基于进化人工神经网络的组合系统滤波技术(英文)

Integrated navigation system filter algorithms based on evolutionary artificial neural networks

在组合系统运用 Kalman 滤波器技术时,准确的系统模型和可靠的观测数据是保证其性能的重要因素,否则将大大降低 Kalman 滤波器的估计精度,甚至导致滤波器发散。为解决上述 Kalman 应用中的实际问题,提出了一种新颖的基于进化人工神经网络技术的自适应 Kalman 滤波器。仿真试验表明该算法可以在系统模型不准确时、甚至外部观测数据短暂中断时,仍能保证 Kalman 滤波器的性能。

The performance of Kalman filter in integrated navigation system depends on accurate system model and reliable observation data. Inaccurate system model or trustless observation data will cause low precision of Kalman filter, and even lead filter to divergence. So a new adaptive Kalman filter based on evolutionary artificial neural networks is used in this system. The algorithm is tested by simulations, and the results indicate that the proposed algorithm can efficiently overcome the shortcomings of traditional Kalman filter with better accuracy.