Simultaneous State and Parameter Estimation with Trajectory Shape Constraints (Poster)
Keyi Li, Gongjian Zhou, Thia Kirubarajan, Jiazhou He · 2019
In some tracking scenarios, the target state is subjected to equality constraints due to external limitations or inherent properties. If the constraints are known a priori, more accurate state estimates can be produced by taking advantage of these additional information in tracking algorithms. In this paper, a new model of the trajectory shape constraint is proposed when the target trajectory is known to be a straightline. The unknown slope and intercept of the straightline are treated as states to be estimated along with the target state. Then, two pseudo-measurements are constructed and augmented into the measurement equation in the filtering process. A trajectory shape constraint augmented state filter (TSC-ASF) is developed to produce constrained state estimates and constraint parameter estimates simultaneously. The nonlinear radar measurements and pseudo-measurements are processed by the converted measurement Kalman filter (CMKF) and unscented Kalman filter (UKF), sequentially. The unscented transform (UT) is employed to initialize the filter. Monte-Carlo simulation results illustrate the effectiveness of the proposed algorithm.