Identification device, identification system, identification method, and program recording medium using a walking waveform of a user

Patent metadata

Field	Value
Publication number	US-12299100-B2
Application number	US-202018015876-A
Country	US
Kind code	B2
Filing date	Jul 22, 2020
Priority date	Jul 22, 2020
Publication date	May 13, 2025
Grant date	May 13, 2025

Abstract

Provided is an identification device for identifying an individual on the basis of gait irrespective of the type of footwear, the identification device comprising a detection unit that detects a walking event on the basis of a walking waveform of a user, a waveform processing unit that normalizes the walking waveform on the basis of the detected walking event and generates a normalized waveform, and an identification unit that identifies the user on the basis of the normalized waveform.

First claim

What is claimed is:

1. An identification device comprising:
one or more memories storing instructions; and
one or more processors configured to execute the instructions to:
detect a walking event from a walking waveform of a user;
normalize the walking waveform based on the detected walking event to generate a normalized waveform;
identify the user based on the normalized waveform;
wherein generating the normalized waveform includes:
normalize a plantar angle walking waveform to generate the normalized waveform used for identifying the user; and
normalize walking waveforms of each of accelerations in three axial directions, angular velocities in the three axial directions, and angles in the three axial directions in accordance with the generated normalized plantar angle walking waveform to generate normalized waveforms of each of the accelerations in the three axial directions, the angular velocities in the three axial directions, and the angles in the three axial directions;
wherein the normalized waveform used for identifying the user comprises at least the generated normalized plantar angle walking waveform and the normalized waveforms of each of the accelerations in the three axial directions, the angular velocities in the three axial directions, and the angles in the three axial directions.

2. The identification device according to claim 1 ,
wherein the one or more processors are configured to execute the instructions to:
detect a first dorsiflexion peak, a first plantarflexion peak, a second dorsiflexion peak, and a second plantarflexion peak from the walking waveform of a plantar angle for two gait cycles;
set a time at a midpoint between a first time of the first dorsiflexion peak and a second time of the first plantarflexion peak as a start point time;
set a time at a midpoint between a third time of the second dorsiflexion peak and a fourth time of the second plantarflexion peak as an end point time;
cut out the walking waveform for one gait cycle from the start point time to the end point time;
divide the cut-out walking waveform for one gait cycle into a first divided waveform from the start point time to the second time, a second divided waveform from the second time to the third time, and a third divided waveform from the third time to the end point time;
normalize each of the first divided waveform, the second divided waveform, and the third divided waveform; and
integrate the normalized first divided waveform, second divided waveform, and third divided waveform to generate the normalized waveform of the plantar angle.

3. The identification device according to claim 2 ,
wherein the one or more processors are configured to execute the instructions to normalize each of the first divided waveform, the second divided waveform, and the third divided waveform in such a way that, in one gait cycle, the first divided waveform has a fraction of 30%, the second divided waveform has a fraction of 40%, and the third divided waveform has a fraction of 30%.

4. The identification device according to claim 1 ,
wherein the one or more processors are configured to execute the instructions to input a feature extracted from the normalized waveform of at least one of accelerations, angular velocities, and angles of an identification target user in three axial directions to a trained model trained using, as training data, a predictor vector including the feature extracted from the normalized waveform of at least one of accelerations, angular velocities, and angles of a registration target user in the three axial directions and an identifier of the registration target user, and identifies the identification target user.

5. The identification device according to claim 4 ,
wherein the one or more processors are configured to execute the instructions to access a database storing authentication information to perform authentication of the user identified by the identification device.

6. The identification device according to claim 5 ,
wherein the one or more processors are configured to execute the instructions to control a control target device according to an authentication result.

7. An identification method executed by a computer, the identification method comprising:
detecting a walking event from a walking waveform of a user;
normalizing the walking waveform based on the detected walking event to generate a normalized waveform; and
identifying the user based on the normalized waveform;
wherein generating the normalized waveform includes:
normalizing a plantar angle walking waveform to generate the normalized waveform used for identifying the user; and
normalizing walking waveforms of each of accelerations in three axial directions, angular velocities in the three axial directions, and angles in the three axial directions in accordance with the generated normalized plantar angle walking waveform to generate normalized waveforms of each of the accelerations in the three axial directions, the angular velocities in the three axial directions, and the angles in the three axial directions;
wherein the normalized waveform used for identifying the user comprises at least the generated normalized plantar angle walking waveform and the normalized waveforms of each of the accelerations in the three axial directions, the angular velocities in the three axial directions, and the angles in the three axial directions.

8. A non-transitory program recording medium recorded with a program for causing a computer to execute:
processing of detecting a walking event from a walking waveform of a user;
processing of normalizing the walking waveform based on the detected walking event to generate a normalized waveform; and
processing of identifying the user based on the normalized waveform;
wherein generating the normalized waveform includes:
normalizing a plantar angle walking waveform to generate the normalized waveform used for identifying the user; and
normalizing walking waveforms of each of accelerations in three axial directions, angular velocities in the three axial directions, and angles in the three axial directions in accordance with the generated normalized plantar angle walking waveform to generate normalized waveforms of each of the accelerations in the three axial directions, the angular velocities in the three axial directions, and the angles in the three axial directions;
wherein the normalized waveform used for identifying the user comprises at least the generated normalized plantar angle walking waveform and the normalized waveforms of each of the accelerations in the three axial directions, the angular velocities in the three axial directions, and the angles in the three axial directions.

Assignees

• Nec Corp

Inventors

• Huang Chenhui
• Fukushi Kenichiro
• Wang Zhenwei
• Nihey Fumiyuki

Classifications

• A61B5/7235

  Details of waveform analysis (detecting specific parameters of the electrocardiograph cycle A61B5/349) · CPC title

• A61B5/112

  Gait analysis · CPC title

• A61B5/7242

  using integration · CPC title

• A61B2562/0219

  Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches · CPC title

• A61B5/6807

  Footwear · CPC title