基于RF-GA的六维力传感器解耦方法

多维力传感器的维间耦合问题严重影响了检测精度的提高。通过设计新型RF-GA(基于遗传算法的改进随机森林算法)解耦方法解决多维力信息的解耦问题,实现提高力传感器检测精度的目标。针对随机森林算法中含有大量子树,但每个子树的预测准度无法保证的问题,利用遗传算法对随机森林的子树进行筛选,保留优质子树,从而提高预测精度。以基于应变检测的六维力传感器为实验对象,将RF-GA算法运用到实际力信息解耦中,并通过解耦实验对RF-GA算法进行验证。与现有解耦算法相比,RF-GA解耦方法具有精度高、解耦时间短的优点,实验结果表明该算法能有效提高多维力传感器的解耦精度。

Decoupling Method of Six-Axis Force Sensor Based on RF-GA

The inter-dimension coupling of multi-axis force sensor seriously affects the improvement of detection precision.A novel random forest algorithm based on genetic algorithm (RF-GA) is proposed to decouple the multi-axis force sensor and improve the measurement performance.The normal random forest contains a large number of subtrees,but the accuracy of each subtree cannot be guaranteed.Therefore,the genetic algorithm (GA) was used to select high-quality subtrees of the random forest to improve prediction accuracy.Taking the six-axis force sensor based on strain detection as the experimental object,the RF-GA was applied to the actual force information decoupling,and was verified by the decoupling experiment.Compared with other traditional decoupling algorithms,RF-GA based on decoupling method has the advantages of high precision and short decoupling time.Experimental results show that this method can improve the decoupling performance.