Year: 2025
Type: article
Abstract: Gesture recognition utilizing flexible strain sensors is widely applied in human–machine interfaces. However, it is limited to strict environmental requirements and inadequate recognition accuracy. Herein, a flexible strain sensor is developed based on microfluidics using a back propagation (BP) neural network for highly accurate gesture recognition. The microfibers are achieved with stretchable polyurethane shells and conductive liquid metal cores in a convenient, safe, and effective way. Further, by combining 3D printing with microfluidics, the flexible strain sensor based on microfiber units is constructed. Because of the great advantages in mechanical and electrical performance of microfiber units, the achieved sensor shows impressive stretching and conductivity. When applied to the finger, the sensor converts the bending deformation into electrical signals directly. Thus, a set of sensors is attached to five fingers to record corresponding signals of different gestures, which can be further analyzed for gesture recognition. To improve the accuracy, flexible strain sensors are combined with a BP neural network. With electrical signals introduced into the input layer as the characteristic values of training set, high‐precision gesture recognition is achieved by updating the weights through training. This strategy has broad application potential in biomedicine, aerospace, and other fields.
Source: Advanced Intelligent Systems
Authors: Yikai Wu, Jinpeng Wang, Jinpeng Wang, Can Wang, Linlin Xia, Jie Wang, Jie Wang
Institution: Nanjing Agricultural University
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,970
Awards: G2955953321, G829825560
OpenAlex ID: W4409517158