Research On Cervical Spondylosis Prevention Action Recognition Based On Acceleration Sensor

With the changes in the working methods of modern society,many workers are required to maintain a fixed head and neck posture for a long time.The incidence of cervical spondylosis is increasing year by year and there is a trend of youth.More and more people are suffering from cervical spondylosis.The recent studies have shown that neck movements have significant effects in the prevention and treatment of cervical spondylosis,so that head and neck movement monitoring and effective head and neck action recognition are of great significance in the prevention of cervical spondylosis.In view of the above problems,the research on cervical spondylosis prevention action recognition based on acceleration sensors was designed and implemented.The specific research contents are as follows: Firstly,the cervical disease prevention action recognition system hardware was designed and implemented.The hardware part of this system includes four working modules: a signal acquisition module where an acceleration sensor is applied,a signal control and processing module,a wireless communication module,and a power module.The hardware is small and can be placed on the user’s head portablely.The acceleration data of the head and neck can be extracted,and the motion data can be comprehensively analyzed by combining different neck action.Secondly,the neck actions that could prevent cervical spondylosis were examined,and 8 kinds of neck actions were determined,which were bow down,head up,left turn,right turn,left flexion,right flexion,left loop,right loop,etc.More than 10,000 independent and continuous acceleration data were acquired according to the 8 neck actions.Thirdly,the acceleration data preprocessing was performed including filtering,gravity removing,signal segmentation and splicing.Subsequently,support vector machine,random forest and fully connected neural network methods were performed to classify and identify cervical spondylosis prevention actions.The final classification accuracy was 89.66%,91.43%,and97.51%,respectively,and the fully connected neural network method was used in the mobile system.Finally,the cervical spondylosis prevention action recognition based on mobile system was developed and implemented.The system is connected to the hardware system via bluetooth module,receiving continuous movement data of the head and neck in real time,and performing motion classification recognition based on the fully connected neural network.The number of effective cervical spondylosis prevention actions are counted and sorted the over a period of time to evaluate the degree of cervical spondylosis prevention.When the amount of exercise is less than the threshold,the user is reminded to adjust the exercise program to prevent the occurrence of cervical spondylosis.In general,the hardware and software of our system are efficient and accurate;it can monitor the movement of the head and neck continuously.The system can display the number of effective cervical spondylosis prevention actions in real time on the smartphone to prompt and guide the user.The system can be applied to the daily self-detection for early prevention of cervical spondylosis,prevent cervical spondylosis caused by long-term office,which is beneficial to the family and the community.