Over-the-horizon radar target tracking using MQP ionospheric modeling

As over-the-horizon radars use ionosphere reflection to detect and track targets, they are faced with a possible appearance of multipaths and difficulties to estimate the target ground location. To solve these problems we propose to use a ionosphere model called multi-quasi-parabolic (MQP) model. Introducing the coefficients of the MQP model into a target tracking algorithm leads to nonlinear evolution and measurement equations. Furthermore we only have estimation, through ionosphere measurements, of these parameters values. To take into account MQP parameters in a target tracking algorithm, we propose two different approaches built on the same algorithm: the unscented Kalman filter. In the first approach, we use a joint unscented Kalman filter and in the second one we use an unscented particle filter applied on both target state space and MQP parameters space (we could call this a joint unscented particle filter). We compare these two approaches and present our conclusions illustrating them by numerical simulations on the French over-the-horizon radar nostradamus.