There is a documented need for improved instrumentation for study of falls, loss of balance, and development of early detection and prevention methods. Current approaches to fall detection simply monitor system outputs (most often body segment velocities and/or velocities). They do not measure system inputs. Ahmed and Ashton-Miller at the University of Michigan, conceptualized a model-reference adaptive controller and failure detection algorithm to imitate the central nervous system (CNS) decision making based on both input and output signals obtained during a challenging whole-body planar balancing task. Conclusive evidence was found that a control error signal anomaly (CEA) could be detected by tracking externally observable physical input and output parameters while healthy subjects perform a challenging balancing task. We will develop body conformable, miniaturized instrumentation for (i) Measuring the ground reaction (input parameters) between the dominant foot and the floor that can be incorporated into a commercial insole with a wireless transmitter, (ii) Measuring the torso kinematics (output parameters) using a stamp-size inertial measurement cube (tri-axial angular rate sensors, accelerometers) that is chest mountable with a wireless transmitter, and (iii) a Waist-mounted central processor with a wireless receiver to monitor the signals and generate a CEA on detecting loss of balance.
