Robot pose recognition method, device and equipment, readable storage medium
By utilizing the robot's external environment information and motion state for feature matching in SLAM localization technology and eliminating erroneous matching points, the problem of inaccurate pose recognition in complex environments is solved, thereby improving the accuracy and robustness of robot localization.
CN116412827BActive Publication Date: 2026-06-19CHINA UNITED NETWORK COMM GRP CO LTD +2
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA UNITED NETWORK COMM GRP CO LTD
- Filing Date
- 2021-12-31
- Publication Date
- 2026-06-19
AI Technical Summary
Technical Problem
Existing SLAM localization technology does not yield ideal pose recognition results in complex environments, such as environmental and lighting changes, and interference from multiple dynamic objects.
Method used
By acquiring information about the robot's external environment and motion state, feature matching is performed using a relative pose transformation matrix to eliminate incorrectly matched feature point combinations, thereby improving the accuracy of pose recognition.
Benefits of technology
It improves the accuracy and robustness of robot pose recognition, reduces the accumulation of positioning errors, and enhances the robot's navigation capabilities in complex environments.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116412827B_ABST
Abstract
This disclosure provides a robot pose recognition method, apparatus, device, and readable storage medium, relating to pose recognition technology. The method includes: acquiring external environment information and motion state of the robot at various times; the external environment information includes external environment features; acquiring the robot's pose at the previous time step, and determining the relative pose transformation matrix between the previous and current time steps based on the current motion state and the previous pose; performing feature matching on the external environment features at the current and previous times steps to obtain a feature matching result; processing the feature matching result based on the relative pose transformation matrix to obtain a processed feature matching result; and determining the current pose based on the processed feature matching result. The solution provided in this disclosure can utilize the robot's current motion state to assist in determining the current pose, thereby improving the accuracy of pose recognition results.
Need to check novelty before this filing date? Find Prior Art