The Research Of Fall Detection Based On Mobile Terminal

The fall is one of the important factors of hazarding the elderly and other specialpopulations. Promptly fall detection and rescue can gain precious time for treatmentand salvage, and it is very important to secure health of people and improve quality ofmedical care. At present, our country has entered the aging society, which presentsaccelerating development tendency as well. So portable, accurate and real-time falldetection method and system urgently need to be researched and developed, in orderto meet wide social requirements.Currently, the domestic and international numerous research institutes anduniversities have studied and experiment fall detection method continuously, butexisting methods generally train One-Class Classification model by feature extractionfrom the normal activity data, such as walk and running. However, the transientfeature information of individual normal activity such as running and down stairs ishighly similar to fall, and with the influence of the noise, which leads to distinguishability of the model is insufficient and can’t meet the requirements of high detectionaccuracy and low false alarm rate.In order to improve the practicability of fall detection method, this paper, bystatistical analyzing the characteristics of fall occurrence, proposes a fall detectionmethod based on activity transition. The research of this paper mainly includes thefollowing four parts:1、Highly accurate continuous activity recognition model. It builds a continuousactivity recognition model to identify the human activity. This stage comparativelyanalyzes the performance of diverse multiple classifier by a large number ofexperiments, and verifies that the multimodal sensors detection result performs betterthan single sensor. It filters feature set in accordance with the feature selectionfunction of WEKA tool, and obtains more robust feature set;2、Automatic segmentation technology for activity transition data. It segmentscontinuous activity sequence according to the result of activity recognition, andobtains activity transition dataset. This paper proposes a automatic segmentationmethod for activity transition data;3、Research and construction of anomaly detection model. It builds anomalydetection model by feature extraction from the normal transition dataset of adjacentactivities. The model detects fall by identifying abnormal activity transition. The activity transition based fall detection method builds feature space by activitytransition dataset, which can filtrate numerous normal activity date, reduce thecomplexity of feature space, and enhance the distinguish ability of model.Comparative experiments of this paper show that classification accuracy of ourmethod is better than the traditional methods, which has certain practical value.4、A fall detection system based on Android phones is designed and implemented.The system can collect real-time sensor data, and real-time monitor the people falls ornot by the fall detection model. If the people fall, mobile phone will alarm withbuzzer sound locally, and obtain the location information of the people by localizationmodule, and automatically edit the fall information and location information to createSMS messages, and then sent it to the specified contact in order to get timely rescue.