Triaxial Sensor Calibration via Spatial Distribution Indicator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing calibration methods for three-axis sensors, such as magnetometers, accelerometers, and gyrometers, face challenges in achieving precise calibration due to poorly distributed measurements, which lead to errors from intrinsic and external sources like scale factor, misalignment, and soft iron interference, requiring significant computing resources for effective correction.
Innovation Solution
A computer-implemented method that calculates a spatial distribution indicator to select an appropriate adjustment method among ellipsoid, sphere, or two-stage adjustments, applying a test transformation to measurements and updating the calibration transformation if the adjusted measurements show improved alignment with a unit sphere, thereby optimizing calibration precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ellipsoid fitting method is used for calibration, then measurement precision is improved, but computing resources are excessively consumed when measurements are poorly distributed
Solution Approach 1:
The patent changes the parameter being optimized by switching between different fitting methods (ellipsoid, sphere, or hybrid) based on the spatial distribution quality of measurements. When measurements are poorly distributed, the system transitions from computationally intensive ellipsoid fitting to simpler sphere fitting or hybrid approaches, thereby reducing computational resource consumption while maintaining acceptable calibration precision.
Solution Approach 2:
The calibration system dynamically selects the appropriate fitting method based on real-time assessment of measurement distribution characteristics. This dynamic adaptation allows the system to optimize between precision and computational cost by choosing the most suitable algorithm for the current measurement conditions rather than always using the most precise but computationally expensive ellipsoid fitting method.
2Adaptability or versatility
If multiple fitting methods are implemented for different measurement distributions, then adaptability is improved, but device complexity increases
Solution Approach 1:
The system manages complexity by parameterizing the selection of fitting methods based on measurement distribution characteristics. Rather than implementing completely separate calibration systems, the patent uses a unified framework that selects among predefined fitting approaches (ellipsoid, sphere, hybrid) based on quantitative assessment of measurement spatial distribution, thereby achieving adaptability without proportionally increasing system complexity.
3Measurement precision
If measurements are poorly distributed, then calibration precision deteriorates, but the sensor can still operate in practical conditions
Solution Approach 1:
The patent addresses the conflict between precision and usability by changing the calibration approach based on measurement distribution quality. When measurements are poorly distributed (common in practical conditions), the system selects simpler fitting methods that are more robust to distribution issues, thereby maintaining acceptable calibration accuracy while enabling operation in realistic scenarios where perfect measurement distribution cannot be guaranteed.
Solution Approach 2:
The patent employs simpler, less computationally intensive fitting methods (such as sphere fitting or hybrid approaches) as fallback options when measurement distribution is insufficient for precise ellipsoid fitting. These simpler methods act as practical alternatives that sacrifice some precision but enable calibration to proceed under less ideal conditions, similar to using a simpler substitute when the optimal solution is not available.
Data Source
Figure 1a~1b
Figure 2
Figure 3a~3b
AI summary
The invention relates to a method for calibrating a tri-axis sensor such as a magnetometer. The method comprises calculation steps (DIS) of a spatial distribution indicator for a set of measurements delivered by the sensor and determination steps (AJS), according to a fitting method, of parameters of a parametric surface for fitting said surface to said set of measurements. The fitting method is selected from several predetermined fitting methods based on said spatial distribution indicator. The predetermined fitting methods may include a fitting method to an ellipsoid, a fitting method to a sphere, and a two-step fitting method consisting of a first fitting to a sphere followed by a second fitting to an ellipsoid whose center coincides with the center of the sphere of the first fitting.