v0.2.0

New outlier suppression scheme works directly on Kalman gain, which also affects the posterior covariance (reducing how much a measurement affects the posterior as it is being suppressed). This is not a substitute for logic outside the Kalman filter to deal with high-conflict scenarios (where the innovation value is far outside the expected distribution). It does however, greatly mitigate the risk of a single measurements from corrupting the state before such logic can be applied.