Microcontroller interface for audio signal processing
Disclosed is a neuromorphic-processing systems including, in some embodiments, a special-purpose host processor operable as a stand-alone host processor; a neuromorphic co-processor including an artificial neural network; and a communications interface between the host processor and the co-processor configured to transmit information therebetween. The co-processor is configured to enhance special-purpose processing of the host processor with the artificial neural network. Also disclosed is a method of a neuromorphic-processing system having the special-purpose host processor and the neuromorphic co-processor including, in some embodiments, enhancing the special-purpose processing of the host processor with the artificial neural network of the co-processor. In some embodiments, the host processor is a hearing-aid processor.
1. A neuromorphic-processing system comprising, comprising:
a special-purpose host processor operable as a stand-alone host processor;
a neuromorphic co-processor including an artificial neural network; and
a communications interface between the host processor and the co-processor configured to transmit information therebetween,
wherein the co-processor is configured to enhance special-purpose processing of the host processor with the artificial neural network;
wherein the host processor is a hearing-aid processor configured to transmit frequency elements or signal spectrum information to the co-processor in the form of Fourier transforms over a serial communications interface as the communications interface, and
wherein the co-processor further includes a demultiplexer configured to demultiplex serial signals from the serial communications interface into parallel signals for a plurality of inputs of the artificial neural network.
2. The neuromorphic-processing system of claim 1 , wherein the communications interface between the host processor and the co-processor is a serial peripheral interface “SPI”) bus or inter-integrated circuit (“I2C”) bus.
3. The neuromorphic-processing system of claim 1 , wherein the co-processor is configured to enhance the special-purpose processing of the hearing-aid processor by providing information to the hearing-aid processor over the communications interface, thereby enabling the hearing-aid processor to selectively suppress noise and enhance desired signals.
4. The neuromorphic-processing system of claim 1 , wherein the co-processor is configured to enhance the special-purpose processing of the hearing-aid processor by providing a frequency mask to the hearing-aid processor over the communications interface, thereby indicating noise frequencies to suppress and signal frequencies to boost.
5. The neuromorphic-processing system of claim 4 , wherein the frequency mask is a set of attenuation or amplification factors corresponding to a set of frequencies to be suppressed or boosted for each of a number of time steps in an audio sample.
6. The neuromorphic-processing system of claim 1 , wherein the artificial neural network is disposed in an analog multiplier array of a plurality of two-quadrant multipliers in a memory sector of the neuromorphic-processing system.
7. The neuromorphic-processing system of claim 1 ,
wherein synaptic weights of the artificial neural network are stored in firmware of the neuromorphic co-processor, and
wherein the firmware is configured for cloud-based updates to update the synaptic weights of the artificial neural network.
8. The neuromorphic-processing system of claim 1 ,
wherein the host processor and the co-processor are embodied in a single monolithic integrated circuit, a stacked die assembly, a multi-chip module, or separate integrated circuits of separate modules, and
wherein neuromorphic-processing system is configured to operate on battery power.
9. A neuromorphic processor, comprising:
one or more interface circuits comprising a demultiplexer configured to demultiplex serial signals into parallel signals;
wherein the serial signals are received from a serial communications interface between the neuromorphic processor and a special-purpose host processor; and
a multi-layered artificial neural network configured to receive the parallel signals from the interface circuits,
wherein the neuromorphic processor is configured to enhance special-purpose processing of the host processor with the artificial neural network;
wherein the host processor is a hearing-aid processor, and
wherein the neuromorphic processor is configured to receive frequency elements or signal spectrum information from the hearing-aid processor in the form of Fourier transforms over the serial communications interface.
10. The neuromorphic processor of claim 9 , wherein the neuromorphic processor is configured to enhance the special-purpose processing of the hearing-aid processor by providing information to the hearing-aid processor over the serial communications interface, thereby enabling the hearing-aid processor to selectively suppress noise and enhance desired signals.
11. The neuromorphic processor of claim 9 , wherein the neuromorphic processor is configured to enhance the special-purpose processing of the hearing-aid processor by providing a frequency mask to the hearing-aid processor over the serial communications interface, thereby indicating noise frequencies to suppress and signal frequencies to boost.
12. The neuromorphic processor of claim 11 , wherein the frequency mask is a set of attenuation or amplification factors corresponding to a set of frequencies to be suppressed or boosted for each of a number of time steps of an audio sample.
13. The neuromorphic processor of claim 9 , wherein the artificial neural network is disposed in an analog multiplier array of a plurality of two-quadrant multipliers in a memory sector of the neuromorphic processor.
14. A method of a neuromorphic-processing system having a special-purpose host processor and a neuromorphic co-processor, comprising:
enhancing special-purpose processing of the host processor with an artificial neural network of the co-processor,
wherein the host processor is operable as a stand-alone host processor; and
transmitting frequency elements or signal spectrum information from the host processor configured as a hearing-aid processor to the co-processor in the form of Fourier transforms over a serial communications interface; and
demultiplexing serial signals from the serial communications interface with a demultiplexer of the co processor into parallel signals for a plurality of inputs of the artificial neural network.
15. The method of claim 14 , wherein enhancing the special-purpose processing of the hearing-aid processor includes providing information to the hearing-aid processor over the serial communications interface, thereby enabling the hearing-aid processor to selectively suppress noise and enhance desired signals.
16. The method of claim 14 , wherein enhancing the special-purpose processing of the hearing-aid processor includes providing a frequency mask to the hearing-aid processor over the serial communications interface, thereby indicating noise frequencies to suppress and signal frequencies to boost.
17. The method of claim 14 , further comprising:
updating synaptic weights of the artificial neural network,
wherein the synaptic weights of the artificial neural network are stored in firmware of the neuromorphic co-processor configured for cloud-based updates.