Covariance Matrix Evaluation in Correlated Type B Measurement Uncertainty Propagation
Evaluating input covariance matrices in Type B propagation prevents underestimating multi-channel variance and ensures valid system uncertainty limits.
Off Diagonal Terms
Mathematical Component
Elements in a square matrix that do not lie on the line from the top left corner to the bottom right corner. Representing the coupling between different axes of a multi-axis sensor involves placing values in the off diagonal terms of a sensitivity matrix. Cross-axis values indicate how a force or acceleration in one direction affects the output of a sensor aligned with a perpendicular direction.
Ideally these values remain at zero for perfectly isolated channels.
Cross Sensitivity
Misalignment between the internal sensing element and the package frame creates non-zero off diagonal terms during laboratory testing. Quantifying these interactions is necessary for high precision navigation and robotic control systems. Large values indicate that the sensor cannot accurately distinguish between signals on different axes.
Alignment Correction
Compensating for alignment errors requires calculating the inverse of the matrix containing the off diagonal terms and applying it to the raw sensor data. This process removes the crosstalk that would otherwise degrade the accuracy of the vector measurement. Automated test equipment measures these terms by applying known inputs in precisely controlled orientations.
Covariance Entry
Statistical analysis uses the off diagonal terms in a covariance matrix to describe the correlation between two different measurement variables. Positive values suggest that the variables increase together while negative values indicate an inverse relationship. Independent variables result in zero values in these positions.