Method and system for recognizing road surface
A road surface recognition method may include extracting first point cloud data in a front road surface region of a vehicle from point cloud data generated by LiDAR of the vehicle, generating a point cloud map by use of a local map and the first point cloud data, determining a predicted driving route of the vehicle based on motion-related data of the vehicle; and generating a road surface height profile of the predicted driving route based on the point cloud map and the predicted driving route.
1. A method for controlling driving of a vehicle, the method comprising:
performing, by a processor in the vehicle, time synchronization of point cloud data generated by use of Light Detection and Ranging (LiDAR) of the vehicle and motion-related data of the vehicle based on a time stamp of the point cloud data to obtain time-synchronized motion-related data of the vehicle;
determining, by the processor, an initial location of the vehicle based on the point cloud data and the time-synchronized motion-related data of the vehicle;
extracting, by the processor, first point cloud data from the point cloud data, the first point cloud data corresponding to a predetermined region with respect to the vehicle;
generating, by the processor, a point cloud map by use of a local map and the first point cloud data, the local map including map data on the predetermined region;
determining, by the processor, a predicted driving route of the vehicle based on the motion-related data of the vehicle;
generating, by the processor, a road surface height profile of the predicted driving route based on the point cloud map and the predicted driving route; and
controlling, by a controller in the vehicle, the driving of the vehicle based on road surface height profile.
2. The method of claim 1 , wherein the generating of the point cloud map includes:
determining transformation information for registration between the first point cloud data and the local map;
transforming the first point cloud data into second point cloud data based on the transformation information; and
generating the point cloud map by merging the second point cloud data into the local map.
3. The method of claim 2 ,
wherein the transformation information includes rotation information related to a roll and rotation information related to a pitch of the vehicle, and Z-axis movement information of the first point cloud data.
4. The method of claim 3 ,
wherein a process of the extracting of the first point cloud data, the generating of the point cloud map, the determining of the predicted driving route, and the generating of the road surface height profile is iterated per time instant, and the transformation information is determined based on a point cloud map generated at a previous time instant.
5. The method of claim 2 , wherein the determining of transformation information includes:
generating a plurality of pieces of point data corresponding to a road surface in the local map; and
generating a corresponding relationship between the plurality of pieces of point data and the first point cloud data.
6. The method of claim 5 ,
wherein the transforming of the first point cloud data into the second point cloud data is performed when a root mean square error (RMSE) between the plurality of pieces of point data and the first point cloud data in the corresponding relationship is less than a predetermined threshold.
7. The method of claim 1 ,
wherein the motion-related data of the vehicle includes a yaw rate of the vehicle, a speed of the vehicle, and a wheel angle of the vehicle.
8. The method of claim 7 ,
wherein the yaw rate is a value obtained by subtracting a pre-calculated bias error value from a yaw rate value of the vehicle obtained during movement of the vehicle, and wherein the pre-calculated bias error value is an average value of yaw rate values of the vehicle obtained while the vehicle is in a stationary state.
9. The method of claim 2 , wherein the generating of the road surface height profile includes:
sampling a plurality of points from the point cloud map along the predicted driving route;
determining an average value of height values of a plurality of points of the second point data within a first range of a predesignated size around each of locations of the plurality of points; and
generating the average value as the road surface height profile of the predicted driving route.
10. The method of claim 9 , further including:
estimating height information of a portion of the predicted driving route in which the road surface height profile is not generated through an interpolation.
11. The method of claim 9 , further including removing noise included in the road surface height profile.
12. A system for controlling driving of a vehicle, the system comprising:
an interface configured for data communication;
a processor electrically connected to the interface; and
a controller configured to control the driving of the vehicle,
wherein the processor is configured to perform:
receiving point cloud data from Light Detection and Ranging (LiDAR) of the vehicle and motion-related data of the vehicle from a sensing device of the vehicle through the interface;
performing time synchronization of the point cloud data and the motion-related data of the vehicle based on a time stamp of the point cloud data to obtain time-synchronized motion-related data of the vehicle;
determining an initial location of the vehicle based on the point cloud data and the time-synchronized motion-related data of the vehicle;
extracting first point cloud data corresponding to a predetermined region with respect to the vehicle from the point cloud data;
generating a point cloud map by use of a local map and the first point cloud data, the local map including map data on the predetermined region;
determining a predicted driving route of the vehicle based on motion-related data of the vehicle; and
generating a road surface height profile of the predicted driving route based on the point cloud map and the predicted driving route, and
wherein the controller is configured to control the driving of the vehicle based on the road surface height profile.
13. The system of claim 12 , wherein the generating of the point cloud map includes:
determining transformation information for registration between the first point cloud data and the local map;
transforming the first point cloud data into second point cloud data based on the transformation information; and
generates the point cloud map by merging the second point cloud data into the local map.
14. The system of claim 13 , wherein the determining of transformation information includes:
generating a plurality of pieces of point data corresponding to a road surface in the local map; and
generating a corresponding relationship between the plurality of pieces of point data and the first point cloud data.
15. The system of claim 14 ,
wherein the transforming of the first point cloud data into the second point cloud data is performed when a root mean square error (RMSE) between the plurality of pieces of point data and the first point cloud data in the corresponding relationship is less than a predetermined threshold.
16. The system of claim 12 , wherein the generating of the road surface height profile includes:
sampling a plurality of points from the point cloud map along the predicted driving route;
determining an average value of height values of a plurality of points of the second point data within a first range of a predesignated size around each of locations of the plurality of points; and
generating the average value as the road surface height profile of the predicted driving route.
17. The system of claim 16 , wherein the processor is configured to further perform:
estimating height information of a portion of the predicted driving route in which the road surface height profile is not generated through an interpolation.
18. The system of claim 17 , wherein the processor is configured to further perform removing noise included in the road surface height profile.