Environment sensing using wireless power transmission and a neural network

1. A method, comprising:

inputting, by a system comprising a processor via a neural network of the system, a reference-subtraction corrected phase sample of sensed environment information; and

based on the sensed environment information, mapping, by the system, an environment, wherein the sensed environment information comprises distance information representing a distance from a power transmitter of the system to an object, and wherein the mapping of the environment comprises:

based on the reference-subtraction corrected phase sample, estimating, via the neural network, the distance from the power transmitter to the object; and

based on the distance, generating a map of the environment comprising the object.

2. The method of claim 1 , wherein the inputting of the reference-subtraction corrected phase sample comprises:

receiving phase data; and

based on the phase data, generating the reference-subtraction corrected phase sample.

3. The method of claim 2 , wherein the generating of the reference-subtraction corrected phase sample comprises:

determining a calibration phase sample;

obtaining, via a wireless receiver device of the system, a collected phase sample; and

subtracting the calibration phase sample from the collected phase sample to obtain the reference-subtraction corrected phase sample.

4. The method of claim 1 , wherein the neural network comprises a convolutional neural network comprising respective layers that are used to extract respective types of information from given data, and wherein the estimating of the distance comprises:

performing, via the convolutional neural network, semantic segmentation to encode the reference-subtraction corrected phase sample to return a recreated sample as an abstract representation for evaluating regions and time-series elements of the recreated sample.

5. The method of claim 4 , further comprising:

in response to the performing of the encoding, applying, by the system via the convolutional neural network, a weighted crop on the recreated sample to generate cropped images.

6. The method of claim 5 , further comprising:

representing, by the system via the convolutional neural network, the cropped images as a single array of units.

7. The method of claim 6 , further comprising:

inputting, by the system via the convolutional neural network, the single array of units into long-short term memory layers that introduce an internal state in the neural network to facilitate evaluation of the time-series elements.

8. The method of claim 7 , wherein the time-series elements comprise a time-series of input phases, and wherein the estimating of the distance comprises:

based on the time-series of input phases, estimating the distance from the power transmitter to the object.

9. A system, comprising:

a processor; and

a memory that stores executable instructions that, when executed by the processor, facilitate performance of operations, comprising:

inputting a reference-subtraction corrected phase sample of sensed environment information; and

based on the sensed environment information, mapping, via a neural network, an environment, wherein the sensed environment information comprises distance information, wherein the neural network comprises a convolutional neural network that is a type of neural network that performs image-type analysis tasks by applying layers of filters to input data comprising the reference-subtraction corrected phase sample to obtain feature data and regress to a corrected output representing the distance information.

10. The system of claim 9 , wherein the operations further comprise:

estimating, via the convolutional neural network using a multi-variate linear regression, the distance information.

11. The system of claim 10 , wherein the estimating of the distance information comprises:

generating, via the convolutional neural network, input images comprising the input data;

for each input image of the input images, generating an abstract corrected representation of a processed sample;

based on the processed sample, generating long-short term memory layers that introduce an internal state in the neural network to facilitate an evaluation of a time-series of input phases; and

based on an evaluation of the time-series of input phases, estimating the distance information.

12. The system of claim 11 , wherein the system comprises a wireless receiver device, and wherein the estimating of the distance information comprises:

estimating a distance from the wireless receiver device to an object.

13. The system of claim 12 , wherein the input data comprises phase information of a reflected beacon signal that has been received by the wireless receiver device.

14. The system of claim 12 , wherein the mapping of the environment comprises:

based on the distance, generating a map of the environment comprising the object.

15. The system of claim 12 , wherein the object comprises an active power receiver, an obstruction, or a living entity.

16. A non-transitory machine-readable medium comprising instructions that, in response to execution, cause a system comprising a processor to perform operations, comprising:

obtaining phase sample data corresponding to wireless power transmission;

correcting the phase sample data based on reference data to obtain corrected sample data; and

applying a neural network to the corrected phase sample data to obtain sensed environment data.

17. The non-transitory machine-readable medium of claim 16 , wherein the operations further comprise:

controlling subsequent wireless power transmission based on the sensed environment data.

18. The non-transitory machine-readable medium of claim 16 , wherein the sensed environment data comprises distance data representing a distance from a power transmitter of the system to an object.

19. The non-transitory machine-readable medium of claim 18 , wherein the operations further comprise mapping an environment corresponding to the sensed environment data, the mapping comprising, based on a reference-subtraction corrected phase sample, estimating, via the neural network, the distance from the power transmitter to the object.

20. The non-transitory machine-readable medium of claim 19 , wherein the mapping further comprises, based on the distance, generating a map of the environment comprising the object.