An Artificial Inteligence System with Wearable Devices for Correction of Inadequate Posture

Symptoms of early degeneration of the spine due to poor posture are increasingly common. Especially the long-time use of smart communication devices has become a daily life style. If inadequate posture can be identified real-time and accurately, warning and correction can be provided by an evaluation system as soon as possible. The introduction of education and training by human engineering will reduce and prevent chronic disease. In order to continuously monitor the posture of individuals but not affect their daily activities, this study is to utilize wearable and tiny sensors arranged on inner clothing systematically. The on-line system can detect real-time data of body motion and physiological parameters. Combining posture and physiological data related to emotion, an algorism of artificial intelligence is used to identify inadequate posture and its long-term impact immediately. Then, it will issue a warning message and double check the completeness of the posture correction. This project, first of all, plans to design and develop a smart clothing which wearable sensors are embedded, including accelerometers and gyroscopes for detecting body movements, as well as blood oxygen and respiration sensors for detecting physiological conditions. Reliability and validity of measurement need to be confirmed. Furthermore, certain groups with different ages and high risks for early degeneration of spine will be invited in our study. The artificial intelligence will summarize the relationship between the discomfort symptoms and data of sensory measurement and subjective scores. The system will be able to exert a substantial warning and follow-up user’s correction. Finally, density and location of sensors will be increased and optimized. The warning and correction functions will be adjusted by the artificial intelligence after accumulating data from more subjects and monitoring time. The results can be compared with that of the control group to prove the feasibility and usability of the system. Because the system continuously records useful data from substantial subjects, classification and warning of inadequate posture can be extended as an effective reference for screening possible candidates with back pain in a bigger population in the long run. Hopefully, our results can be used to predict the early degeneration of the spine in potential groups without symptoms and meet the goal for health management at home. It will benefit users with correction of posture and relief of emotional disorders. Reduction of premature degeneration of spine can save costs for health care in our community.