Volterra级数改进的Tolles-Lawson磁干扰补偿方法

Volterra series-improved Tolles-Lawson magnetic interference compensation method

摘要:

地磁导航作为一种全天候、多时段、全自主的导航方式，需要精确的实现干扰磁场的建模与补偿；针对无人机航空磁测数据时间序列混沌性的特点，提出了一种考虑非线性动态磁场的航磁补偿方法。该方法首先使用Tolles-Lawson(T-L)模型对航空磁测数据进行初步的补偿；再使用互信息法和Cao氏法确定了无人机非线性动态磁场的延迟时间和嵌入维数，经过相空间重构，输入Volterra级数模型，使用递推最小二乘法对Volterra级数模型的核函数进行辨识；最后对T-L补偿后的结果进行二次补偿。通过实际飞行数据的验证，该方法大大地提高补偿的精度，且具有较好的鲁棒性和泛化性，改善比超过了30。

Abstract:

navigation, as an all-weather, multi-period, and fully autonomous navigation method, requires precise modeling and compensation of interference magnetic fields. Considering the chaotic characteristics of UAV aeromagnetic data time series, this study proposes a compensation method for nonlinear dynamic magnetic fields. First, the Tolles-Lawson (T-L) model is used for preliminary compensation of aeromagnetic data. Then, the delay time and embedding dimension of the UAV""s nonlinear dynamic magnetic field are determined using the mutual information method and Cao""s method. After phase space reconstruction, the Volterra series model is applied, and its kernel functions are identified using the recursive least squares method. Finally, secondary compensation is performed on the results of the T-L model. Validation with real flight data shows that this method significantly improves compensation accuracy, with robustness and generalization capabilities, achieving an improvement of over 30.