基于遗传算法和Elman神经网络的接触式探头动态特性补偿

Dynamic Compensation for Contact Probe Based on Genetic Algorithm and Elman Neural Network

动态特性不理想是接触式探头系统动态测量误差的重要来源,严重制约探头测量速度和精度的提升。提出一种基于遗传算法优化Elman神经网络的探头动态特性补偿方法。针对微纳米接触式探头,采用遗传算法优化Elman神经网络的方法对其动态响应输出信号进行了补偿,使用自适应递推最小二乘方法辨识出补偿前后的探头系统动态模型。探头系统的动态测量不确定度由补偿前的77.8 nm减小至12.1 nm。遗传算法具有较好的全局搜索能力,克服了Elman神经网络容易陷入局部极值的缺陷,该动态补偿方法具有较快的网络训练速度和较高的动态补偿精度。仿真分析及不确定度评定结果都验证了该方法的有效性。

The undesired dynamic characteristic is an important source for dynamic measurement errors of contact probe systems,which greatly restricts the improvement of measurement speed and precision.A dynamic compensation method based on genetic algorithm(GA)and Elman neural network(ENN)is presented to compensate the dynamic characteristics of probes.The genetic algorithm is used to optimize the ENN method to dynamically compensate the output signal of dynamic response.The adaptive recursive least-square method is used to identify the dynamic models of the probe system before and after compensation.The dynamic measurement uncertainty of the probe system is reduced from 77.8 nm to 12.1 nm.The global search ability of GA is utilized to overcome ENN’s shortcoming of easy convergence to the local extreme values.This method has fast network training speed and high dynamic compensation precision.The effectiveness of this method is verified by the simulation analysis and the uncertainty evaluation results.