A new constant gain Kalman filter based on TP model transformation

A constant gain Kalman filter is proposed in this paper concerned with the problem of the complexity and large calculation in nonlinear system estimation. With the introduction of linear parameter varying (LPV) model and tensor product (TP) model transformation method, the nonlinear system is represented by a linear polytopic model. The transformation directly leads to the reduction of the conservativeness for the linear polytopic model gained by the parameter bounds method and avoids solving infinite number of linear matrix inequalities (LMIs). Moreover, a constant gain filter is developed based on the EKF and robust H2 filtering, which greatly reduces the calculation number. Finally, an example is employed to illustrate the effectiveness of the proposed filter.