Aircraft System Identification Research Articles

Estimation of aerodynamic parameters using neural artificial bee colony fusion algorithm for moderate angle of attack using real flight data

Aircraft system identification aims to estimate the aerodynamic force and moment coefficients utilizing intelligent modeling and parametric identification methodologies. Classical methods like output, filter, and equation error methods apply extensively as parametric approaches. In contrast, machine learning approaches like Artificial Neural Networks (ANN), Adaptive Neuro-Fuzzy Inference Systems (ANFIS), etc., are alternatives to model-based methods. This work presents a novel aerodynamic parameters estimation technique that fuses two biologically inspired optimization techniques, (i) the Artificial Bee Colony (ABC) optimization and (ii) ANN for an actual aircraft while incorporating system and measurement uncertainty. The fusion of ABC and ANN imparts the ability to address sensor noise challenges associated with system identification and parameter estimation. Comparison of the proposed method’s results with the benchmark techniques like Least Square, Filter Error, and Neural Gauss Methods using recorded flight data of the ATTAS (DLR German Aerospace Centre) and HANSA-3 (IIT Kanpur) aircrafts established its adequacy and efficacy. Furthermore, the capability of the proposed hybrid method to extract stability and control variables from the stable aircraft kinematics is shown even with insufficient information in its data history.

Advancements in the Theory and Practice of Flight Vehicle System Identification

System identification methods have played an essential role in the research and industry projects at the Institute of Flight System Dynamics of the Technical University of Munich. Besides the application of the established system identification approaches to multiple aircraft, novel methods have been developed at the institute to deal with the challenges associated with fixed- and rotary-wing aircraft system identification in the time and frequency domains. This paper provides an overview of these new developments, as well as practical application examples from the past and ongoing projects at the institute. After a brief introduction to the work at the institute, the paper describes the new advancements in the fields of time domain system identification and optimal input design by introducing optimal control methods. It continues with an alternative problem formulation for applying frequency domain system identification and parameter estimation methods to a novel flight control design and tuning approach. Additional topics from the past and ongoing research at the institute such as novel methods and practical findings in rotorcraft system identification and flight path reconstruction for different applications will be discussed and referenced.

Flight-Test System Identification Techniques and Applications for Small, Low-Cost, Fixed-Wing Aircraft

This paper provides an overview of flight-test system identification methods applied in the Virginia Tech Nonlinear Systems Laboratory that focus on modeling small, inexpensive, fixed-wing aircraft controlled by a ground-based pilot. The general aircraft system identification approach is outlined with details provided on the flight-test facilities, experiment design methods, instrumentation systems, flight-test operations, data processing techniques, and model identification methods enabling small aircraft flight dynamics model development. Specific small aircraft system identification challenges are overcome, including low-cost control surface servo-actuators and instrumentation systems, as well as a greater sensitivity to atmospheric disturbances and limited piloting cues. Four recent system identification research advancements using the general system identification process are featured, including application of uncorrelated pilot inputs for remotely piloted aircraft, aero-propulsive model development for propeller aircraft, spin aerodynamic model development, and nonlinear dynamic modeling without mass properties information. Although this paper provides a summary of several research efforts, the core system identification approach is presented with sufficient detail to allow the methods to be readily adapted to other research efforts leveraging small, low-cost aircraft.

Advances and Modern Applications of Frequency-Domain Aircraft and Rotorcraft System Identification

This paper provides an overview of advances and applications of aircraft and rotorcraft flight dynamics system identification at the U.S. Army Combat Capabilities Development Command Aviation & Missile Center over the past 20 years. An overarching topic is how frequency-domain system identification has evolved to aid modeling work of modern aircraft and rotorcraft configurations. The paper starts with a brief overview of the frequency-response method for system identification. Subsequently, the paper highlights and demonstrates several new advances using flight-test data from a conventional helicopter, fixed-wing aircraft, advanced coaxial compound rotorcraft, and a small multirotor unmanned aerial vehicle. The first advancement is the use of identification results to improve simulation model fidelity. The next advancement is the Joint Input-Output Method for identification of aircraft with highly correlated inputs. Next, the paper discusses the advancement of flexible constraints, which provide the ability to directly determine physical parameters from flight-test data. The paper then discusses a methodology to identify lower-order linearized dynamic inflow models from higher-order free wake aerodynamic models. Finally, the paper discusses model stitching, which is the technique of combining or stitching together individual linear models and trim data for discrete flight conditions to produce a continuous, full flight-envelope real-time stitched simulation model.

Unstable tilt-rotor maximum likelihood wavelet-based identification from flight test data

PurposeThe purpose of this paper is to present the methodology that was used to perform system identification of a dynamically unstable tilt-rotor from flight test data. The method incorporated wavelet transform into the maximum likelihood principle formulation, emphasizing both time and frequency responses. Using wavelets allowed to additionally filter noise in the data, and this increased the estimation quality. This approach did not require measurement and process noise modeling in contrast to the Kalman filter usage for parameter estimation.Design/methodology/approachIn the study, lateral-directional stability and control derivatives of an unstable tiltrotor in hover were estimated. This was performed by applying the maximum likelihood output error method. The estimated model response was decomposed using the Mallat pyramid and matched to wavelet coefficients obtained directly from measurements. In addition, a coherence-based weighting function was used to put more emphasis on the most reliable data. For comparison, the same set of data was used to identify a model with the same structure using the maximum likelihood principle with an incorporated Kalman filter.FindingsIt was found that maximum likelihood principle and wavelet transform allowed for estimating aerodynamic coefficients of a dynamically unstable aircraft. The estimation was performed with high accuracy.Practical implicationsThe designed method can be used for system identification of unstable aircraft and when additional noise is present (e.g. when noise due to turbulence was observable during the flight test or higher noise levels were present in the sensors data).Originality/valueThe paper presents verification of a wavelet-based maximum likelihood principle output error method using flight test data.

Retrospective and Recent Examples of Aircraft and Rotorcraft System Identification at DLR

Aircraft system identification has a five-decades-long tradition at German Aerospace Center (DLR). Over the last two decades, the research covered various topics related to system identification of fixed- and rotary-wing aircraft, nonconventional applications and atmospheric effects, the development of new flight-test procedures for system identification purposes, and specific aircraft model enhancements and corresponding parameter estimation. Comprehensive tools were developed that support this research and can be applied to a variety of different problems and types of vehicles. The paper starts with a short description of the different system identification methods used at DLR and the corresponding tools. The discussion of flight-test procedures and maneuver design as well as sensor fusion and flight-path reconstruction provides information on how to optimize the flight tests for system identification and to arrive at a consistent flight-test database. The examples for fixed-wing aircraft provide information on identification including abnormal conditions such as icing and interaction with atmospheric disturbances as well as modeling of structural mechanics and loads. The identification of high-order rotorcraft models that account for rotor and engine dynamics and even structural modes is discussed, and the identification of rotor mast moments as well as the identification of non-physics-based models and their integration into physics-based models are also covered. A final section shows that system identification can also be used to derive models for gyroplanes and parachutes as well as to derive control equivalent turbulence input models and to estimate complex wind field geometries. Thus, a broad overview of possible applications of system identification is given.

Advances in Aircraft System Identification at NASA Langley Research Center

Advances in aircraft system identification at NASA Langley Research Center are discussed. The relevant time period includes the years since the last summary paper of this kind, which was published in the Journal of Aircraft in 2005. Research advances were achieved in flight test experiment design, frequency-domain modeling, real-time autonomous global modeling, rapid simulation development and updating, dynamic modeling in turbulence, flight data corrections, model uncertainty characterization, and aeroelastic modeling using distributed sensing. Possible future developments in the field are identified.

Direct data driven safety control for aircraft flight system-Part 1: the theory case

Purpose This new paper aims to combine the recent new contributions about direct data driven control and other safety property to form an innovative direct data driven safety control for aircraft flight system. More specifically, within the framework of direct data driven strategy, the collected data are dealt with to get the identified plant and designed controller. After reviewing some priori information about aircraft flight system, a closed loop system with the unknown plant and controller simultaneously is considered. Data driven estimation is proposed to identify the plant and controller only through the ratios of two correlation functions, computed from the collected data. To achieve the dual missions about perfect tracking and safety property, a new notion about safety controller is introduced. To design this safety controller, direct data driven safety controller is proposed to solve one constrain optimization problem. Then the authors apply the Karush–Kuhn–Tucker (KKT) optimality conditions to derive the explicit safety controller. Design methodology approach First, consider one closed loop system corresponding to aircraft flight system with the unknown plant and feed forward controller, data driven estimation is used to identify the plant and feed forward controller. This identification process means nonparametric estimation. Second, to achieve the perfect tracking one given transfer function and guarantee the closed loop output response within one limited range simultaneously, safety property is introduced. Then direct data driven safety control is proposed to design the safety controller, while satisfying the dual goals. Third, as the data driven estimation and direct data driven safety control are all formulated as one constrain optimization problem, the KKT optimality conditions are applied to obtain the explicit safety controller. Findings Some aircraft system identification and aircraft flight controller design can be reformulated as their corresponding constrain optimization problems. Then through solving these constrain optimization problems, the optimal estimation and controller are yielded, while satisfying our own priori goals. First, data driven estimation is proposed to get the rough estimation about the plant and controller. Second, data driven safety control is proposed to get one safety controller before our mentioned safety concept. Originality/value To the best of the authors’ knowledge, some existing theories about nonparametric estimation and tube model predictive control are very mature, but few contributions are applied in practice, such as aircraft system identification and aircraft flight controller design. This new paper shows the new theories about data driven estimation and data driven safety control on aircraft, being corresponded to the classical nonparametric estimation and tube model predictive control. Specifically, data driven estimation gives the rough estimations for the aircraft and its feed forward controller. Furthermore, after introducing the safety concept, data driven safety control is introduced to achieve the desired dual missions with the combination of KKT optimality conditions.

System Identification Approach for eVTOL Aircraft Demonstrated Using Simulated Flight Data

This paper describes a system identification method for electric vertical takeoff and landing (eVTOL) aircraft. The approach merges fixed-wing and rotary-wing modeling techniques with new strategies to develop a modeling method for eVTOL vehicles using flight-test data. The eVTOL aircraft system identification approach is demonstrated through application to the NASA Langley Aerodrome No. 8 tandem tilt-wing, distributed electric propulsion aircraft using a high-fidelity flight dynamics simulation. Orthogonal phase-optimized multisine inputs are applied to each control surface and propulsor at numerous flight conditions throughout the flight envelope to collect informative flight data. An aero-propulsive model is identified at each flight condition using the equation-error method in the frequency domain. The local model parameters are then blended to create a global model across the nominal flight envelope. The identified models are shown to provide a good fit to modeling data with good prediction capability. The methodology is developed with a discussion of unique eVTOL vehicle aerodynamic characteristics and practical strategies intended to inform future flight-based system identification efforts for eVTOL aircraft.

Synthesis cascade estimation for aircraft system identification

PurposeThe purpose of this paper extends the authors’ previous contributions on aircraft system identification, such as open loop identification or closed loop identification, to cascade system identification. Because the cascade system is one special network system, existing in lots of practical engineers, more unknown systems are needed to identify simultaneously within the statistical environment with the probabilistic noises. Consider this problem of cascade system identification, prediction error method is proposed to identify three unknown systems, which are parameterized by three unknown parameter vectors. Then the cascade system identification is transferred as one parameter identification problem, being solved by the online subgradient descent algorithm. Furthermore, the nonparametric estimation is proposed to consider the general case without any parameterized process. To make up the identification mission, model validation process is given to show the asymptotic interval of the identified parameter. Finally, simulation example confirms the proposed theoretical results.Design/methodology/approachFirstly, aircraft system identification is reviewed through the understanding about system identification and advances in control theory, then cascade system identification is introduced to be one special network system. Secondly, for the problem of cascade system identification, prediction error method and online subgradient decent algorithm are combined together to identify the cascade system with the parameterized systems. Thirdly from the point of more general completeness, another way is proposed to identify the nonparametric estimation, then model validation process is added to complete the whole identification mission.FindingsThis cascade system corresponds to one network system, existing in lots of practice, such as aircraft, ship and robot, so it is necessary to identify this cascade system, paving a way for latter network system identification. Parametric and nonparametric estimations are all studied within the statistical environment. Then research on bounded noise is an ongoing work.Originality/valueTo the best of the authors’ knowledge, research on aircraft system identification only concern on open loop and closed loop system identification, no any identification results about network system identification. This paper considers cascade system identification, being one special case on network system identification, so this paper paves a basic way for latter more advanced system identification and control theory.

Multi-Axis Inputs for Identification of a Reconfigurable Fixed-Wing UAV

Designing a reconfiguration system for an aircraft requires a good mathematical model of the object. An accurate model describing the aircraft dynamics can be obtained from system identification. In this case, special maneuvers for parameter estimation must be designed, as the reconfiguration algorithm may require to use flight controls separately, even if they usually work in pairs. The simultaneous multi-axis multi-step input design for reconfigurable fixed-wing aircraft system identification is presented in this paper. D-optimality criterion and genetic algorithm were used to design the flight controls deflections. The aircraft model was excited with those inputs and its outputs were recorded. These data were used to estimate stability and control derivatives by using the maximum likelihood principle. Visual match between registered and identified outputs as well as relative standard deviations were used to validate the outcomes. The system was also excited with simultaneous multisine inputs and its stability and control derivatives were estimated with the same approach as earlier in order to assess the multi-step design.

Generalized Linear Quadratic Control for a Full Tracking Problem in Aviation.

In this paper, the full tracking problem in aircraft system identification and control is presented. Time domain output error method with maximum likelihood principle was used to perform system identification. The linear quadratic regulator (LQR)-based approach has been used for solving aviation full tracking problems in aviation. It has been shown that the generalized nonlinear LQR control is able to handle such problems even in case of inaccurate measurements and in the presence of moderate disturbances provided that the model of an aircraft is properly identified.

Aircraft parameter estimation using ELM network

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.

Practical Aspects of Frequency-Domain Approaches for Aircraft System Identification

Practical aspects of frequency-domain approaches for aircraft system identification are explained and demonstrated. Topics related to experiment design, flight data analysis, and dynamic modeling are included. For demonstration purposes, simulated time-series data and simulated flight data from an F-16 nonlinear simulation with realistic noise are used. This enables detailed evaluations of the techniques and results, because the true characteristics of the data and aircraft dynamics are known for the simulated data. Analytical techniques and practical considerations are examined for the finite Fourier transform, nonparametric frequency response estimation, parametric modeling in the frequency domain, experiment design for frequency-domain modeling, data analysis and modeling in the frequency domain, and real-time calculations. Flight data from a subscale jet transport aircraft are used to demonstrate some of the techniques and technical issues.

Nonlinear aeroelastic modeling of aircraft using support vector machine method

PurposeThis paper aims to propose a nonlinear model for aeroelastic aircraft that can predict the flight parameters throughout the investigated flight envelopes.Design/methodology/approachA system identification method based on the support vector machine (SVM) is developed and applied to the nonlinear dynamics of an aeroelastic aircraft. In the proposed non-parametric gray-box method, force and moment coefficients are estimated based on the state variables, flight conditions and control commands. Then, flight parameters are estimated using aircraft equations of motion. Nonlinear system identification is performed using the SVM network by minimizing errors between the calculated and estimated force and moment coefficients. To that end, a least squares algorithm is used as the training rule to optimize the generalization bound given for the regression.FindingsThe results confirm that the SVM is successful at the aircraft system identification. The precision of the SVM model is preserved when the models are excited by input commands different from the training ones. Also, the generalization of the SVM model is acceptable at non-trained flight conditions within the trained flight conditions. Considering the precision and generalization of the model, the results indicate that the SVM is more successful than the well-known methods such as artificial neural networks.Practical implicationsIn this paper, both the simulated and real flight data of the F/A-18 aircraft are used to provide aeroelastic models for its lateral-directional dynamics.Originality/valueThis paper proposes a non-parametric system identification method for aeroelastic aircraft based on the SVM method for the first time. Up to the author’s best knowledge, the SVM is not used for the aircraft system identification or the aircraft parameter estimation until now.

Mechanical design and dynamics system identification of two-section Flapping Wing Aircraft

The mechanical design and system identification of the Flapping Wing Aircraft (FWA) has always been a problem. In this paper, we design a two-section FWA through the study of bird flight, where the wing structure is divided into the inner wing and outer wing, which can imitate the flight of birds. When flapping upward, the wings can be folded to reduce resistance, and when flapping downward, the wings can be expanded to increase lift. This flying method can improve the flight efficiency. The aircraft we designed has achieved flexible flight under human operation. Aiming at the complex problem of dynamic modeling of FWA, this paper proposes a linear system model by analyzing the force of the aircraft.We use the collected flight data to train the system model. The training of the model uses the gradient descent method to minimize the loss function, and the accuracy of the model is verified with the test data. The experimental results show that the system model can accurately calculate the acceleration, velocity, angle velocity and orientation of the aircraft based on the current control input.

Kalman-filter based online system identification of fixed-wing aircraft in upset condition

Online system identification became an integral part of the design process for aerodynamic parameter estimation with the technological progress. This paper presents two Kalman-filter based online system identification (SID) techniques for estimating aerodynamic parameters of fixed-wing aircraft in upset condition like stall. Unlike existing SID ones, the proposed methods first include aerodynamic characteristics directly in the aircraft dynamics, i.e. variation of aerodynamic derivatives or flow separation point, associated with the upset condition. Then, the conventional or unscented Kalman filter is applied in real time to obtain optimal estimates of the aerodynamic parameters under consideration. The proposed methods are tested with real flight data sets of several aircraft to demonstrate their effectiveness and superiority to a recently proposed method.

Quadrotor aircraft intelligent system identification experiment design

Aircraft prototyping and modeling is usually associated with resource expensive techniques and significant post flight analysis. The NASA Learn-To-Fly concept targets the replacement of the conventional ground-based aircraft model development and prototyping approaches with an efficient real time paradigm. The work presented herein describes the development of an intelligent excitation input design technique that determines excitation frequencies based on predefined rotational motion dynamic model. The input design is then evaluated on quadcopter unmanned aircraft that utilizes the new multisine input design. In order to minimize flight excursions without compromising the modeling capabilities, multisine input power spectrum is optimized based on the vehicle’s frequency response. The proposed methodology emphasizes excitation of modal frequencies which yields flight data rich information content. The generated optimized multisine input design is utilized for a quadcopter aircraft system identification and the performance is compared to conventional uniform amplitudes design. Simulation results show highly accurate model estimation in all identification results in addition to reduction of induced perturbations and power consumption. Additionally, the generated model prediction capabilities are not compromised after power spectrum optimization. Overall, the proposed technique introduces an efficient and intelligent system identification experiment design that can minimize the time and effort spent during excitation input design.

Flight dynamics modeling of elastic aircraft using signal decomposition methods

To improve the precision and accuracy of the flight dynamic models for elastic aircraft, this paper provides a novel method that extracts observable flight modes from flight test data and uses them in the identification process. For this purpose, a gray-box time-domain method is employed with the nonlinear ARX structure and the Levenberg–Marquardt parameter estimation technique. In the proposed method, the components of the flight parameters are extracted by two signal decomposition techniques, namely the singular spectrum analysis and empirical mode decomposition. These components are inputted into the identification process. Flight test data of the active aeroelastic wing is examined by the proposed method. The results indicate that both the singular spectrum analysis-based and empirical mode decomposition-based identification processes have desired performances in dealing with nontrained flight conditions. Thus, these methods may be utilized as an interpolation method to estimate the aircraft flight dynamics within the flight envelope. However, the empirical mode decomposition outperforms the singular mode decomposition because it is more significant in decomposing flight parameters based on the frequency content. Hence, the empirical mode decomposition may be a better tool to be employed for the aeroelastic aircraft system identification.

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

This paper aims to improve the performance of artificial neural networks used for the aircraft system identification by taking flight dynamic characteristics into consideration. In the proposed method, flight dynamic modes are recognized, isolated, and inputted individually to feed-forward neural networks. This method has several advantages such as being adaptive, involving all observable modes in the identification process, considering interactions between longitudinal and lateral-directional modes, and reducing noise effects. Simulated and real flight data of the HARV aircraft at high-angle of attack maneuvers are employed to train the neural networks and evaluate them. Results demonstrate improved accuracy and generality of the proposed method in comparison with the conventional ones.