Development ofaHumanAirbagSystemforFallProtection UsingMEMS MotionSensing Technology
Guangyi Shi, Cheung-Shing Chan, Yilun Luo, Guanglie Zhang, Philip H. W. Leong · 2006
This paperdescribes thedevelopment ofa human airbag system whichisdesigned toreduce theimpact force from falls. A MicroInertial Measurement Unit(gIMU), basedon MEMS accelerometers andgyrosensors isdeveloped asthe motionsensing partofthesystem. A recognition algorithm is usedforreal-time falldetermination. Withthealgorithm, a microcontroller integrated withthegIMU candiscriminate falling-down motion fromnormalhumanmotions andtrigger an airbag system whenafall occurs. Ourairbag system isdesigned tohavefastresponse withmoderate inputpressure, i.e., the experimental response timeisless than0.3second under0.4MPa. Inaddition, wepresent ourprogress onusingSupport Vector Machine(SVM)training together withthegIMUtobetter distinguish falling andnormalmotions. Experimental results show thatselected eigenvector setsgenerated from200 experimental datasetscanbeaccurately separated intofalling andother motions.