Contact Location Estimation fromaNonlinear ArrayofPressure Sensors
Amaya Arcelus, Megan Holtzman, I. Veledar, Rafik A Goubran, Heidi Sveistrup, Paulette Guitard · 2008
Inphysical medicine andrehabilitation, itisimportant tobeabletocollect information regarding apatient's behavior and range ofmobility throughout their daily activities. Grabbarsareused widely inthehomesofindividuals withmobility impairments sotheir usage while performing physical tasks canprovide valuable informa- tion astotheindividual's physical status. Thispaperexplores the extraction oflocation information forforces applied toagrabbar embedded withanonlinear pressure sensor array oflowspatial res- olution. Itfirst describes theinstrumentation ofthegrabbarandthe calibration procedure. Itthen investigates three methods ofestimating thecontact location; asimple centroid, apercentage-based lookup ta- bleandanartificial neural network. Results ofthethree methods are reported based ondatacollectedfrom differentforces andcontact lo- cations applied along thebarTheartificial neural network proves to bethemostsuccessful method ofestimating thepoints ofcontact, by mostaccurately modeling thenonlinearities inthesystem.