US 12,404,158 B2
Local map updating method and apparatus, and electronic device
Fan Zheng, Shenzhen (CN); Cheng Zhang, Shenzhen (CN); Yujie Lu, Shenzhen (CN); Feng Zhang, Shenzhen (CN); and Jiang Liu, Shenzhen (CN)
Assigned to VisionNav Robotics USA Inc., Acworth, GA (US)
Filed by VisionNav Robotics USA Inc., Acworth, GA (US)
Filed on Jan. 31, 2025, as Appl. No. 19/042,556.
Claims priority of application No. 202410174186.6 (CN), filed on Feb. 7, 2024.
Prior Publication US 2025/0250151 A1, Aug. 7, 2025
Int. Cl. B66F 9/075 (2006.01); B66F 9/06 (2006.01); G01S 17/89 (2020.01)
CPC B66F 9/0755 (2013.01) [B66F 9/063 (2013.01); G01S 17/89 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A local map updating method, performed by an intelligent robot, wherein the intelligent robot comprises a point cloud collector, a processor and a memory, and the method comprises:
acquiring, by the point cloud collector, key-frame point cloud data, wherein the key-frame point cloud data comprises point cloud coordinate points;
matching, by the processor, the key-frame point cloud data with a prior map stored in the memory to determine target voxels, in the prior map, corresponding to the point cloud coordinate points; wherein the prior map comprises voxels and a respective point cloud probability distribution corresponding to each of the voxels; each of the target voxels comprises a first map coordinate point and a second map coordinate point, wherein the first map coordinate point is obtained by converting the point cloud coordinate point corresponding to the target voxel, and the second map coordinate point is a coordinate point comprised in the target voxel in the prior map;
computing, by the processor, a first probability distribution parameter value corresponding to a first target voxel according to the first map coordinate point comprised in the first target voxel, wherein the first target voxel is any one of the target voxels;
fusing, by the processor, the first probability distribution parameter value with a second probability distribution parameter value comprised in the point cloud probability distribution corresponding to the first target voxel, to obtain a target probability distribution parameter value, wherein the second probability distribution parameter value is obtained based on the second map coordinate point comprised in the first target voxel;
updating, by the processor, the point cloud probability distribution corresponding to the first target voxel according to the target probability distribution parameter value; and
positioning, by the processor, the intelligent robot based on the updated point cloud probability distribution.