基于滑动窗滤波与卡尔曼滤波的信息融合

Application of Window Space and Kalman Filter in Information Fusion

波音737MAX-8客机空难事故分析是:电脑系统是根据一个攻角传感器的错误数据在运行,机头越高,攻角越大,飞机失速坠毁[1]。为增强飞行器的安全性、可靠性,通常在机上采用传感器信号多余度设计。常规的“三判二”方法无法综合有效地利用余度信号,因此本文针对独立不相干的信号,采用窗函数滤波器、置信度动态加权平均方法、卡尔曼滤波技术,提出并设计了一种既能满足时域离散系统、过滤信号中的高斯噪声,又能诊断并屏蔽失效传感器、人工主动选择,将多路数据信号融合一体的方法。

The analysis of two air crashes of Boeing 737MAX-8 airliner is: the computer system is running according to the wrong data of an angle of attack sensor. In order to enhance the safety and reliability of flight,the sensor signal redundancy design is usually adopted on the aircraft. The conventional “three-judgment-two”method cannot comprehensively and effectively utilize the redundant signals. Therefore, this paper proposes and designs a method for independent and incoherent signals using window function filter, dynamic weighted average method of confidence and Kalman filter technology. It can not only satisfy the time-domain discrete system, filter the Gaussian noise in the signal, but also diagnose and shield the failed sensor, manually select and integrate the multi-channel data signal into one.