Reconfigurable event driven hardware using reservoir computing for monitoring an electronic sensor and waking a processor

Patent metadata

Field	Value
Publication number	US-9541982-B2
Application number	US-201313749854-A
Country	US
Kind code	B2
Filing date	Jan 25, 2013
Priority date	Jan 25, 2013
Publication date	Jan 10, 2017
Grant date	Jan 10, 2017

Abstract

The present inventors have recognized that proper utilization of reconfigurable event driven hardware may achieve optimum power conservation in energy constrained environments including a low power general purpose primary processor and one or more electronic sensors. Aspects of neurobiology and neuroscience, for example, may be utilized to provide such reconfigurable event driven hardware, thereby achieving energy-efficient continuous sensing and signature reporting in conjunction with the one or more electronic sensors while the primary processor enters a low power consumption mode. Such hardware is event driven and operates with extremely low energy requirements.

First claim

What is claimed is:

1. An energy constrained system comprising:
(a) a primary processor having a low power consumption mode, the primary processor being configured to respond to a plurality of trigger signatures, each trigger signature classifying one or more predetermined sensed conditions, by
(i) exiting the low power consumption mode; and
(ii) invoking a software routine to initiate an action specific to a trigger signature received by the primary processor;

(b) an electronic sensor adapted to continuously provide a sensor output that reflects a presently sensed condition; and
(c) reconfigurable event driven hardware coupled between the primary processor and the electronic sensor, the reconfigurable event driven hardware receiving the sensor output, the reconfigurable event driven hardware being implemented using reservoir computing including:
(i) a capturing element configured to capture a temporal behavior of the sensor output; and
(ii) a classifying element configured to classify an output of the capturing element including the temporal behavior to identify a trigger signature,

wherein the primary processor is operable to configure the reconfigurable event driven hardware to detect the plurality of trigger signatures,
wherein the reconfigurable event driven hardware is, adapted to continuously monitor the sensor output to detect a trigger signature from among the plurality of trigger signatures, and
wherein the reconfigurable event driven hardware is further adapted to communicate the trigger signature to the primary processor without communicating the one or more predetermined sensed conditions, in response to a detection of a trigger signature while the primary processor is in the low power consumption a mode.

2. The system of claim 1 , wherein the reservoir computing includes a Liquid State Machine comprised of a plurality of leaky-integrate-and-fire neurons.

3. The system of claim 2 , wherein the reservoir computing further includes a Multi-Layered Perceptron Network comprised of a plurality of leaky-integrate-and-fire neurons.

4. The system of claim 3 , wherein the Multi-Layered Perceptron Network detects the trigger signature.

5. The system of claim 3 , wherein the Liquid State Machine and the Multi-Layered Perceptron Network are digitally implemented and operate in a clocked manner.

6. The system of claim 1 , wherein the reservoir computing includes a Multi-Layered Perceptron Network comprised of a plurality of leaky-integrate-and-fire neurons.

7. The system of claim 1 , wherein the reservoir computing comprises of a plurality of leaky-integrate-and-fire neurons, and the reconfigurable event driven hardware includes a reconfigurable event driven interconnect in communication with the primary processor for configuring the plurality of leaky-integrate-and-fire neurons.

8. The system of claim 1 , wherein the primary processor deactivates the reconfigurable event driven hardware and continuously monitors the sensor output for the trigger signature when the primary processor is not in the low power consumption mode.

9. The system of claim 1 , wherein the electronic sensor is an accelerometer or a gyroscope.

10. The system of claim 1 , wherein the electronic sensor is an electrocardiograph or an electroencephalograph.

11. The system of claim 10 , wherein the reconfigurable event driven hardware is configured to continuously monitors slow wave activity of the brain.

12. The system of claim 1 , wherein the primary processor configures the reconfigurable event driven hardware via an application program interface.

13. The system of claim 1 , wherein the electronic sensor and the reconfigurable event driven hardware are contained in an enclosure separate from the primary processor.

14. The system of claim 1 , wherein the primary processor and the reconfigurable event driven hardware are contained in an enclosure separate from the electronic sensor.

15. An apparatus for use in an energy constrained environment comprising:
a reconfigurable event driven hardware implemented using reservoir computing for coupling between a primary processor having a low power consumption mode and an electronic sensor with a sensor output that reflects a presently sensed condition, the reservoir computing including:
(i) a capturing element configured to capture a temporal behavior of the sensor output; and
(ii) a classifying element configured to classify an output of the capturing element, including the temporal behavior to identify a trigger signature from among a plurality of trigger signatures, each trigger signature classifying one or more predetermined sensed conditions,
wherein the reconfigurable event driven hardware is configurable by the primary processor to detect the plurality of trigger signatures,
wherein the reconfigurable event driven hardware is adapted to continuously monitor the sensor output to detect a trigger signature from among the plurality of trigger signatures, and
wherein the reconfigurable event driven hardware is further adapted to communicate the trigger signature to the primary processor without communicating the one or more redetermined sensed conditions in response to a detection of a trigger signature while the primary processor is in the low power consumption mode.

16. The apparatus of claim 15 , wherein the reservoir computing includes a Liquid State Machine comprised of a plurality of leaky-integrate-and-fire neurons.

17. The apparatus of claim 15 , wherein the reservoir computing includes a Multi-Layered Perceptron Network comprised of a plurality of leaky-integrate-aid-fire neurons.

18. The apparatus of claim 15 , wherein the electronic sensor is an electroencephalograph.

Assignees

• Wisconsin Alumni Res Found

Inventors

• Lipasti Mikko H
• Hashmi Atif G
• Nere Andrew
• Tononi Giulio

Classifications

• G06F1/3209

  Monitoring remote activity, e.g. over telephone lines or network connections · CPC title

• G06F1/3293 Primary

  by switching to a less power-consuming processor, e.g. sub-CPU · CPC title

• G06F1/3215

  Monitoring of peripheral devices · CPC title

• Y02B60/121

  Cross-Sectional Technologies · mapped topic

• G06F1/3206 Primary

  Monitoring of events, devices or parameters that trigger a change in power modality · CPC title