Combined Use of FSR Sensor Array and SVM Classifier for Finger Motion Recognition Based on Pressure Distribution Map

Nan Li,Dapeng Yang,Li Jiang,Hong Liu,Hegao Cai
DOI: https://doi.org/10.1016/s1672-6529(11)60095-4
IF: 2.995
2012-01-01
Journal of Bionic Engineering
Abstract:For controlling dexterous prosthetic hand with a high number of active Degrees of Freedom (DOF), it is necessary to reliably extract control volitions of finger motions from the human body. In this study, a large variety of finger motions are discriminated based on the diversities of the pressure distribution produced by the mechanical actions of muscles on the forearm. The pressure distribution patterns corresponding to the motions were measured by sensor array which is composed of 32 Force Sensitive Resistor (FSR) sensors. In order to map the pressure patterns with different finger motions, a multiclass classifier was designed based on the Support Vector Machine (SVM) algorithm. The multi-subject experiments show that it is possible to identify as many as seventeen different finger motions, including individual finger motions and multi-finger grasping motions, with the accuracy above 99% in the in-session validation. Further, the cross-session validation demonstrates that the performance of the proposed method is robust for use if the FSR array is not reset. The results suggest that the proposed method has great application prospects for the control of multi-DOF dexterous hand prosthesis.
What problem does this paper attempt to address?