Vehicle Navigation Filtering for Stable Multi-Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation methods using linear and extended Kalman filters struggle to effectively combine inertial measurements with GPS and odometer-based measurements, leading to sub-optimal results and instability due to feedback loops that depend on the coordinate system used.
Innovation Solution
A hybrid navigation method utilizing an invariant Kalman filter that combines inertial, GPS, and odometer measurements by incorporating a system of error variables and virtual observations to suppress feedback, allowing for precise navigation by integrating all available measurements in a single filter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If an extended Kalman filter is used to handle non-linear state estimation, then the filter can operate with non-linear models, but feedback loops are introduced that cause instability and depend on the coordinate system used
Solution Approach 1:
The patent transforms the error variables from standard coordinates to a parameterized form that eliminates the feedback dependency. By changing the mathematical representation of errors (from eT, eV, eX to a parameterized form), the filter achieves stability while maintaining non-linear model handling capability.
Solution Approach 2:
The patent introduces an intermediate transformation step that mediates between the non-linear model requirements and the stability requirement. The error parameterization acts as an intermediary that decouples the feedback loop from the coordinate system dependencies.
2Ease of operation
If separate filters are used for position sensors and odometry sensors, then each filter can independently process its measurements, but the measurements cannot be combined and the result is sub-optimal
Solution Approach 1:
The patent merges the separate position and odometry filters into a single unified filter that processes all measurements simultaneously. This combination allows optimal integration of GPS position data and odometer velocity data, achieving superior navigation accuracy compared to independent filters.
3Stability of the object's composition
If traditional error variables are used in invariant filters, then the feedback can be suppressed for specific sensor combinations, but it is not possible to combine inertial measurements, position measurements and odometer-based velocity measurements in a single filter
Solution Approach 1:
The patent creates a universal error parameterization that works for multiple sensor combinations simultaneously. The same parameterized error variables enable the filter to handle inertial measurements, position measurements, and odometer-based velocity measurements in a single unified framework, providing multi-functionality.
Data Source
AI summary
Method, navigation device and computer program product for assisting with the navigation of a vehicle equipped with a navigation device, comprising the following steps: acquiring (201) a priori values of kinematic variables of the navigation device, determining (202) respective current values of kinematic variables of the navigation device and a current uncertainty matrix representative of an uncertainty of the respective current values of the kinematic variables, based on respective previous values of the kinematic variables, determining (203) a correction, updating (204) the respective current values of the kinematic variables based on the correction and on the current uncertainty matrix.