Self-Calibrating Inertial Measurement for MEMS Sensor Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inertial navigation systems using MEMS sensors suffer from errors due to aging and require time-consuming recalibration processes, which affect measurement accuracy.
Innovation Solution
A self-calibrating inertial measurement system that includes a sensor cluster with multiple inertial sensors and a processing engine, dynamically self-calibrating parameters such as scale factor, cross-axis sensitivity, and bias by combining individual sensor outputs and applying corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If repeated recalibration of MEMS sensors is performed to counter aging effects, then measurement accuracy is improved, but time consumption and operational complexity increase
Solution Approach 1:
The system performs self-calibration using redundant sensor clusters and processing engines that automatically detect divergent outputs and apply corrections without external intervention. The processing engine compares outputs from multiple sensors, identifies outliers using decision rules, and dynamically adjusts calibration parameters during operation, enabling the system to service itself without removing sensors or using external equipment.
Solution Approach 2:
The system performs preliminary calibration adjustments by dynamically updating calibration parameters based on real-time sensor output comparisons. The processing engine continuously monitors sensor outputs and applies corrective actions before significant drift occurs, maintaining accuracy proactively rather than waiting for scheduled recalibration events.
2Measurement precision
If repeated recalibration of MEMS sensors is performed to counter aging effects, then measurement accuracy is improved, but device complexity and operational burden increase
Solution Approach 1:
The system performs self-calibration using redundant sensor clusters and processing engines that automatically detect divergent outputs and apply corrections without external intervention. The processing engine compares outputs from multiple sensors, identifies outliers using decision rules, and dynamically adjusts calibration parameters during operation, enabling the system to service itself without removing sensors or using external equipment.
Solution Approach 2:
The system divides the calibration function into separate processing engine modules that handle specific tasks: comparing sensor outputs, detecting divergences, and applying corrections. This segmentation allows the complex calibration process to be broken into manageable automated steps, reducing operational burden while maintaining accuracy.
3Measurement precision
If dynamic self-calibration is performed during operation to correct sensor drift, then measurement accuracy is improved, but processing complexity increases
Solution Approach 1:
The system divides the calibration function into separate processing engine modules that handle specific tasks: comparing sensor outputs, detecting divergences, and applying corrections. This segmentation allows the complex calibration process to be broken into manageable automated steps, reducing operational burden while maintaining accuracy.
Solution Approach 2:
The processing engine continuously compares sensor outputs and uses feedback from divergence detection to dynamically adjust calibration parameters. The system monitors sensor performance in real-time and applies corrective feedback loops that automatically update scale factors, cross-axis sensitivities, and biases based on observed output variations, maintaining accuracy through continuous adaptive correction.
Data Source
AI summary
An inertial measurement system comprising at least one sensor cluster comprising a plurality of inertial sensors for sampling at least one of acceleration and angular velocity of said at least one sensor cluster with respect to each axis in a plurality of axes of a reference frame, and for producing individual outputs associated with said at least one of acceleration and angular velocity, at least three of said inertial sensors said sampling with respect to each same respective said axis; and a processing engine for receiving said individual outputs, combining said individual outputs to yield respective combined outputs, detecting which of said individual outputs diverges from at least one of their inter-comparison, and its respective combined output, according to a decision rule, said processing engine configured to dynamically self-calibrate a parameter that includes individual scale factor of those said inertial sensors whose said individual outputs were detected to diverge.