Presence sensing is one of those things that is easy to do badly 
Honestly, I have yet to run into a system that does not generate false reports (either positive or negative). I've found two technologies to be very useful in this area: passive infrared and doppler/x-band radar. The Sharp distance sensors are very nice and easy to work with, but this is really not the problem that they are meant to solve (they are really for detecting presence/movement in the 0 - 14" range, give or take) . Ditto for ultrasonic sensors. Yes, you can achieve a limited sort of motion/presence sensing with these technologies, but IMHO they are the wrong tools for the job given the requirements that you've stated.
Parallax makes a nice radar sensor that is great for a 8' - 30' range: http://www.parallax.com/StoreSearchResults/tabid/768/txtSearch/x-band/List/0/SortField/4/ProductID/606/Default.aspx; I've also had decent luck with cheaper models purchased via EBay: http://www.ebay.com/itm/130512020245?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649#ht_3211wt_1114
Check out the Zilog Z-pir sensors as well - about $6.50 from mouser http://www.mouser.com/ProductDetail/ZiLOG/ZEPIR0AAS02MODG/?qs=sGAEpiMZZMv2WiXdFcmpiDB8nPmOn4Po. These are microprocessor based, low power, and painfully simple to integrate with arduino or other uC projects. They support direction sensing as well (did the person move from left to right or the other way around?) which is fairly accurate.
The ping sensors are actually quite nice. Just don't put them in a metal enclosure unless you fancy a migraine (my old ears are still sensitive to high frequencies, so I may be an outlier in this case).