Design a MEMS switch for fall detection

Falling is considered to be one of the most important reasons for injuries sustained by the elderly like the hip fracture. It can even lead to death due to delayed detection. In this thesis, we will study and analyze human falling from a dynamic view and provide a model of a proposed MEMS falling detection sensor.;In the first part, we model the human body using one- and two-link models. In each model, the links are modeled falling from a vertical to a horizontal position as an inverted pendulum. We analyze the acceleration of the falling links for different positions and determine the normal and tangential acceleration components.;In the second part, we provide experimental fall data from real falls using a tri-axial accelerometer placed at three major positions: hip, chest and head for different kinds of falls and Activities of Daily Living (ADL). The collected data of acceleration is analyzed and then the proper threshold of acceleration for each position is extracted to recognize a fall when the acceleration exceeds certain thresholds in the normal and tangential direction.;In the third part, we use nonlinear lumped parameters with a single degree of freedom spring-mass-damper system model to simulate a proposed MEMS switch for falling. The concept of the switch is based on the triggering of pull-in instability in electrostatically actuated cantilever beam due to the acceleration signal from falling.;We create a Fast Fourier transform (FFT) for each fall signal to show that all dynamic fall signals are within the quasi static regime, and hence the Squeeze Film Damping (SQFD) phenomenon doesn't have any effect on the system. We examine the response of our system under the effect of the fall signal to choose proper design parameters that result in the best performance. The effect of the self-weight of the sensor has been tested using ANSYS software and it is found to have a maximum effect of 9.45% of the total gap width. Two switches are connected together in series to detect a fall event. One of the switches will trigger the normal acceleration of the fall, while the other switch will trigger the tangential acceleration. When both switches are closed, the fall alarm will go off.. In the last part, we make a few suggestions for future work.