Temperature Compensation for a Six-Axis Force/Torque Sensor Based on the Particle Swarm Optimization Least Square Support Vector Machine for Space Manipulator

Abstract

The output of a six-axis force/torque sensor (F/T sensor) not only varies with force or torque but also is affected by ambient temperature. In order to check the effects of temperature drift on the measurement precision of the sensor, this paper carried out experiments to obtain the quantitative results without loading the force and torque. In detail, three methods, including the least square method, radial basis function neural network, and least square support vector machine (LSSVM), were used to achieve the temperature compensation, showing that the LSSVM has the obvious advantage. However, for obtaining the optimization parameters of the LSSVM model, the particle swarm optimization (PSO) was adopted. Experimental results imply that the F/T sensor compensated by the PSO LSSVM has higher measurement precision and temperature stability.