Visual Odometry Buffer Tuning for Mixed-Reality Pose Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mixed-reality systems face challenges in accurately determining the pose of a device relative to its environment, particularly when data conflicts arise due to movement, affecting the display of holograms.
Innovation Solution
The system determines a level of depth uncertainty for a 3D feature represented in images, ejects an estimated state from a buffer to modify available data, and tunes representative measurements to preserve uncertainty, ensuring robust depth computation and pose estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple estimated states are stored in a buffer to improve pose estimation accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the uncertainty values associated with each estimated state based on temporal characteristics. By modifying the uncertainty parameter rather than the buffer structure itself, the system maintains multiple states for improved accuracy while managing complexity through adaptive parameter tuning rather than structural complexity
Solution Approach 2:
The patent implements preliminary action by pre-calculating and storing uncertainty values for each estimated state before they are needed for final pose determination. This advance preparation allows the system to efficiently manage multiple states in the buffer without increasing real-time processing complexity, as the uncertainty characteristics are determined in advance
2Loss of substance
If estimated states are ejected from the buffer to reduce memory usage, then loss of substance is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent applies local quality by differentiating between various estimated states based on their specific uncertainty characteristics and temporal relevance. Instead of uniformly managing all states, the system selectively retains or ejects states based on their local quality metrics (uncertainty values and temporal properties), allowing memory-efficient management while preserving the most valuable states for accurate depth uncertainty calculation
Solution Approach 2:
The patent uses parameter changes by dynamically adjusting the uncertainty values of remaining states when states are ejected from the buffer. This compensation through parameter modification ensures that the overall measurement precision is maintained even as the buffer size is reduced, addressing the trade-off between memory usage and accuracy
3Reliability
If representative measurements are tuned to preserve uncertainty after state ejection, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent implements feedback by continuously monitoring the depth uncertainty levels after state ejection and using this information to tune the representative measurements of remaining states. This feedback loop ensures reliability is maintained by automatically adjusting measurements based on the actual uncertainty introduced by buffer management, rather than using complex predetermined tuning algorithms
Solution Approach 2:
The patent applies self-service by enabling the measurement system to automatically adjust and tune its own representative measurements based on observed uncertainty levels. The system serves itself by detecting when uncertainty increases due to state ejection and autonomously compensating through measurement tuning, reducing the need for external complex control mechanisms
Data Source
AI summary
Techniques for identifying a 3D feature represented in images in a buffer are disclosed. The images comprise a first and second image, and the images provide data that is usable to determine a depth for the 3D feature. A 2D feature point that is an observation of the 3D feature is identified. A pseudo measurement is computed for the 2D feature point. This measurement includes a depth and an uncertainty value. A first joint uncertainty is determined for the depth using the pseudo measurement. The first image is ejected from the buffer, resulting in a reduction to an amount of the data that is usable to determine the depth. The pseudo measurement is tuned until a resulting second joint uncertainty, which is based on the pseudo measurement, is within a threshold level of similarity to the first joint uncertainty.