Truncated randomized unscented Kalman filter for interval constrained state estimation

The paper deals with state estimation of nonlinear stochastic dynamic systems with constraints imposed on the state. The constraints are considered in the form of a generally nonlinear inequality. Such constrained state estimation problems frequently appear in tracking application where kinematic or geometry constraints often arise. In the paper a truncated randomized unscented Kalman filter is developed that is built on the algorithm of the randomized unscented Kalman filter and a probability density function truncation technique. The proposed filter achieves quality estimates with low computational costs. The proposed filter is illustrated in two numerical tracking examples.