第二次审稿人意见回来了，遇到两个审稿人，一个已经搞定，第二个审稿人又提了一堆问题，以下是他的审稿意见开头：Actually the calibration using a new G matrix with G= F_1^T(ff^T)^(-1), with two different matrices for different load directions, takes away the modelling of the structural stiffnesses, and replaces them with a new matrix that takes care of the linear mapping between individual sensor values and the 6-dimensional force acting on the sensor, if I am correct. And, I like this way of doing it, learning the stiffnesses or correlations and cross-correlations from data of know load forces and branch outuput values. However, it should be maybe more clearly noted that now you are actually getting rid of all the other elastic and geometric modelling mathematics presented, and keeping only the part related to different load directions (signs) having different matrix components. Thus the role of the elastic modelling (?) in the article could be decreased. you are just comparing in the experiments the residual errors after fitting two different calibration models
实际上第一个审稿人意见我回复了，但是这个审稿人说我没回复，郁闷死了，而且第一次这个审稿人就提了23个问题，这次提了25个，有几个问题是第一次提了认真答复的，但是还是在这次意见中提问，真是无语了。而且这个审稿人说的好多话感觉很口语话，理解需要读半天才能明白，请大家看看这个审稿人说的什么意思。

 返回小木虫查看更多