Flight test results of Observer/Kalman Filter Identifi[|#12#|]cation of the Pegasus unmanned vehicle
Flight test results of Observer/Kalman Filter Identifi[|#12#|]cation of the Pegasus unmanned vehicle
复制标题
Observer/Kalman Filter Identifi[| 的飞行测试结果
DOI：
--
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
J. Valasek
中科院分区：
文献类型：
--
作者：
Timothy Woodbury;Frank Arthurs;J. Valasek
飞行试验是获得精确的、局部线性的、非线性飞机动力学动态模型的首选方法。本文采用观测器/卡尔曼滤波辨识方法辨识了无人机纵向和横向/航向解耦线性动力学模型。该方法是一种时域技术，它从已知的输入和输出数据样本中提取离散的输入-输出映射。该方法是为光测试，包括仪器，测量和数据后处理技术，如非线性估计的细节。进行了多项八项试验，并给出了纵向和横向/方向动力学的试验示例，包括模型选择过程。通过比较测量和模型预测输出与八次测试的测量输入，验证了所识别的线性模型对非线性对象的保真度。均方误差和Theil信息系数被用作精度度量。结果表明，由八次试验结果再现的线性模型是巡航修正中飞机非线性动力学的可接受的表示。
Flight testing is the preferred means of obtaining accurate, locally linear, dynamic models of nonlinear aircraft dynamics. In this paper, decoupled longitudinal and lateral/directional linear dynamic models of an unmanned air vehicle are identied using the Observer/Kalman Filter Identication method. This method is a time-domain technique that identies a discrete input-output mapping from known input and output data samples. The method is developed for ight testing, including details of instrumentation, measurements, and data post-processing techniques such as nonlinear estimation. Multiple ight tests were conducted, and experimental examples for longitudinal and lateral/directional dynamics are presented, including the model selection process. Fidelity of the identied linear models to the nonlinear plant is validated by comparing measured and model predicted outputs with measured inputs from ight test. Mean squared errors and the Theil information coecient are used as accuracy metrics. Results presented in the paper demonstrate that the linear models reproduced from ight test results are acceptable representations of the nonlinear aircraft dynamics in the cruise conguration.