Visual-Inertial Pose Estimation With Feature Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation technologies, such as GPS-based systems, face challenges in urban and indoor environments due to poor signal reception and inability to detect obstacles, and suffer from sensor errors and limited precision in close-quarter navigation.
Innovation Solution
A vision-aided inertial navigation system that processes visual data from cameras and inertial sensor data using an Extended Kalman filter to estimate pose and localization information, with computational complexity linearly related to the number of tracked features, allowing for high-precision pose estimation in large-scale environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS-based navigation is used, then global positioning is achieved, but signal reception fails in urban and indoor environments
Solution Approach 1:
The patent introduces visual features and inertial sensors as intermediary elements to bridge the gap when GPS signals are unavailable. The system uses camera-captured visual features and inertial measurement unit data as mediators to continue navigation operations in urban canyons and indoor environments where satellite signals are blocked or degraded.
Solution Approach 2:
The system transitions from relying solely on GPS satellite signal parameters to using visual feature parameters and inertial sensor parameters. By changing the fundamental measurement parameters from radio frequency signals to visual and inertial data, the navigation system maintains operation across diverse environments including those with poor GPS reception.
2Measurement precision
If traditional navigation systems are used, then basic positioning is provided, but close-quarter navigation and obstacle detection fail
Solution Approach 1:
The navigation system segments the navigation function into multiple independent components: global positioning (when available), visual feature tracking, and inertial navigation. This segmentation allows each component to specialize in specific tasks, with the visual-inertial system providing detailed close-quarter navigation and obstacle detection while GPS provides broad positional context when available.
Solution Approach 2:
The system adds the visual dimension to traditional navigation by incorporating camera data for detecting obstacles and environmental features. This dimensional addition transforms the navigation capability from three-dimensional spatial positioning to four-dimensional navigation that includes visual obstacle awareness and environmental context.
3Measurement precision
If more visual features are tracked, then pose estimation accuracy improves, but computational complexity increases
Solution Approach 1:
The patent replaces traditional computationally intensive visual odometry algorithms with an inertial navigation system augmented by visual features. The inertial measurement unit provides direct acceleration and orientation data that requires minimal processing, substituting the complex mechanical vision processing with simpler inertial sensing while maintaining or improving accuracy.
Solution Approach 2:
The visual features serve multiple functions simultaneously: they provide positioning references, enable pose estimation, and can be used for navigation guidance. This multi-functionality means that tracking visual features does not add separate computational loads for each function, as the same feature data serves multiple navigation purposes, reducing overall system complexity.
Data Source
AI summary
Localization and navigation systems and techniques are described. An electronic device comprises a processor configured to maintain a state vector storing estimates for a position of the electronic device at poses along a trajectory within an environment along with estimates for positions for one or more features within the environment. The processor computes, from the image data, one or more constraints based on features observed from multiple poses of the electronic device along the trajectory, and computes updated state estimates for the position of the electronic device in accordance with the motion data and the one or more computed constraints without computing updated state estimates for the features for which the one or more constraints were computed.