I am interested in representing the performance/consistency of my Kalman filter in a single plot. I would like to compare the norm of the estimate error against 3$\sigma$ error. I would also like the plot to capture how the uncertainty of the estimate changes with time. When I am plotting the performance of a single state, I plot abs(error) and 3$\sigma$ where 3$\sigma$ is the sqrt of the appropriate diagonal entry of the cov mat.
I could plot there mahalanobis distance vs the 3$\sigma$ chi squared distance but this would not capture how the uncertainty of the kf changes with time. I've played around with scaling both of these terms by functions of the cov but I haven't found anything satisfying yet.
Is there a standard way to capture estimate consistency/performance of a multi-d kf in a single plot?