RKHS Based State Estimator for Radar Sensor in Indoor Application

For the estimation of targets’ states (location, velocity,
and acceleration) from nonlinear radar measurements,
usually, the improved version of well known Kalman filter: extended
Kalman filter (EKF) and unscented Kalman filter (UKF)
are used. However, EKF and UKF approximates the nonlinear
measurement function either by Jacobian or using sigma points.
Consequently, because of the approximation of the measurement
function, the EKF and UKF cannot achieve high estimation
accuracy. The potential solution is to replace the approximation
of nonlinear measurement function with its estimate, obtained in
high dimensional reproducing kernel Hilbert space (RKHS). An
ample amount of research has been done in this direction, and the
combined filter is termed RKHS based Kalman filter. However,
there is a shortage of literature dealing with estimating the
dynamic state of the target in an indoor environment using RKHS
based Kalman filter. Therefore, in this paper, we propose the use
of RKHS based Kalman filter for indoor application. Specifically,
we validate the suitability of the RKHS based Kalman filtering
approach using simulations performed over three different target
motion models.