论文部分内容阅读
针对飞行模拟器的时间延迟补偿问题,采用卡尔曼滤波和随机逼近的方法,提出了自适应遗忘因子卡尔曼补偿法和自适应遗忘因子随机逼近补偿法,这2种方法通过预测偏差动态调节遗忘因子的大小,可以提高预测精度并减小预测偏差;在2种补偿法的收敛矩阵初始值中引入调节参数,避免了初始值发生奇异现象;并且与改进的McFarland补偿方法进行了比较.结果表明,这2种方法的预测偏差分别小于相对应的改进的McFar-land补偿方法的预测偏差,且自适应遗忘因子卡尔曼补偿法是4种方法中最优的补偿方法.
Aiming at the time delay compensation problem of flight simulator, Kalman filter and random approximation method are used to propose adaptive forgetting factor Kalman compensation and adaptive forgetting factor stochastic approximation compensation. These two methods are dynamically forgotten by predicting deviation The size of the factor can improve the prediction accuracy and reduce the prediction error. The adjustment parameters are introduced into the initial values of the two kinds of compensation methods to avoid the singularity of the initial value. The results are compared with the improved McFarland compensation method. , Respectively. The predicted deviations of the two methods are less than those of the corresponding improved McFar-land compensation method, respectively. And the adaptive forgetting factor Kalman compensation method is the best among the four methods.