Temperature compensation for six-dimension force/torque sensor based on Radial Basis Function Neural Network

Not only output of the six-dimension force/torque sensor changes with force or torque, but also be susceptible to ambient temperature, thus limiting measurement accuracy of the sensor. In order to overcome the above drawbacks of six-dimension force/torque sensor, this paper proposes a temperature compensation method based on Radial Basis Function (RBF) Neural Network. Compared with the conventional least squares method (LSM), RBF Neural Network has advantage obviously in compensating temperature drift for output nonlinear problems. Therefore, this method can eliminate the influence temperature drift of the sensor effectively. Examples show that the six-dimension force/torque sensor compensated by RBF has higher measurement precision and temperature stability.