Comparison of IMM and robust filters in impulsive noise environments

Impulsive noise environments require the modification of standard filters to perform acceptably. The Kalman filter (KF) is notoriously sensitive to frequent outliers and model mismatches resulting from the presence of impulsive noise. We present several alternatives to handle the presence of these frequent outliers and show that a simple interacting multiple model (IMM) is in general sufficient to achieve low mean squared error (MSE) performance.