Granted Patent B2

US 12,613,328 · App. 17/865,667 · Granted Apr 28, 2026

Inventors: Jaehwa Lee (Seoul, KR); Dongkyu Park (Incheon, KR)

Assignee: Samsung Electronics Co., Ltd.

G01S13/726G01S7/415G01S13/89G01S13/931

Abstract

Claims (75)

1 . A processor-implemented method with grid map generation, comprising:

generating position information of a moving object corresponding to a first time step based on a position sensor of the moving object;

generating detection information of nearby objects present around the moving object corresponding to the first time step based on a radio detection and ranging (radar) sensor of the moving object;

selecting a still object in a moving range of the moving object from among the nearby objects, based on the position information and the detection information;

generating an updated point cloud by updating a point cloud determined based on the radar sensor in a previous time step of the first time step, based on the position information and on detection information of the still object comprised in the detection information of the nearby objects, wherein the detection information includes relative position information of the still object with respect to the position of the moving object; and

generating a grid map based on an occupancy probability for each grid of the updated point cloud,

wherein the generating of the updated point cloud comprises:

transforming coordinates of a first point cloud determined in the previous time step based on the position information;

generating a second point cloud corresponding to the still object, based on detection information of the still object determined in the first time step; and

generating the updated point cloud corresponding to the first time step based on the transformed first point cloud and the second point cloud,

wherein the detection information comprises relative position information of the nearby objects with respect to a position of the radar sensor, and

wherein the generating of the updated point cloud further comprises transforming relative position information of the still object comprised in the detection information with respect to a position of the moving object, based on a relative position relationship of the radar sensor and the moving object.

2 . The method of claim 1 , wherein the

transforming of the coordinates of the first point cloud is with respect to a position of the moving object in the first time step determined based on the position information;

and

wherein the generating of the updated point cloud comprises accumulating the second point cloud with the transformed first point cloud.

3 . The method of claim 2 , further comprising removing the first point cloud determined in the previous time step.

4 . The method of claim 1 , wherein the selecting of the still object comprises:

filtering out nearby objects positioned at a height greater than or equal to a preset first threshold value corresponding to a height of the moving object, based on relative position information of the nearby objects comprised in the detection information; and

selecting the still object from among the filtered nearby objects, based on velocity information of the moving object comprised in the position information and relative velocity information of the nearby objects comprised in the detection information.

5 . The method of claim 1 , wherein the selecting of the still object comprises:

filtering out nearby objects of which an angle formed with the moving object is greater than or equal to a preset second threshold value, based on relative position information of the nearby objects comprised in the detection information; and

selecting the still object from among the filtered nearby objects, based on velocity information of the moving object comprised in the position information and relative velocity information of the nearby objects comprised in the detection information.

6 . The method of claim 1 , wherein the generating of the updated point cloud further comprises:

generating a region of interest (ROI) of traveling of the moving object based on the position information of the moving object; and

filtering out a point cloud corresponding to a still object that is not comprised in the ROI among point clouds corresponding to still objects selected through the selecting.

7 . The method of claim 1 , wherein the generating of the grid map further comprises generating a moving path of the moving object corresponding to the still object based on the grid map.

8 . The method of claim 7 , wherein the generating of the moving path of the moving object comprises generating the moving path of the moving object that detours a space occupied by the still object, based on the grid map.

9 . The method of claim 1 , wherein the generating of the position information of the moving object comprises determining any one or any combination of any two or more of velocity information of the moving object, yaw rate information of the moving object, and global positioning system (GPS) information of the moving object corresponding to the first time step.

10 . The method of claim 1 , wherein the generating of the detection information of the nearby objects comprises determining either one or both of relative position information of the nearby objects and relative velocity information of the nearby objects with respect to the radar sensor corresponding to the first time step.

11 . The method of claim 1 , wherein the point cloud is generated based on a signal sensed by the radar sensor, and comprises a set of one or more points corresponding to the sensed signal in a coordinate system having a position of the moving object as a reference point.

12 . A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, configure the one or more processors to perform the method of claim 1 .

13 . An apparatus with grid map generation, comprising:

one or more processors configured to:

generate position information of a moving object corresponding to a first time step based on a position sensor of the moving object;

generate detection information of nearby objects present around the moving object corresponding to the first time step based on a radio detection and ranging (radar) sensor of the moving object;

select a still object in a moving range of the moving object from among the nearby objects based on the position information and the detection information;

generate an updated point cloud determined based on the radar sensor in a previous time step of the first time step, based on the position information and detection information of the still object comprised in the detection information of the nearby objects, wherein the detection information includes relative position information of the still object with respect to the position of the moving object; and

generating a grid map based on an occupancy probability for each grid of the updated point cloud,

wherein the generating of the updated point cloud comprises:

transform coordinates of a first point cloud determined in the previous time step based on the position information;

generate a second point cloud corresponding to the still object, based on detection information of the still object determined in the first time step; and

generate the updated point cloud corresponding to the first time step based on the transformed first point cloud and the second point cloud,

wherein the detection information comprises relative position information of the nearby objects with respect to a position of the radar sensor, and

wherein for the generating of the updated point cloud, the one or more processors are configured to transform relative position information of the still object comprised in the detection information with respect to a position of the moving object, based on a relative position relationship of the radar sensor and the moving object.

14 . The apparatus of claim 13 , wherein

the transformation of the coordinates of the first point cloud is with respect to a position of the moving object in the first time step determined based on the position information;

and

wherein the generation of the updated point cloud comprises accumulating the second point cloud with the transformed first point cloud.

15 . The apparatus of claim 13 , wherein, for the selecting of the still object, the one or more processors are configured to:

filter out nearby objects positioned at a height greater than or equal to a preset first threshold value corresponding to a height of the moving object, based on relative position information of the nearby objects comprised in the detection information; and

select the still object from among the filtered nearby objects, based on velocity information of the moving object comprised in the position information and relative velocity information of the nearby objects comprised in the detection information.

16 . The apparatus of claim 13 , wherein, for the selecting of the still object, the one or more processors are configured to:

filter out nearby objects of which an angle formed with the moving object is greater than or equal to a preset second threshold value, based on relative position information of the nearby objects comprised in the detection information; and

select the still object from among the filtered nearby objects, based on velocity information of the moving object comprised in the position information and relative velocity information of the nearby objects comprised in the detection information.

17 . The apparatus of claim 13 , wherein, for the generating of the updated point cloud, the one or more processors are configured to:

generate a region of interest (ROI) of traveling of the moving object based on the position information of the moving object; and

filter out a point cloud corresponding to a still object that is not comprised in the ROI among point clouds corresponding to still objects selected through the selecting.

18 . The apparatus of claim 13 , wherein, for the generating of the grid map, the one or more processors are configured to generate a moving path of the moving object corresponding to the still object based on the grid map.

19 . A method with grid map generation, comprising:

generating a first point cloud of a stationary object using a radio detection and ranging (radar) sensor at a first position;

transforming coordinates of the first point cloud based on a second position of the radar sensor;

generating a second point cloud of the stationary object using the radar sensor at the second position,

generating an updated point cloud of the stationary object based on the second point cloud and the transformed first point cloud; and

generating a grid map based on an occupancy probability for each grid of the updated point cloud,

wherein the generating of the updated point cloud comprises:

transforming the coordinates of the first point cloud generated in the first position, based on the stationary object;

generating the second point cloud corresponding to the stationary object, based on detection information of the stationary object generated in the second position,

wherein the detection information includes relative position information of the stationary object with respect to the second position of the radar sensor; and

generating the updated point cloud corresponding to the second position based on the transformed first point cloud and the second point cloud,

wherein the generating of the first point cloud comprises generating the first point cloud in response to determining that a difference between an absolute value of a relative velocity of the stationary object and an absolute value of the velocity of the radar sensor is less than or equal to a preset threshold value.

20 . The method of claim 19 , wherein the radar sensor is a moving object.

21 . The method of claim 19 , wherein the generating of the first point cloud comprises generating the first point cloud in response to determining that either one or both of:

a height of the stationary object is less than or equal to a preset first threshold value; and

an angle between a direction of movement of the radar sensor and a direction from the radar sensor to the stationary object is less than or equal to a preset second threshold value.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (14)