Reconfigurable event driven hardware using reservoir computing for monitoring an electronic sensor and waking a processor
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.
1. A method for analyzing a sensory stream from an electronic sensor in an energy-constrained environment, the method comprising:
(a) configuring a reconfigurable event driven hardware in communication with a plurality of electronic sensory to detect a trigger signature from a sensory stream of a sensor to be monitored from among the plurality of electronic sensors as indicated by an application, wherein the reconfigurable event driven hardware includes reservoir computing implementing:
(i) a capturing element configured to capture a temporal behavior of the sensory stream; and
(ii) a classifying element configured to classify an output of the capturing element to identify the trigger signature, and
wherein an interconnect allows configuring the capturing element and the classifying element into structures defined for the trigger signature, the structures each comprising a plurality of neurons having defined neuron parameters;
(b) placing a processor in a low power consumption mode while the reconfigurable event driven hardware monitors the sensory stream to detect the trigger signature;
(c) after step (b), upon the reconfigurable event driven hardware detecting the trigger signature, communicating details of the trigger signature to the processor, and invoking the application in response to the communication; and
(d) reconfiguring the reconfigurable event driven hardware to detect a different trigger signature from the sensory stream during runtime of the application in response to a chance from the application.
2. The method of claim 1 , further comprising using the processor to configure the reconfigurable event driven hardware to detect the trigger signature.
3. The method of claim 2 , further comprising the reconfigurable event driven hardware continuously monitoring the sensory stream to detect the trigger signature while the processor is in the low power consumption mode.
4. The method of claim 2 , further comprising configuring the reconfigurable event driven hardware via an Application Program Interface (API).
5. The method of claim 1 , wherein the reservoir computing comprises a Liquid State Machine (LSM) comprised of a plurality of leaky-integrate-and-fire neurons.
6. The method of claim 5 , wherein the reservoir computing further comprises a Multi-Layered Perceptron Network (MLPN) comprised of a plurality of leaky-integrate-and-fire neurons.
7. The method of claim 6 , further comprising implementing the capturing element using the LSM and the classifying using the MLPN.
8. The method of claim 1 , further comprising, after step (c), deactivating the reconfigurable event driven hardware and using the processor for continuously monitoring the sensory stream.
9. The method of claim 1 , wherein the electronic sensor is at least one of an electrocardiograph and an electroencephalograph, and further comprising the configuring the reconfigurable event driven hardware to continuously monitor slow wave activity of a brain.
10. The method of claim 1 , further comprising providing the electronic sensor and the reconfigurable event driven hardware in a first enclosure and providing the processor in a second enclosure separate from the first enclosure.
11. The method of claim 1 , further comprising providing the processor and the reconfigurable event driven hardware in a first enclosure and providing the electronic sensor in a second enclosure separate from the first enclosure.
12. A system for analyzing a sensory stream in an energy-constrained environment comprising:
plurality of electronic sensors producing sensory streams;
a processor having a low power consumption mode; and
a reconfigurable event driven hardware in communication with the plurality of electronic sensors and the processor, the reconfigurable event driven hardware being configured to detect a trigger signature from a sensory stream of a sensor to be monitored from among the plurality of electronic sensors as indicated by an application, wherein the reconfigurable event driven hardware includes reservoir computing implementing:
(i) a capturing element configured to capture a temporal behavior of the sensory stream; and
(ii) a classifying element configured to classify an output of the capturing element to identify the trigger signature, and
wherein an interconnect allows configuring the capturing element and the classifying element into structures defined for the trigger signature, the structures each comprising a plurality of neurons having defined neuron parameters,
wherein, upon the reconfigurable event driven hardware detecting the trigger signature, the reconfigurable event driven hardware is operable to communicate details of the trigger signature to the processor while the processor is in the low power consumption mode, and invoke the application in response to the communication, and
wherein the processor is operable to reconfigure the reconfigurable event driven hardware to detect a different trigger signature from the sensory stream during runtime of the application in response to a change from the application.
13. The system of claim 12 , wherein the processor configures the reconfigurable event driven hardware to detect the trigger signature.
14. The system of claim 13 , wherein the reconfigurable event driven hardware continuously monitors the sensory stream to detect the trigger signature while the processor is in the low power consumption mode.
15. The system of claim 12 , wherein the reservoir computing comprises a Liquid State Machine (LSM) comprised of a plurality of leaky-integrate-and-fire neurons.
16. The system of claim 15 , wherein the reservoir computing further comprises a Multi-Layered Perceptron Network (MLPN) comprised of a plurality of leaky-integrate-and-fire neurons.
17. The system of claim 16 , wherein the capturing element is implemented by the LSM and the classifying is implemented by the MLPN.
18. The system of claim 12 , wherein the electronic sensor is at least one of an electrocardiograph and an electroencephalograph and the reconfigurable event driven hardware is configured to continuously monitor slow wave activity of a brain.
19. The system of claim 12 , wherein the reconfigurable event driven hardware and one of the electronic sensor and the processor is provided in a first enclosure, and wherein one of the electronic sensor and the processor not provided in the first enclosure is provided in a second enclosure separate from the first enclosure.