迎角快速校准及误差对控制系统的影响研究

摘要

基于迎角传感器数据校准,目前所采用的线性插值算法对于测量区域的数值处理存在精度丢失的问题,采用三次曲线拟合以及神经网络逼近的方法实现对已有测量数据的精确映射,避免了由于二次映射引起的精度损失,同时提高了对于未测量区域的预测能力,保证了控制系统的跟踪性能.将校准方法通过两余度迎角校准系统进行了验证,并将校准系统加入动态逆飞行控制系统进行仿真,结果显示了算法的精确性和可靠性.%Linear interpolation algorithm for angle of attack sensor data calibration has a problem of loss precision for handling the measurement area,cubic curve fitting and neural network approximation methods was used to achieve accurate mapping. The loss of precision due to the secondary mapping is avoid,the nolinear predictive ability of the unmeasured region was upgraded,also the tracking performance of the control system of was ensured.Each calibration method was validated and compared in the Dual-redundancy angle of attack calibration system,also the simulation result within Dynamic-Inversion flight control system shows the accuracy and reliability of the algorithm.