Kalman Filter Multipath Modeling for Deep GPS Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deeply integrated inertial navigation systems and GPS face challenges in maintaining accurate navigation solutions due to kinematic inconsistencies caused by GPS signal multipath, which can lead to erroneous estimates as the Kalman filter loses track under multipath conditions.
Innovation Solution
Incorporating a multipath error model into the Kalman filter, using either a random walk or a Gauss-Markov process to model the multipath states, allowing the system to continue providing accurate estimates by accounting for multipath errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the Kalman filter closely matches GPS signal dynamics with vehicle dynamics, then the navigation accuracy is improved, but the system becomes vulnerable to multipath errors causing kinematic inconsistencies
Solution Approach 1:
The error model is segmented into multiple components: a long-term navigation error state and a short-term multipath error state. This segmentation allows the Kalman filter to separately model and handle different types of errors, maintaining navigation accuracy while preventing multipath from causing track loss.
Solution Approach 2:
A short-term error state is introduced as an intermediary between the GPS measurements and the navigation solution. This intermediary absorbs the multipath errors, preventing them from directly affecting the navigation accuracy while allowing the long-term error model to maintain track stability.
2Device complexity
If the Kalman filter uses standard deep integration without multipath modeling, then the system complexity is kept low, but erroneous navigation solutions occur under multipath conditions
Solution Approach 1:
The error model is made dynamic by introducing a time-varying short-term error state that adapts to multipath conditions. This dynamic component allows the filter to respond to changing multipath environments without requiring a complete restructuring of the filter architecture.
Solution Approach 2:
The error covariance matrix parameters are adjusted to account for multipath effects. By changing the parameters of the existing Kalman filter structure rather than redesigning the entire system, the solution maintains relative simplicity while improving reliability under multipath conditions.
3Productivity
If the Kalman filter ignores multipath errors, then the computational load is reduced, but the navigation solution becomes erroneous under multipath conditions
Solution Approach 1:
Instead of modeling all possible error sources, the solution applies partial action by focusing only on the short-term multipath error component. This selective approach improves navigation accuracy under multipath conditions without the computational burden of modeling every possible error source in detail.
Data Source
AI summary
A state is added to a Kalman filter to model GPS multipath errors. The multipath states may be modeled as either a random walk model or a Gauss-Markov process. The choice of the model depends on the characteristics of the multi-path error and the GPS receiver. Adding this state to the Kalman filter to model multipath improves the navigation system's robustness when operating as a deeply integrated system when multipath is present.