Recognition system of acupuncture manipulations based on an array PVDF tactile sensor and machine learning

In learning and evaluation of acupuncture manipulations, there are lack of quantitative physical parameters on exertion strength, duration and direction of acupuncture technique at present. Based on the tactile parameters collected during and lifting-thrusting of needling, a kind of array polyvinylidene fluoride (PVDF) tactile sensor was designed. Followed by, a window segmentation method for tactile signal was proposed and the time domain features of the window were extracted. Finally, an identification method of acupuncture manipulation based on FCM (Fuzzy C-Means) was constructed. Through the experiment, it was proved that this sensor can effectively identify the four kinds of basic acupuncture manipulations, i.e. reinforcing by twirling and rotating, reducing by twirling and rotating, reinforcing by lifting and thrusting and reducing by lifting and thrusting and it was conductive to the quantification and dissemination of acupuncture manipulations.