在弹载等高动态环境下组合导航系统状态方程具有强非线性,且各状态相互耦合影响,传统的扩展卡尔曼滤波(E KF)算法因忽略高阶项相互影响,其模型线性化展开会导致模型不准确引起导航精度下降;无迹卡尔曼滤波(UKF)算法能有效避免引入线性化误差,却存在因组合导航系统维数过高引起大量粒子递推滤波计算复杂而影响算法实时性的问题.为此,针对发射惯性系下弹载组合导航系统对滤波算法高实时性和高精确性的要求,设计了一种简化UKF(SUKF)算法,SUKF算法通过对导航系统的状态参数直接进行建模估计,解决了传统UKF算法实时性差的问题,同时继承了传统UKF算法无需模型一阶线性化展开的优点,提高了导航系统的精度.算法仿真结果表明,SUKF算法有效提高了系统解算的实时性和滤波精度,非常适合用于实际工程系统.
Each state of integrated navigation system is strongly nonlinear and coupled with each other un- der missile-borne high-dynamic environment. Since the traditional extended Kalman filter algorithm ignors the interaction of higher order terms, its linear expansion may teads to its inaccuracy and the decrease in navigation precision. UKF algorithm can effectively avoid bringing a linearization error, but the real-time of the algorithm can be affected by the complex filtering calculation for a large number of particles in high-dimensional integrated navigation system. A simplified unscented Kalman filter (SUKF) algorithm is proposed for the requirement of high real-time and accuracy for the missile-borne integrated navigation system in launch inertial coordinate system. SUKF can be used to directly estimates the state parameters of the navigation system to achieve the real-time and improve the navigation precision without linearization extension in traditional EKF. The simulation results show that SUKF effectively improves the real-time and filtering precision of algorithm, so SUKF is very suitable for the actual engineering system.