Device and method for generating geomagnetic sensor based location estimation model using artificial neural networks
AssigneeKorea University Research and Business Foundation
InventorsLynn Choi, Ho Jun Jang
Life of this patent
Estimating the term…
Abstract
A device that generates a location estimation model is provided. The location estimation model generator device includes a map generator part configured to generate a magnetic field map, which includes magnetic field values corresponding respectively to the coordinates of an indoor space; a data generator part configured to generate learning data by implementing the magnetic field map; and a learning part configured to generate a location estimation model by artificial neural network (ANN) learning implementing the learning data.
Claims13
A device that generates a location estimation model, the device comprising:
- a map generator part configured to generate a magnetic field map, the magnetic field map including magnetic field values corresponding respectively to coordinates of an indoor space;
- a data generator part configured to generate virtual paths of movements of a pedestrian in the indoor space, and generate learning data by using the magnetic field map;
- and
- a learning part configured to generate the location estimation model by artificial neural network (ANN) learning using the learning data,
- wherein the learning part iterates learning until an output result of the ANN achieves a target error rate, and
- wherein a total number of epochs decreases as a total number of hidden nodes increases, until the target error rate is achieved.
A method, comprising:
- generating a magnetic field map from magnetic field values corresponding to coordinates of an indoor space by using linear interpolation, the generating of the magnetic field map performed by a map generator part of a location estimation model generator device;
- generating learning data including coordinates of each point included in a virtual path of a pedestrian by using the magnetic field map, the generating of the learning data performed by a data generator part of the location estimation model generator device;
- and
- generating a location estimation model by recurrent neural network learning using the learning data, the generating of the location estimation model performed by a learning part of the location estimation model generator device,
- wherein the generating of the learning data comprises iterating learning until an output result of the recurrent neural network achieves a target error rate, and
- wherein a total number of epochs decreases as a total number of hidden nodes increases, until the target error rate is achieved.
A location estimation device, comprising:
- a magnetic field sensor part configured to measure continuous three-dimensional magnetic field values in predetermined intervals as the magnetic field values change according to a movement of a user within an indoor space;
- a location estimator part configured to estimate a current location of the user by inputting the continuous three-dimensional magnetic field values measured by the magnetic field sensor part into a learned artificial neural network (ANN);
- and
- a location output part configured to output the current location of the user estimated by the location estimator part,
- wherein learning of the ANN is implemented until an output of the ANN achieves a target error rate, and
- wherein a total number of epochs decreases as a total number of hidden nodes increases, until the target error rate is achieved.
Dependencies are inferred from the claim text.