Visual Inertial Odometry Heading Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for estimating the position and orientation of electronic devices, such as smartphones and wearables, face challenges in accuracy and reliability due to biases in magnetometer readings and drift errors in visual inertial odometry, especially when used in augmented reality applications.
Innovation Solution
The system employs a combination of GNSS receivers, accelerometers, gyroscopes, image sensors, and machine learning filters to correlate image data with mapping information, adjust magnetometer outputs, and fuse visual inertial odometry with GNSS data to improve heading estimates and reduce drift errors, using architectures like the visual inertial odometry module and extended Kalman filters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If magnetometer readings are used for heading estimation, then device orientation can be determined, but biases in magnetometer readings reduce accuracy
Solution Approach 1:
The patent combines magnetometer readings with visual inertial odometry (VIO) data to estimate device heading. By merging these two independent measurement sources, the system compensates for magnetometer biases through VIO-based heading corrections, thereby improving overall heading estimation accuracy while mitigating the unreliability of magnetometer readings alone.
Solution Approach 2:
The patent introduces visual inertial odometry as an intermediary system that mediates between magnetometer readings and final heading estimation. The VIO system processes image data and device motion information to generate independent heading estimates, which then serve to correct and stabilize the magnetometer-based heading, reducing the direct impact of magnetometer biases.
2Measurement precision
If visual inertial odometry is used for position estimation, then device position can be tracked, but drift errors accumulate over time
Solution Approach 1:
The patent implements feedback mechanisms where GNSS position data is used to correct and reset accumulated drift errors in visual inertial odometry. When GNSS measurements are available, the system compares the VIO-derived position with GNSS position and applies corrections to eliminate drift, thereby maintaining long-term position estimation accuracy.
Solution Approach 2:
The system performs preliminary drift compensation by periodically resetting the VIO position estimate using GNSS measurements before drift errors can significantly accumulate. This preliminary correction action prevents large drift errors from developing, maintaining position accuracy over extended periods.
3Measurement precision
If multiple sensors are combined for positioning, then accuracy improves, but system complexity increases
Solution Approach 1:
The patent employs an extended Kalman filter that serves multiple functions: it fuses data from magnetometers, visual inertial odometry, and GNSS receivers; it performs state estimation for position, velocity, and orientation; and it handles drift compensation and bias correction. This multi-functional approach consolidates complex sensor fusion operations into a unified algorithmic framework, improving positioning accuracy while managing system complexity through functional integration.
Data Source
AI summary
A device implementing a system for estimating device location includes at least one processor configured to receive a first estimated position of the device at a first time. The at least one processor is further configured to capture, using an image sensor of the device, images during a time period defined by the first time and a second time, and determine, based on the images, a second estimated position of the device, the second estimated position being relative to the first estimated position. The at least one processor is further configured to receive a third estimated position of the device at the second time, and estimate a location of the device based on the second estimated position and the third estimated position.