating a frequency for the bin;
determining whether the calculated frequency for the bin has deviated by a specified amount from a mean frequency for the bin by using a Sequential Probability Ratio Test (SPRT), wherein the SPRT is applied to each bin in the histogram; and
if so,
determining that the sensor is observing a real event; and
transmitting sensor data for the real event to a receiver.
6. The computer-readable storage medium of claim 5 , wherein the SPRT includes at least one of:
a positive-mean test, which tests whether the mean of a time-series for the inferential variable is above a reference level;
a negative-mean test, which tests whether the mean of the time-series for the inferential variable is below a reference level;
a nominal-variance test, which tests whether the variance of the time-series for the inferential variable is proportional to a scale factor;
a inverse-variance test, which tests whether the variance of the time-series for the inferential variable is proportional to the inverse of the scale factor;
a positive first-difference test, which tests whether the time-series for the inferential variable is increasing;
a negative first-difference test, which tests whether the time-series for the inferential variable is decreasing;
a positive first-difference variance test, which tests whether the first difference of the variance of the time-series is increasing; or
a negative first-difference variance test, which tests whether the first difference of the variance of the time-series is decreasing.
7. The computer-readable storage medium of claim 5 , wherein prior to determining whether the calculated frequency has deviated by the specified amount from the mean frequency, the method further comprises determining the mean frequency and a variance for the quantized value by:
observing a sequence of quantized values during a training phase, wherein the sensor observes background variations during the training phase;
for each quantized value in the sequence of quantized values,
updating a bin of the histogram that is associated with the quantized value to indicate that the quantized value was observed; and
calculating a mean frequency and a variance for the bin.
8. The computer-readable storage medium of claim 5 , wherein the sensor is located on a smart mote, which measures and preprocesses physical variables, and transmits the resulting data wirelessly to a receiver.
9. An apparatus that uses statistical techniques to selectively transmit data from a sensor, comprising:
the sensor, which produces a sequence of quantized values;
a processor which is configured to:
for each quantized value in the sequence of quantized values,
update a bin of a histogram that is associated with the quantized value to indicate that the quantized value was observed;
calculate a frequency for the bin;
determine whether the calculated frequency for the bin has deviated by a specified amount from a mean frequency for the bin by using a Sequential Probability Ratio Test (SPRT), wherein the SPRT is applied to each bin in the histogram; and
if so, determine that the sensor is observing a real event; and
a transmitter, wherein if the distribution indicates that the sensor is observing a real event, the transmitter is configured to transmit sensor data for the real event to a receiver.
10. The apparatus of claim 9 , wherein the SPRT includes at least one of:
a positive-mean test, which tests whether the mean of a time-series for the inferential variable is above a reference level;
a negative-mean test, which tests whether the mean of the time-series for the inferential variable is below a reference level;
a nominal-variance test, which tests whether the variance of the time-series for the inferential variable is proportional to a scale factor;
a inverse-variance test, which tests whether the variance of the time-series for the inferential variable is proportional to the inverse of the scale factor;
a positive first-difference test, which tests whether the time-series for the inferential variable is increasing;
a negative first-difference test, which tests whether the time-series for the inferential variable is decreasing;
a positive first-difference variance test, which tests whether the first difference of the variance of the time-series is increasing; or
a negative first-difference variance test, which tests whether the first difference of the variance of the time-series is decreasing.
11. The apparatus of claim 9 , wherein prior to determining whether the calculated frequency has deviated by the specified amount from the mean frequency, the processor is configured to determine the mean frequency and a variance for the quantized value by:
observing a sequence of quantized values during a training phase, wherein the sensor observes background variations during the training phase;
for each quantized value in the sequence of quantized values,
updating a bin of the histogram that is associated with the quantized value to indicate that the quantized value was observed; and
calculating a mean frequency and a variance for the bin.
12. The apparatus of claim 9 , wherein the sensor is located on a smart mote, which measures and preprocesses physical variables, and transmits the resulting data wirelessly to a receiver.