SMFEKF-IMM algorithm for maneuvering target tracking with correlated noise

The interacting multiple model (IMM) estimator based on the Kalman and extended Kalman filters can reduce the model uncertainty, but can not eliminate the state bias. The accuracy of target state estimates was improved with a tracking algorithm for maneuvering targets with correlated noise which combines IMM with the suboptimal multiple fading extended Kalman filter (SMFEKF). This algorithm makes the residual error approximately orthogonal by introducing a suboptimal fading factor matrix. This approach reduces the effect of the correlated noise so that the maneuvering target can be fit by dynamically adjusting the gain and noise levels based on the changing of the residual error. Simulation results demonstrate the effectiveness of the proposed approach.