Adaptive sensor sampling for power efficient context aware inferences
Disclosed is a system, apparatus, computer readable storage medium, and method to perform a context inference for a mobile device. In one embodiment, a data processing system includes a processor and a storage device configurable to store instructions to perform a context inference for the data processing system. Data may be received from at least a first sensor, and a first classification of the data from the sensor may be performed. Confidence for the first classification can be determined and a second sensor can be activated based on a determination that the confidence fails to meet a confidence threshold. A data sample classification from the activated second sensor may be classified jointly with the data from first sensor.
1. A method for performing context inference for a mobile device comprising:
receiving a data sample from at least a first sensor;
performing a first classification of the data sample from the first sensor;
determining a confidence value for the first classification, wherein determining the confidence value comprises:
detecting a change in classification from a prior data sample classification, of a prior data sample from the first sensor, to the first data sample classification, wherein the change in classification matches a change in classification predetermined to trigger a confidence failure;
activating a second sensor based on a determination the confidence value of the first classification fails to meet a confidence threshold; and
performing a second classification of a second data sample from the activated second sensor jointly with the data from the at least first sensor.
2. The method of claim 1 , further comprising:
determining a confidence value for the second classification;
activating a third sensor based on a determination that the confidence value fails to meet a confidence threshold; and
performing a third classification of a third data sample.
3. The method of claim 1 , wherein receiving data from at least a first sensor comprises:
receiving data from two or more sensors.
4. The method of claim 1 , wherein the confidence value is based on at least one of: a probability of correctly classifying the first data sample, a comparison of the first data sample to a data sample in a historical data set, and a difference between a nearest classification and a next nearest classification.
5. The method of claim 1 , wherein the first sensor is part of a first sensor subsystem and the second sensor is part of a second sensor subsystem, and wherein activating the second sensor further comprises:
powering up the second sensor subsystem, and
communicating between the first and second sensor subsystems.
6. A machine readable non-transitory storage medium containing executable program instructions which cause a mobile device to perform a method for performing context inference for the mobile device, the method comprising:
receiving a data sample from at least a first sensor;
performing a first classification of the data sample from the first sensor;
determining a confidence value for the first classification, wherein determining the confidence value comprises:
detecting a change in classification from a prior data sample classification, of a prior data sample from the first sensor, to the first data sample classification, wherein the change in classification matches a change in classification predetermined to trigger a confidence failure;
activating a second sensor based on a determination the confidence value of the first classification fails to meet a confidence threshold; and
performing a second classification of a second data sample from the activated second sensor jointly with the data from the at least first sensor.
7. The machine readable non-transitory storage medium of claim 6 , further comprising:
determining a confidence value for the second classification;
activating a third sensor based on a determination that the confidence value fails to meet a confidence threshold; and
performing a third classification of a third data sample.
8. The machine readable non-transitory storage medium of claim 6 , wherein receiving data from at least a first sensor comprises:
receiving data from two or more sensors.
9. The machine readable non-transitory storage medium of claim 6 , wherein the confidence value is based on at least one of: a probability of the classification occurring, a comparison of the first data sample to a data sample in a historical data set, and a difference between a nearest classification and a next nearest classification.
10. The machine readable non-transitory storage medium of claim 6 , wherein the first sensor is part of a first sensor subsystem and the second sensor is part of a second sensor subsystem, and wherein activating the second sensor further comprises:
powering up the second sensor subsystem, and
communicating between the first and second sensor subsystems.
11. A data processing device comprising:
a processor;
a storage device coupled to the processor and configurable for storing instructions, which, when executed by the processor cause the processor to:
receive a data sample from at least a first sensor;
perform a first classification of the data sample from the first sensor;
determine a confidence value for the first classification, wherein to determine the confidence value comprises:
detect a change in classification from a prior data sample classification, of a prior data sample from the first sensor, to the first data sample classification, wherein the change in classification matches a change in classification predetermined to trigger a confidence failure;
activate a second sensor based on a determination the confidence value of the first classification fails to meet a confidence threshold; and
perform a second classification of a second data sample from the activated second sensor jointly with the data from the at least first sensor.
12. The data processing device of claim 11 , further comprising instructions to:
determine a confidence value for the second classification;
activate a third sensor based on a determination that the confidence value fails to meet a confidence threshold; and
perform a third classification of a third data sample.
13. The data processing device of claim 11 , wherein receiving data from at least a first sensor further comprises instructions to:
receive data from two or more sensors.
14. The data processing device of claim 11 , wherein the confidence value is based on at least one of: a probability of the classification occurring, a comparison of the first data sample to a data sample in a historical data set, and a difference between a nearest classification and a next nearest classification.
15. The data processing device of claim 11 , wherein the first sensor is part of a first sensor subsystem and the second sensor is part of a second sensor subsystem, and wherein activating the second sensor further comprises instructions to:
power up the second sensor subsystem, and
communicate between the first and second sensor subsystems.
16. An apparatus for performing context inference for a mobile device comprising:
means for receiving data from at least a first sensor;
means for performing a first classification of a data sample from the first sensor;
means for determining a confidence value for the first classification, wherein determining the confidence value comprises:
detecting a change in classification from a prior data sample classification, of a prior data sample from the first sensor, to the first data sample classification, wherein the change in classification matches a change in classification predetermined to trigger a confidence failure;
means for activating a second sensor based on a determination the confidence value of the first classification fails to meet a confidence threshold; and
means for performing a second classification of a second data sample from the activated second sensor jointly with the data from the at least first sensor.
17. The apparatus of claim 16 , further comprising:
means for determining a confidence value for the second classification;
means for activating a third sensor based on a determination that the confidence value fails to meet a confidence threshold; and
means for performing a third classification of a third data sample.
18. The apparatus of claim 16 , wherein receiving data from at least a first sensor comprises:
receiving data from two or more sensors.
19. The apparatus of claim 16 , wherein the confidence value is based on at least one of: a probability of the classification occurring, a comparison of the first data sample to a data sample in a historical data set, and a difference between a nearest classification and a next nearest classification.
20. The apparatus of claim 16 , wherein the first sensor is part of a first sensor subsystem and the second sensor is part of a second sensor subsystem, and wherein activating the second sensor further comprises:
means for powering up the second sensor subsystem, and
means for communicating between the first and second sensor subsystems.