Modified Kalman Filter for Satellite Attitude Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current satellite communication systems face computational intensity issues with traditional 8-state Kalman filters for attitude correction, particularly when secondary attitude sensors provide irregular and infrequent measurements, leading to inefficient processor throughput and increased error propagation.
Innovation Solution
A modified Kalman filter approach that partitions 8x8 matrices into smaller block matrices (3x3, 3x2, 2x3, and 2x2) reduces unnecessary calculations and performs covariance propagation in a single step using analytically derived equations, leveraging body-reference frame equations for more efficient attitude error corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a conventional 8-state Kalman filter is used to combine attitude measurements from inertial sensors and secondary APS, then attitude measurement accuracy is improved, but processor throughput requirements increase beyond available flight computer capabilities
Solution Approach 1:
The patent segments the 8x8 Kalman filter matrices into smaller block sub-matrices (3x3, 3x2, 2x3, and 2x2 blocks). This segmentation reduces the computational complexity of matrix operations while maintaining the filter's ability to produce accurate attitude measurements. The divided matrices can be processed more efficiently by flight computers with limited processing power.
Solution Approach 2:
The patent pre-calculates certain matrix components and stores them in lookup tables before flight operations. By performing these calculations in advance on the ground, the flight computer only needs to retrieve and combine pre-computed values during operation, significantly reducing real-time processing requirements while maintaining measurement accuracy.
2Reliability
If the Kalman filter is executed multiple times to handle irregular and infrequent APS measurements, then attitude error correction is maintained, but processor throughput is wasted without new information available
Solution Approach 1:
The patent implements a dynamic execution strategy where the Kalman filter is only executed when new APS measurements are actually available. The system dynamically adjusts its processing schedule based on the arrival of new data, avoiding unnecessary executions when no new information is present. This maintains reliable error correction while eliminating wasted processor throughput.
3Measurement precision
If full 8x8 matrix calculations are performed in real-time, then complete attitude state estimation is achieved, but computational intensity exceeds flight computer processing capabilities
Solution Approach 1:
The patent divides the complex 8x8 matrix operations into smaller block sub-matrices (3x3, 3x2, 2x3, and 2x2 blocks). This segmentation reduces the computational complexity of each individual operation while preserving the overall estimation accuracy through proper combination of the block results.
Solution Approach 2:
The patent pre-calculates and stores certain matrix components in lookup tables before flight operations begin. This preliminary action transfers computational burden from the flight computer to ground-based preprocessing, reducing real-time computational complexity while maintaining complete attitude state estimation accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods, systems, and computer-readable media are described herein for using a modified Kalman filter to generate attitude error corrections. Attitude measurements are received from primary and secondary attitude sensors of a satellite or other spacecraft. Attitude error correction values for the attitude measurements from the primary attitude sensors are calculated based on the attitude measurements from the secondary attitude sensors using expanded equations derived for a subset of a plurality of block sub-matrices partitioned from the matrices of a Kalman filter, with the remaining of the plurality of block sub-matrices being pre-calculated and programmed into a flight computer of the spacecraft. The propagation of covariance is accomplished via a single step execution of the method irrespective of the secondary attitude sensor measurement period.