In the MSc programme, a master student at the Control and Simulation (C&S) group eventually needs to choose a specific specialisation for the MSc thesis. To make a well-informed decision about what electives to choose and what thesis topic to work on, it is first important to understand how C&S is structured.
The C&S group is typically characterised by the following four knowledge clusters: 1) Aerospace Guidance, Navigation, and Control (AGNC), 2) Aerospace Human-Machine Systems (AHMS), 3) Communication, Navigation & Surveillance in Air Traffic Management (CNS/ATM), and 4) Micro (Unmanned) Aerial Vehicles (MAV).
Control disciplines & courses
The C&S group is all about disciplines found in many facets of control. The main control disciplines focused on at C&S are:
Manual Control
Overview
Manual Control Cybernetics aims to analyse and model human perception and neuromuscular control dynamics in order to design manual control interfaces that make the control task safer, easier and more fun to use. The cybernetics theory is closely related to modern (automatic) control and identification techniques, so to prepare for a graduation project in this field it is important that apart from the solid C&S core background, you followed at least one elective course on system identification. Thus, the manual control cybernetics discipline requires quite a mathematical background in order to be able to analyse and interpret data obtained from human-in-the-loop experiments featuring the Citation, SIMONA or the HMI-Lab. Examples of cybernetics projects that students can do are (primarily in aerospace, but also automotive domains):
Advancing the cybernetics theory (pursuit, preview, learning and adaptation)
Human visual and vestibular self-motion perception models
Flight simulator fidelity studies
Handling qualities, fly-by-wire system design
Neuromuscular system dynamics, biodynamic feed-through
Tele-operation and remote control (e.g., of UAVs)
Haptic interfaces (in collaboration with 3ME)
Recommended/Required electives:
Manual Control & Cybernetics (AE4319)
Modern System Identification of Aerospace Vehicles (AE4320)
Real-time Distributed Flight and Space Simulation (AE4323)
The automatic control discipline is related to mechatronics, which is a multi-disciplinary engineering field of self-regulating mechanisms consisting of mechanical, electronic, and computer systems. Note that the main difference between automatic control and autonomous control is that inputs are “provided” (by humans) to control systems, whereas autonomous systems “acquire” inputs by their own. The aerospace automatic control discipline aims at developing new generations of advanced and fault-tolerant flight control systems and is primarily driven by questions such as: “how can we make the control of aerial vehicles (either manned or unmanned) more robust and flexible against system failures and damages?” together with “how can we reduce the reliance on knowledge about the (damaged) aerial vehicle to make this possible?” The various approaches adopted in this discipline heavily rely on novel mathematical tools and practices grounded in (complex) systems and control theories. The simulation and application platforms for the automatic control discipline can include the Citation, SIMONA or the Flight Arena to investigate new theories and concepts for manned and unmanned aerial vehicles, respectively. Examples of automatic control projects that students can do are:
Nonlinear dynamics, optimisation, and control
Adaptive and reconfigurable flight control
Helicopter flight dynamics and handling qualities
State estimation (sensor fusion) and distributed control
Recommended/required electives:
Nonlinear and Adaptive Flight Control (AE4311)
Autonomous Flight of Micro Air Vehicles (AE4317)
Modern System Identification of Aerospace Vehicles (AE4320)
In the supervisory control discipline the focus is on designing advanced decision support systems for operators in work environments (such as cockpits and air traffic control centres) where many of the human tasks have been automated. For supervisory control, the human mind (rather than its neuromuscular properties) is considered as a highly adaptive and intelligent control system. Backed by elementary principles grounded in both control theory and psychology, the goal is to find the right information, show it at the right time and in the right way such that humans can make smart decisions and respond robustly to system anomalies. In particular, principles from Cognitive Systems Engineering (CSE) and Ecological Interface Design (EID) are applied, both powerful methods to analyse the work domain and create interfaces that show the operator the primary means-ends relationships, facilitating decision making and aiming for superior situation awareness. Note that this control discipline is less mathematical than cybernetics and automatic control. However, the ecological approach that we take is one of the hardest (but also the best) approach to develop radically innovative interfaces (we have many examples of that!), and the experimental evaluation of these interfaces or automation can be more difficult than the often straightforward tracking experiments in manual control cybernetics and automatic control. Examples of supervisory control topics that we study are (in all transport domains):
Ecological flight deck design, the cockpit of the future
Ecological warning systems (terrain, traffic, weather…)
Ecological automation design (e.g., for UAVs, single-pilot cockpits)
Future ATC/ATM controller workspace design
Trust, acceptance, transparency and conformance of automation
Novel interfaces such as touch screens
Recommended/required electives:
Supervisory Control & Cognitive Systems (AE4318)
Air Traffic Management (AE4321)
Real-time Distributed Flight and Space Simulation (AE4323)
The autonomous control discipline aims at studying the sensors, systems, and algorithms enabling fully automatic flight operations in current and future airspace environments. The focus in this control discipline is much less on humans and the control of individual aircraft, but much more on the airspace system as a whole. One branch of this control discipline specifically focuses on studying (emergent) traffic patterns and analysing and developing requirements for (future) communication, navigation, and surveillance infrastructures and systems, where the goal is to create a resilient airspace environment capable of absorbing system perturbations and anomalies without compromising the safety and efficiency of the air transportation system as a whole.
The other branch focuses on designing (small) autonomous air vehicles for various real-world applications, such as remote surveillance tasks and aiding search-and-rescue operations. To realise that, this branch essentially integrates most of the knowledge, tools, and practices found in automatic control. This means that if you want to turn your design into an operational UAV (e.g., to test it in the Flight Arena), you need some knowledge on advanced control and sensor fusion techniques. Examples of autonomous control applications that we study are:
Autonomous operations with Airborne Separation Assistance Systems
Sense and avoid systems in the cockpit, on the ground and for UAVs
Intelligent control through reinforcement learning
Vision-based autonomous operations
UAV swarm coordination techniques
Data-mining and big data applications, e.g., ADS-B
Complexity of traffic patterns and traffic flows
Scientific foundations in current and future air traffic management and trajectory planning
Recommended/required electives:
UAV branch: Nonlinear and Adaptive Control (AE4311); Autonomous Flight of Micro Air Vehicles (AE4317)
In a nutshell, these control disciplines are ordered by their decreasing involvement of humans in the control loop. That is, the manual control discipline focuses primarily on studying, identifying, and modelling the human as a neuromuscular control system in order to make informed design and implementation decisions about handling qualities of dynamic systems and simulator fidelity. In automatic control, the capabilities of computer technology are explored to act as clever and fault-tolerant servo mechanisms that close (inner) control loops and automatically stabilize and guide aircraft, but leave the final decision-making and navigation tasks to the human operator. In supervisory control, the human mind (rather than its neuromuscular properties) is considered as an intelligent and highly adaptive control system that can make smart decisions on when and how to intervene when the automation fails. In autonomous control, the focus is on the analysis and design of intelligent computer-controlled systems (and their of functional requirements) featuring very limited or no human involvement, for example, autonomous sense-and-avoid systems.
Application platforms
After choosing a particular control discipline, the theory can then be applied to a specific simulation and/or application platform featuring various facilities/laboratories:
Unmanned aerial vehicles: MAV-Lab and the Flight Arena
Air traffic management: ATM-Lab
Finding a MSc thesis assignment
To start your graduation project, you contact one (or more) of the principle investigators associated with each subcluster. This also means that we generally do not accept thesis topics that students come up with themselves. Additionally, an internship can almost never be extended into a thesis topic within our group. Although this may sound harsh, we generally like our students to continue with the work that has been done by other students, and also like to “link” your project to one of the many running projects of our PhD students. From our 20+ years of experience we know that this has many advantages!
First of all, we know a lot about the subject, which makes it easier for us to define a project that can be done using a realistic time schedule of 9 months full time work. Second, much of the materials have been sorted out, the literature is well known, many things are already programmed in MATLAB/Python/C++/Java and/or the simulators, etc. So you don’t need to start from scratch and be programming all the time before you can do something nice and exciting (like doing an experiment with real pilots!). Third, with a PhD student near-by, you will have a ‘supervisor’ (more of a companion) that you can work with together, someone who knows all of the issues, the software, the simulator, etc. All three advantages lead to high-quality work, a project that you will be proud of when you have finished, a project that you did not believe you were capable of doing before, a project that will finish your 5+ years at TU Delft in style. The graduation project, you will see, is the best learning experience you will ever have!
Finally, for all assignments that we have the rule is: first come, first serve. You can only “claim” a particular subject or project when you are capable of starting it very soon. Note that for many students it is often surprising to see what we do at C&S, as the courses that we give explain the basics of our field and only briefly touch upon the (often more exciting) projects that we do. It is therefore very important to carefully explore all possibilities, and think by yourself what exactly do you want to learn, what motivates you to work, do you like theoretical or experimental work, or both, do you want to design something new, or test an interface or device or theory experimentally? On all possible scales and dimensions of projects (e.g., theory/practice, experiment/design, etc.) you have a lot to say. In the end, it is you who decide what project you will start working on, not us!
MSc thesis timeline
In C&S, the first year of the MSc is about doing all courses (core, profile, and electives). In the second year (or Q4 of the first year), you generally start with the internship. After completing all courses (except Research Methodologies and the Literature Survey) and the internship, you can pick an available thesis topic in one of the control disciplines. Then, the timeline (and associated milestones and deliverables) of the MSc thesis is as follows:
Thesis printing arrangement
The C&S department will pay for three copies of every student’s final MSc thesis. This is done in cooperation with the professional print shop located in the Faculty of Architecture (CSinBKCity). In order to get your copies paid for, please visit the Stabilo page on BrightSpace.
Need help or more information?
To get support and advice in organising your courses (list of electives) and possibilities in graduation projects, please contact the C&S track coordinator dr ir Clark Borst.
