论文部分内容阅读
对于带不确定噪声方差的多传感器单通道自回归滑动平均(ARMA)信号系统,当观测噪声中包含白噪声和一个自回归滑动平均(ARMA)有色观测噪声时,通过增广状态方法把ARMA信号系统模型转化为状态空间模型.应用加权最小二乘法和极大极小鲁棒估计准则,基于带噪声方差保守上界的最坏保守系统,提出了鲁棒加权观测融合稳态Kalman信号预报器.对于噪声方差的所有可能的不确定性,它们的实际预报误差方差保证有相应的最小上界.应用Lyapunov方程方法,证明了局部和加权观测融合稳态Kalman信号预报器的鲁棒性和鲁棒精度关系.通过一个仿真例子验证了所提出理论结果的正确性和有效性.
For multi-sensor single-channel autoregressive moving average (ARMA) signal system with uncertain noise variance, when the observed noise contains white noise and an autoregressive moving average (ARMA) colored observation noise, the ARMA signal The system model is transformed into the state space model.Weighted robust least squares (LSLS) and maximally minimal robust estimation criteria are used to construct the robust weighted observation fusion steady-state Kalman signal predictor based on the worst conservative system with upper bound of noise variance. For all possible uncertainties of noise variance, their actual forecast error variance is guaranteed to have a corresponding minimum upper bound.Applying the Lyapunov equation method, the robustness and robustness of the locally and weighted observations fusion steady-state Kalman signal predictor are proved The relationship between accuracy and accuracy is verified by a simulation example to verify the validity and validity of the proposed theoretical result.