为解决惯性传感器在磁场干扰和大加速度干扰坏境下姿态解算精度下降的问题，提出一种基于决策树自适应的互补滤波姿态解算方法。针对陀螺仪特性，对其进行零偏补偿处理。将处理后的角速度数据与加速度数据构成静态检测单元，实现了静态情况和动态情况的准确分类。在动态情况下，由加速度数据和磁场数据组成的决策树通过分析外界干扰程度来自主调节系统误差增益，进而达到抗干扰目的，有效提高姿态解算精度。实验结果表明：静态情况下姿态解算不受磁场干扰，动态情况下能够有效避免磁场和大加速度干扰。

In order to solve the problem of the attitude estimation accuracy decline of the inertial sensor in the environment of magnetic field interference and large acceleration interference, a complementary filter attitude estimation method based on decision tree adaptive was proposed. Aiming at the zero-degree offset characteristic of the gyroscope, it performed zero-degree offset processing. The processed angular velocity data and acceleration data constituted a static detection unit, which realized accurate classification of static and dynamic situations. In the dynamic situation, the decision tree composed of acceleration data and magnetic field data can automatically adjust the system error gain by analyzing the degree of external interference, so as to achieved the purpose of anti-interference and effectively improve the accuracy of attitude estimation. The experimental results show that the attitude estimation is not affected by the magnetic field in the static state, and can effectively avoid the magnetic field interference and the large acceleration interference in the dynamic state.