High-Sensitivity and Wide-Range Resistance Measurement Based on Self-Balancing Wheatstone Bridge and Gated Recurrent Neural Network

Abstract

A Wheatstone bridge and the following amplifier are widely used for measuring resistance. But for the measurement of small resistance change, the gain of the amplifier must be very high, which can reduce measurement range and introduce obvious baseline drift into the measured signal inevitably. In this study, a self-balancing Wheatstone bridge is designed based on a digitally controlled potentiometer (DCP). During the initial balance of the Wheatstone bridge, the DCP is adjusted until the output of the bridge is close to zero volts. When the bridge works for the measurement of dynamic resistance, the DCP is readjusted to compensate the remarkable change in the resistance, and the output of the amplifier is thus limited in the input voltage range of the ADC. At the same time, the signal sampled by the ADC is processed by the least-squares fitting to extend the measurement range. In addition, an attention-based gated recurrent unit (AGRU) is proposed to remove the drift in the signal. The effectiveness, stability, and accuracy of the proposed method have been demonstrated through multiple force measurement experiments on the instrument of a robot assisted minimally invasive surgery (RMIS) system.