中文摘要：

捷联惯导系统（SINS）中卡尔曼滤波的运算时间与系统阶次的三次方成正比，滤波失去实时性，达不到捷联惯导系统的快速性要求。通过研究捷联惯导系统、最小二乘支持向量机和卡尔曼滤波，提出既能保证准确性又能提高实时性的支持向量机初始对准算法。用Matlab软件仿真的结果表明，提出的捷联惯导初始对准算法是有效的。

英文摘要：

As Kalman filter computation time is proportional to the cube of the system order, the filter will lose real time. An effective algorithm to improve real-time filtering is prosed based on studying the least squares support vector machine, （SINS） and Kahnan filter. Matlab software simulation results showthat the initial alignment algorithm is effective and feasible.