针对加速度传感器动态模型的研究中,不考虑加速度传感器的非线性特性而将其简化为一个二阶线性系统,导致模型精度不高的问题,本文提出了一种基于混合核SVM的加速度传感器非线性动态建模方法。该方法引入线性核函数和径向基核函数构成的混合核函数,利用赤池信息量准则确定模型阶数,通过调节混合核的权重系数,建立加速度传感器的非线性动态模型。实验结果表明,该方法所建立的加速度传感器模型比二阶线性模型精度更高、泛化性能更好。 In the research of dynamic model of accelerometer, the nonlinearity of accelerometer is reduced to a second-order linear system, which leads to the problem of low accuracy of the model. In this paper, a hybrid SVM-based acceleration sensor nonlinear Dynamic modeling method. The method introduces a hybrid kernel function composed of linear kernel function and radial basis function, determines the order of the model by using the Akaike information criterion, and establishes the nonlinear dynamic model of the acceleration sensor by adjusting the weight coefficient of the hybrid kernel. The experimental results show that the acceleration sensor model established by this method is more accurate and has better generalization performance than the second-order linear model.