Estimation of target lateral acceleration for Augmented Proportional Navigation

Abstract

Extending the ordinary Proportional Navigation control law (PN) for a homing missile to the so called Augmented Proportional law (APN) can reduce miss distances, e.g. if there are limitations in realizable lateral acceleration for the missile. However, APN requires knowledge about the target's lateral acceleration. This paper is an investigation and comparison of four different algorithms for estimation of the lateral acceleration of the target for a homing missile. These algorithms are: Kalman estimator An ordinary linear Kalman filter with states describing relative position, velocity, and target acceleration. Two Stage Kalman estimator A linear Kalman filter which basically contains states only for relative position and velocity, whose innovations are used to separately estimate the target acceleration. Interactive Multiple Model estimator, IMM Two Kalman filters, each one working on its own system model. One of the models contains only position and velocity states, the other one also a target acceleration state. Probabilities for each model to be true are continuously maintained, and the target acceleration is estimated as the second model's acceleration estimate, weighted with the corresponding probability. Interactive Acceleration Compensation estimator, I AC A combination of the Two-Stage and the IMM algorithm. The purpose is to get an algorithm with smaller computational requirements than IMM.