Compressed-Domain Star Tracking for Low-Memory Attitude Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current star-tracking systems require large amounts of memory and processing power to accurately determine the attitude of an object based on starfield images, which can be costly and power-intensive, especially in radiation-hardened environments.
Innovation Solution
The implementation of compressive sampling techniques to capture and process starfield image data in a compressed domain, reducing memory and processing requirements while maintaining accuracy, by identifying star positions and estimating attitude using fewer measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional star-tracking systems acquire and store large numbers of starfield images in memory, then measurement precision of attitude is improved, but memory requirements and processing power increase prohibitively
Solution Approach 1:
The patent extracts only the essential information from starfield images by applying compressive sampling techniques that capture the most significant features while discarding redundant data. This allows attitude estimation to be performed on compressed representations rather than full-resolution images, dramatically reducing memory requirements while maintaining measurement precision.
Solution Approach 2:
The system changes the parameter representation of starfield data by transforming images into a compressed domain using specific sampling matrices and reconstruction algorithms. This parameter transformation enables the same attitude information to be extracted from fewer, more efficiently encoded measurements.
2Measurement precision
If traditional systems process and average multiple starfield images to improve quality, then measurement precision is improved, but processing power and energy consumption increase
Solution Approach 1:
The patent applies compressive sampling as a preliminary action before full image processing. By pre-processing the starfield data through compression and noise filtering in the compressed domain, the system eliminates the need for computationally intensive averaging of multiple full-resolution images, significantly reducing energy consumption while maintaining noise reduction quality.
3Reliability
If radiation-hardened memory and circuits are used to protect against computational errors, then reliability is improved, but cost and processing limitations increase
Solution Approach 1:
The patent applies partial processing by performing attitude estimation directly on compressed data without fully reconstructing the original images. This partial action approach reduces the computational burden on rad-hard systems, allowing them to operate within their limited processing capabilities while maintaining reliability through error-correcting codes and robust algorithms designed for compressed sensing.
4Measurement precision
If full-resolution starfield images are stored and processed, then measurement precision is improved, but input/output bandwidth and memory requirements increase
Solution Approach 1:
The patent creates a compressed copy of the starfield data that preserves the essential information needed for attitude estimation. This compressed representation serves as a surrogate for the full-resolution images, enabling accurate star position determination with dramatically reduced data storage volume and I/O bandwidth requirements.
Data Source
AI summary
In general, in one embodiment, a starfield image as seen by an object is analyzed. Compressive samples are taken of the starfield image and, in the compressed domain, processed to remove noise. Stars in the starfield image are identified and used to determine an attitude of the object.