Variance Measurement Using High-Pass Filter for Navigation Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for data fusion in navigation systems, such as those using Kalman filters, rely on estimating variance from a model, which can be inflexible and resource-intensive, especially when measurement signal variance is unknown or variable.
Innovation Solution
A method that determines noise variance in measurement signals using a high-pass filter, independent of the Kalman filter model, allowing for adaptive filter performance and reduced resource usage, by filtering the measurement signal and then using a low-pass filter to calculate variance based on the filtered signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive Kalman filters are used to adjust variances for measurement signals with unknown or variable variance, then filter performance is improved, but complex matrix operations are required to estimate variances
Solution Approach 1:
The patent extracts the variance measurement function from the Kalman filter itself and implements it as a separate, independent measurement process. By using a high-pass filter to suppress the DC component and directly measuring the variance of the filtered signal, the system obtains measurement variance without requiring complex matrix operations within the Kalman filter framework. This separation simplifies the overall system while maintaining adaptive performance.
Solution Approach 2:
The measurement signal itself serves to determine its own variance through the high-pass filtering process. Rather than requiring external estimation or complex internal Kalman filter operations, the signal's own fluctuations after DC suppression provide direct variance information. This self-service approach eliminates the need for complex matrix operations while maintaining reliability.
2Adaptability or versatility
If variance estimation is performed within the Kalman filter, then model-based estimation is achieved, but adaptation to changes in measurement signal is slow due to inherent inertia
Solution Approach 1:
The patent segments the variance determination process from the Kalman filter operation. By implementing variance measurement as an independent parallel process using high-pass filtering, the system eliminates the sequential dependency and inherent inertia of model-based estimation. This segmentation allows variance information to be obtained immediately from the measurement signal without waiting for Kalman filter iterations, enabling rapid adaptation to changing conditions.
3Measurement precision
If model-based variance estimation is used in Kalman filters, then statistical filtering is achieved, but resource-intensive operations are required
Solution Approach 1:
The patent replaces the complex mechanical/computational system of model-based variance estimation with a simpler signal processing approach. By substituting the resource-intensive matrix operations with straightforward high-pass filtering and variance calculation, the system achieves the same measurement precision with significantly reduced hardware resource expenditure and energy consumption.
Data Source
Figure 1
Figure 2
Figure 3a
AI summary
The invention relates to a method (500) for measuring the variance in a measurement signal, comprising the following steps: filtering (501) the measurement signal by means of a high-pass filter (HP) in order to obtain a filtered measurement signal; determining (502) the variance by using the filtered measurement signal.