不同传感器具有不同的非线性。为提高传感器数据重构精度,本文给出两种数据重构方法。对于不是特别严重的非线性,采用最小二乘拟合与径向基函数残差插值进行融合重构,可以在增加有限计算量条件下提高数据的近似精度;对于非线性较严重的传感器,为兼顾局部特性,采用移动最小二乘法进行数据重构,它通过全局近似向局部近似的转化,同样使重构结果具有满意的近似精度。选用两种不同传感器进行实验,结果表明两种方法均行之有效。
Aiming to different nonlinearity,two improved methods for sensors' data reconstruction are proposed.The first method is for the slight nonlinearity,it can be impleted by using the least squares and the radical basis function interpolation to make the results more precise under the limited increament of computation.The second is for serious nonlinearity,and the sensor data reconstraction is based on the moving least squares to convert the approximation from general into locality,and its result is also satisfied.