Equidistant interpolation with outliers

Consider a 2 dimensional curve based on GPS data. That is, each point consists of lat long coordinates and we addiontally have a time stamp. How to create a curve that "resembles" the original but that is formed by equidistant points? The method should be robust with respect to outliers.

Answer 1

[Deepak Meena]

Hi,

You can do the following :

• Use the fit fucntion with ROBUST name pair argument to get a fit
• Get the equidistant x points as

Xq = linspace(a,b,no);

• Get the data at those points using feval function :

Yq = feval(f,Xq);% f is the fit 

Now plot the data [Xq,Yq] , we will get a curve that "resembles" the original but that is formed by equidistant points

Thanks