Environment recognition system

Patent metadata

Field	Value
Publication number	US-9779312-B2
Application number	US-201615007276-A
Country	US
Kind code	B2
Filing date	Jan 27, 2016
Priority date	Jan 30, 2015
Publication date	Oct 3, 2017
Grant date	Oct 3, 2017

Abstract

Provided is a system capable of further reducing risk such as a contact between a moving body such as a vehicle and a traffic participant present around the moving body. According to an environment recognition system ( 1 ) of the present invention, a database ( 10 ) stores each of a plurality of reference symbol strings describing the state of an environmental element constituting each of a plurality of scenes assumed to be around the moving body. A first arithmetic processing element ( 11 ) detects a scene around the moving body and generates a symbol string describing the state of the environmental element constituting the detected scene. A second arithmetic processing element ( 12 ) evaluates similarity between the symbol string and each of the plurality of reference symbol strings stored in the database ( 10 ).

First claim

What is claimed is:

1. An environment recognition system comprising:
a database which stores each of a plurality of reference symbol strings describing a state of an environmental element constituting each of a plurality of scenes assumed to be around a moving body;
a first arithmetic processing element which detects a scene around the moving body and generates a symbol string describing a state of an environmental element constituting the scene; and
a second arithmetic processing element which evaluates similarity between the symbol string generated by the first arithmetic processing element and each of the plurality of reference symbol strings stored in the database,
wherein, in a case where a difference in similarity to the symbol string generated by the first arithmetic processing element between a first reference symbol string and a second reference symbol string among the plurality of reference symbol strings is equal to or lower than a threshold value, wherein the first reference symbol string holds a first place in ranking of the similarity and a second reference symbol string holds a second place in the ranking of the similarity, the second arithmetic processing element selects a reference symbol string which has been selected most frequently in a selection history for a latest specified number of times or a reference symbol string selected at a previous time among the first reference symbol string and the second reference symbol string.

2. The environment recognition system according to claim 1 , wherein the second arithmetic processing element selects a reference symbol string having a highest current value of the similarity or a highest cumulative value for latest specified number of times as a current reference symbol string and calculates potential risk of the moving body based on at least some symbols constituting the selected reference symbol string according to a data table stored in the database or a specification rule defined by a function.

3. The environment recognition system according to claim 2 , wherein the second arithmetic processing element evaluates the potential risk of the moving body based on a symbol describing at least one of a type, position, and behavior of a traffic participant as the environmental element as the at least some symbols.

4. The environment recognition system according to claim 2 , wherein:
the first arithmetic processing element extracts a location on a deep side when viewed from the moving body, as a caution needed location, out of objects present in a movable region of the moving body or adjacent thereto in the scene; and
the second arithmetic processing element evaluates the potential risk of the moving body based on at least one of positions and a number of caution needed locations extracted by the first arithmetic processing element in addition to the selected reference symbol string.

5. The environment recognition system according to claim 2 , further comprising a third arithmetic processing element which generates at least one execution command among notification to an operator of the moving body or a person as the moving body or to traffic participants present around the moving body, acceleration suppression of the moving body, steering control, and changing speed of the moving body, depending on a degree of potential risk to the moving body evaluated by the second arithmetic processing element.

6. The environment recognition system according to claim 1 , wherein the second arithmetic processing element evaluates the similarity to the current symbol string generated by the first arithmetic processing element, only with respect to some reference symbol strings whose similarity to a previous symbol string, which has been generated by the first arithmetic processing element, is equal to or higher than a predetermined value or is within a predetermined ranking, among the plurality of reference symbol strings.

7. An environment recognition method comprising:
a first arithmetic processing step for detecting a scene around a moving body and then generating a symbol string describing a state of an environmental element constituting the scene; and
a second arithmetic processing step for evaluating similarity between the symbol string generated in the first arithmetic processing step and each of a plurality of reference symbol strings describing a state of an environmental element constituting each of a plurality of scenes assumed to be around the moving body stored in a database; and
a third arithmetic processing step for, in a case where a difference in similarity to the symbol string generated in the first arithmetic processing step between a first reference symbol string and a second reference symbol string among the plurality of reference symbol strings is equal to or lower than a threshold value, wherein the first reference symbol string holds a first place in ranking of the similarity and a second reference symbol string holds a second place in the ranking of the similarity, selecting a reference symbol string which has been selected most frequently in a selection history for a latest specified number of times or a reference symbol string selected at a previous time among the first reference symbol string and the second reference symbol string.

Assignees

• Honda Motor Co Ltd

Inventors

• Sakagami Yoshiaki
• Osada Taku
• Ishikawa Kensuke

Classifications

• G06V10/768

  using context analysis, e.g. recognition aided by known co-occurring patterns · CPC title

• G06F17/30268

  Physics · mapped topic

• G06K9/00805 Primary

  Physics · mapped topic

• G06K9/72

  Physics · mapped topic

• G06V20/58 Primary

  Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads · CPC title