Method for generating accurate lane level maps

• US 9,384,394 B2
• Filed: 10/31/2013
• Issued: 07/05/2016
• Est. Priority Date: 10/31/2013
• Status: Active Grant

• Current Assignee: Toyota Jidosha Kabushiki Kaisha (Toyota Motor Corporation)
• Original Assignee: Toyota Motor Engineering & Manufacturing North America Incorporated (Toyota Motor Corporation)
• Inventors: Joshi, Avdhut S., James, Michael R.
• Primary Examiner(s): Bella, Matthew
• Assistant Examiner(s): SHIN, SOO JUNG
• Application Number:
  US14/069,069
• Publication Number:
  US 20150117723A1
• Time in Patent Office: 978 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G01C 21/20 Instruments for performing navigational calculations G01C21/24, G01C21/26 take precedence
• 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: G06V 30/422 Technical drawings; Geographical maps
• CPC Class Codes: G09B 29/106 using electronic means

Abstract

A method for generating accurate lane level maps based on course map information and Lidar data captured during the pass of a sensor carrying vehicle along a road. The method generates accurate lane estimates including the center of each lane, the number of lanes, and the presence of any bicycle paths and entrance and exit ramps using a computer-implemented method where the course map data and the Lidar data are subjected to particle filtering.

Claims

1. A method for determining road lane characteristics of a road, the method comprising:

obtaining Lidar data of road markers on the road, and extracting road lane marker data from the Lidar data;

accessing road map data;

determining road segment node locations from the road map data, wherein the nodes are evenly spaced at an interval defined by δ

_o/2;

searching for line segments lateral to the nodes, wherein the searching is performed over a rectangle defined by vectors of a length δ

_o/2;

tying the Lidar generated road marker data to respective road segment node locations;

applying particle filtering, using a dual particle filter, to the road lane marker data and road segment node locations to estimate road lane characteristics including at least one of a bicycle lane, a road exit, and a road entrance adjacent to the road; and

generating a road lane map including the road lane characteristics.

2. The method of claim 1 wherein the step of applying particle filtering comprises:

using both a regular particle filter and the dual particle filter.

3. The method of claim 1 further comprising:

using both a regular particle filter and the dual particle filter to detect;

(1) a transition of one lane of a road into at least two side-by-side road lanes, or (2) a transition of two side-by-side road lanes into one road lane.

4. The method of claim 1 further comprising:

refining the road lane marker data by use of a GraphSlam algorithm.

5. The method of claim 1 wherein the step of tying the Lidar generated road marker data to the road segment node locations comprises:

generating each of the road segment node locations one at a time;

determining side boundaries of a road lane at the one node;

forming a set of all line segments within the rectangle;

filtering the set of line segments within the rectangle;

clustering line segments to the lane markers;

computing a mean offset distance;

grouping lane markers longitudinally; and

calculating binary features of the Lidar generated marker data and map nodes.

6. The method of claim 1 wherein the road lane characteristics further include at least one of a number of lanes, a width of the lanes, and a center of each lane.

7. The method of claim 1 wherein the step of obtaining Lidar data of road markers comprises:

equipping a vehicle with a GPS locator, a rotating laser beam, and a reflected laser beam detector;

moving the vehicle along one lane of the road; and

detecting the intensity of the laser beam reflected from reflective road markers on all lanes of the road as the vehicle moves along the one lane of the road.

8. A method of generating a road lane map, the method comprising:

extracting road lane marker data from Lidar data obtained by a vehicle moving along one lane of a road;

tying the extracted road marker data to respective road segment node locations obtained from prior road map data, wherein the nodes are evenly spaced at an interval defined by δ

o, and a search for line markers lateral to the nodes is performed over a rectangle defined by vectors of a length δ

o/2;

using both a regular particle filter and a dual particle filter applied to the road lane marker data or the road segment node locations to detect lane characteristics including at least one of;

(1) a transition of one lane of a road into at least two side-by-side road lanes, and (2) a transition of two side-by-side road lanes into one road lane; and

generating a road lane map depicting the detected lane characteristics.

9. The method of claim 8 further comprising:

using particle filtering to detect at least one of a bicycle lane, a road exit, and a road entrance adjacent to the one lane of the road.

10. The method of claim 8, further comprising:

refining the road lane marker data by use of a GraphSlam algorithm.

11. The method of claim 8 further comprising:

detecting at least one of a number of lanes, a width of the lanes, and a center of each lane of the road.