A function I wrote last night was fairly flexible in terms of detecting whitespace, and even took into account the pesky non-breaking spaces / zero-width spaces further up the Unicode alphabet.

The benefit here was being able to isolate and identify specific Unicode indices based on their subrange.

<?php
    //    Returns TRUE if the ASCII value of $string matches a registered whitespace character.
    //    *    This includes non-breaking spaces, zero-width spaces, and any unicode values below 32.
    //    *    $string: Character to identify. If string extends past one character, the value
    //        is truncated and only the initial character is examined.
    function is_whitespace($string){
        //    Return FALSE if passed an empty string.
        if($string == "") return FALSE;
        
        $char    =    ord($string);
        
        //    Control Characters
        if($char < 33)                        return TRUE;
        if($char > 8191 && $char < 8208)    return TRUE;
        if($char > 8231 && $char < 8240)    return TRUE;
        
        //    Additional Characters
        switch($char){
            case 160:    // Non-Breaking Space
            case 8287:    // Medium Mathematical Space
                return TRUE;
                break;
        }
        return FALSE;
    }
?>

thanks to gardnerjohng, but a had some problems with non-breaking spaces in this function.

I added 2 more cases for this: 
case 0xC2
case 0xA0

After this modification non-breaking spaces in my test code were successfully detected.