LOW POWER NEUROMORPHIC VOICE ACTIVATION SYSTEM AND METHOD

• US 20170229117A1
• Filed: 02/06/2017
• Published: 08/10/2017
• Est. Priority Date: 02/05/2016
• Status: Active Grant

• Current Assignee: Brainchip, Inc. (BrainChip Holdings Ltd.)
• Original Assignee: Brainchip, Inc. (BrainChip Holdings Ltd.)
• Inventors: Elkhatib, Mouna, Oros, Nicolas Yvan, van der Made, Peter AJ
• Granted Patent: US 10,157,629 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 3/167 Audio in a user interface, e.g. using voice commands for navigating, audio feedback
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/049 Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
• CPC Class Codes: G06N 3/063 using electronic means
• CPC Class Codes: G06N 3/088 Non-supervised learning, e.g. competitive learning
• CPC Class Codes: G06N 3/10 Interfaces, programming languages or software development kits, e.g. for simulating neural networks
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 25/30 using neural networks
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10

Abstract

The present invention provides a system and method for controlling a device by recognizing voice commands through a spiking neural network. The system comprises a spiking neural adaptive processor receiving an input stream that is being forwarded from a microphone, a decimation filter and then an artificial cochlea. The spiking neural adaptive processor further comprises a first spiking neural network and a second spiking neural network. The first spiking neural network checks for voice activities in output spikes received from artificial cochlea. If any voice activity is detected, it activates the second spiking neural network and passes the output spike of the artificial cochlea to the second spiking neural network that is further configured to recognize spike patterns indicative of specific voice commands. If the first spiking neural network does not detect any voice activity, it halts the second spiking neural network.

Claims

1. A system for controlling an electronic device through voice control commands, the system comprising:

a hierarchical arrangement of a first spiking neural network and a second spiking neural network, said first spiking neural network receives output spikes from an artificial cochlea, said first spiking neural network configured to recognize and learn spike patterns, in the output spikes, that indicates presence of a voice activity in the output spikes, and said first spiking neural network activates said second spiking neural network from a low energy halted or off state, if any voice activity is detected, in the output spikes, by said first spiking neural network.

2. The system of claim 1, wherein the second spiking neural network performs a cognitive and learning function aiming to detect the spike patterns, in the output spikes from the artificial cochlea, indicative of specific voice commands.

3. The system of claim 1, wherein the electronic device is a portable wireless device such as a mobile phone, a laptop, a smartphone and the like.

4. The system of claim 1, wherein the device includes one or more digital microphones;

one or more decimation filters coupled to the digital microphones and to the artificial cochlea; and

the hierarchical arrangement of the first spiking neural network and the second spiking neural network.

5. The system of claim 1, wherein the first spiking neural network and the second spiking neural network comprise a plurality of artificial neurons interconnected with a plurality of synapses.

6. The system of claim 1, wherein the second spiking neural network configured to recognize spike patterns indicative of specific voice commands, comprises:

a means for receiving spikes from the first spiking neural network representative of the voice activity;

a means for associating received spikes, by activating neurons within the second spiking neural network, by triggering synaptic activity derived from training; and

a means for producing an output that is representative of at least one voice command.

7. The system of claim 1, the second spiking neural network is in halted or OFF state, when the spike pattern indicative of any voice activity is not detected by the first spiking neural network.

8. The system of claim 1, wherein the system is configured to reject background noise by selectively rejecting certain bandwidths of an input signal within the artificial cochlea.

9. The system of claim 1, wherein the system is a hard wired digital logic circuit running at a low clock speed.

10. The system of claim 1 further includes a “

labeling and decode”

logic to generate interrupt to a digital signal processor based on the activity of the first and second spiking neural networks.

11. A method for allowing a user to control an electronic device by means of voice commands, the method comprising:

receiving, an output of parallel spikes, by a first spiking neural network from an artificial cochlea;

detecting, a presence of a voice activity in the output spikes, by the first spiking neural network; and

activating, a second spiking neural network, by the first spiking neural network, when a voice activity is detected.

12. The method as claimed in claim 11 further comprises:

passing the output spikes to the second spiking neural network; and

recognizing spike patterns in the output spikes, by the second spiking neural network, that indicates at least one specific voice command in the output spikes.

13. The method as claimed in claim 11, wherein the artificial cochlea receives a pulse code modulation signal from one or more decimation filters and produces the output of parallel spikes to send to the first spiking neural network.

14. The method as claimed in claim 13, wherein said one or more decimation filters receive a pulse density modulation signal from one or more microphones to convert the pulse density modulation signal into the pulse code modulation signal;

and said one or more digital microphones receive and convert the voice signal, spoken by a user, into the pulse density modulation signal.

15. The method of claim 11, wherein the first spiking neural network halts the second spiking neural network when no voice activity is detected.

16. The method of claim 11, wherein the artificial cochlea is configured as a digital emulation of a biological cochlea.

17. The method of claim 11, wherein the first spiking neural network and the second spiking neural network include a plurality of artificial neurons interconnected by a plurality of digital synapses.

18. The method of claim 11, wherein the electronic device is a portable wireless device such as a mobile, a laptop, a smartphone and the like.