Privacy preserving sensor apparatus

1. A privacy preserving sensor apparatus, comprising:

a microphone that is configured to output a signal that is representative of a spoken utterance of a user; and

feature extraction circuitry that is electrically coupled to the microphone, wherein the feature extraction circuitry is configured to extract a set of features from the signal, a voice recognition system is configured to use the set of features to recognize a first word in a predefined set of words when the first word is included in the spoken utterance, and the voice recognition system is not configured to use the set of features to recognize a second word that is outside the predefined set of words when the second word is included in the spoken utterance, wherein the feature extraction circuitry is configured to discard the signal immediately responsive to extracting the set of features from the signal, and further wherein neither the signal nor any features other than the set of features are provided to the voice recognition system.

2. The privacy preserving sensor apparatus of claim 1 , wherein the microphone and the feature extraction circuitry are co-located on a semiconductor chip.

3. The privacy preserving sensor apparatus of claim 2 , wherein the microphone is a micro electro-mechanical system microphone that is embedded in a silicon chip.

4. The privacy preserving sensor apparatus of claim 1 , further comprising a wireless transceiver that is configured to transmit the set of features extracted by the feature extraction circuitry to a computing device that executes the voice recognition system.

5. The privacy preserving sensor apparatus of claim 1 , wherein a computing system is unable to reconstruct the signal based solely upon the set of features.

6. The privacy preserving sensor apparatus of claim 1 , further comprising a tamper-resistant housing that houses the microphone and the feature extraction circuitry.

7. The privacy preserving sensor apparatus of claim 1 , wherein the feature extraction circuitry is a pre-configured application specific integrated circuit.

8. The privacy preserving sensor apparatus of claim 1 , further comprising:

second feature extraction circuitry that is configured to extract a second set of features from the signal output by the microphone, wherein only one of the feature extraction circuitry or the second feature extraction circuitry is electrically coupled to the microphone at any one point in time.

9. The privacy preserving sensor apparatus of claim 8 , further comprising:

a mode selector circuit that couples one of the feature extraction circuitry or the second feature extraction circuitry to the microphone and decouples the other of the feature extraction circuitry or the second feature extraction circuitry from the microphone.

10. The privacy preserving sensor apparatus of claim 1 , further comprising the voice recognition system.

11. A method performed at a privacy preserving sensor apparatus, the method comprising:

generating, by a microphone, a signal that is representative of a spoken utterance of a user;

extracting first features from the signal, the first features are configured for use by a voice recognition system to recognize that the spoken utterance includes a first word in a predefined set of words, the first features are not configured for use by the voice recognition system to recognize that the spoken utterance includes a second word that is outside the predefined set of words;

providing the first features to the voice recognition system responsive to extracting the first features from the signal, wherein the privacy preserving sensor apparatus refrains from providing the voice recognition system with either the signal or second features that are configured for use by the voice recognition system to recognize that the spoken utterance includes the second word; and

discarding the signal immediately responsive to extracting the first features from the signal.

12. The method of claim 11 , wherein the privacy preserving sensor apparatus comprises the voice recognition system, the method further comprising:

recognizing that the spoken utterance of the user includes the first word based upon the first features extracted from the signal.

13. The method of claim 11 , wherein providing the first features to the voice recognition system comprises transmitting a second signal to an access point, wherein the second signal comprises data that is indicative of the first features, and further wherein the access point is in network communication with a computing device that executes the voice recognition system.

14. The method of claim 13 , further comprising:

prior to transmitting the second signal, encrypting the first features, wherein the second signal includes the encrypted first features.

15. A privacy preserving sensor apparatus, comprising:

a microphone;

a processor; and

memory that stores instructions that, when executed by the processor, cause the processor to perform acts comprising:

extracting a set of features from a signal output by the microphone, wherein the signal is representative of a spoken utterance of a user, wherein a voice recognition system is configured to use the set of features to recognize a first word in a predefined set of words when the first word is included in the spoken utterance, and further wherein the voice recognition system is not configured to use the set of features to recognize a second word outside the predefined set of words when the second word is included in the spoken utterance; and

discarding the signal immediately responsive to extracting the set of features from the signal, wherein the voice recognition system is prevented from receiving the signal or any features other than those in the set of features, and further wherein the signal is unable to be reconstructed based solely upon the set of features.

16. The privacy preserving sensor apparatus of claim 15 , the acts further comprising recognizing that the spoken utterance includes the first word based upon the set of features extracted from the signal.

17. The privacy preserving sensor apparatus of claim 15 , the acts further comprising causing the set of features to be transmitted to a computing device that executes the voice recognition system.

18. The privacy preserving sensor apparatus of claim 15 , wherein the voice recognition system is configured to use the set of features to recognize a plurality of words when the plurality of words are included in the spoken utterance, wherein the first word is included in the plurality of words.

19. The privacy preserving sensor apparatus of claim 15 , further comprising a tamper-resistant housing that houses the microphone, the processor, and the memory.