Early detection of falls with multidimensional sensors

DESCRIPTION (provided by applicant): The long-term objective of the project is to develop and commercialize an intelligent fall detector which can detect a fall in its early phase, well before the individual's hip, arm or head hit the ground Unlike any existing fall detectors which detect the fall after the impact, the new detector will be able to trigger a protective mechanism, e g and inflatable hip pads (and/or wrist pads, head pad) to protect the person from injuries due to the fall. The first application of the intelligent fall detection technology will be the Smart HipSaver This device includes the fall detector and the airbag-like inflatable hip pads which are concealed in comfortable and aesthetically appealing underpants The system will provide the elderly and others who are at the risk of falls with the protection against hip fracture, while affording them the freedom to extend their daily activities outside the specially designed hospital or nursing home rooms, such as the smart room, or rooms with dually stiff floor. The system will be expected to receive much higher user acceptance than the existing hip protectors because the inflatable underpants can be made to look and feel much like regular underpants when not inflated. The key to this system is the intelligent fall detector The goals for the fall detector are 1) The detector must be highly reliable in detecting actual falls, 2) the detector must detect a fall in its early phase, and leave sufficient time for any protective mechanisms to be put in place, 3) the detector should not give false alarms during the user's normal activities. In this proposed study (SBIR Phase I) the feasibility of the fall detector will be investigated with an integrated multidimensional sensor, which includes three linear accelerometers, and three angular rate sensors The fundamental requirements of the sensor, data acquisition and computational hardware and software, and most importantly the multi-dimensional event space discrimination algorithms of fall detection will be determined