Abstract:The two-step filter has been combined with a modified Sage-Husa time-varying measurement noise statistical estimator, which is able to estimate the covariance of measurement noise on line, to generate an adaptive two-step filter. In many practical applications such as the bearings-only guidance, some model parameters and the process noise covariance are also unknown a priori. Based on the adaptive two-step filter, we utilize multiple models in the first-step filtering as well as in the time update of the second-step filtering to handle the uncertainties of model parameters and process noise covariance. In each timestep of the multiple model filtering, probabilistic weights punishing the estimates of first-step state from different models, and their associated covariance matrices are acquired according to Bayes' rule. The weighted sum of the estimates of first-step state and that of the associated covariance matrices are extracted as the ultimate estimate and covariance of the first-step state, and are used as measurement information for the measurement update of the second-step state. Thus there is still only one iteration process and no apparent enhancement of computation burden. A motion tracking sliding-mode guidance law is presented for missiles with non-negligible delays in actual acceleration. This guidance law guarantees guidance accuracy and is able to enhance observability in bearings-only tracking. In bearings-only cases, the multiple model adaptive two-step filter is applied to the motion tracking sliding-mode guidance law, supplying relative range, relative velocity, and target acceleration information. In simulation experiments satisfactory filtering and guidance results are obtained, even if the filter runs into unknown target maneuvers and unknown time-varying measurement noise covariance, and the guidance law has to deal with a large time lag in acceleration.

Interactive Multiple Model Filter for Tracking a Pursuer With Proportional Navigation Guidance Law

Applying a Novel Extended Kalman Filter to Missile–target Interception with APN Guidance Law: A Benchmark Case Study

Using Interacting Multiple Model Particle Filter to Track Airborne Targets Hidden in Blind Doppler

Novel multiple-model probability hypothesis density filter for multiple maneuvering targets tracking

Current Statistical Model Probability Hypothesis Density Filter For Multiple Maneuvering Targets Tracking

Imm Algorithm Based On H-Infinity Filter For Maneuvering Target Tracking

IMM Algorithm Based on H∞ Filter for Maneuvering Target Tracking

Bearings-Only Tracking of Maneuvering Target Based on IMM-UKF Algorithm

Adaptive interacting multiple model algorithm for high maneu-vering target tracking

Multiple Model Adaptive Two-Step Filter and Motion Tracking Sliding-Mode Guidance for Missiles with Time Lag in Acceleration

Tracking for near space target based on IMM algorithm

A novel robust interacting multiple model filter for manoeuvering target tracking

Interacting multiple model algorithm based on H∞ filter

Novel Interacting Multiple Model Filter for Uncertain Target Tracking Systems Based on Weighted Kullback–Leibler Divergence

Maneuvering Target Tracking Method Based on IMM-UKF

Adaptive interacting multiple model particle filter based on a dual-pattern bat algorithm for maneuvering target tracking

Three-dimensional Underwater Dynamic Target Tracking Based on Adaptive Interactive Multi-model Algorithm

Application of Adaptive Weighted Strong Tracking Unscented Kalman Filter in Non-Cooperative Maneuvering Target Tracking

Interacting Multiple Model Particle-type Filtering Approaches to Ground Target Tracking.

An IMM-UF algorithm for tracking highly maneuvering target

Maneuvering Target Tracking Using IMM-PF with Doppler-Aided Measurement