Method and system of crowd- sourced magnetic fingerprinting with neural network re-training

1. A method, executed in a processor of a server computing device, of crowd-sourced magnetic fingerprinting for mobile device indoor navigation with neural network re-training, the method comprising:

determining, in the processor, at a plurality of locations, a set of magnetic input parameters in accordance with a magnetic infrastructure profile of at least a portion of an indoor area, the processor implementing an input layer of a neural network, the set of magnetic input parameters providing a magnetic feature input to the input layer of the neural network;

receiving, from a mobile device positioned at the first location, a set of measured magnetic parameters at respective ones of the plurality of locations;

computing a route segment similarity factor for the mobile device based at least in part on the measured magnetic parameters profile of a corresponding route segment accessible from the magnetic fingerprint dataset;

if the route segment similarity factor exceeds a threshold similarity factor, adding the set if measured magnetic parameters as measured by the mobile device to the magnetic fingerprint dataset;

for at least a second location along the route, receiving at least a second set of magnetic measured parameters of the mobile device;

computing, at an output layer of the neural network implemented by the processor, an error matrix based on comparing an initial matrix of weights associated with the at least a first neural network layer representing the magnetic feature input based on the magnetic fingerprint dataset to a magnetic feature output in accordance with the at least a second set of measured parameters; and

recursively adjusting the initial weights matrix by backpropagation to diminish the error matrix until the magnetic feature output matches the at least a second set of magnetic measured parameters within a threshold percentage value of the at least second set of magnetic measured parameters.

2. The method of claim 1 wherein the set of magnetic input parameters and the initial weights matrix are determined in the processor of the server computing device in accordance with execution of the magnetic infrastructure profile.

3. The method of claim 1 wherein the set of magnetic measured parameters comprises a magnetic field strength vector.

4. The method of claim 1 wherein the set of magnetic measured parameters comprises a magnetic dip angle.

5. The method of claim 1 wherein the set of magnetic measured parameters comprises a magnetic direction.

6. The method of claim 1 further recursively adjusting the initial weights matrix by backpropagation only after a predetermined duration of time in which the route segment similarity factor exceeds a threshold similarity factor.

7. The method of claim 1 wherein the neural network is one of a recurrent neural network and a convolution neural network.

8. The method of claim 7 wherein the neural network comprises the convolution neural network, wherein the at least a first neural network layer corresponds to the set of magnetic input parameters for a magnetic field component.

9. The method of claim 1 wherein the backpropagation comprises a backward propagation of errors in accordance with the error matrix as computed at the output layer, the errors being distributed backwards throughout the weights of the at least one neural network layer.

10. A server computing system for magnetic fingerprinting of an indoor area, the server computing system comprising:

a processor; and

a memory including instructions executable in the processor to:

determine, in the processor, at a plurality of locations, a set of magnetic input parameters in accordance with a magnetic infrastructure profile of at least a portion of an indoor area, the processor implementing an input layer of a neural network, the set of magnetic input parameters providing a magnetic feature input to the input layer of the neural network;

receive, from a mobile device positioned at the first location, a set of measured magnetic parameters at respective ones of the plurality of locations;

compute a route segment similarity factor for the mobile device based at least in part on the measured magnetic parameters profile of a corresponding route segment accessible from the magnetic fingerprint dataset;

if the similarity factor exceeds a threshold similarity factor, add the set of measured magnetic parameters as measured by the mobile device to the magnetic fingerprint dataset;

for at least a second location along the route, receive at least a second set of magnetic measured parameters of the mobile device;

compute, at an output layer of the neural network implemented by the processor, an error matrix based on comparing an initial matrix of weights associated with the at least a first neural network layer representing the magnetic feature input based on the magnetic fingerprint dataset to a magnetic feature output in accordance with the at least second set of measured parameters; and

recursively adjust the initial weights matrix by backpropagation to diminish the error matrix until the magnetic feature output matches the at least a second set of magnetic measured parameters within a threshold percentage value of the at least second set of magnetic measured parameters.

11. The server computing system of claim 10 wherein the set of magnetic input parameters and the initial weights matrix are determined in the processor of the server computing device in accordance with execution of the magnetic infrastructure profile.

12. The server computing system of claim 10 wherein the set of magnetic measured parameters comprises a magnetic field strength vector.

13. The server computing system of claim 10 wherein the set of magnetic measured parameters comprises a magnetic dip angle.

14. The server computing system of claim 10 wherein the set of magnetic measured parameters comprises a magnetic direction.

15. The server computing system of claim 10 further recursively adjusting the initial weights matrix by backpropagation only after a predetermined duration of time in which the route segment similarity factor exceeds a threshold similarity factor.

16. The server computing system of claim 10 wherein the neural network is one of a recurrent neural network and a convolution neural network.

17. The server computing system of claim 16 wherein the neural network comprises the convolution neural network, wherein the at least a first neural network layer corresponds to the set of magnetic input parameters for a magnetic field component.

18. The server computing system of claim 10 wherein the backpropagation comprises a backward propagation of errors in accordance with the error matrix as computed at the output layer, the errors being distributed backwards throughout the weights of the at least one neural network layer.