机器人位姿估计方法、机器人点云地图构建方法及装置
By acquiring the ground and elevation orientation angles from laser point cloud data, a Manhattan descriptor is generated. Combined with ground and elevation constraints, distributed pose estimation is performed, which solves the problem of low accuracy in robot pose estimation and point cloud map in indoor environments, and achieves more efficient global point cloud map construction.
CN116973883BActive Publication Date: 2026-07-17CAPITAL NORMAL UNIVERSITY
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CAPITAL NORMAL UNIVERSITY
- Filing Date
- 2023-07-26
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In indoor environments, robots struggle to determine their pose using visual features, resulting in poor pose estimation accuracy and low point cloud map precision.
Method used
By acquiring the ground and elevation principal orientation angles from the laser point cloud data, updating the reference orientation angles, generating the Manhattan descriptor, and combining the ground and elevation constraints to perform distributed pose estimation, a robot point cloud map is constructed.
Benefits of technology
It improves the pose estimation accuracy and point cloud map accuracy of robots in indoor environments, and enhances the efficiency and accuracy of global point cloud map acquisition.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116973883B_ABST
Abstract
本申请提供了一种机器人位姿估计方法、机器人点云地图构建方法及装置，在所谓机器人位姿估计方法中，提出通过机器人针对当前环境获得的激光点云数据中地面点云数据及立面点云数据，确定当前环境对应的地面主方向角、立面主方向角、地面基准主方向角以及立面基准主方向角，以及提出使用上述参数生成曼哈顿描述符，如此，通过此曼哈顿描述符，构建机器人在当前环境对应的地面约束以及立面约束，以及基于曼哈顿描述符在地面约束及立面约束下，对机器人进行分布式位姿估计，从而有效解决了现有技术尚且存在的约束不足的问题，提高了所获得的机器人在当前环境对应的运动位姿的准确性。
Need to check novelty before this filing date? Find Prior Art