Aerodynamic angle estimation: comparison between numerical results and operative environment data

Several architectures exist to measure aerodynamic angles based on physical sensors. As far as unmanned aerial vehicle (UAV) is concerned, traditional systems hardly comply with reliability and redundancy requirements due to size and weight limitations. A patented virtual sensor, based on artificial neural network (ANN) techniques, named smart-air data, attitude and heading reference system (Smart-ADAHRS) has been investigated as a good estimator for aerodynamic angles in simulated environment. This paper focuses on flight testing procedures in operative environment and data processing for the Smart-ADAHRS validation with real data. As many factors interfere during the generation of the ANN training set, an accurate choice and integration of the flight test instrumentation (FTI) system components becomes crucial. A comprehensive description has been included about the FTI equipment and its influence on the neural network performance. Differences between numerical simulation and operative environment data are detailed as final aim of this work. At the end, feasible solutions are suggested to solve the typical gap between virtual and real scenario, both in terms of data analysis and neural network architecture.

[1]  G. Hardier,et al.  Aerodynamic model inversion for virtual sensing of longitudinal flight parameters , 2013, 2013 Conference on Control and Fault-Tolerant Systems (SysTol).

[2]  Mario Lavorgna,et al.  Low Cost Virtual Pressure Sensor , 2004 .

[3]  Rajkumar Thirumalainambi,et al.  Training data requirement for a neural network to predict aerodynamic coefficients , 2003, SPIE Defense + Commercial Sensing.

[4]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[5]  Sheldon M. Ross,et al.  Introduction to Probability and Statistics for Engineers and Scientists , 1987 .

[6]  Marc Rauw,et al.  FDC 1.2 - A Simulink Toolbox for Flight Dynamics and Control Analysis , 2001 .

[7]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[8]  Bernard Etkin,et al.  Dynamics of flight , 1959 .

[9]  J L Castro,et al.  Neural networks with a continuous squashing function in the output are universal approximators , 2000, Neural Networks.

[10]  Rolf Isermann,et al.  Trends in the Application of Model Based Fault Detection and Diagnosis of Technical Processes , 1996 .

[11]  J. Anderson,et al.  Fundamentals of Aerodynamics , 1984 .

[12]  Manuela Battipede,et al.  Air Data Virtual Sensor: a Data-Driven Approach to Identify Flight Test Data Suitable for the Learning Process , 2019 .

[13]  Manuela Battipede,et al.  Neural Networks for Air Data Estimation: Test of Neural Network Simulating Real Flight Instruments , 2012, EANN.

[14]  Rolf Isermann.  Model-based fault-detection and diagnosis - status and applications § , 2004 .

[15]  Manuela Battipede,et al.  Survey on a neural network for non linear estimation of aerodynamic angles , 2017, 2017 Intelligent Systems Conference (IntelliSys).

[16]  Manuela Battipede,et al.  Sensitivity Analysis of a Neural Network based Avionic System by Simulated Fault and Noise Injection , 2018 .

[17]  Manuela Battipede,et al.  Development of Neural Networks for Air Data Estimation: Training of Neural Network Using Noise-Corrupted Data , 2011 .

[18]  Ihab Samy,et al.  Fault Detection and Flight Data Measurement , 2011 .

[19]  Manuela Battipede,et al.  Joseph covariance formula adaptation to Square-Root Sigma-Point Kalman filters , 2017 .

[20]  G. N. Fouskitakis,et al.  Aircraft angle-of-attack virtual sensor design via a functional pooling narx methodology , 2003, 2003 European Control Conference (ECC).

[21]  J.J. Gertler,et al.  Survey of model-based failure detection and isolation in complex plants , 1988, IEEE Control Systems Magazine.

[22]  Manuela Battipede,et al.  Advantages of Neural Network Based Air Data Estimation for Unmanned Aerial Vehicles , 2017 .

[23]  Cédric Seren,et al.  Model–Based Techniques for Virtual Sensing of Longitudinal Flight Parameters , 2015, Int. J. Appl. Math. Comput. Sci..

[24]  Federico Rossi,et al.  Mnemosine Mark-V: the Fifth Generation of an Ultra Light Machine-Dedicated FTI System , 2016 .

[25]  Marcelo Embiruçu,et al.  Fault Detection and Diagnosis in dynamic systems using Weightless Neural Networks , 2017, Expert Syst. Appl..

[26]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[27]  Manuela Battipede,et al.  Ill-conditioned problems improvement adapting Joseph covariance formula to non-linear Bayesian filters , 2016 .

[28]  J. Tukey,et al.  Variations of Box Plots , 1978 .

[29]  Donald H. Lenschow.  Vanes for sensing incidence angles of the air from an aircraft , 1971 .

[30]  Gilles Pagès,et al.  Approximations of Functions by a Multilayer Perceptron: a New Approach , 1997, Neural Networks.

[31]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[32]  Robert Babuska,et al.  Virtual Sensor for the Angle-of-Attack Signal in Small Commercial Aircraft , 2006, 2006 IEEE International Conference on Fuzzy Systems.

[33]  T. Rajkumar,et al.  Prediction of Aerodynamic Coefficients Using Neural Networks for Sparse Data , 2002, FLAIRS.

[34]  Manuela Battipede,et al.  Novel Neural Architecture for Air Data Angle Estimation , 2013, EANN.