惯性导航系统非线性初始对准的LS-SVM方法研究

由陀螺仪和加速度计等惯性传感器组成的惯性导航系统,在进入导航状态之前必须进行初始对准.根据支持向量机强大的非线性映射能力,建立了基于最小二乘支持向量机(LS-SVM)的捷联惯导系统非线性初始对准方法.LS-SVM避开了经典支持向量机求解时的复杂优化运算,通过求解一组线性方程组就可以得到唯一的全局最优解,所以算法的复杂度大大降低,能更好的满足工程应用中的实时性要求.针对方位初始失准角为大角度时的捷联惯导系统非线性误差模型进行仿真分析,并在相同条件下与卡尔曼滤波方法作比较,结果表明LS-SVM在初始对准中的有效性和可行性.

Nonlinear initial alignment of strapdown inertial navigation system using LS-SVM

Before being in the state of navigation, initial alignment of the inertial navigation system which is composed of gyroscope and accelerometer, is necessary. Based on the powerful nonlinear mapping ability of support vector machine, the nonlinear initial alignment of strap-down inertial navigation system using least square support vector machine (LS-SVM) is studied. LS-SVM gets the exclusive global optimization solution by solving a set of linear equations, avoid quadratic programming for classical support vector machines, the complexity of the algorithm is obviously reduced, can subject to the real time needs of engineering. Simulations had been done for the nonlinear initial alignment of strap-down inertial navigation system while the azimuth is big, and compared with Kalman filter under the same condition, the results verify the ability of LS-SVM in initial alignment.