Robot Pose Correction Using Depth Mapping and VIO Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot pose estimation methods, particularly in monitoring systems, suffer from measurement drift, leading to inaccurate and unreliable pose calculations over time, which can result in incorrect robot operations and potential collisions or damage.
Innovation Solution
The method involves obtaining and comparing images from different locations to detect feature points, generating depth data using epipolar processes and scale factors, and updating the robot's pose estimation to prevent drift, utilizing a combination of visual inertial odometry and depth data for accurate navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If visual inertial odometry is used to track robot pose over time, then the robot can maintain continuous pose estimation, but measurement drift causes the pose estimation to become less accurate and reliable over time
Solution Approach 1:
The system implements feedback by continuously comparing the robot's predicted pose (from VIO) with the actual pose (estimated from depth data and feature point comparisons). When discrepancies are detected, the system updates the robot's actual pose using the estimate, thereby correcting drift and improving long-term reliability while maintaining measurement precision.
Solution Approach 2:
The system performs preliminary action by capturing and storing images with identifying information (depth data, feature points, location data) during the mapping phase. These pre-captured images serve as reference data that enable accurate pose estimation during navigation without requiring continuous depth sensing, thus preventing drift accumulation.
2Measurement precision
If depth information is captured and stored for every location, then accurate pose estimation can be performed, but the data storage requirements and system complexity increase
Solution Approach 1:
The system performs preliminary action by capturing depth information, feature points, and location data during an initial mapping phase when the robot first visits locations. This pre-captured reference data is stored for later use, enabling accurate pose estimation during navigation without requiring continuous depth sensing hardware or processing, thus reducing system complexity while maintaining precision.
Solution Approach 2:
The system creates a copy of the environment in the form of stored images with embedded depth data, feature points, and location information. This digital copy serves as a reference map that can be repeatedly queried for pose estimation without requiring physical presence or active sensing at each location, reducing both hardware complexity and data storage requirements compared to capturing continuous depth information.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy of robot pose correction and navigation, reducing the likelihood of collisions, property damage, and personal injury by providing more precise mapping and destination finding capabilities.
Implementation Method 1
a process may use a change of position from one location to another indicated by visual inertial odometry (VIO) measurement or other measurement processes to generate a scale factor
Implementation Method 2
generating the depth data for the feature points uses an epipolar process and a scale factor
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for estimating a robot pose. One of the methods includes the actions of obtaining two or more images captured at two or more locations on a property; detecting feature points at positions within two or more images including first feature points in the first image and second feature points in the second image; comparing the positions of the first feature points in the first image to positions of the second feature points in the second image; obtaining data indicating the two or more locations on the property; comparing the two or more locations; and generating depth data for the feature points for use by a robot navigating the property.