Since you are convinced that there is no simpler way to achieve what you want, here is the method described in Loren Shure's article (see link posted in Dan's comment), applied to your example:
%// Remove overlapping fields from first struct
s_merged = rmfield(s_old, intersect(fieldnames(s_old), fieldnames(s_new)));
%// Obtain all unique names of remaining fields
names = [fieldnames(s_merged); fieldnames(s_new)];
%// Merge both structs
s_merged = cell2struct([struct2cell(s_merged); struct2cell(s_new)], names, 1);
Note that this slightly improved version can handle arrays of structs, as well as structs with overlapping field names (this is what I believe you call collision).
