Open-loop architecture. We show the method used to acquire the pressure data from the external surface of the flexible wing skin, using incorporated Kulite pressure sensors and the instrumentation of the morphing controller. The acquired pressure data are analyzed through fast Fourier transforms to detect the magnitude of the noise in the surfaceairflow.Subsequently,thedataare filteredbymeansofhigh-pass filtersandprocessedbycalculatingtheroot mean square of the signal to obtain a plot diagram of the noise in the airflow. This signal processing is necessary to remove the inherent noise electronically induced from the Tollmien-Schlichting waves, which are responsible for triggeringthetransitionfromlaminartoturbulent flow.The flexibleskinisrequiredtomorphtheshapeoftheairfoil throughtwoactuationpointstoachieve anoptimizedairfoil shapebasedonthe theoretical flowconditionssimilar to thosetestedinthewindtunnel. Twoshapememoryalloyactuators withanonlinearbehaviordrivethedisplacement ofthetwocontrolpointsofthe flexibleskintowardtheoptimizedairfoilshape.Eachoftheshapememoryactuatorsis activatedbyapowersupplyunitandcontrolledusingtheSimulink/MATLAB®softwarethroughaself-tuningfuzzy controller. The methodology and the results obtained during the wind-tunnel test proved that the concept and validityof thesystemin real timearediscussed in thispaper. Real-time acquisitionandsignalprocessing of pressure data are needed for further development of the closed-loop controller to obtain a fully automatic morphing wing system.

The present paper describes the design of a hybrid actuation control concept, a fuzzy logic proportional-integral-derivative plus a conventional on-off controller, for a new morphing mechanism using smart materials as actuators, which were made from shape memory alloys (SMA). The research work described here was developed for the open loop phase of a morphing wing system, whose primary goal was to reduce the wing drag by delaying the transition (from laminar to fully turbulent flows) position toward the wing trailing edge. The designed controller drives the actuation system equipped with SMA actuators to modify the flexible upper wing skin surface. The designed controller was also included, as an internal loop, in the closed loop architecture of the morphing wing system, based on the pressure information received from the flexible skin mounted pressure sensors and on the estimation of the transition location.The controller’s purposes were established following a comprehensive presentation of the morphing wing system architecture and requirements. The strong nonlinearities of the SMA actuators’ characteristics and the system requirements led to the choice of a hybrid controller architecture as a combination of a bi-positional on-off controller and a fuzzy logic controller (FLC). In the chosen architecture, the controller would behave as a switch between the SMA cooling and heating phases, situations where the output current is 0A or is controlled by the FLC.In the design phase, a proportional-integral-derivative scheme was chosen for the FLC. The input-output mapping of the fuzzy model was designed, taking account of the system’s error and its change in error, and a final architecture for the hybrid controller was obtained. The shapes chosen for the inputs’ membership functions were s -function, π-function, and z -function, and product fuzzy inference and the center average defuzzifier were applied (Sugeno).

The paper presents the numerical and experimental validation of a hybrid actuation control concept – fuzzy logic proportional-integral-derivative (PID) plus conventional on-off – for a new morphing wing mechanism, using smart materials made of shape memory alloy (SMA) as actuators. After a presentation of the hybrid controller architecture that was adopted in the Part 1, this paper focuses on its implementation, simulation and validation.The PID on-off controller was numerically and experimentally implemented using the Matlab/Simulink software. Following preliminary numerical simulations which were conducted to tune the controller, an experimental validation was performed. To implement the controller on the physical model, two programmable switching power supplies (AMREL SPS100-33) and a Quanser Q8 data acquisition card were used. The data acquisition inputs were two signals from linear variable differential transformer potentiometers, indicating the positions of the actuators, and six signals from thermocouples installed on the SMA wires. The acquisition board’s output channels were used to control power supplies in order to obtain the desired skin deflections. The experimental validation utilised an experimental bench test in laboratory conditions in the absence of aerodynamic forces, and a wind-tunnel test for different actuation commands. Simultaneously, the optimised aerofoils were experimentally validated with the theoretically-determined aerofoils obtained earlier. Both the transition point real time position detection and visualisation were realised in wind tunnel tests.

Two main factors, including regression accuracy and adversarial attack robustness, of six trajectory prediction models are measured in this paper using the traffic flow management system (TFMS) public dataset of fixed-wing aircraft trajectories in a specific route provided by the Federal Aviation Administration. Six data-driven regressors with their desired architectures, from basic conventional to advanced deep learning, are explored in terms of the accuracy and reliability of their predicted trajectories. The main contribution of the paper is that the existence of adversarial samples was characterized for an aircraft trajectory problem, which is recast as a regression task in this paper. In other words, although data-driven algorithms are currently the best regressors, it is shown that they can be attacked by adversarial samples. Adversarial samples are similar to training samples; however, they can cause finely trained regressors to make incorrect predictions, which poses a security concern for learning-based trajectory prediction algorithms. It is shown that although deep-learning-based algorithms (e.g., long short-term memory (LSTM)) have higher regression accuracy with respect to conventional classifiers (e.g., support vector regression (SVR)), they are more sensitive to crafted states, which can be carefully manipulated even to redirect their predicted states towards incorrect states. This fact poses a real security issue for aircraft as adversarial attacks can result in intentional and purposely designed collisions of built-in systems that can include any type of learning-based trajectory predictor.