ROAD SHAPE LEARNING APPARATUS

• US 20110218724A1
• Filed: 02/18/2011
• Published: 09/08/2011
• Est. Priority Date: 03/04/2010
• Status: Active Grant

• Current Assignee: DENSO Corporation
• Original Assignee: DENSO Corporation
• Inventors: IIDA, Masahiro
• Granted Patent: US 8,583,366 B2
• Time in Patent Office: Days
• US Class Current: 701/70
• CPC Class Codes: B60W 2556/50 of positioning data, e.g. GPS [Global Positioning System] data
• CPC Class Codes: B60W 40/072 Curvature of the road
• CPC Class Codes: G01C 21/30 Map- or contour-matching
• CPC Class Codes: G01C 21/3819 Road shape data, e.g. outline of a route
• CPC Class Codes: G01C 21/3837 Data obtained from a single source
• CPC Class Codes: G01C 21/3844 Data obtained from position sensors only, e.g. from inertial navigation
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

When determining that a subject vehicle traveled a curve, a road shape learning apparatus amends an entrance coordinate, a middle coordinate, and an exit coordinate of the curve based on predetermined amendment values to thereby obtain a post-amendment entrance coordinate, a post-amendment middle coordinate, and a post-amendment exit coordinate. The road shape learning apparatus computes a radius of a circular arc which passes through the post-amendment entrance coordinate, the post-amendment middle coordinate, and the post-amendment exit coordinate and designates the computed radius as a curvature radius of the curve.

Claims

1. A road shape learning apparatus for a vehicle, the apparatus comprising:

a map data storage device configured to store map data;

a position acquisition device configured to acquire position information relative to a position of the vehicle;

a position extraction section configured to, when it is determined that the vehicle traveled a curve,

specify an entrance, a middle, and an exit of the curve based on the map data, and

extract an entrance coordinate, a middle coordinate, and an exit coordinate which indicate the positions of the vehicle corresponding to the entrance, the middle, and the exit, respectively, from the position information relative to the vehicle acquired by the position information acquisition device;

a position amendment section configured to obtain a post-amendment entrance coordinate, a post-amendment middle coordinate, and a post-amendment exit coordinate, which are obtained by amending the entrance coordinate, the middle coordinate, and the exit coordinate using amendment values which are predetermined depending on a travel tendency;

a curvature radius computation section configured to

compute a radius of a circular arc passing through the post-amendment entrance coordinate, the post-amendment middle coordinate, and the post-amendment exit coordinate and

designate the computed radius as a curvature radius of the curve; and

a learned data storage device configured to store the designated curvature radius as learned data associated with the map data.

2. The road shape learning apparatus according to claim 1,

the position amendment section being further configured to

perform a curve direction determination as to whether the curve is a left-hand curve or a right-hand curve based on the position information relative to the vehicle acquired by the position acquisition section, and

designate an amendment value responding to an out-in-out tendency of the travel tendency based on either one of the left-hand curve and the right-hand curve, which is determined by the performed curve direction determination.

3. The road shape learning apparatus according to claim 2,

the position amendment section being further configured to

define an amendment value relative to each of the entrance coordinate and the exit coordinate as H1,

define an amendment value relative to the middle coordinate as H2,

the position amendment section being further configured to, with respect to a right-hand traffic,

designate a relation of H1 >

H2 when the curve is determined to be the left-hand curve relative to a heading direction of the vehicle, and

designate a relation of H1 <

H2 when the curve is determined to be the right-hand curve relative to a heading direction of the vehicle,

the position amendment section being further configured to, with respect to a left-hand traffic,

designate a relation of H1 <

H2 when the curve is determined to be the left-hand curve relative to a heading direction of the vehicle, and

designate a relation of H1 >

H2 when the curve is determined to be the right-hand curve relative to a heading direction of the vehicle.

4. The road shape learning apparatus according to claim 1,

the position amendment section being further configured to designate the amendment value as a value corresponding to a distance from a center in a road width direction.

5. The road shape learning apparatus according to claim 1, further comprising:

a travel track storage device configured to store the position information relative to the vehicle acquired by the position acquisition device as a travel track; and

a travel track amendment section configured to amend the travel track stored in the travel track storage section,

the travel track amendment section being further configured to, with respect to the travel track stored in the travel track storage section,

apply a parallel translation to a road portion before entering the entrance of the curve while applying a parallel translation to a road portion after exiting the exit of the curve, amending the travel track.

6. The road shape learning apparatus according to claim 1, further comprising:

a travel track storage device configured to store the position information relative to the vehicle acquired by the position acquisition device as a travel track; and

a travel track amendment section configured to amend the travel track stored in the travel track storage section,

the travel track amendment section being further configured to, with respect to the travel track stored in the travel track storage section,

interpolate using a curve

an interval between the straight line portion and the post-amendment entrance coordinate in a road portion of the travel track before entering the curve, and

an interval between the straight line portion and the post-amendment exit coordinate, in a road portion of the travel track after exiting the curve.

7. The road shape learning apparatus according to claim 1, further comprising:

a control data generation section configured to generate control data to control a behavior of the vehicle based on the learned data.

8. The road shape learning apparatus according to claim 7,

the control data generation section being further configured to generate control data to control a velocity of the vehicle.

9. The road shape learning apparatus according to claim 1,

the position amendment section being further configured to

assume an out-in-out travel tendency, in which the vehicle travels a lane of the curve to generate a travel track forming a circular arc having a radius greater than a radius of a circular arc generated by a travel track generated by traveling the lane of the curve continuously along on a center of the lane in a width direction,

perform (i) a curve direction determination and (ii) a traffic direction determination, based on the map data and the acquired position information relative to the vehicle,

the curve direction determination being as to whether the vehicle travels a curve in a curve direction of either a left-hand curve or a right-hand curve,

the traffic direction determination being as to whether the vehicle travels a left-hand traffic or a right-hand traffic, and

designate an amendment value responding to the out-in-out travel tendency, with respect to each curve direction of each traffic direction, based on results of the performed curve direction determination and the traffic direction determination.