基于Stewart平台六维力传感器的分区静态标定方法

基于Stewart平台的六维力传感器具有结构紧凑、刚度大、量程宽等特点,它在工业机器人、空间站对接等领域具有广泛的应用前景。好的标定方法是正确使用传感器的基础。由于基于Stewart平台的六维力传感器是一个复杂的非线性系统,所以采用常规的线性标定方法必将带来较大的标定误差从而影响其使用性能。标定的实质是,由测量值空间到理论值空间的映射函数的确定过程。由函数逼近理论可知,当只在已知点集上给出函数值时,可用多项式或分段多项式等较简单函数逼近待定函数。基于上述思想,本文将整个测量空间划分为若干连续的子测量空间,再对每个子空间进行线性标定,从而提高了整个测量系统的标定精度。实验分析结果表明了该标定方法有效。

Owing to the excellent characteristic of compact structure, high stiffness and wide-ranged measurement, the six-axis force sensor based on the stewart platform gets its widely application in many research areas such as industrial robot and spacecraft’s docking process. The realization of high-precision measurement for the sensor bears a key problem of calibration method. Since the six-axis force sensor based on the platform of STEWART has complexity as a nonlinear system, its performance is decreased for the calibration error introduced when the traditional methods of calibration are adopted. This paper analyses the essentiality of calibration, which is a process of determining the mapping function from measure space to theory space. The undetermined function can be approximated, based on the function approximation theory, with some simple functions such as polynomial or piecewise polynomial when the function values are known on the defined point set. Based on the analysis above, the paper proposed a novel multi-point linear calibration algorithm. The algorithm, adopting the one order linear function as piecewise polynomial, determines the mapping function of the six-axis force sensor from measure space to theory space by treating the measure values as function variable and actual loaded values as the corresponding function values. The experiment results demonstrate that the proposed method of calibration can improve the precision in measurement and be realized in real-time, so the algorithm is practical in the practice application of measurement.

国家自然科学基金(50275143)资助课题。