Nonlinear aircraft system identification using artificial neural networks enhanced by empirical mode decomposition

Aerodynamic parameter estimation involves modelling of force and moment coefficients and computation of stability and control derivatives from recorded flight data. This problem is extensively studied in the past using classical approaches such as output error, filter error and equation error methods. An alternative approach to these model based methods is the machine learning such as artificial neural network. In this paper, radial basis function neural network (RBF NN) is used to model the lateral-directional force and moment coefficients. The RBF NN is trained using k-means clustering algorithm for finding the centers of radial basis function and extended Kalman filter for obtaining the weights in the output layer. Then, a new method is proposed to obtain the stability and control derivatives. The first order partial differentiation is performed analytically on the radial basis function neural network approximated output. The stability and control derivatives are computed at each training data point, thus reducing the post training time and computational efforts compared to hitherto delta method and its variants. The efficacy of the identified model and proposed neural derivative method is demonstrated using real time flight data of ATTAS aircraft. The results from the proposed approach compare well with those from the other.

https://publications.drdo.gov.in/ojs/index.php/dsj/article/download/11843/6184

Aerodynamic Parameters Estimation Using Radial Basis Function Neural Partial Differentiation Method

In this paper, aerodynamic parameters have been estimated using neuro-fuzzy-based novel approach (ANFIS-Delta). ANFIS-Delta is an extension of a feed-forward neural network based Delta method. This...

ANFIS-Delta method for aerodynamic parameter estimation using flight data:

Purpose




The purpose of this paper is to build a neural model of an aircraft from flight data and online estimation of the aerodynamic derivatives from established neural model.




Design/methodology/approach




A neural model capable of predicting generalized force and moment coefficients of an aircraft using measured motion and control variable is used to extract aerodynamic derivatives. The use of neural partial differentiation (NPD) method to the multi-input-multi-output (MIMO) aircraft system for the online estimation of aerodynamic parameters from flight data is extended.




Findings




The estimation of aerodynamic derivatives of rigid and flexible aircrafts is treated separately. In the case of rigid aircraft, longitudinal and lateral-directional derivatives are estimated from flight data. Whereas simulated data are used for a flexible aircraft in the absence of its flight data. The unknown frequencies of structural modes of flexible aircraft are also identified as part of estimation problem in addition to the stability and control derivatives. The estimated results are compared with the parameter estimates obtained from output error method. The validity of estimates has been checked by the model validation method, wherein the estimated model response is matched with the flight data that are not used for estimating the derivatives.




Research limitations/implications




Compared to the Delta and Zero methods of neural networks for parameter estimation, the NPD method has an additional advantage of providing the direct theoretical insight into the statistical information (standard deviation and relative standard deviation) of estimates from noisy data. The NPD method does not require the initial value of estimates, but it requires a priori information about the model structure of aircraft dynamics to extract the flight stability and control parameters. In the case of aircraft with a high degree of flexibility, aircraft dynamics may contain many parameters that are required to be estimated. Thus, NPD seems to be a more appropriate method for the flexible aircraft parameter estimation, as it has potential to estimate most of the parameters without having the issue of convergence.




Originality/value




This paper highlights the application of NPD for MIMO aircraft system; previously it was used only for multi-input and single-output system for extraction of parameters. The neural modeling and application of NPD approach to the MIMO aircraft system facilitate to the design of neural network-based adaptive flight control system. Some interesting results of parameter estimation of flexible aircraft are also presented from established neural model using simulated data as a novelty. This gives more value addition to analyzing the flight data of flexible aircraft as it is a challenging problem in parameter estimation of flexible aircraft.

Aircraft neural modeling and parameter estimation using neural partial differentiation

System identification is fundamental to any observed dynamic process. This overview presentation brings out why and what system identification is as applied to flight vehicles. A historical background is briefly provided, tracing the developments to the 18th Century and early flight experiments to 1920s. This is followed by highlighting the unified Quad-M approach to flight vehicle system identification that has evolved over the last three decades. Several examples are presented, covering: 1) generation of complete aerodynamic databases, 2) estimation of nonlinear models for control surface effectiveness and stall hysteresis, 3) aerodynamic effects due to landing gear, and 4) Air drop. Some aspects of model validation are brought out. Finally, diversified applications to other flight vehicles are briefly elaborated to demonstrate the wide domain of applicability. The talk is concluded by indicating the possible future directions of work.

A Historical and Prospective Look at Aerodynamic Modeling and System Identification from Flight Data

In this paper, application of neural networks combined with partial differentiation of the neural outputs has been discussed to estimate lateral-directional flight stability and control parameter. A neural model capable of predicting generalized force and moment coefficients using measured motion and control variables can be employed to extract aerodynamic parameters from flight data. The Neural Partial Differentiation method is used for this purpose. The estimated results are compared with the parameter estimates obtained from Output Error Method. The validity of estimates has been verified by the model validation method, wherein the estimated model response is matched with the flight-test data that are not used for estimating the parameter.

Lateral-directional aerodynamics parameter estimation using neural partial differentiation

In a fairly complex system like aircraft, modeling and parameter estimation plays a crucial role in determining its stability and control characteristics.

Aircraft System Identification

Aircraft Aerodynamic Parameter Estimation from Flight Data Using Neural Partial Differentiation

The application of neural networks combined with partial differentiation of the neural outputs is discussed in this chapter to estimate lateral-directional flight stability and control derivatives from flight data. Primary investigation is carried out with simulated data and results are found to be encouraging to apply with flight data.

Vikalp Dongare

Papers

Lateral-directional aerodynamics parameter estimation using neural partial differentiation

Aircraft neural modeling and parameter estimation using neural partial differentiation

Aircraft System Identification

Aircraft Aerodynamic Parameter Estimation from Flight Data Using Neural Partial Differentiation

Identification of Aircraft Lateral-Directional Derivatives