Research And Design Of Patient Fall Detection System In Hospital Corridor

With the advancement of science and technology and the development of the times,hospitals have also greatly improved their humanized services.However,due to the limited number of medical staff in hospitals and the large number of patients,how to safely consider the travel of patients in the corridors,Over-occupation of the time of medical staff,there are contradictions in these two aspects.Therefore,it is of great significance to be able to accurately determine the fall of a corridor patient and to promptly provide assistance.This paper analyzes the current research status of fall detection and finds that it is basically a study of fall detection for the elderly,and there are very few systems designed specifically for patients.Therefore,combined with the efforts of predecessors,considering the portability of wearable testing equipment,a fall detection system for patients in hospital corridors was designed.Firstly,some data collection methods of fall detection are analyzed,the attitude model of the human body is established,the system scheme is designed,the device is selected according to the scheme,the hardware data collection circuit is designed,and the function of the circuit is programmed.The hardware circuit is mainly divided into two parts,including the data acquisition and transmission of the child nodes and the data reception of the central node.The child node mainly collects the acceleration and angular velocity information of the human body through the sensor,and transmits it to the central node through the wireless module.The center saves the data and prepares for the fall detection algorithm.Secondly,the analysis of the data is performed using the pose data saved by the center node.Due to the waveform jitter during data acquisition,some noise is doped,so the data is filtered.Because the data acquisition time is different,the attitude data is a time-discrete sequence of different lengths,so the data is intercepted.The feature vector is extracted from the intercepted time window to establish the feature vector space.Considering that if there are too many feature vectors,the fall algorithm will run for too long.In order to reduce the running time of the classification algorithm,data reduction is performed,and the fall algorithm is prepared on the premise of ensuring that the information is not lost.Finally,the data is classified.Since the support vector machine works well for fall detection,it is used for fall classification.Considering that the effect of single support vector machine classification is not optimal,the accuracy of classification is improved by parameter optimization.The grid optimization,genetic algorithm and particle swarm optimization algorithm are used to optimize the parameters of the support vector machine respectively.The accuracy is 90%,95% and 95% respectively.By comparison,the accuracy is improved by 93.3% compared with the simple use of support vector machine.The paper has 44 pictures,15 tables and 53 references.