State estimation apparatus, state estimation method, and integrated circuit

• US 10,586,154 B2
• Filed: 03/24/2017
• Issued: 03/10/2020
• Est. Priority Date: 03/25/2016
• Status: Active Grant

• Current Assignee: Megachips Corporation
• Original Assignee: Megachips Corporation
• Inventors: Hasegawa, Hiromu
• Primary Examiner(s): Turner, Brian
• Application Number:
  US15/468,748
• Publication Number:
  US 20170278006A1
• Time in Patent Office: 1,082 Days
• CPC Class Codes: G06N 3/088 Non-supervised learning, e.g. competitive learning

Abstract

Provided is a state determination apparatus that appropriately performs pattern classification processing and/or pattern determination processing even when a map generated by the SOM technique includes discontinuous image regions (e.g., split image regions). In the state determination apparatus, the matching processing unit obtains adaptability data indicating a correlation degree between template data indicating a predetermined state and the SOM output data. The state determination unit determines a state of an input data. This allows for appropriately performing pattern classification processing and/or pattern determination processing even when a map generated by the SOM technique includes discontinuous image regions (e.g., split image regions).

Claims

1. A movement state determination system comprising:

an acceleration sensor configured to output an analog signal corresponding to movement of the acceleration sensor;

input circuitry configured to receive the analog signal output from the acceleration sensor, convert the received analog signal to digital signal, and transmit the digital signal;

feature vector obtaining circuitry configured to receive the digital signal transmitted from the input circuitry, perform frequency transform on the received digital signal to obtain feature vector data, and transmit the obtained feature vector data;

selector circuitry configured to receive the feature vector data transmitted from the feature vector obtaining circuitry, receive a control signal from a controller, transmit the feature vector data to (i) mapping conversion circuitry or (ii) self organizing map (SOM) processing circuitry depending on the control signal;

the mapping conversion circuitry configured to receive the feature vector data transmitted from the selector circuitry and obtain SOM output data by mapping the feature vector data into a space whose dimension differs from a dimension of the feature vector data;

matching processing circuitry configured to obtain adaptability data indicating a correlation degree between template data indicating a predetermined movement state and the SOM output data when a distribution of the SOM output data is discontinuous; and

movement state determination circuitry configured to

identify one of a plurality of candidate movement states of the acceleration sensor based on the adaptability data; and

output a signal indicating the identified one of the plurality of candidate movement states corresponding to the analog signal output by the acceleration sensor,

wherein

the acceleration sensor is configured to be attached to an object and the analog signal output by the acceleration sensor corresponds to movement of the object, and

the object is a human body and the analog signal output by the acceleration sensor corresponds to movement of the human body.

2. The movement state determination system of claim 1, wherein

the plurality of candidate movement states include at least a walking state, a running state and a standing state.