Sample efficiency is a critical metric in intelligent control systems as it directly influences the feasibility and effectiveness of learning-based approaches. This paper presents the study of how Randomized Ensemble Double Q-Learning (REDQ), a sample-efficient model free algorithm, can be used in flight control applications. Three controllers were developed for: pitch, roll and combined biaxial attitude tracking tasks and tested on a high fidelity Cessna Citation 550 model. For each control task, three agents were trained offline: two using REDQ enhanced Soft Actor Critic (SAC) architectures and one using a standard SAC architecture for comparison. REDQ agents showed statistically significant improvements in sample efficiency during initial learning. Average accurate tracking convergence (error < 1◦) occurred within 5,500 training steps for pitch, 6,400 for roll and 11,500 for biaxial control. The gains in sample efficiency were shown to have drawbacks in learning stability and robustness when deviated too far from nominal conditions.

As the aviation industry faces growing environmental and societal pressures, novel concepts such as the Flying-V emerge. Nevertheless, its unconventional configuration poses unique challenges in terms of stability and control, highlighting the pressing need for advanced control systems development. A gap in the literature arises concerning the lack of system robustness in the presence of uncertainties and the absence of robust controllers designed for the aircraft. Hence, the current study aims to increase the maturity of the flight control system of the Flying-V concept by implementing a C* longitudinal controller, while guaranteeing robustness stability and performance against uncertainty, adequate performance in the presence of disturbances and measurement noise, and compliance with Level 1 handling qualities. The Flying-V model is implemented in MATLAB/Simulink which served as the foundation for the controllers synthesis. Three designs are developed, which include a continuous time, a modified continuous time, and a modified continuous time multi-modeling controllers. These are conducted within the standard and signal-based H-infinity mixed sensitivity frameworks. Extensive analysis of the controllers is performed in terms of stability assessments, linear and nonlinear time domain simulations, and uncertainty sensitivity. Results demonstrate that the feedback of a combined signal demand such as the C* parameter provides, for a single input multiple-output system, a balanced disturbance rejection at the plant outputs. Conclusions are drawn in terms of the feedback controller structure, highlighting that high gain is necessary at low frequencies for disturbance rejection and roll-off at high frequencies allows control signal reduction and measurement noise attenuation. Moreover, taking into consideration the discretization effects of the flight computer in the design phase improves considerably the stability margins of the controllers. Lastly, it is verified that the proposed controller structures which are designed in the linear domain perform satisfactorily in the nonlinear simulation model and comply with the requirements defined. The multi-modeling controller proved successful in terms of robustness and performance across the flight envelope. Hence, we conclude that implementing this H-infinity optimization process for the design is a promising and viable method to guarantee robustness against uncertainties, disturbances, and measurement noise, as well as compliance with Level 1 HQ. The present work naturally paves the way for recommendations for future work. Some of these recommendations include the extension to lateral-directional control laws, gain-scheduling to guarantee the requirements across the flight envelope, maturity of the aerodynamic data, and pilot assessments of the handling qualities.

Quad-planes combine hovering and vertical takeoff and landing capability with fast and efficient forward flight. Regular Quad-planes with dedicated pusher motor can be subject to gust disturbances, and are not well-equipped to deal with actuator faults. Dual-axis Tilt-Rotor quad-planes are more maneuverable due to their overactuation. This also increases their gust resilience and allows them to hover statically after actuator failures. The vehicle in this paper uses an Incremental Nonlinear Dynamic Inversion (INDI ) controller, combined with a nonlinear Sequential Quadratic Programming (SQP) Control Allocation (CA ) algorithm, which can also find hover solutions in the case of actuator failures. We investigate both a combined allocation of linear and angular accelerations, as well as a cascaded allocation scheme. Due to large required changes in roll and pitch angles, the cascaded approach is selected in this research. Introduction of a tertiary control effort term, separation of attitude and actuator command optimization and a simulated Fault Detection and Identification ( FDI) mechanism led to repeated successful recovery from a motor failure in hover. Position tracking was demonstrated under failure in the recon- figured flight condition. Index Terms- Tilt-rotor, dual-axis tilt, quad-plane, FTC, over- actuated, control allocation

In nature, birds can intelligently adapt their wing shapes to their environment. This paper aims to replicate this capability by designing an online data-driven aerodynamic performance optimization framework for an unconventional morphing aircraft. Compared to state-of-the-art methods, the proposed framework efficiently finds global optima with reduced computational load when addressing time-varying, nonlinear, and non-convex problems. It also demonstrates enhanced adaptability to unforeseen scenarios. In the event of a sudden actuator fault, the algorithm can automatically detect the fault, adapt the onboard data-driven model, and continue performing optimization and trimming tasks using the remaining healthy actuators. Additionally, the paper addresses the optimal number of actuators within a morphing surface, considering the tradeoff between optimization performance and the weight penalty. High-fidelity simulations demonstrate that through active morphing, the proposed framework achieves drag reductions of 1.9–4.9 % during cruise and up to 12.6 % at higher operational lift coefficients (due to heavier weight and lower speed), resulting in an overall drag reduction of 2.98 % over a typical flight cycle, which corresponds to fuel savings of approximately 150 kg/h. This research represents a significant advancement in sustainable aviation, contributing to reduced fuel consumption, lower emissions, and improved fault tolerance for next-generation aircraft.

The Flying-V is a tailless, V-shaped flying-wing type aircraft that promises to offer significant increases in aerodynamic efficiency. Due to its configuration, the Flying-V faces some control and stability related issues. These include limited control authority, pitch break tendencies and non-ideal handling qualities. To enhance the handling qualities of the Flying-V, Incremental Nonlinear Dynamic Inversion (INDI)-based flight control systems have been proposed. INDI, a sensor-based alternative to conventional Nonlinear Dynamic Inversion (NDI), is rooted in the principle of feedback linearization. Unlike NDI, INDI does not depend heavily on accurate on-board models (OBM), thereby offering increased robustness to aerodynamic uncertainties. However, singular perturbations—such as time delays, aeroelastic effects, and additional unmodeled or unknown dynamics—have been identified as challenges for INDI-based control laws. Various strategies have been explored to improve the overall robustness of INDI-based flight control systems, including outer-loop tuning and inversion loop augmentation strategies. In this research a multi-loop µ-optimal approach for designing robust inversion-based flight control laws is explored for the design of an explicit model-following pitch-rate control system for a short-period approximation of the Flying-V’s longitudinal dynamics. The design problem takes into account both regular and singular perturbations. To assess the robust stability and performance of the proposed control systems, a structured singular value analysis was performed. It was concluded that a multi-loop synthesis approach is capable of achieving better robust stability and performance levels when compared to either strictly inner-loop or outer-loop synthesis. As such, it can be concluded that multi-loop synthesis approaches are best capable of leveraging the robustness functionalities of multi-loop inversion-based control systems.

To gain more insight into the performance of state-of-the-art Static Output Feedback (SOF) controller synthesis methods for H∞-control, quantitative comparisons are made between Lyapunov methods and well-known established non-smooth optimization methods, i.e. Hinfstruct and HIFOO. Three methods were deemed to be the most promising to compete and were bundled into one toolbox named SOFHi. The algorithms were extended to incorporate structured SOF and a variant of SOFHi was proposed to significantly improve upon the computational efficiency of the original implementation. Extensive comparisons show that SOFHi was able to compete with the established non-smooth methods and even able to significantly outperform one of them. Lastly, an elaborate flight control benchmark example is given to showcase the effectiveness of the algorithms, which involves the design of a gain-scheduled normal acceleration Control Augmentation System (CAS) for a highly maneuverable fighter aircraft.

Climate change poses a significant global challenge and the aviation industry, a major contributor to greenhouse gas emissions, must address its environmental impact. As the demand for air travel continues to rise, transitioning to sustainable forms of aviation becomes crucial. This paper focuses on the challenges and opportunities associated with implementing full electric aircraft at regional airports, using Rotterdam the Hague Airport (RTHA) asa case study. The research aims to determine the optimal infrastructure sizing of RTHA and provide insights into the integration of electric aviation. A novel mixed integer linear programming (MILP) model is proposed to optimize the charging schedule and determine the required number of charging stations and electricity grid capacity. The study considers different scenarios of electric aviation uptake, aircraft types, charger types, and spatial aspects of charging locations. The findings offer comprehensive recommendations for regional airport infrastructure planning and address uncertainties through sensitivity analysis.

Even though Deep Reinforcement Learning (DRL) techniques have proven their ability to solve highly complex control tasks, the opaqueness and inexplicability associated with these solutions many times stops them from being applied to real flight control applications. In this research, reward decomposition explanations are used to tackle this issue and augment DRL end-user explainability. A reward decomposition-based DRL controller is deployed in a longitudinal state-space model of the Cessna Citation 500 aircraft, and it is assessed on two attitude flight control tasks. Furthermore, a new explanation type called Dominant Reward eXplanations (DRX) is presented, which allows users to obtain more global insights than the ones generated by Reward Difference eXplanations (RDX). Results show that the explanations produced lead to straightforward and intuitive insights about the controller’s behaviour, capable of improving end-user explainability. Moreover, a small analysis seems to indicate that the decomposed method has similar performance to the one obtained without reward decomposition, however, training time increases considerably. To the author’s best knowledge, this is the first application of reward decomposition explanations to the flight control domain.

Reinforcement Learning is being increasingly applied to flight control tasks, with the objective of developing truly autonomous flying vehicles able to traverse highly variable environments and adapt to unknown situations or possible failures. However, the development of these increasingly complex models and algorithms further reduces our understanding of their inner workings. This can affect the safety and reliability of the algorithms, as it is difficult or even impossible to determine which are their failure characteristics and how they will react in situations never tested before. It is possible to remedy this lack of understanding through the development of eXplainable Artificial Intelligence and eXplainable Reinforcement Learning methods like SHapley Additive Explanations. In this thesis, this tool is used to analyze the strategy learnt by an Actor-Critic Incremental Dual Heuristic Programming controller architecture when presented with a series of pitch rate and roll rate tracking tasks in a variety of non-linear flying conditions, which include flying close to the stall regime, experiencing non-linear reference signals, reaching the control surface deflection limits, and flying with large sideslip angles. This same controller architecture has been previously explored with the same analysis tool but limited to the nominal linear flight regime, and it was observed that the controller learnt linear control laws, even though its Artificial Neural Networks should be able to approximate any function. This thesis shows that, even in the non-linear flight regime, it is still more optimal for this controller architecture to learn quasi-linear control laws, although it seems to continuously modify the linear slopes as if it was an extreme case of the gain scheduling technique. Additionally, a more complex Reinforcement Learning architecture based on the Soft Actor-Critic algorithm is also explored with the same analysis tool, to demonstrate its usability in the presence of non-linear control laws and to improve our understanding of this offline state-of-the-art algorithm.

The quasi-steady model (QSM) was developed to calculate the power output of a soft-wing airborne wind energy (AWE) system operated in pumping cycles. The comparison with experimental data shows that the original formulation of the model underestimates the power output. The model does not take inertial forces into account and highly simplifies the reel-out flight path. In this work, these assumptions are critically reviewed and their impact on the simulation results assessed.

The effect of inertial forces is analytically described for a helical flight path with fixed turning radius around the downwind axis. The derived analytical solution for this flight path verifies the formulation of inertial forces in spherical coordinates proposed in earlier studies. The proposed equations are modified and added to the QSM. Both, the analytically developed set of equations and the QSM with added inertia are compared for the helical flight case. It is observed that at larger cone angles where the tether length is rather short, but the turning radius is large, the inertial effect increases the tether force and the power output. With small cone angles this effect reverses. Based on the example case, the impact of the inertial effects on the output power computed with the QSM is limited to 1-2% for lightweight soft kites.

In the original formulation of the QSM the figure-of-eight maneuvers are not resolved. Instead, the tether is reeled out at representative constant values for the elevation, azimuth and course angles. QSM output based on real flight path coordinates is used as a benchmark to calculate the accuracy of reel-out path simplifications. The power output error of resolving reel-out flight with constant angles is on average 4.1% when compared to the reference simulations. An alternative flight path representation is developed, fitting a known figure-of-eight, lemniscate, to measured asymmetric reel-out flight paths. By flying the fitted figure in the QSM, the output power error caused by this path representation is estimated to be only 1.7% in comparison to the reference simulations. It is recommended to use this fitting method based on a class function/shape function transformation technique (CST), to parametrize the reel-out flight path, despite the added complexity.

The study has shown that the accuracy improvement by including the inertial effects in the QSM is marginal if the model is limited to lightweight soft wing simulation. As a follow-up study the inertial effects should be tested with bigger kites and heavier rigid-wing systems. The adjusted CST fitting method together with the inertial equations enable these maneuvering flight simulations in the QSM.

To meet the demanding requirements on the environmental impact of aircraft, radically new aircraft concepts need to be developed. Within the NOVAIR project, Royal Netherlands Aerospace Centre (NLR) tests these new concepts on a Scaled flight demonstrator (SFD). Using an SFD allows for testing of dynamic and flight physical behavior of new aircraft concepts, which is difficult with more theoretical methods. Furthermore, by using an SFD, the risks associated with full-scale testing in terms of cost and time are minimized. One of the new concepts developed for the SFD is Distributed electric propulsion (DEP). Here, the two jet engines are replaced by six electric propellers. As these can be used actively for control, this results in an over-actuated system. These propellers interact with the aerodynamics of the wing resulting in Propulsion airframe interaction (PAI) effects. Although using the PAI effects for control has the potential to improve capabilities, efficiency and robustness of the aircraft, research into controllers using these effects actively is limited.

This thesis, therefore, presents a new control method including control allocation for the DEP-SFD aircraft, based on the nonlinear Incremental nonlinear dynamic inversion (INDI) controller. INDI enables controlling the nonlinear dynamics of the DEP aircraft over the complete flight envelope with one controller. By feeding back real-time sensor measurements, robustness to modeling errors and external disturbances is increased. Using the Incremental nonlinear control allocation (INCA) method, the full control authority of the DEP can be used. This technique enables taking into account nonlinear allocation relations and control effector interactions, while solving the control allocation real-time, which is key in actively using the PAI effects for control. The INCA method is used for two performance improvements: tracking performance and propeller power efficiency. To compensate for actuator dynamics, this thesis implements an Model predictive control (MPC) controller which results in improved tracking and higher efficiency. The performance of the controller was analyzed in simulation, where a reference square signal input on the roll angle was applied, while minimizing the sideslip angle and maintaining a steady altitude and velocity. The INCA controller with MPC is compared to a conventional INDI controller, showing a significant decrease in rise time from 2.46 s to 0.703 s with minimal tracking error. Furthermore, the effective bandwidth of the system was increased from 0.186 Hz to 0.663 Hz and the power consumption reduced by 6.3%. Modeling uncertainties, external disturbances and a propeller fault were introduced to verify the robustness of the controller. Finally, the reference altitude and velocity were varied, demonstrating controller performance over a large part of the flight envelope.

This thesis reports on the research and design of a real-time Time-Varying Model Predictive Control (TVMPC) scheme to stabilize a tilt-rotor aircraft with four independently tilting rotors. The aircraft design stems from a prototype system constructed by the drone technology start-up Avy.
First, a motivation is presented for implementation of model-based flight control techniques. Consequently, TVMPC is introduced as a middle ground between Adaptive Model Predictive Control (MPC) and fully nonlinear MPC, capturing model variation over the control horizon while retaining the computational efficiency of quadratic optimization algorithms. In a simulation of a simple nonlinear system, TVMPC outperforms fixed-model and adaptive MPC using only a small set of extra models.
Secondly, the tilt-rotor system model is developed. Models for rotor torque and thrust responses and stepper motor movement are introduced and calibrated using experiments, leading to a full nonlinear model of 21 states, 8 control inputs and 8 box constraints. Further envisioned experimental work and its preparations such as aerodynamic testing and tracked constrained flight are briefly mentioned.
The implementation of TVMPC is tested in simulation. Five scenarios compare the effectiveness of time-varying control versus adaptive and fixed-model MPC, with TVMPC not outperforming fixed-model control in horizontal vehicle orientations, but improving on oscillatory behavior shown by adaptive control. TVMPC outperforms the other two controllers in a past 90 degree pitch scenario (partial backflip).
In parallel, the setup of Robot Operating System (ROS) in conjunction with a Hardwarein-the-Loop (HIL) experiment is presented to test tilt-rotor control in real-time. In closed loop, the system was stabilized for limited degrees of freedom using a combination of PID controllers, validating the framework and providing more opportunities for research.
It is concluded that TVMPC improves on the predictions of adaptive MPC, but is prone to oscillatory behavior. Only in highly nonlinear situations, fixed-model MPC is outperformed. For real-time, embedded results, more research still needs to be performed.

In this work interval analysis is applied to the thirteen control effector Innovative Control Effectors model to find its trim set. The method to find trim states is based on interval box consistency. At low speed the method is capable of finding interval enclosures of single trim points with a high accuracy if the minimum number of required control effectors is used. At higher speeds the found accelerations are larger. When looking for a full trim set the method finds continuous bounds on the control effectors for the entire input range in one run. This is a good demonstration of the advantages that interval analysis has over conventional methods that generally can only find one trim point at a time. The found bounds are a maximum of 1 deg wide for each control effector, but despite this the remaining accelerations can be up to 0.5 m/s^2 for linear accelerations and up to 10 deg/s^2 for rotational accelerations. Because of these large accelerations the found solutions are not acceptable as trim conditions. On the other hand the potential that interval analysis has as a trimming method is demonstrated, since continuous bounds on trim sets have been found in a single run. This is a feat that no other trimming method has yet accomplished. Further research is needed to exploit the full potential of interval trim methods so that the results can be used for other purposes such as flight envelope prediction.

Glacial isostatic adjustment is the viscoelastic response of the earth to ice sheet loading. Ice dynamic models employ glacial isostatic adjustment to accurately simulate bedrock uplift. In previous studies, ice dynamic models use only a constant relaxation time to predict bedrock uplift. In this study, the aim is to improve an ice dynamic model by using a laterally varying bedrock relaxation time.

Shear wave velocity data suggest a sharp increase in relaxation time between West and East Antarctica. Viscosity fields based on shear wave velocity data were converted to a set of relaxation time fields that are compatible with the ice dynamic model. These relaxation time fields were used in an ice dynamic simulation for the Antarctic Ice Sheet.

It was found that largest differences between the different relaxation time models occur from 20 000 years ago up to PD. The model with a high relaxation time average showed a decrease in ice volume of 22.0%, whereas the model with a low relaxation time average, decreased by 17.4%. It was found that the largest differences between the models occurred at the WAIS due to the strong interaction of sea level and bedrock uplift. Furthermore, it was found that a reduction of ice sheet wavelength led to relaxation time decrease, locally up to 50%. This appeared to be a stabilising mechanism for decreasing ice sheets, but the effective retained ice volume by this mechanism remains unknown.

From 2019 onwards, airline pilots will be required to follow stall training in simulators. A major open research question is which level of model fidelity is required for effective training. As part of an effort to answer this question, a stall model of a Cessna Citation II aircraft is developed from specifically-gathered flight test data. The model is based on Kirchoff’s theory of flow separation. During identification, the nonlinear and linear parameters of the model are estimated in separate steps. This distinction enables the application of a semi-objective model structure selection method using multivariate orthogonal functions. It is shown that stall-related effects should be included in the model equations for lift, drag, and pitch moment. The model parameters are accurately estimated, and model output is validated around a flight condition of 5,500 m altitude. The developed methodology is concluded to be well-suited for stall model identification of small business jets.