Identification of the Relationships Between Noncontact Capacitive Sensing Signals and Continuous Grasp Forces: Preliminary Study.

Enhao Zheng,Qining Wang,Hong Qiao
DOI: https://doi.org/10.1109/embc.2018.8513251
2018-01-01
Abstract:This study explores the relationships between noncontact capacitive sensing signals and continuous grasp forces. It is a crucial step towards the volitional control of robotic systems based on the noncontact sensing approach. We firstly designed a measurement system including the capacitive sensing front-ends, the grasp force sensor, the signal sampling circuits and the graphic user interface. The capacitive sensing front-end was specifically designed for human forearm signal sampling, which was worn outside of the clothes. After implementation of the system, we carried out experiments on five healthy subjects, and the sensing bands were customized with their arm shapes. The grasp force and the capacitance signals were record simultaneously when the subjects gradually increased the force according to instruction. Linear regression and quadratic regression were used to evaluate the regulated signals. For each subject, at least one channel of capacitance signals were linear correlated to the normalized grasp force with R 2 ≥ 0.85. We found there was inter-subject similarity on the capacitance-force relationships. Cross validation on grasp force estimation with capacitance signals were also carried out, and the average relative estimation error was about 18%. The results proved the feasibility of the noncontact capacitive sensing method for human joint force estimation.
What problem does this paper attempt to address?