Aircraft parameter estimation using ELM network

2020 ◽  
Vol 92 (6) ◽  
pp. 895-907
Author(s):  
Hari Om Verma ◽  
Naba Kumar Peyada
Keyword(s):  
Neural Network ◽  
Parameter Estimation ◽  
Mathematical Formulation ◽  
Estimation Methods ◽  
Content Type ◽  
Flight Data ◽  
The Real ◽  
Aerodynamic Parameters ◽  
Hidden Layer ◽  
Error Method

Purpose The purpose of this paper is to investigate the estimation methodology with a highly generalized cost-effective single hidden layer neural network. Design/methodology/approach The aerodynamic parameter estimation is a challenging research area of aircraft system identification, which finds various applications such as flight control law design and flight simulators. With the availability of the large database, the data-driven methods have gained attention, which is primarily based on the nonlinear function approximation using artificial neural networks. A novel single hidden layer feed-forward neural network (FFNN) known as extreme learning machine (ELM), which overcomes the issues such as learning rate, number of epochs, local minima, generalization performance and computational cost, as encountered in the conventional gradient learning-based FFNN has been used for the nonlinear modeling of the aerodynamic forces and moments. A mathematical formulation based on the partial differentiation is proposed to estimate the aerodynamic parameters. Findings The real flight data of longitudinal and lateral-directional motion have been considered to estimate their respective aerodynamic parameters using the proposed methodology. The efficacy of the estimates is verified with the results obtained through the conventional parameter estimation methods such as the equation-error method and filter-error method. Originality/value The present study is an outcome of the research conducted on ELM for the estimation of aerodynamic parameters from the real flight data. The proposed method is capable to estimate the parameters in the presence of noise.

Aerodynamic coefficients modeling using Levenberg–Marquardt algorithm and network

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Zhigang Wang ◽  
Aijun Li ◽  
Lihao Wang ◽  
Xiangchen Zhou ◽  
Boning Wu
Keyword(s):  
Neural Network ◽  
Particle Swarm ◽  
Particle Swarm Algorithm ◽  
Aerodynamic Parameter ◽  
Content Type ◽  
Flight Data ◽  
Output Error ◽  
Aerodynamic Derivatives ◽  
Levenberg Marquardt ◽  
Error Method

Purpose The purpose of this paper is to propose a new aerodynamic parameter estimation methodology based on neural network and output error method, while the output error method is improved based on particle swarm algorithm. Design/methodology/approach Firstly, the algorithm approximates the dynamic characteristics of aircraft based on feedforward neural network. Neural network is trained by extreme learning machine, and the trained network can predict the aircraft response at (k + 1)th instant given the measured flight data at kth instant. Secondly, particle swarm optimization is used to enhance the convergence of Levenberg–Marquardt (LM) algorithm, and the improved LM method is used to substitute for the Gauss Newton algorithm in output error method. Finally, the trained neural network is combined with the improved output error method to estimate aerodynamic derivatives. Findings Neither depending on the initial guess of the parameters to be estimated nor requiring numerical integration of the aircraft motion equation, the proposed algorithm can be used for unstable aircraft and is successfully applied to extract aerodynamic derivatives from both simulated and real flight data. Research limitations/implications The proposed method requires iterative calculation and can only identify parameters offline. Practical implications The proposed method is successfully applied to estimate aircraft aerodynamic parameters and can also be used as a new algorithm for other optimization problems. Originality/value In this study, the output error method is improved to reduce the dependence on the initial value of parameters and expand its application scope. It is applied in aircraft aerodynamic parameter identification together with neural network.

Parameter estimation of UAV from flight data using neural network

2018 ◽  
Vol 90 (2) ◽  
pp. 302-311 ◽  
Author(s):  
Dhayalan R. ◽  
Subrahmanyam Saderla ◽  
Ajoy Kanti Ghosh
Keyword(s):  
Parameter Estimation ◽  
Design Methodology ◽  
Maximum Likelihood Method ◽  
Estimation Method ◽  
Likelihood Method ◽  
Estimation Methods ◽  
Process Noise ◽  
Content Type ◽  
Flight Data ◽  
Aerial Vehicle

Purpose The purpose of this paper is to present the application of the neural-based estimation method, Neural-Gauss-Newton (NGN), using the real flight data of a small unmanned aerial vehicle (UAV). Design/methodology/approach The UAVs in general are lighter in weight and their flight is usually influenced by the atmospheric winds because of their relatively lower cruise speeds. During the presence of the atmospheric winds, the aerodynamic forces and moments get modified significantly and the accurate mathematical modelling of the same is highly challenging. This modelling inaccuracy during parameter estimation is routinely treated as the process noise. Furthermore, because of the limited dimensions of the small UAVs, the measurements are usually influenced by the disturbances caused by other subsystems. To handle these measurement and process noises, the estimation methods based on neural networks have been found reliable in the manned aircrafts. Findings Six sets of compatible longitudinal flight data of the designed UAV have been chosen to estimate the parameters using the NGN method. The consistency in the estimates is verified from the obtained mean and the standard deviation and the same has been validated by the proof-of-match exercise. It is evident from the results that the NGN method was able to perform on a par with the conventional maximum likelihood method. Originality/value This is a partial outcome of the research carried out in estimating parameters from the UAVs.

Parameter estimation of aircraft using extreme learning machine and Gauss-Newton algorithm

2019 ◽  
Vol 124 (1272) ◽  
pp. 271-295 ◽  
Author(s):  
H. O. Verma ◽  
N. K. Peyada
Keyword(s):  
Neural Network ◽  
Extreme Learning Machine ◽  
Computational Cost ◽  
Feedforward Neural Network ◽  
Estimation Methods ◽  
Flight Data ◽  
Regression Methods ◽  
Directional Motion ◽  
Aerodynamic Parameters ◽  
Learning Machine

ABSTRACTThe research paper addresses the problem of estimating aerodynamic parameters using a Gauss-Newton-based optimisation method. The process of the optimisation method lies on the principle of minimising the residual error between the measured and simulated responses of the system. Usually, the simulated response is obtained by integrating the dynamic equations of the system, which is found to be susceptible to the initial values, and the integration method. With the advent of the feedforward neural network, the data-driven regression methods have been widely used for identification of the system. Among them, a variant of feedforward neural network, extreme learning machine, which has proven the performance in terms of computational cost, generalisation, and so forth, has been addressed to predict the responses in the present study. The real flight data of longitudinal and lateral-directional motion have been considered to estimate their respective aerodynamic parameters. Furthermore, the estimates have been validated with the values of the classical estimation methods, such as the equation-error and filter-error methods. The sample standard deviations of the estimates demonstrate the effectiveness of the proposed method. Lastly, the proof-of-match exercise has been conducted with the other set of flight data to validate the estimated parameters.

Aircraft parameter estimation using a new filtering technique based upon a neural network and Gauss-Newton method

2009 ◽  
Vol 113 (1142) ◽  
pp. 243-252 ◽  
Author(s):  
N. K. Peyada ◽  
A. K. Ghosh
Keyword(s):  
Neural Network ◽  
Parameter Estimation ◽  
Dynamic Model ◽  
Newton Method ◽  
Initial Conditions ◽  
Estimation Method ◽  
Wind Tunnel Test ◽  
Neural Model ◽  
Estimation Methods ◽  
Flight Data

Abstract A new parameter estimation method based upon neural network is proposed. The method proposed here uses feed forward neural networks to establish a neural model that could be used to predict subsequent time histories given the suitable measured initial conditions. The proposed neural model would not represent a generic flight dynamic model. The neural model in this case develops point to point fitting of the input and the output data. Thus, it could at best be referred to as flight dynamic model in restricted sense. Gauss-Newton method is then used to obtain optimal values of the aerodynamic parameters by minimising a suitable defined error cost function. The method has been validated using longitudinal and lateral-directional flight data of various test aircraft. The results thus obtained were compared with those obtained through wind tunnel test, or those obtained using Maximum likelihood and/or Filter error methods. Unlike, most of the parameter estimation methods, the proposed method does not require a prior description of the model. It also bypasses the requirement of solving equations of motion. This feature of the proposed method may have special significance in handling flight data of an unstable aircraft.

Bio-inspired neural network with application to license plate recognition: hysteretic ELM approach

2016 ◽  
Vol 36 (2) ◽  
pp. 172-178 ◽  
Author(s):  
Liang Chen ◽  
Leitao Cui ◽  
Rong Huang ◽  
Zhengyun Ren
Keyword(s):  
Neural Network ◽  
Activation Function ◽  
Neural System ◽  
Learning System ◽  
License Plate ◽  
Sigmoid Function ◽  
License Plate Recognition ◽  
Content Type ◽  
Hidden Layer ◽  
Automatic License Plate Recognition

Purpose This paper aims to present a bio-inspired neural network for improvement of information processing capability of the existing artificial neural networks. Design/methodology/approach In the network, the authors introduce a property often found in biological neural system – hysteresis – as the neuron activation function and a bionic algorithm – extreme learning machine (ELM) – as the learning scheme. The authors give the gradient descent procedure to optimize parameters of the hysteretic function and develop an algorithm to online select ELM parameters, including number of the hidden-layer nodes and hidden-layer parameters. The algorithm combines the idea of the cross validation and random assignment in original ELM. Finally, the authors demonstrate the advantages of the hysteretic ELM neural network by applying it to automatic license plate recognition. Findings Experiments on automatic license plate recognition show that the bio-inspired learning system has better classification accuracy and generalization capability with consideration to efficiency. Originality/value Comparing with the conventional sigmoid function, hysteresis as the activation function enables has two advantages: the neuron’s output not only depends on its input but also on derivative information, which provides the neuron with memory; the hysteretic function can switch between the two segments, thus avoiding the neuron falling into local minima and having a quicker learning rate. The improved ELM algorithm in some extent makes up for declining performance because of original ELM’s complete randomness with the cost of a litter slower than before.

Estimation of lateral-directional aerodynamic derivatives from flight data using conventional and neural based methods

2014 ◽  
Vol 118 (1210) ◽  
pp. 1453-1479 ◽  
Author(s):  
R. Kumar ◽  
A. K. Ghosh
Keyword(s):  
Mathematical Model ◽  
Parameter Estimation ◽  
A Priori ◽  
Estimation Methods ◽  
Control Input ◽  
Feed Forward Neural Network ◽  
Additional Advantage ◽  
Flight Data ◽  
Aerodynamic Derivatives ◽  
Different Types

Abstract The paper presents the estimation of lateral-directional aerodynamic derivatives (parameters) using conventional and neural based methods from real flight data of Hansa-3 aircraft. The conventional methods such as least squares (LS) and maximum likelihood (ML) require an a priori postulation of mathematical model to estimate the parameters. Whereas the neural-based method such as Neural-Gauss-Newton (NGN) is an algorithm that utilises feed forward neural network and Gauss-Newton optimisation to estimate the parameters and does not require any a priori postulation of mathematical model or solution of equations of motion. In the paper, the LS, ML and NGN methods are applied to lateral-directional flight data in order to estimate parameters. The results obtained in terms of lateral-directional aerodynamic derivatives are reasonably accurate to establish LS, ML and NGN as parameter estimation methods along with NGN method having an additional advantage of non-requirement of a priori mathematical model. The paper also highlights the effect of different types of control inputs on parameter estimation. For this, three types of control inputs were used to generate real flight data. The ailerons and rudder were deflected in the first, the ailerons were deflected while keeping rudder at trim condition in the second and the rudder was deflected while keeping ailerons at trim condition in the third type of control input to generate the real flight data. The paper presents the effect of three different types of control inputs in terms of aerodynamic parameters estimated through conventional and neural based methods using flight data generated through these inputs.

Longitudinal parameter estimation from real flight data of unmanned cropped delta flat plate configuration

2016 ◽  
Vol 4 (1) ◽  
pp. 2-22 ◽  
Author(s):  
Subrahmanyam Saderla ◽  
Dhayalan R ◽  
Ajoy Kanti Ghosh
Keyword(s):  
Parameter Estimation ◽  
Maximum Likelihood ◽  
Rectangular Cross Section ◽  
Flight Test ◽  
Least Square ◽  
Likelihood Method ◽  
Content Type ◽  
Flight Data ◽  
Aerodynamic Model ◽  
Estimated Parameters

Purpose – The purpose of this paper is to describe the longitudinal aerodynamic characterization of an unmanned cropped delta configuration from real flight data. In order to perform this task an unmanned configuration with cropped delta planform and rectangular cross-section has been designed, fabricated, instrumented and flight tested at flight laboratory in Indian Institute of Technology Kanpur (IITK), India. Design/methodology/approach – As a part of flight test program a real flight database, through various maneuvers, have been generated for the designed unmanned configuration. A dedicated flight data acquisition system, capable of onboard logging and telemetry to ground station, has been used to record the flight data during these flight test experiments. In order to identify the systematic errors in the measurements, the generated flight data has been processed through data compatibility check. Findings – It is observed from the flight path reconstruction that the obtained biases are negligible and the scale factors are almost close to unity. The linear aerodynamic model along with maximum likelihood and least-square methods have been used to perform the parameter estimation from the obtained compatible flight data. The lower values of Cramer-Rao bounds obtained for various parameters has shown significant confidence in the estimated parameters using maximum likelihood method. In order to validate the aerodynamic model used and to increase the confidence in the estimated parameters a proof-of-match exercise has been carried out. Originality/value – The entire work presented is original and all the experiments have been carried out in Flight laboratory of IITK.

Solving two-dimensional linear partial differential equations based on Chebyshev neural network with extreme learning machine algorithm

2020 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Min Liu ◽  
Muzhou Hou ◽  
Juan Wang ◽  
Yangjin Cheng
Keyword(s):  
Neural Network ◽  
Partial Differential Equations ◽  
Extreme Learning Machine ◽  
Two Dimensional ◽  
Content Type ◽  
Linear Partial Differential Equations ◽  
Chebyshev Neural Network ◽  
Learning Machine ◽  
Hidden Layer ◽  
Linear Pdes

Purpose This paper aims to develop a novel algorithm and apply it to solve two-dimensional linear partial differential equations (PDEs). The proposed method is based on Chebyshev neural network and extreme learning machine (ELM) called Chebyshev extreme learning machine (Ch-ELM) method. Design/methodology/approach The network used in the proposed method is a single hidden layer feedforward neural network. The Kronecker product of two Chebyshev polynomials is used as basis function. The weights from the input layer to the hidden layer are fixed value 1. The weights from the hidden layer to the output layer can be obtained by using ELM algorithm to solve the linear equations established by PDEs and its definite conditions. Findings To verify the effectiveness of the proposed method, two-dimensional linear PDEs are selected and its numerical solutions are obtained by using the proposed method. The effectiveness of the proposed method is illustrated by comparing with the analytical solutions, and its superiority is illustrated by comparing with other existing algorithms. Originality/value Ch-ELM algorithm for solving two-dimensional linear PDEs is proposed. The algorithm has fast execution speed and high numerical accuracy.

Aircraft neural modeling and parameter estimation using neural partial differentiation

2018 ◽  
Vol 90 (5) ◽  
pp. 764-778 ◽  
Author(s):  
Majeed Mohamed ◽  
Vikalp Dongare
Keyword(s):  
Parameter Estimation ◽  
Simulated Data ◽  
Neural Model ◽  
Flexible Aircraft ◽  
Content Type ◽  
Flight Data ◽  
Stability And Control ◽  
Aerodynamic Derivatives ◽  
Aircraft System ◽  
And Control

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.

Sign in / Sign up

Export Citation Format

Share Document