Design and Calibration of a Dual-Frame Force and Torque Sensor

In this study, a dual-frame six-axis force and torque (F/T) sensor is proposed. The dual-frame F/T sensor consists of inner and outer elastic parts for simultaneously measuring Fx, Fy, and Tzas well as Tx, Ty, and Fz. Load-strain equations were derived according to beam theory and structure theory. These equations and optimization methods were used to establish the parameter design process of the F/T sensor. A calibration machine is proposed to generate a large number of multiaxis training and testing F/T sets for sensor calibration. The calibration results from the least square method and neural network (NN) model revealed the coupling and nonlinearity of the multiaxis F/T sensor. The dual-frame structure reduced the coupling effect between the inner-frame and outer-frame measuring F/Ts. Therefore, the calibration process could be simplified using two tri-axis F/T sets with NN models. The calibration process was verified using six-axis testing F/T sets, and a mean calibration error of 0.25% was observed.