Including State Excitation in the Fixed-Interval Smoothing Algorithm andImplementation of the Maneuver Detection Method Using Error Residuals

摘要

The effects of the state excitation matrix Q underlined sub k in the smoothingroutine of an extended Kalman filter is investigated. A new algorithm to derive the Q underlined sub k matrix is also developed. In addition, the accuracy of the filter was substantially improved by implementing a new maneuver detection technique. Several tracking scenarios are simulated and analyzed for noise free and noisy cases and statistical data are obtained for the maneuver detection technique. The program codes are included as appendices.