GPS Velocity Integration in Visual-Inertial Odometry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Visual-inertial odometry (VIO) systems in autonomous vehicles face challenges such as unbounded error growth due to measurement noise and limited visual features, leading to inaccurate positioning and orientation, especially in scenarios like highway driving where inertial sensor measurements are not sufficiently excited and visual features are sparse.
Innovation Solution
The integration of GPS velocity into the VIO system using an extended Kalman filter to refine bias, scale, and misalignment estimates, and provide pose information, which limits error growth and maintains accurate positioning and orientation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If visual-inertial odometry is used for positioning and orientation, then the system can operate without GPS, but error grows unbounded due to measurement noise and limited visual features
Solution Approach 1:
The system uses GPS velocity measurements as feedback to correct drift in the visual-inertial odometry system. The GPS velocity is integrated into the VIO state estimator, providing periodic corrections that prevent unbounded error growth while maintaining the system's ability to operate without GPS for positioning and orientation.
Solution Approach 2:
The patent introduces GPS velocity as an intermediary measurement that bridges the visual-inertial system and the global reference frame. This intermediary provides absolute position information that constrains the drift inherent in relative visual-inertial odometry, allowing the system to maintain accuracy over extended periods.
2Measurement precision
If GPS velocity is integrated into VIO to reduce drift, then positioning accuracy improves, but system complexity increases
Solution Approach 1:
The system merges GPS velocity measurements with visual-inertial odometry data in a unified state estimation framework. By combining these sensors and their measurements into a single integrated system using an extended Kalman filter, the patent achieves improved velocity and position accuracy without requiring separate processing pipelines, thus managing complexity through consolidation.
3Reliability
If multiple sensors are integrated for sensor fusion, then robustness improves, but computational resources required increase
Solution Approach 1:
The system applies partial sensor fusion by integrating only the essential GPS velocity measurements with visual-inertial odometry, rather than fusing all available sensor data. This selective approach provides sufficient robustness improvement while avoiding the excessive computational burden of complete sensor fusion, thus optimizing the trade-off between reliability and energy consumption.
Data Source
AI summary
A method performed by an electronic device is described. The method includes determining a predicted velocity relative to Earth corresponding to a first epoch using a camera and an inertial measurement unit (IMU). The method also includes determining, using a Global Positioning System (GPS) receiver, a GPS velocity relative to Earth. The method further includes determining a difference vector between the predicted velocity and the GPS velocity. The method additionally includes refining a bias estimate and a scale factor estimate of IMU measurements proportional to the difference vector. The method also includes refining a misalignment estimate between the camera and the IMU based on the difference vector. The method further includes providing pose information based on the refined bias estimate, the refined scale factor, and the refined misalignment estimate.