Fall Detection Algorithm Based on Wearable Design

As the world's aging problem intensifies, monitoring the physiological status of the elderly has become a hot issue of social concern. As an important method, body health monitoring based on wearable sensors has the characteristics of simplicity, efficiency, and low power consumption. Aiming at the data differentiation problem caused by different physiological data of different groups of people, a new feature space construction method is proposed, which introduces the concept of feature importance to minimize the complexity and calculation delay of feature calculation. In addition, a feedforward neural network is constructed to achieve secondary direction judgment after a fall to balance algorithm complexity and overall judgment accuracy. Simulation results show that compared with traditional neural network models, this algorithm has advantages in classification accuracy and model complexity, and has the potential to be deployed on small devices.