阅读说明：本技术 低带宽联网系统中的音频处理 (Audio processing in low bandwidth networked systems ) 是由 杰里米·佩恩 托默·阿马里利奥 于 2018-08-01 设计创作，主要内容包括：本公开一般涉及一种用于检测在低带宽网络上发送的输入音频信号内的激活短语的系统。所述系统可以使用两阶段激活短语检测过程。首先,可以包括用于检测输入音频信号的多个麦克风的感测设备可以检测包括候选激活短语的输入音频信号。其次,感测设备可以将输入音频信号的记录发送到客户端设备,以确认输入音频信号包括激活短语。(The present disclosure relates generally to a system for detecting an activation phrase within an input audio signal transmitted over a low bandwidth network. The system may use a two-stage activation phrase detection process. First, a sensing device, which may include multiple microphones for detecting an input audio signal, may detect the input audio signal including a candidate activation phrase. Second, the sensing device may send a recording of the input audio signal to the client device to confirm that the input audio signal includes the activation phrase.)

1. A system for detecting an activation phrase in a remote device, comprising:

a natural language processor component executed by a first client device to:

receiving a first instance of a first input audio signal detected by a first microphone of a sensing device;

parsing the first instance of the first input audio signal to identify a first candidate activation phrase in the first instance of the first input audio signal;

determining that the first candidate activation phrase does not comprise a predetermined activation phrase;

receiving a second instance of the first input audio signal detected by a second microphone of the sensing device;

parsing the second instance of the first input audio signal to identify a second candidate activation phrase in the second instance of the first input audio signal;

determining that the second candidate activation phrase comprises the predetermined activation phrase; and

an interface of the first client device to:

based on determining that the second candidate activation phrase includes the predetermined activation phrase, sending an audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal to a data processing system, the data processing system including a second natural language processor component for identifying a request in at least one of the first instance of the first input audio signal and the second instance of the first input audio signal.

2. The system of claim 1, comprising:

the interface to send, from the first client device to the sensing device, a request for the second instance of the first input audio signal based on a determination that the first candidate activation phrase does not include the predetermined activation phrase.

3. The system of claim 1 or 2, comprising:

the natural language processor component to:

receiving a first instance of a second input audio signal detected by the first microphone of the sensing device;

parsing the first instance of the second input audio signal to identify a third candidate activation phrase;

determining that the third candidate activation phrase comprises the predetermined activation phrase; and

the interface to send a request for a third input audio signal to the sensing device based on determining that the third candidate activation phrase comprises the predetermined activation phrase.

4. The system of any preceding claim, comprising:

the natural language processor component to:

receiving a first instance of a second input audio signal detected by the first microphone of the sensing device;

parsing the first instance of the second input audio signal to identify a third candidate activation phrase;

determining that the third candidate activation phrase comprises the predetermined activation phrase; and

the interface to terminate receiving the second instance of the second input audio signal based on determining that the third candidate activation phrase comprises the predetermined activation phrase.

5. The system of any preceding claim, comprising:

the interface is configured to establish a bluetooth connection between the first client device and the sensing device.

6. The system of any preceding claim, comprising the natural language processor component to:

receiving a third instance of the first input audio signal from a sensor of the first client device;

parsing the third instance of the first input audio signal to identify a third candidate activation phrase in the third instance of the first input audio signal; and

determining that the first input audio signal includes the predetermined activation phrase based at least on the third candidate activation phrase and the second candidate activation phrase.

7. The system of any preceding claim, comprising:

the natural language processor component to:

receiving a first instance of a second input audio signal detected by a first microphone of a second sensing device;

parsing the first instance of the second input audio signal to identify a third candidate activation phrase in the first instance of the second input audio signal;

determining that the third candidate activation phrase does not comprise a predetermined activation phrase;

receiving a second instance of the second input audio signal detected by the first microphone of the second sensing device, the second instance of the second input audio signal having a lower compression rate than the first instance of the second input audio signal;

parsing the second instance of the second input audio signal to identify a fourth candidate activation phrase in the second instance of the second input audio signal;

determining that the fourth candidate activation phrase comprises the predetermined activation phrase; and

the interface of the first client device to:

based on determining that the fourth candidate activation phrase includes the predetermined activation phrase, sending at least one of the first instance of the second input audio signal and the second instance of the second input audio signal to the data processing system including the second natural language processor component to identify a second request in at least one of the first instance of the second input audio signal and the second instance of the second input audio signal.

8. A system for transmitting data in a voice activated network, comprising:

a first microphone of a sensing device for receiving a first instance of a first input audio signal;

a second microphone of the sensing device to receive a second instance of the first input audio signal;

a natural language processor component executed by the sensing device to parse the first instance of the first input audio signal to identify an activation phrase;

an interface of the sensing device to send, at a first point in time, the first instance of the first input audio signal to a client device based on the identification of the activation phrase in the first instance of the first input audio signal, the client device including a second natural language processor component;

the interface of the sensing device to send the second instance of the first input audio signal to the client device at a second point in time after the first point in time; and

the interface of the sensing device to send an audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal to the client device based on an acknowledgement message from the client device of the identification of the activation phrase in the second instance of the first input audio signal.

9. The system of claim 8, comprising the interface to:

at the first point in time, sending the first instance of the first input audio signal to the client device at a first compression level; and

at the second point in time, sending the second instance of the first input audio signal to the client device at a second compression level that is lower than the first compression level.

10. The system of claim 8 or 9, comprising the interface to:

at the first point in time, sending the second instance of the first input audio signal to the client device at a first compression level; and

at the second point in time, sending the first instance of the first input audio signal and the second instance of the input audio signal to the client device at a second compression level that is lower than the first compression level.

11. The system of any of claims 8, 9 or 10, comprising the interface to:

at the second point in time, sending the second instance of the first input audio signal to the client device based on a confirmation message that the activation phrase is not in the first instance of the input audio signal.

12. The system of any of claims 8 to 11, comprising the interface to:

at the second point in time, sending the second instance of the first input audio signal to the client device prior to receiving a confirmation message that the activation phrase is not in the first instance of the input audio signal.

13. The system of claim 12, comprising:

the interface to terminate the sending of the second interface of the first input audio signal based on an acknowledgement message of the activation phrase in the first instance of the input audio signal.

14. The system of any of claims 8 to 13, comprising the interface to:

establishing a Bluetooth connection with the client device;

transmitting the first instance of the first input audio signal and the second instance of the first input audio signal over the Bluetooth connection.

15. A method of transmitting data in a voice activated network, comprising:

receiving, by a first microphone of a sensing device, a first instance of a first input audio signal;

receiving, by a second microphone of the sensing device, a second instance of the first input audio signal;

parsing, by a natural language processor component executed by the sensing device, the first instance of the first input audio signal to identify an activation phrase;

sending, by an interface of the sensing device, the first instance of the first input audio signal to a client device at a first point in time based on the identification of the activation phrase in the first instance of the first input audio signal, the client device comprising a second natural language processor component;

sending, by the interface of the sensing device, the second instance of the first input audio signal to the client device at a second point in time after the first point in time; and

sending, by the interface of the sensing device, an audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal based on a confirmation message from the client device of the identification of the activation phrase in the second instance of the first input audio signal.

16. The method of claim 15, comprising:

sending, by the interface, the first instance of the first input audio signal to the client device at the first point in time at a first compression level; and

sending, by the interface, the second instance of the first input audio signal to the client device at a second compression level lower than the first compression level at the second point in time.

17. The method according to claim 15 or 16, comprising:

sending, by the interface, the second instance of the first input audio signal to the client device at the first point in time at a first compression level; and

sending, by the interface, the first instance of the first input audio signal and the second instance of the input audio signal to the client device at a second compression level lower than the first compression level at the second point in time.

18. The method of any of claims 15 to 17, comprising:

sending, by the interface, the second instance of the first input audio signal to the client device at the second point in time based on a confirmation message that the activation phrase is not in the first instance of the input audio signal.

19. The method of any of claims 15 to 18, comprising:

sending, by the interface, the second instance of the first input audio signal to the client device at the second point in time prior to receiving a confirmation message that the activation phrase is not in the first instance of the input audio signal.

20. The method of claim 19, comprising:

terminating, by the interface, the sending of the second interface of the first input audio signal based on an acknowledgement message of the activation phrase in the first instance of the input audio signal.

21. A system for detecting an activation phrase in a remote device, comprising:

a natural language processor component executed by a first client device to:

receiving a first instance of a first input audio signal detected by a first microphone of a sensing device;

parsing the first instance of the first input audio signal to identify a first candidate activation phrase in the first instance of the first input audio signal;

determining that the first candidate activation phrase does not comprise a predetermined activation phrase;

receiving a second instance of the first input audio signal detected by a second microphone of the sensing device;

parsing the second instance of the first input audio signal to identify a second candidate activation phrase in the second instance of the first input audio signal;

determining that the second candidate activation phrase comprises the predetermined activation phrase; and

an interface of the first client device to:

based on determining that the second candidate activation phrase includes the predetermined activation phrase, sending at least one of the first instance of the first input audio signal and the second instance of the first input audio signal to a data processing system comprising a second natural language processor component to identify a request in at least one of the first instance of the first input audio signal and the second instance of the first input audio signal.

22. The system of claim 21, comprising:

the interface to send, from the first client device to the sensing device, a request for the second instance of the first input audio signal based on a determination that the first candidate activation phrase does not include the predetermined activation phrase.

23. The system of claim 21, comprising:

the natural language processor component to:

receiving a first instance of a second input audio signal detected by the first microphone of the sensing device;

parsing the first instance of the second input audio signal to identify a third candidate activation phrase;

determining that the third candidate activation phrase comprises the predetermined activation phrase; and

the interface to send a request for a third input audio signal to the sensing device based on determining that the third candidate activation phrase comprises the predetermined activation phrase.

24. The system of claim 21, comprising:

the natural language processor component to:

receiving a first instance of a second input audio signal detected by the first microphone of the sensing device;

parsing the first instance of the second input audio signal to identify a third candidate activation phrase;

determining that the third candidate activation phrase comprises the predetermined activation phrase; and

the interface to terminate receiving the second instance of the second input audio signal based on determining that the third candidate activation phrase comprises the predetermined activation phrase.

25. The system of claim 21, comprising:

the interface is configured to establish a bluetooth connection between the first client device and the sensing device.

26. The system of claim 21, comprising the natural language processor component to:

receiving a third instance of the first input audio signal from a sensor of the first client device;

parsing the third instance of the first input audio signal to identify a third candidate activation phrase in the third instance of the first input audio signal; and

determining that the first input audio signal includes the predetermined activation phrase based at least on the third candidate activation phrase and the second candidate activation phrase.

27. The system of claim 21, comprising:

the natural language processor component to:

receiving a first instance of a second input audio signal detected by a first microphone of a second sensing device;

parsing the first instance of the second input audio signal to identify a third candidate activation phrase in the first instance of the second input audio signal;

determining that the third candidate activation phrase does not comprise a predetermined activation phrase;

receiving a second instance of the second input audio signal detected by the first microphone of the second sensing device, the second instance of the second input audio signal having a lower compression rate than the first instance of the second input audio signal;

parsing the second instance of the second input audio signal to identify a fourth candidate activation phrase in the second instance of the second input audio signal;

determining that the fourth candidate activation phrase comprises the predetermined activation phrase; and

an interface of the first client device to:

based on determining that the fourth candidate activation phrase includes the predetermined activation phrase, sending at least one of the first instance of the second input audio signal and the second instance of the second input audio signal to the data processing system including the second natural language processor component to identify a second request in at least one of the first instance of the second input audio signal and the second instance of the second input audio signal

28. A system for transmitting data in a voice activated network, comprising:

a first microphone of a sensing device to receive a first instance of a first input audio signal and a first instance of a second input audio signal;

a second microphone of the sensing device to receive a second instance of the first input audio signal and a second instance of the second input audio signal;

a natural language processor component executed by the sensing device to parse the first instance of the first input audio signal to identify an activation phrase;

an interface of the sensing device to send the first instance of the first input audio signal to a client device at a first point in time based on identification of the activation phrase in the first instance of the first input audio signal, the client device including a second natural language processor component;

the interface of the sensing device to send the second instance of the first input audio signal to the client device at a second point in time after the first point in time; and

the interface of the sensing device to send the first instance of the second input audio signal to the client device based on a confirmation message from the client device of the identification of the activation phrase in the second instance of the first input audio signal.

29. The system of claim 28, comprising the interface to:

at the first point in time, sending the first instance of the first input audio signal to the client device at a first compression level; and

at the second point in time, sending the second instance of the first input audio signal to the client device at a second compression level that is lower than the first compression level.

30. The system of claim 28, comprising the interface to:

at the first point in time, sending the second instance of the first input audio signal to the client device at a first compression level; and

at the second point in time, sending the first instance of the first input audio signal and the second instance of the input audio signal to the client device at a second compression level that is lower than the first compression level.

31. The system of claim 28, comprising the interface to:

sending the second instance of the first input audio signal to the client device at the second point in time based on a confirmation message that the activation phrase is not in the first instance of the input audio signal.

32. The system of claim 28, comprising the interface to:

sending the second instance of the first input audio signal to the client device at the second point in time prior to receiving a confirmation message that the activation phrase is not in the first instance of the input audio signal.

33. The system of claim 32, comprising:

the interface to terminate the sending of the second interface of the first input audio signal based on an acknowledgement message of the activation phrase in the first instance of the input audio signal.

34. The system of claim 28, comprising the interface to:

establishing a Bluetooth connection with the client device;

transmitting the first instance of the first input audio signal and the second instance of the first input audio signal over the Bluetooth connection.

35. A method for transmitting data in a voice activated network, comprising:

receiving, by a first microphone of a sensing device, a first instance of a first input audio signal and a first instance of a second input audio signal;

receiving, by a second microphone of the sensing device, a second instance of the first input audio signal and a second instance of the second input audio signal;

parsing, by a natural language processor component executed by the sensing device, the first instance of the first input audio signal to identify an activation phrase;

sending, by an interface of the sensing device, the first instance of the first input audio signal to a client device at a first point in time based on the identification of the activation phrase in the first instance of the first input audio signal, the client device comprising a second natural language processor component;

sending, by the interface of the sensing device, the second instance of the first input audio signal to the client device at a second point in time after the first point in time; and

sending, by the interface of the sensing device, the first instance of the second input audio signal to the client device based on a confirmation message from the client device of the identification of the activation phrase in the second instance of the first input audio signal.

36. The method of claim 35, comprising:

sending, by the interface, the first instance of the first input audio signal to the client device at the first point in time at a first compression level; and

transmitting, by the interface, the second instance of the first input audio signal to the client device at a second compression level lower than the first compression level at the second point in time.

37. The method of claim 35, comprising:

sending, by the interface, the second instance of the first input audio signal to the client device at the first point in time at a first compression level; and

sending, by the interface, the first instance of the first input audio signal and the second instance of the input audio signal to the client device at a second compression level that is lower than the first compression level at the second point in time.

38. The method of claim 35, comprising:

sending, by the interface, the second instance of the first input audio signal to the client device at the second point in time based on a confirmation message that the activation phrase is not in the first instance of the input audio signal.

39. The method of claim 35, comprising:

sending, by the interface, the second instance of the first input audio signal to the client device at the second point in time prior to receiving a confirmation message that the activation phrase is not in the first instance of the input audio signal.

40. The method of claim 39, comprising:

terminating, by the interface, the sending of the second interface of the first input audio signal based on an acknowledgement message of the activation phrase in the first instance of the input audio signal.

Background

The networked devices may process the audio-based signals. The ability of the device to process the audio-based signal may be based on the quality of the audio-based signal. High quality audio-based signals may have a relatively large file size. Packet-based or other excessive network transmission of network traffic data between computing devices may prevent the computing devices from properly processing audio-based signals, performing operations related to the audio-based signals, or responding to the audio-based signals in a timely manner.

Disclosure of Invention

In accordance with at least one aspect of the present disclosure, a system for detecting an activation phrase in a remote device may include a natural language processor component executed by a first client device. The system may receive a first instance of a first input audio signal detected by a sensing device. The system may parse the first instance of the first input audio signal to identify a first candidate activation phrase in the first instance of the first input audio signal. The system may determine that the first candidate activation phrase does not comprise the predetermined activation phrase. The system may receive a second instance of the first input audio signal obtained by the sensing device. The system may parse the second instance of the first input audio signal to identify a second candidate activation phrase in the second instance of the first input audio signal. The system may determine that the second candidate activation phrase comprises a predetermined activation phrase. The system may include an interface to transmit an audio signal associated with at least one of a first instance of the first input audio signal and a second instance of the first input audio signal based on a determination that the second candidate activation phrase comprises the predetermined activation phrase. The data processing system may include a second natural language processor component to identify a request in at least one of the first instance of the first input audio signal and the second instance of the first input audio signal.

According to at least one aspect of the present disclosure, a system for transmitting data in a voice activated network may include a client device for receiving a first instance of a first input audio signal. The system may include a client device to receive a second instance of the first input audio signal. The system may include a natural language processor component executed by the client device to parse a first instance of the first input audio signal to identify the activation phrase. The system may include an interface of the client device to transmit a first instance of the first input audio signal to the data processing system at a first point in time based on the identification of the activation phrase in the first instance of the first input audio signal. The data processing system may include a second natural language processor component. The interface of the client device may transmit a second instance of the first input audio signal to the data processing system at a second point in time after the first point in time. The interface of the client device may send an audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal to the data processing system based on a confirmation message from the data processing system identifying the activation phrase in the second instance of the first input audio signal.

According to at least one aspect of the present disclosure, a method of transmitting data in a voice-activated network may include receiving, by a client device, a first instance of a first input audio signal. The method may include obtaining, by the client device, a second instance of the first input audio signal. The method may include a natural language processor component executed by the client device parsing a first instance of the first input audio signal to identify an activation phrase. The method can comprise the following steps: based on the identification of the activation phrase in the first instance of the first input audio signal, the interface of the client device sends the first instance of the first input audio signal to the data processing system at a first point in time. The data processing system may include a second natural language processor component. The method may include transmitting, by the interface of the client device, a second instance of the first input audio signal to the data processing system at a second point in time after the first point in time. The method may include sending, by an interface of the client device, an audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal to the data processing system based on the identified confirmation message of the activation phrase in the second instance of the first input audio signal from the data processing system.

According to at least one aspect of the disclosure, the system for detecting an activation phrase in a remote device may include a natural language processor component executed by a first client device. The system may receive a first instance of a first input audio signal detected by a first microphone of a sensing device. The system may parse the first instance of the first input audio signal to identify a first candidate activation phrase in the first instance of the first input audio signal. The system may determine that the first candidate activation phrase does not comprise the predetermined activation phrase. The system may receive a second instance of the first input audio signal detected by a second microphone of the sensing device. The system may parse the second instance of the first input audio signal to identify a second candidate activation phrase in the second instance of the first input audio signal. The system may determine that the second candidate activation phrase comprises a predetermined activation phrase. The system may include an interface to send at least one of a first instance of the first input audio signal and a second instance of the first input audio signal to the data processing system based on a determination that the second candidate activation phrase comprises the predetermined activation phrase. The data processing system may include a second natural language processor component to identify a request in at least one of the first instance of the first input audio signal and the second instance of the first input audio signal.

According to at least one aspect of the present disclosure, a system for transmitting data in a voice-activated network may include a first microphone of a client device for receiving a first instance of a first input audio signal and a first instance of a second input audio signal. The system may include a second microphone of the client device for receiving a second instance of the first input audio signal and a second instance of the second input audio signal. The system may include a natural language processor component executed by the client device to parse a first instance of the first input audio signal to identify the activation phrase. The system may include an interface of the client device to transmit, at a first point in time, a first instance of a first input audio signal to the data processing system based on identification of an activation phrase in the first instance of the first input audio signal. The data processing system may include a second natural language processor component. The interface of the client device may transmit a second instance of the first input audio signal to the data processing system at a second point in time after the first point in time. The interface of the client device may send the first instance of the second input audio signal to the data processing system based on a confirmation message from the data processing system identifying the activation phrase in the second instance of the first input audio signal.

According to at least one aspect of the present disclosure, a method of transmitting data in a voice-activated network may include receiving, by a first microphone of a client device, a first instance of a first input audio signal and a first instance of a second input audio signal. The method may include receiving, by a second microphone of the client device, a second instance of the first input audio signal and a second instance of the second input audio signal. The method may include a natural language processor component executed by the client device parsing a first instance of the first input audio signal to identify an activation phrase. The method can comprise the following steps: at a first point in time, an interface of the client device sends a first instance of a first input audio signal to a data processing system based on recognition of an activation phrase in the first instance of the first input audio signal. The data processing system may include a second natural language processor component. The method may include transmitting, by the interface of the client device, a second instance of the first input audio signal to the data processing system at a second point in time after the first point in time. The method may include sending, by the interface of the client device, the first instance of the second input audio signal to the data processing system based on a confirmation message from the data processing system identifying the activation phrase in the second instance of the first input audio signal.

Each aspect may optionally include one or more of the following features. The system may include first and second microphones. A first instance of the first input audio signal may be detected by a first microphone and a second instance of the first input audio signal may be detected by a second microphone. A first instance of a second input audio signal may be received and a second instance of the second input audio signal may be received. The audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal may be at least one of the first instance of the second input audio signal and the second instance of the second input audio signal. Alternatively, the audio signal associated with at least one of the first instance of the first input audio signal and the second instance of the first input audio signal may be at least one of the first instance of the first input audio signal and the second instance of the first input audio signal or a portion thereof.

The interface may send a request for a second instance of the first input audio signal from the first client device to the sensing device based on determining that the first candidate activation phrase does not include the predetermined activation phrase. The natural language processor component may: receiving a first instance of a second input audio signal detected by a first microphone of a sensing device; parsing the first instance of the second input audio signal to identify a third candidate activation phrase; determining that the third candidate activation phrase comprises the predetermined activation phrase; and the interface may send a request for a third input audio signal to the sensing device based on determining that the third candidate activation phrase comprises the predetermined activation phrase. The natural language processor component may: receiving a first instance of a second input audio signal detected by a first microphone of a sensing device; parsing the first instance of the second input audio signal to identify a third candidate activation phrase; determining that the third candidate activation phrase comprises the predetermined activation phrase; and the interface may terminate reception of the second instance of the second input audio signal based on determining that the third candidate activation phrase comprises the predetermined activation phrase. The interface may establish a bluetooth connection between the first client device and the sensing device. The natural language processor component may: receiving a third instance of the first input audio signal from a sensor of the first client device; parsing the third instance of the first input audio signal to identify a third candidate activation phrase in the third instance of the first input audio signal; and determining that the first input audio signal contains the predetermined activation phrase based on at least the third candidate activation phrase and the second candidate activation phrase. The natural language processor component may: receiving a first instance of a second input audio signal detected by a first microphone of a second sensing device; parsing the first instance of the second input audio signal to identify a third candidate activation phrase in the first instance of the second input audio signal; determining that the third candidate activation phrase does not comprise the predetermined activation phrase; receiving a second instance of a second input audio signal detected by a first microphone of a second sensing device, the second instance of the second input audio signal having a lower compression rate than the first instance of the first input audio signal; parsing the second instance of the second input audio signal to identify a fourth candidate activation phrase in the second instance of the second input audio signal; determining that the fourth candidate activation phrase comprises the predetermined activation phrase; the interface of the first client device may: based on a determination that the fourth candidate activation phrase comprises the predetermined activation phrase, sending at least one of the first instance of the second input audio signal and the second instance of the second input audio signal to a data processing system comprising a second natural language processor component to identify a second request in at least one of the first instance of the second input audio signal and the second instance of the second input audio signal.

The interface may be: at a first point in time, sending a first instance of a first input audio signal to a client device at a first compression level; and at a second point in time, sending a second instance of the first input audio signal to the client device at a second compression level lower than the first compression level. The interface may transmit a second instance of the first input audio signal to the client device at a first point in time at a first compression level; and at a second point in time, transmitting the first instance of the first input audio signal and the second instance of the input audio signal to the client device at a second compression level that is lower than the first compression level. The interface may be: at a second point in time, a second instance of the first input audio signal is sent to the client device based on a confirmation message that the activation phrase is not in the first instance of the input audio signal. The interface may be: sending a second instance of the first input audio signal to the client device at a second point in time before receiving a confirmation message that the activation phrase is not in the first instance of the input audio signal. The interface may terminate transmission of the second interface of the first input audio signal based on an acknowledgement message of the activation phrase in the first instance of the input audio signal. The interface may be: establishing a Bluetooth connection with a client device; the first instance of the first input audio signal and the second instance of the first input audio signal are transmitted over a bluetooth connection.

These and other aspects and embodiments are discussed in detail below. The foregoing information and the following detailed description include illustrative examples of various aspects and embodiments, and provide an overview or framework for understanding the nature and character of the claimed aspects and embodiments. The accompanying drawings provide an illustration and a further understanding of the various aspects and embodiments, and are incorporated in and constitute a part of this specification.

Drawings

The drawings are not intended to be drawn to scale. Like reference numbers and designations in the various drawings indicate like elements. For purposes of clarity, not every component may be labeled in every drawing. In the figure:

FIG. 1 illustrates an example system for detecting an activation phrase in an input audio signal transmitted in a low bandwidth network.

Fig. 2 shows a top view of the vehicle and shows the interior compartment of the vehicle shown in fig. 1.

FIG. 3 illustrates a block diagram of an example method for detecting an activation phrase in a networked system with limited bandwidth.

FIG. 4 is a block diagram of an example computer system.

Detailed Description

The following are more detailed descriptions of various concepts related to and embodiments of methods, apparatus, and systems for multi-modal transmission of packet data in a computer network environment based on voice activated data packets. The various concepts introduced above and discussed in more detail below may be implemented in any of a number of ways.

The present disclosure relates generally to a system for detecting an activation phrase within an input audio signal transmitted over a low bandwidth network. For example, one or more devices of the system may be communicatively coupled together via a bluetooth connection. The system may use a two-stage activation phrase detection process. First, a sensing device, which may include multiple microphones for detecting an input audio signal, may detect the input audio signal with a candidate activation phrase. When determining whether the input audio signal includes an activation phrase, the language processing of the sensing device may be inaccurate and trigger false positives. Second, the sensing device may send a recording of the input audio signal to the client device to confirm that the input audio signal includes the activation phrase. Due to the limited bandwidth of the link between the sensing device and the client device, sending both recordings by the sensing device at the same time introduces a delay when the data transfer is complete.

To reduce the delay of receiving an acknowledgement from the client device, the sensing device may transmit a first recording of the input audio signal to the client device. If the client device determines that an activation phrase is present in the first record, other records made by the sensing device may be discarded and not sent to the client device, thereby saving bandwidth. If the client device is unable to recognize the activation phrase in the first recording, the client device may receive a second recording from the sensing device. The client device may then process the second recording to confirm whether the input audio signal includes the activation phrase. Thus, the client device may use the recordings made by the sensing device to provide improved detection of the activation phrase without introducing delays due to additional information sent between the sensing device and the client device.

FIG. 1 illustrates an example system 100 for detecting an activation phrase in an input audio signal transmitted in a low bandwidth network. The system 100 may include at least one data processing system 102, one or more client devices 128, and one or more sensing devices 142 that may be located in a vehicle 140, such as an automobile. The sensing device 142 may be referred to as a remote device because the sensing device 142 may be located remotely from the client device 128 in order to detect and transmit the input audio signal to the client device 128.

The data processing system 102 may include an interface 104. The data processing system 102 may include a Natural Language Processor (NLP) component 106 for parsing an audio-based input, such as an input audio signal. The data processing system 102 may include an interface management component 108 for detecting and managing interfaces of other devices in the system 100. The data processing system 102 may include an audio signal generator component 110 for generating an audio-based signal. Data processing system 102 may include a direct action Application Programming Interface (API) 112. The data processing system 102 can include a response selector component 114 for selecting a response to an audio-based input signal. The data processing system 102 may include a data repository 118, wherein the data processing system 102 may store parameters 120, policies 122, response data 124, and templates 126. The client device 128 and the sensing device 142 may include and execute instances of components of the data processing system 102. In some implementations, the client device 128 and the sensing device 142 may each comprise an instance of the data processing system 102 or any component thereof.

The system 100 may also include one or more client devices 128. The client device 128 may include a sensor 130, a speaker 132(1), an interface 134, and a transducer 136 (1). The client device 128 may execute an instance of the NLP component 106. The system 100 may also include one or more data providers 138. The system 100 may include one or more vehicles 140. Each vehicle 140 may include one or more sensing devices 142. Sensing device 142 may include a first sensor 130(1), a second sensor 130(2) (which may be generically referred to as sensor 130), a speaker 132(2), an interface 134, and a transducer 136 (2). The sensing devices 142 can each execute an instance of the NLP component 106. The client device's speaker 132(1) and the sensing device's speaker 132(2) may be the same type of speaker or different types of speakers. The client device's speaker 132(1) and the sensing device's speaker 132(2) may be collectively referred to as speaker 132. The transducer 136(1) of the client device and the transducer 136(2) of the sensing device may be the same type of transducer or different types of transducers. Transducer 136(1) of the client device and transducer 136(2) of the sensing device may be collectively referred to as transducer 136.

The components of system 100 may communicate over a network 116. One or more of the client devices 128 may be located within an instance of the vehicle 140. For example, the client device 128 may be a mobile phone of a driver driving the vehicle 140. One or more client devices 128 may be remote from the vehicle 140. For example, after the driver parks and leaves the vehicle 140 for work, the driver's mobile phone is away from the vehicle 140. The interface 104 of the client device 128 may establish a connection with a sensing device 142 within the vehicle 140 via bluetooth when within a predetermined proximity of the vehicle 140. For example, the portion of the network 116 between the client device 128 and the sensing device 142 may be a bluetooth connection.

The network 116 may include a computer network such as the internet, a local, wide area, metropolitan area, or other area network, an intranet, a satellite network, other computer networks such as voice or data mobile telephone communication networks, and combinations thereof. The network 116 may be used by the data processing system 102, the client device 128, and the sensing device 142 to access information resources of the data provider 138, such as web pages, web sites, domain names, uniform resource locators. For example, the data processing system 102 may access a data provider 138 via the network 116, the data provider 138 providing weather data for a particular location, such as a location associated with the vehicle 140.

The network 116 may include, for example, a point-to-point network, a broadcast network, a wide area network, a local area network, a telecommunications network, a data communications network, a computer network, an ATM (asynchronous transfer mode) network, a SONET (synchronous optical network) network, an SDH (synchronous digital hierarchy) network, a wireless network, or a wired network, and combinations thereof. The network 116 may include a wireless link, such as a bluetooth link, an infrared channel, or a satellite band. The topology of the network 116 may include a bus, star, or ring network topology. The network 116 may include a mobile telephone network using any one or more protocols for communicating between mobile devices, including advanced mobile phone protocol ("AMPS"), time division multiple access ("TDMA"), code division multiple access ("CDMA"), global system for mobile communications ("GSM"), general packet radio service ("GPRS"), or universal mobile telecommunications system ("UMTS"). Different types of data may be sent via different protocols, or the same type of data may be sent via different protocols.

The client device 128 and the sensing device 142 may each include at least one logical device, such as a computing device having a processor, for communicating with each other and with the data processing system 102 via the network 116. The sensing device 142 may communicate with the data processing system 102 via the client device 128. For example, the sensing device 142 may transmit the input audio signal to the client device 128 via a bluetooth connection, and the client device 128 may transmit the input audio signal to the data processing system 102.

The client device 128 and the sensing device 142 may include instances of any of the components described with respect to the data processing system 102, such as the NLP component 106. The client device 128 and the sensing device 142 may comprise instances of the data processing system 102.

The client device 128 may include a desktop computer, laptop computer, tablet computer, personal digital assistant, smart phone, mobile device, portable computer, thin client computer, virtual server, speaker-based digital assistant, or other computing device. The client device 128 may be a mobile phone of the user. An instance of a digital assistant executed by the client device 128 may receive an input audio signal from a user. The input audio signal may include a voice-based or audio-based request. For example, the user may generate the utterance "Ok, play the next song. The client device 128 may detect the input audio signal and send the input audio signal to the data processing system 102, and the data processing system 102 may generate a response or action based on the request within the input audio signal. .

Sensing device 142 may be a remote sensing device. The sensing device 142 may be remote from the client device 128, but within a predetermined distance of the client device 128. The predetermined distance may be a distance over which a bluetooth connection between the client device 128 and the sensing device 142 may be established. The sensing device 142 may include one or more sensors 130 for detecting input audio signals generated by a user. The sensing device 142 may send instances of the input audio signal captured by the sensing device's sensors 130 to the client device 128 for processing. In the event that the sensor 130 of the client device is unable to detect the input audio signal, the sensor 130 of the sensing device may detect the input audio signal. For example, the sensing device 142 may be positioned within an open area and may detect the input audio signal when the client device 128 is, for example, in a user's pocket and is unable to detect the input audio signal.

As shown in fig. 1, the sensing device 142 may be configured as a vehicle accessory. For example, the sensing device 142 may be a bluetooth enabled charging device. The sensing device 142 may be a power adapter that adapts the output of the cigarette lighter of the vehicle to one or more USB ports. For example, the sensing device 142 may convert 12V or 24V of a cigarette lighter of a vehicle to 5V and 1A or 5V and 2.1A. The sensing device 142 may generate other output voltages and currents, such as a voltage between about 5V and about 10V. The sensing device 142 may include one or more processors that may execute an instance of the data processing system 102 or any of its components. The sensing device 142 may be removable from the vehicle 140. For example, the sensing device 142 may be a removable power adapter that may be inserted into and then removed from a cigarette lighter of a vehicle. The sensing device 142 may be a permanent component of the vehicle 140. For example, the sensing device 142 may be a component of a vehicle head unit. The vehicle 140 may be an automobile, truck, motorcycle, or other transportation device.

Sensing device 142 may be configured as a remote sensing device for non-vehicular environments. For example, the sensing device 142 may be a sensor array that may include a plurality of sensors 130. The sensor array may be located in a room in the user's home to capture and transmit input audio signals to, for example, a speaker-based digital assistant located in a different room in the user's home.

The client device 128 and the sensing device 142 may include at least one sensor 130, at least one transducer 136, at least one audio driver, and at least one speaker 132. The sensor 130 may include a microphone or an audio input sensor. The sensors 130 may also include one or more of a GPS sensor, a proximity sensor, an ambient light sensor, a temperature sensor, a motion sensor, an accelerometer, or a gyroscope. The transducer 136 may convert the audio input into an electronic signal. The audio driver may include scripts or programs that are executed by one or more processors of the client device 128 or the sensing device 142 to control the speaker 132. The speaker 132 may present audio signals by converting electrical signals into audible waves.

The client device 128 and the sensing device 142 may be associated with an end user who inputs a voice query as an input audio signal to the client device 128 or the sensing device 142. The user may register the sensing device 142 with the user's client device 128 by, for example, completing a bluetooth pairing procedure to pair the sensing device 142 with the client device 128. The user may input a voice query to the client device 128 via the sensing device 142. In response to the input audio signal, the user may receive audio output in the form of computer-generated speech from the client device 128 or the sensing device 142. The audio output may be generated by the data processing system 102. The client device 128 may also receive an action data structure to perform a predetermined function or action in response to the input audio signal. The interface 134 may receive or provide data messages to the direct action API112 of the data processing system 102 and enable communication between components of the system 100. The client device 128 may also include a user interface that enables a user to interact with the components of the system 100.

The sensing device 142 may transmit one or more instances of the input audio signal to the client device 128. The client device 128 may forward the input audio signal to the data processing system 102, which may satisfy the request that the NLP component 106 of the data processing system parse from the input audio signal. To conserve bandwidth and power consumption of the sensing device 142 and the client device 128, the sensing device 142 may transmit the input audio signal to the client device 128 only when the sensing device 142 determines that the input audio signal includes an activation phrase. The activation phrase may be a phrase, term, wake-up word, or hotword that the user includes at the beginning of the request to indicate to the client device 128 or sensing device 142 that the input audio signal is directed to the client device 128 or sensing device 142. For example, in an input audio signal "Ok device, change the song," the phrase "Ok device" may be an activation phrase.

The sensing device 142 may include a first sensor 130(1) and a second sensor 130 (2). When the user generates an input audio signal comprising a voice query or request, the first sensor 130(1) may detect and record the voice query or request as a first instance of the input audio signal, and the second sensor 130(2) may detect and record the voice query or request as a second instance of the input audio signal.

The sensing device 142 may include an instance of the NLP component 106, which may interpret the input audio signals received by the first sensor 130(1) and the second sensor 130 (2). The NLP component 106 of the sensing device 142 can be an optimized instance of the NLP component 106 that is configured to run on a device with a lesser amount of computing resources, such as the sensing device 142. For example, rather than being able to resolve any words in the input audio signal, an instance of the NLP component 106 of the sensing device can be configured to detect a limited number of words, such as an activation phrase. If the NLP component 106 of the sensing device determines that the instance of the input audio signal detected by one of the sensors 130 includes an activation phrase, the sensing device 142 can send at least one of the instances of the input audio signal to the client device 128. The client device 128 may have greater computing power than the sensing device 142. The NLP component 106 of the client device can parse the input audio signal to confirm whether the input audio signal includes the activation phrase. If the input audio signal includes an activation phrase, the client device 128 may send the input audio signal to the data processing system 102 for parsing and implementation by the data processing system 102.

The client device 128 may establish a low bandwidth connection with the sensing device 142 to receive instances of the input audio signal detected by the sensing device 142. The low bandwidth connection may be a bluetooth connection. Sending the first and second instances of the input audio signal detected by sensors 130(1) and 130(2) may take several seconds (due to the low bandwidth connection), which may result in a delay in confirming whether the input audio signal includes an activation phrase. To more quickly confirm whether the input audio signal includes an activation phrase by the client device 128, and to allow for low bandwidth connections, the sensing device 142 may send instances of the input audio signal to the client device 128 serially or at different compression levels.

For example, and in response to determining, by the sensing device 142, that the input audio signal includes an activation phrase, the sensing device 142 may send one of the detected instances of the input audio signal to the client device 128. After completing the sending of the first instance of the input audio signal to the client device 128, the sensing device 142 may send a second instance of the input audio signal to the client device 128. For example, the sensing device 142 may serially transmit instances of the input audio signal to the client device 128. In response to the client device 128 receiving the first instance of the input audio signal, the NLP component 106 of the client device can begin processing the first instance of the input audio signal to confirm whether the input audio signal includes the activation phrase. The client device 128 may receive the second instance of the input audio signal when the NLP component 106 of the client device processes the first instance of the input audio signal. If the NLP component 106 confirms the presence of the activation phrase in the first instance of the input audio signal, the client device 128 may not process the second instance of the input audio signal. If the NLP component 106 fails to confirm the presence of the activation phrase in the first instance of the input audio signal, the client device 128 may process the second instance of the input audio signal through its NLP component 106 to determine whether the second instance of the input audio signal comprises the input audio signal.

In response to a determination by the sensing device 142 that the input audio signal includes the activation phrase, the sensing device 142 may send two instances of the input audio signal to the client device 128 at a low quality level or a relatively high compression state. Once the low-quality versions of the first and second instances of the input audio signal are transmitted to the client device 128, the sensing device 142 may transmit higher-quality (e.g., less compressed) versions of the first and/or second instances of the input audio signal to the client device 128. If the client device 128 determines that one or more instances of the input audio signal include an activation phrase, the client device 128 may send the input audio signal or a request therein to the data processing system 102 for implementation by the data processing system 102.

The data processing system 102 of the system may include at least one server having at least one processor. For example, the data processing system 102 may include a plurality of servers located in at least one data center or server farm. The data processing system 102 may determine the request and a trigger key associated with the request from the audio input signal. Based on the request and the trigger key, the data processing system 102 may generate or select response data. The response data may be audio-based or text-based. For example, the response date may include one or more audio files that, when rendered, provide an audio output or sound wave. The data within the response data may also be referred to as content items. In addition to audio content, the response data may include other content (e.g., text, video, or image content).

The data processing system 102 may include a number of logically grouped servers and facilitate distributed computing techniques. A logical group of servers may be referred to as a data center, a server farm, or a machine farm. The servers may be geographically dispersed. The data center or machine farm may be managed as a single entity, or the machine farm may include multiple machine farms. The servers within each machine farm may be heterogeneous: one or more servers or machines may operate in accordance with one or more types of operating system platforms. The data processing system 102 may include servers stored in a data center in one or more high-density rack systems and associated storage systems located, for example, in an enterprise data center. In this manner, the data processing system 102 with integrated servers may improve system manageability, data security, physical security of the system, and system performance by locating servers and high performance storage systems on a localized high performance network. Centralizing all or some of the data processing system 102 components, including servers and storage systems and coupling them with advanced system management tools allows for more efficient use of server resources, which saves power and processing requirements and reduces bandwidth usage. Each component of the data processing system 102 may each include at least one processing unit, server, virtual server, circuit, engine, agent, device, or other logic device, such as a programmable logic array configured to communicate with the data repository 118 and with other computing devices.

The data processing system 102 may include a data repository 118. The data repository 118 may include one or more local or distributed databases and may include a database management system. The data repository 118 may include a computer data store or memory and may store one or more parameters 120, one or more policies 122, response data 124, and templates 126, among other data. The parameters 120, policies 122, and templates 126 may include information such as rules regarding voice-based sessions between the client device 128, the data processing system 102, and the sensing device 142. The response data 124 may include content items or associated metadata for audio output and input audio messages, which may be part of one or more communication sessions with the client device 128.

Applications, scripts, programs, or other components associated with the data processing system 102 may be installed at the client device 128 or the sensing device 142. The application may enable the client device 128 or sensing device 142 to communicate input audio signals (and other data) to the interface 104 of the data processing system 102. The application may enable the client device 128 and the sensing device 142 to drive components of the client device 128 and the sensing device 142 to render the output audio signal.

The NLP component 106 of the data processing system can receive an input audio signal. The data processing system 102 may receive an input audio signal from the client device 128 or the sensing device 142. The data processing system 102 may receive an input audio signal from the sensing device 142 via the client device 128. The first device may execute the NLP component 106, and the NLP component 106 may receive the input audio signal from the second device. For example, the sensing device 142 can receive an input audio signal and send it to the NLP component 106 of the client device 128.

The NLP component 106 may convert the input audio signal into recognized text by comparing the input audio signal to a stored set of representative audio waveforms and selecting the closest match. A representative waveform may be generated over a large number of input audio signals. Once the input audio signal is converted to recognized text, the NLP component 106 can match the text to words associated with the action or output audio signal (e.g., via a learning stage).

From the input audio signal, the NLP component 106 can identify at least one request or at least one trigger keyword corresponding to the request. The request may indicate an intent or subject of the input audio signal. The trigger key may indicate the type of action that may be taken. For example, the NLP component 106 may parse the input audio signal to identify at least one request to open a vehicle window or to skip to a next audio file in a music playlist. The trigger key may comprise at least one word, phrase, root or partial word or derivative indicating an action to be taken. The input audio signal may include an activation phrase or term, such as "go" or "ok".

The response selector component 114 can obtain information from the data repository 118, where it can be stored as part of the response data 124. The response selector component 114 can query the data repository 118 to select or otherwise identify, for example, a response phrase or content item from the response data 124.

The audio signal generator component 130 may generate or otherwise obtain an output signal that includes the content item. The data processing system 102 can execute the audio signal generator component 110 to generate or create an output signal corresponding to a content item or request. For example, upon satisfaction of a request, The signal generator component 110 can generate an audio output signal comprising The phrase "The action wascompleted".

The interface 104 may be a data interface, a hardware interface, a software interface, or a network interface that enables the components of the system 100 to communicate with one another. The interface 104 of the data processing system 102 may provide or transmit one or more data packets including motion data structures, audio signals, or other data to the client device 128 or the sensing device 142 via the network 116. For example, the data processing system 102 may provide output signals from the data repository 118 or from the audio signal generator 110 to the client device 128. The data processing system 102 may also instruct the client device 128 or the sensing device 142 to perform the function indicated in the action data structure by the data packet transmission. The output signals may be obtained, generated, transformed into or transmitted as one or more data packets (or other communication protocols) from the data processing system 102 (or other computing device) to the client device 128 or the sensing device 142.

The direct action API112 of the data processing system 102 may generate an action data structure based on, for example, a request. The action data structure may include data or instructions for performing the specified action to satisfy the request. The action data structure may be a JSON-formatted data structure or an XML-formatted data structure.

The action data structure may include information for completing the request. For example, the action data structure may be an XML (extensible markup language) or JSON (JavaScript object notation) formatted data structure that includes attributes for completing or otherwise fulfilling a request. The attributes may include a location of the vehicle 140, a location of the client device 128, an authorization level of a user associated with the client device 128, a vehicle identifier, an interface identifier, a vehicle status, or a request status. The request state may include one or more attributes that should be satisfied before the action is completed. For example, request "Ok, change the song," the request state may have { requestor: authorized, passenger). In this example, the requestor (e.g., the user generating the input audio signal) should be explicitly authorized to change the song or should be a passenger in the vehicle.

The direct action API112 may retrieve the template 126 from the repository 118 to determine which fields or attributes are to be included in the action data structure. The direct action API112 may determine the necessary parameters and may package the information into an action data structure. The direct action API112 may retrieve content from the repository 118 to obtain information about the attributes of the data structure.

The direct action API112 may populate the fields with data from the input audio signal. The direct action API112 may also populate the fields with data from the data provider 138, the client device 128, or the sensing device 142. The direct action API112 may prompt the user to provide additional information when filling in a field. The template 126 may be standardized for different types of actions, such as playing media files through a head unit of the vehicle, responding to messages, and performing functions within the automobile. The action data structure may be initially generated by the direct action API112 executed by the remote data processing system 102. The remote data processing system 102 may send the action data structure to the client device 128, and the client device 128 may add fields and attributes to the action data structure.

The direct action API112 may obtain response data 124 (or parameters 120 or policies 122) from the data repository 118 and data that the end user agrees to receive from the client device 128 to determine location, time, user accounts, logistics or other information to order the car from the car sharing service. The response data 124 (or parameters 120 or policies 122) may be included in an action data structure. When the content included in the action data structure includes end-user data for authentication, the data may be passed through a hash function prior to storage in the data repository 118. Using the direct action API112, the data processing system 102 may receive a subscription to communicate with the service provider computing device 160 to complete the conversion by, in this example, engaging in automobile sharing.

Fig. 2 shows a top view of the vehicle 140 and shows the interior compartment of the vehicle 140. The interior compartment of the vehicle 140 may include a plurality of seats 200. A user 202 may be seated in at least one chair 200. The user 202 may be associated with (e.g., own) the client device 128. The client device 128 may include a sensor 130. The interior of the vehicle 140 may include a sensing device 142. The sensing device 142 may include a first sensor 130(1) and a second sensor 130 (2). Each sensor 130 may be a microphone. The sensing device 142 may be a power adapter or charging device that converts power from the vehicle 140 to a power level that may be consumed by the client device 128. The client device 128 may establish a communication link with the sensing device 142 through a wireless connection, such as bluetooth.

The user 202 may generate the request in the form of an input audio signal 204. The input audio signal 204 may be recorded or detected by the sensors 130 of the sensing device 142 and the sensors 130 of the client device 128. The first sensor 130(1) of the sensing device may record the input audio signal 204 as a first instance of the input audio signal 204, the second sensor 130(2) may record the input audio signal 204 as a second instance of the input audio signal, and the sensor 130 of the client device may record the input audio signal 204 as a third instance of the input audio signal 204. The client device 128 may be at a location where the quality of the input audio signal 204 received by the client device 128 is low compared to the sensing device 142. For example, the client device 128 may be located in the user's pocket or in a center console of the vehicle 140. Given the relative position of each sensor 130 within the vehicle 140, each sensor 130 may record or detect a slightly different version of the input audio signal 204.

Referring also to fig. 1, wherein the NLP component 106 component of the sensing device 142 can parse instances of the input audio signal 204 detected by the first and second sensors 130. If the NLP component 106 detects or identifies the presence of an activation phrase (e.g., "Ok, device") or a candidate activation phrase, the sensing device 142 may send an instance of the input audio signal 204 to the client device 128. The client device 128, which may have greater computational resources than the sensing device 142, may process an instance of the input audio signal 204 with the NLP component 106 of the client device to confirm the presence of the activation phrase in the input audio signal 204. If the client device 128 confirms the presence of the activation phrase, the client device 128 may send the input audio signal to the data processing system 102 for processing.

The sensing device 142 can send the first instance to the client device 128 at a first point in time, for example, at a point in time when the sensing device 142 detects the presence of an activation phrase in the input audio signal 204. If the NLP component 106 of the client device fails to determine that the activation phrase is in the input audio signal 204, the client device 128 can send a message to the sensing device 142 requesting the sensing device 142 to send a second instance of the input audio signal 204 recorded by the sensing device 142. The sensing device 142 may transmit a second instance of the input audio signal 204 to the client device 128 at a second point in time. The second point in time may be a point in time after the sensing device 142 completes transmission of the first instance of the input audio signal 204 to the client device 128.

Upon receipt, the NLP component 106 of the client device can process the second instance of the input audio signal 204 for the presence of the activation phrase. The client device 128 may determine whether the input audio signal 204 includes the activation phrase based on the second instance of the input audio signal 204 or a combination of the first and second instances of the input audio signal 204.

Prior to receiving the request for the second instance of the input audio signal 204 from the client device 128, the sensing device 142 may send the second instance of the input audio signal 204 to the client device 128. For example, upon completion of sending the first instance of the input audio signal 204 to the client device 128, the sensing device 142 may automatically begin sending the second instance of the input audio signal 204 to the client device 128 at the second point in time. When the sensing device 142 automatically begins sending the second instance of the input audio signal 204 to the client device 128, if the client device 128 confirms that the activation phrase is in the first instance of the input audio signal 204, the client device 128 may send a termination message to the sensing device 142 because the client device 128 does not need the second instance of the activation phrase to confirm the presence of the activation phrase in the input audio signal 204.

The client device 128 may also use the instances of the input audio signal 204 detected by the sensor 130 of the client device to confirm whether the activation phrase is present in the input audio signal 204. The client device 128 may use the instance of the input audio signal 204 recorded by the client device 128 in conjunction with either the first instance of the input audio signal 204 or the second instance of the input audio signal 204.

FIG. 3 illustrates a block diagram of an example method 300 for detecting an activation phrase in a networked system with limited bandwidth. The steps shown on the left side of the block diagram may be performed by the sensing device 142 and the steps shown on the right side of the block diagram may be performed by the client device 128.

The method 300 may include receiving a first instance of an input audio signal (ACT 302) and receiving a second instance of the input audio signal (ACT 304). The method 300 may include identifying an activation phrase (ACT 306). The method 300 may include sending a first instance of an input audio signal (ACT 308). The method 300 may include receiving, by a client device, a first instance of an input audio signal (ACT 310). The method 300 may include parsing a first instance of an input audio signal (ACT 312). Method 300 may include determining whether the first instance of the input audio signal includes an activation phrase (act 314). The method 300 may include sending a second instance of the input audio signal by the sensing device (ACT 316) and receiving the second instance of the input audio signal by the client device (ACT 318). The method 300 may include parsing a second instance of the input audio signal (ACT 320). The method 300 may include sending a first instance of a second input audio signal by the sensing device (ACT 322). If the client device recognizes an active phrase in the input audio signal at ACT 314, the method 300 may include terminating the transmission (ACT 324). The method 300 may include sending, by the client device, a first instance of a second input audio signal (ACT 326).

The method 300 may include receiving a first instance of an input audio signal (ACT 302) and receiving a second instance of the input audio signal (ACT 304). Referring also to fig. 1 and 2, the first and second instances of the input audio signal may be received by, for example, first and second microphones (e.g., sensor 130) of sensing device 142. The input audio signal may be an utterance made by a user. The first and second microphones may detect the input audio signal when the volume of the input audio signal exceeds a predetermined threshold.

The method 300 may include identifying an activation phrase (ACT 306). The sensing device 142 may include the NLP component 106. The sensing device 142 may have limited computing capabilities and the NLP component 106 may be configured to only detect the presence of an activation phrase within the input audio signal. The threshold for accurately identifying an activation phrase in an incoming input audio signal may be low. For example, the NLP component 106 of the sensing device may have a high false alarm rate. The activation phrases detected by the sensing device 142 may be referred to as candidate activation phrases.

The method 300 may include sending a first instance of an input audio signal (ACT 308). The sensing device 142 may send a first instance of the input audio signal to the client device 128. The client device 128 may establish a bluetooth connection between the sensing device 142 and the client device 128 over which the first instance of the input audio signal (and other data) is transmitted. The sensing device 142 can send the first instance of the input audio signal to the client device 128 at a first point in time, for example, when the NLP component 106 of the sensing device detects an activation phrase in the first instance of the input audio signal.

The sensing device 142 can send the first instance of the input audio signal to the client device 128 for the client device 128 to confirm whether the activation phrase is in the input audio signal. The sensing device 142 may generate multiple copies of the first instance of the input audio signal. Different copies may be generated by applying different compression levels to the first instance of the input audio signal. For example, the sensing device 142 may generate: a first copy of a first instance of an input audio signal that is highly compressed and has a relatively small file size but low sound quality; and a second copy of the first instance of the input audio signal, which is less compressed, has a relatively large file size but a high sound quality. The sensing device 142 may send a highly compressed (smaller file size) copy of the first instance of the input audio signal to the client device 128 at a first point in time. The sensing device 142 may transmit a second copy of the first instance of the input audio signal to the client device 128 at a second point in time, e.g., in response to a request from the client device 128.

The method 300 may include receiving a first instance of an input audio signal (ACT 310). The client device 128 may receive the first instance of the input audio signal via a bluetooth connection established between the sensing device 142 and the client device 128. The method 300 may include parsing a first instance of an input audio signal (ACT 312). The NLP component 106 of the client device 128 can parse the input audio signal and identify one or more candidate activation phrases. The candidate activation phrase can be a phrase or term that the NLP component 106 determines has a likelihood above a predetermined threshold of being an activation phrase. The candidate activation phrase may be an activation phrase recognized by the sensing device 142. For example, the sensing device 142 may send only a portion of the first instance of the input audio signal that is a cut-out portion of the input audio signal that includes a predetermined recorded amount of candidate activation phrases before and after the candidate activation phrase.

Method 300 may include identifying an activation phrase (act 314). The NLP component 106 of the client device can parse the received instance of the input audio signal to determine or confirm whether the activation phrase is within the input audio signal. For example, the NLP component 106 can determine whether the candidate activation phrase is an activation phrase required to initiate a speech-based query or request. If the activation phrase is not within the input audio signal, the method 300 may continue to ACT 316, and if the NLP component 106 determines that the activation phrase is within the input audio signal, the method 300 may continue to ACT 324.

The method 300 may include identifying an activation phrase within a third instance of the input audio signal. A third instance of the input audio signal may be recorded by a microphone or other sensor 130 of the client device. The ACT 314 of determining whether the input audio signal includes an activation phrase may include determining whether any of the first or third instances of the input audio signal includes an activation phrase. The NLP component 106 of the client device may analyze the first and third instances of the input audio signal, respectively. For example, the NLP component 106 can independently determine whether an instance of the input audio signal includes an activation phrase. If the NLP component 106 determines that any of the instances of the input audio signal are determined to include an activation phrase, the NLP component 106 can determine that the input audio signal includes an activation phrase. The NLP component 106 may analyze the first and third instances of the input audio signal together. For example, the NLP component 106 generates a confidence score for whether each instance of the input audio signal includes an activation phrase, and determines whether the input audio signal includes an activation phrase based on a combination or average of the confidence scores associated with each instance of the input audio signal.

The method 300 may include sending a second instance of the input audio signal by the sensing device (ACT 316). The sensing device 142 may transmit the second instance of the input audio signal to the client device 128 over a bluetooth connection established between the sensing device 142 and the client device 128. The second instance of the input audio signal may be an instance of the input audio signal detected or recorded at the sensing device 142 by a microphone or other sensor 130 that is different from the other microphone or sensor 130 that detected or recorded the first instance of the input audio signal. The sensing device 142 may transmit a second instance of the input audio signal to the sensing device 142 at a second point in time. The second point in time may be subsequent to the first point in time. The second point in time may be when the sending of the first instance of the input audio signal to the client device 128 is complete (e.g., after the ACT 308 is complete). For example, upon completion of sending the first instance of the input audio signal to the client device 128, the sensing device 142 may automatically send the second instance of the input audio signal to the client device 128.

In response to the request from the client device 128, the sensing device 142 may send the second instance of the input audio signal to the client device 128. For example, in the ACT 314, when the client device 128 determines that the activation phrase is not within (or cannot be detected within) the first instance of the input audio signal, the client device 128 may generate and send a request to the sensing device 142 for a second instance of the input audio signal. The client device 128 may send a request for a second instance of the input audio signal over a bluetooth network established between the sensing device 142 and the client device 128.

When the sensing device 142 includes only a single microphone, the second instance of the input audio signal may be a copy of the first instance of the input audio signal, but with a higher audio quality. For example, the copy sent at the ACT 318 may be compressed less than the first instance of the input audio signal sent at the ACT 308.

The method 300 may include parsing a second instance of the input audio signal (ACT 320). The NLP component 106 of the client device can parse the second instance of the input audio signal to identify a candidate activation phrase in the second instance of the input audio signal. If the NLP component 106 of the client device determines that one of the candidate activation phrases is a predetermined activation phrase, the client device 128 can send a second input audio signal to the data processing system 102. The second input audio signal may be detected or recorded by the sensing device 142 or any of the sensors 130 of the client device 128. The second input audio signal may be a continuation of the first input audio signal. For example, the second input audio signal may be or may include a user utterance immediately following the activation phrase.

The method 300 may include sending a first instance of a second input audio signal by the sensing device (ACT 322). The first instance of the second input audio signal may be an input audio signal that the client device 128 receives and transmits to the data processing system 102 based on detection of an activation phrase in the first or second instance of the first input audio signal.

The second input audio signal may be an input audio signal detected or recorded by one of the sensors 130 after the first input audio signal is detected. A portion of the second input audio signal may overlap a portion of the first input audio signal. The first and second input audio signals may be generated from a single input audio signal. The first input audio signal may comprise an activation phrase and the second input audio signal may comprise a request. For example, the input audio signal may be "Ok, skip to the next song". In this example, the first input audio signal may include at least the activation phrase "Ok", and the second input audio signal may include at least the request "skip to the next song". To conserve power and bandwidth, the sensing device 142 may send only a portion of the input audio signal that includes the activation phrase (e.g., the first input audio signal) to the client device 128. The sensing device 142 may not send the portion including the requested input audio signal (e.g., the second input audio signal) until the sensing device 142 receives a confirmation message from the client device that the first input audio signal includes the activation phrase. In some implementations, the sensing device 142 can first transmit one or more instances of the first input audio signal, which may include an activation phrase, and then automatically begin transmitting the second input audio signal to the client device 128 once transmission of the first input audio signal to the client device 128 is complete.

Returning to the recognition step at the ACT 314, if the client device 128 recognizes an activation phrase in the first instance of the input audio signal, the client device 128 may terminate the transmission between the sensing device 142 and the client device 128. The termination of the transmission between the sensing device 142 and the client device 128 may be optional. As described above, upon completion of sending the first instance of the input audio signal to the client device 128, the sensing device 142 may automatically send the second instance of the input audio signal to the client device 128. The client device 128 may generate and send a termination message to the sensing device 142 that may cause the sensing device 142 to stop sending the second instance of the input audio signal to the client device 128. Terminating the transmission of the second instance of the input audio signal to the client device 128 may include the client device 128 ceasing to receive the second instance of the input audio signal. For example, the sensing device 142 may continue to transmit the second instance of the second input audio signal, but the client device 128 may reject receiving the second instance of the input audio signal or perform no further processing on the second instance of the input once the second instance of the input audio signal is received. Not performing further processing may include not processing the second instance of the input audio signal with the NLP component 106 of the client device.

The method 300 may include sending a first instance of a second input audio signal (ACT 326). The method 300 may include transmitting, by the client device 128, the first instance of the second input audio signal to the data processing system 102. The client device 128 may receive a first instance of a second input audio signal at the ACT 322. The client device 128 may send the first instance of the second input audio signal to the data processing system 102 in response to detecting or confirming the presence of the activation phrase in at least one of the first or second instances of the input audio signal. The second input audio signal may include a request that can be parsed by the NLP component 106 of the data processing system and fulfilled (at least in part) by the data processing system 102.

Fig. 4 is a block diagram of an example computer system 400. Computer system or computing device 400 may include or be used to implement system 100 or components thereof, such as data processing system 102. Computing system 400 includes a bus 405 or other communication component for communicating information, and a processor 410 or processing circuit coupled to bus 405 for processing information. Computing system 400 may also include one or more processors 410 or processing circuits coupled to the bus for processing information. Computing system 400 also includes a main memory 415, such as a Random Access Memory (RAM) or other dynamic storage device, coupled to bus 405 for storing information and instructions to be executed by processor 410. The main memory 415 may be or include the data store 118. Main memory 415 also may be used for storing location information, temporary variables, or other intermediate information during execution of instructions by processor 410. Computing system 400 may also include a Read Only Memory (ROM) or other static storage device coupled to bus 405 for storing static information and instructions for processor 410. A storage device 425 such as a solid state device, magnetic disk or optical disk may be coupled to bus 405 for persistently storing information and instructions. The storage device 425 may include the data repository 118 or be part of the data repository 118.

Computing system 400 may be coupled via bus 405 to a display 435, such as a liquid crystal display or active matrix display, for displaying information to a user. An input device 430, such as a keyboard including alphanumeric and other keys, may be coupled to bus 405 for communicating information and command selections to processor 410. Input device 430 may include a touch screen display 435. Input device 430 may also include a cursor control, such as a mouse, a trackball, or cursor direction keys for communicating direction information and command selections to processor 410 and for controlling cursor movement on display 435. The display 435 may be part of the data processing system 102, the client computing device 128, or other components of fig. 1, for example.

The processes, systems, and methods described herein may be implemented by computing system 400 in response to processor 410 executing an arrangement of instructions contained in main memory 415. Such instructions may be read into main memory 415 from another computer-readable medium, such as storage device 425. Execution of the arrangement of instructions contained in main memory 415 causes the computing system 400 to perform the illustrative processes described herein. One or more processors in a multi-processing arrangement may also be employed to execute the instructions contained in main memory 415. Hardwired circuitry may be used in place of or in combination with software instructions and the systems and methods described herein. The systems and methods described herein are not limited to any specific combination of hardware circuitry and software.

Although an example computing system has been described in fig. 4, the subject matter including the operations described in this specification can be implemented in other types of digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them.

For the cases where the system discussed herein collects or may make use of personal information about a user, the user may be provided with an opportunity to control whether programs or features may collect personal information (e.g., information about the user's social network, social activities, the user's preferences, or the user's location), or whether or how to receive content from a content server or other data processing system that may be more relevant to the user. In addition, certain data may be anonymized in one or more ways prior to storage or use, in order to remove personally identifiable information when generating the parameters. For example, the identity of the user may be anonymous so that no personal identity information can be determined for the user, or the geographic location of the user may be generalized when location information is obtained (e.g., to a city, zip code, or state level), so that no particular location of the user can be determined. Thus, the user may control how information is collected about him or her and used by the content server.

The subject matter and the operations described in this specification can be implemented in digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. The subject matter described in this specification can be implemented as one or more computer programs (e.g., one or more circuits of computer program instructions) encoded on one or more computer storage media for execution by, or to control the operation of, data processing apparatus. Alternatively or additionally, the program instructions may be encoded on an artificially generated propagated signal (e.g., a machine-generated electrical, optical, or electromagnetic signal that is generated to encode information for transmission to suitable receiver apparatus for execution by the data processing apparatus). The computer storage medium may be or be included in a computer-readable storage device, a computer-readable storage substrate, a random or serial access memory array or device, or a combination of one or more of them. Although a computer storage medium is not a propagated signal, a computer storage medium can be a source or destination of computer program instructions encoded in an artificially generated propagated signal. The computer storage medium may also be or be included in one or more separate components or media (e.g., multiple CDs, disks, or other storage devices). The operations described in this specification may be implemented as operations performed by data processing apparatus on data stored on one or more computer-readable storage devices or received from other sources.

The terms "data processing system," "computing device," "component," or "data processing apparatus" include various devices, apparatuses, and machines for processing data, including by way of example programmable processors, computers, systems on a chip, or multiple systems or combinations of the foregoing. The apparatus can comprise special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit). The apparatus can include, in addition to hardware, code that creates an execution environment for the computer program in question, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or a combination of one or more of them. The apparatus and execution environment may implement a variety of different computing model infrastructures, such as web services, distributed computing, and grid computing infrastructures. The components of system 100 may include or share one or more data processing apparatuses, systems, computing devices, or processors.

A computer program (also known as a program, software application, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, object, or other unit suitable for use in a computing environment. The computer program may correspond to a file in a file system. A computer program can be stored in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub programs, or portions of code). A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network.

The processes and logic flows described in this specification can be performed by one or more programmable processors (e.g., components of data processing system 102) executing one or more computer programs to perform actions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit). Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media and memory devices, including by way of example: semiconductor memory devices (e.g., EPROM, EEPROM, and flash memory devices); magnetic disks (e.g., internal hard disks or removable disks); magneto-optical disks; and CD ROM and DVD-ROM disks. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

The subject matter described herein can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a client computer having a graphical user interface), or a web browser through which a user can interact with an implementation of the subject matter described in this specification, or a combination of one or more such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks ("LANs") and wide area networks ("WANs"), internetworks (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks).

A computing system, such as system 100 or system 400, may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network, such as network 116. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. In some implementations, the server sends data (e.g., data packets representing content items) to the client device (e.g., for purposes of displaying data to and receiving user input from a user interacting with the client device). Data (e.g., results of user interactions) generated at the client device may be received at the server from the client device (e.g., received by the data processing system 102 from the client computing device 128 or the sensing device 142).

Although operations are depicted in the drawings in a particular order, these operations need not be performed in the particular order shown or in sequential order, and all illustrated operations need not be performed. The actions described herein may be performed in a different order.

In all embodiments, the separation of various system components need not be separate, and the described program components may be included in a single hardware or software product. For example, the NLP component 110 may be a single component, an app or program, or a logical device with one or more processing circuits, or part of one or more servers of the data processing system 102.

Having now described some illustrative embodiments, it will be apparent that the foregoing is illustrative and not limiting, having been presented by way of example. In particular, although many of the examples presented herein involve specific combinations of method acts or system elements, those acts and those elements may be combined in other ways to accomplish the same objectives. Acts, elements and features discussed in connection with one implementation are not intended to be excluded from a similar role in other implementations.

The phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting. The use of "including," "comprising," "having," "containing," "involving," "characterized by," and variations thereof herein, is meant to encompass the items listed thereafter, equivalents thereof, and other items, as well as alternative embodiments that consist exclusively of the items listed thereafter. In one embodiment, the systems and methods described herein consist of one, each combination of more than one, or all of the described elements, acts, or components.

Any reference to an embodiment or element or act of the systems and methods referred to herein in the singular may also encompass embodiments comprising a plurality of such elements, and any plural reference to any implementation or element or act herein may also encompass embodiments comprising only one element. References in the singular or plural form are not intended to limit the presently disclosed systems or methods, their components, acts or elements to a single or multiple configurations. A reference to any action or element based on any information, action, or element may include an implementation in which the action or element is based, at least in part, on any information, action, or element.

Any embodiment disclosed herein may be combined with any other embodiment or examples, and references to "an embodiment," "some embodiments," or "one embodiment," etc., are not necessarily mutually exclusive and are intended to indicate that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment or example. The terms used herein do not necessarily all refer to the same embodiment. Any embodiment may be included in or exclusively combined with any other embodiment in any manner consistent with aspects and embodiments disclosed herein.

References to "or" may be construed as inclusive such that any term described using "or" may refer to any term in the singular, more than one, and all of the described terms. For example, a reference to "at least one of a 'and' B" may include only "a", only "B", and both "a" and "B". These references, used in conjunction with "including" or other open-ended terms, may include additional items.

Where technical features in the figures, detailed description or any claims are followed by reference signs, the reference signs are included to increase the intelligibility of the figures, detailed description and claims. Accordingly, neither the reference signs nor their absence should have any limiting effect on the scope of any claim elements.

The systems and methods described herein may be embodied in other specific forms without departing from the characteristics thereof. The foregoing embodiments are illustrative and not limiting of the described systems and methods. The scope of the systems and methods described herein is, therefore, indicated by the appended claims rather than by the foregoing description, and all changes that come within the meaning and range of equivalency of the claims are intended to be embraced therein.