基于柔性压阻式压力传感器的手写字母识别OA

Flexible pressure sensors face great challenges in the fields of human-computer interaction and handwriting recognition,and have important application prospects.In order to improve the recognition accuracy and capture the characteristic information of hand-written characters in human-computer interaction,a handwritten letter recognition technique for flexible pressure sensors based on the residual network(ResNet)model is proposed.A flexible piezoresistive pressure sensor with high sensitivity(3.22％kPa-1),fast response speed(110 ms),and good stability is prepared by using a cost-effective solution impregnation and drying method.By handwrit-ing 26 English letters on the surface of the sensor,2 600 sets of voltage time series data output from the sensor are collected as experi-mental samples.On this basis,a residual network model is established to classify and recognise the 26 English letters applied to the sen-sor surface,and the average recognition accuracy reaches 97.88％.The experimental results show that the flexible pressure sensor de-signed and prepared is able to accurately sense and detect the contact patterns of different English letters.The ResNet model construc-ted has excellent generalisation ability,and demonstrates high efficient application potential in the study of handwritten character recog-nition by flexible pressure sensor.