# US 2020/0401152 A1 · Self-location estimation method

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0401152 A1
- Document ID: US20200401152A1
- Status: Published application
- Kind: A1
- Application number: 16/975,259
- Filed: 2019-01-28
- Published: 2020-12-24
- Earliest claimed: 2018-04-20
- Assignees: Honda Motor Co., Ltd.
- Applicants: Honda Motor Co., Ltd.
- Inventors: Yuji Hasegawa; Sango Matsuzaki
- Web page: https://patentgrep.com/patent/US20200401152A1

## CPC

- G05D 1/0221: {involving a learning process}
- G01C 15/002: {Active optical surveying means (optical plumbing G01C15/105)}
- G01C 21/08: involving use of the magnetic field of the earth
- G01C 21/20: Instruments for performing navigational calculations (G01C21/24, G01C21/26 take precedence)
- G01C 22/00: Measuring distance traversed on the ground by vehicles, persons, animals or other moving solid bodies, e.g. using odometers, using pedometers
- G05D 1/0251: {extracting 3D information from a plurality of images taken from different locations, e.g. stereo vision (stereoscopic image analysis H04N13/00; depth recovery from images G06T7/593)}
- G05D 1/0259: {using magnetic or electromagnetic means}
- G05D 1/0272: {comprising means for registering the travel distance, e.g. revolutions of wheels (measuring distance traversed on the ground by vehicles, e.g. using odometers G01C22/00)}
- G05D 1/24: Arrangements for determining position or orientation
- G05D 1/2435: {Extracting 3D information}
- G05D 1/247: using signals provided by artificial sources external to the vehicle, e.g. navigation beacons
- G05D 1/644: Optimisation of travel parameters, e.g. of energy consumption, journey time or distance
- G06N 3/04: Architecture, e.g. interconnection topology
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06T 7/74: {involving reference images or patches}
- G06T 2207/20084: Artificial neural networks [ANN]
- G05D: SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G01C: MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY (measuring liquid level G01F; radio navigation, determining distance or velocity by use of propagation effects, e.g. Doppler effects, propagation time, of radio waves, analogous arrangements using other waves G01S)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

The present invention provides a self-location estimation method including: a first step of estimating the self-location of a moving body (1) from the detection information of a plurality of sensors (5) to (8) by using a plurality of algorithms (11) to (13); a second step of determining a weighting factor for each algorithm from one or more state quantities A, B and C, which are obtained by estimation processing for each of a plurality of algorithms, by using a trained neural network (14); and a third step of identifying, as the self-location of the moving body (1), a location obtained by synthesizing the self-locations, which have been estimated by the algorithms, by using weighting factors.

## Claims

### Claim 1

A self-location estimation method for a moving body, comprising:
a first step of estimating a self-location of a moving body from detection information of a plurality of sensors by each of a plurality of algorithms using the detection information of one or more sensors among the plurality of sensors;
a second step of determining a weighting factor for each of the plurality of algorithms by inputting, to a trained neural network, one or more state quantities which are obtained for each of the plurality of algorithms by estimation processing of each algorithm and which are state quantities associated with certainty of the self-location estimated by each algorithm, and then determining the weighting factor for each of the plurality of algorithms by the neural network based on the input state quantities; and
a third step of identifying, as a self-location of the moving body, a location obtained by synthesizing the self-locations estimated by each of the plurality of algorithms by using the determined weighting factors.

### Claim 2 (depends on 1)

The self-location estimation method for a moving body according to claim 1,
wherein the plurality of sensors include a camera that images an external world of the moving body,
an A-th algorithm that is one algorithm among the plurality of algorithms is an algorithm configured to detect a feature point from a captured image of the camera and to sequentially perform processing for matching the detected feature point with a feature point detected from an image captured earlier than the captured image in order to estimate the self-location of the moving body, and
the state quantity related to the A-th algorithm includes a state quantity that indicates a ratio of a quantity of feature points mapped by the matching to a total quantity of feature points detected from the captured image.

### Claim 3 (depends on 1)

The self-location estimation method for a moving body according to claim 1,
wherein the plurality of sensors include a distance measuring sensor that measures a distance to an external object of the moving body,
a B-th algorithm that is one algorithm among the plurality of algorithms is an algorithm which estimates the self-location of the moving body by using distance measurement data obtained by the distance measuring sensor and a particle filter, and
the state quantity related to the B-th algorithm includes a covariance generated in the process of estimating the self-location of the moving body by the B-th algorithm.

### Claim 4 (depends on 1)

The self-location estimation method for a moving body according to claim 1,
wherein the plurality of sensors include a magnetic sensor that detects magnetism acting on the moving body,
a C-th algorithm that is one of the plurality of algorithms is an algorithm which estimates the self-location of the moving body by using magnetism detection data obtained by the magnetic sensor and a particle filter, and
the state quantity related to the C-th algorithm includes a covariance generated in the process of estimating the self-location of the moving body by the C-th algorithm.