a=importdata(filename);
d=char(a)
Okay, so d is a character vector or character array -- in particular, d is not a file id. You did [fileID, errmsg]= fopen(filename); but you never use the fileID other than to check feof() once on it and discard the results of the check.
Latitude = textscan(d,'<trkpt lat="%f" lon="%f">\n');
You are passing the characters in d to textscan. It is valid for textscan purposes if d is a 2D array of characters: textscan will take the characters in linear index order in such a case (which would read along columns, not across rows).
textscan() will skip leading whitespace in most cases, and then will try to literally match what is in d against '<' 't' 'r' k' 'p' 't' and so on in consecutive positions. If the first non-whitespace does not match the literal characters <trkpt lat=" then textscan() will stop reading and return empty. textscan() never searches through the text looking for matching places: it always starts at the current position and tries to match the next format item. The current position for a character array is the beginning of the array.
for Longitude = textscan(d, 'long="%f">\n');
Reminder that when textscan() works, it returns a cell array of values. If textscan() were able to find multiple occurances of that pattern in a row, with there being only one % format, it would return a cell array that contained a single entry, and the single entry would be a numeric column vector.
You would be on better grounds if you were reading from a file instead of from a character array, but you would be much better off using something like regexp() to parse the text.
Given the structure of the file, possibly you could use xmlread() or readstruct()
