Published articles

During and after my Master thesis I authored and co-authored three articles. Two for a scientific journals (IEEE) and one for a Dutch professional magazine (Mechatronica & Machinebouw). While it took me a lot of time writing these articles and I do think that the current publish or perfish system is completly wrong – can write a whole blog about this or maybe read this article – I am glad I took up the challenge to write them next to my thesis.

Writing a scientific article is completely different compared to writing your thesis. For your thesis you have numerous pages to explain your story for a scientific article you only have a few. As a result, you will have to be very precise in your writing and choose the words carefully. It has to be very clear, precise and short. This can be very challenging when your story touches a lot of subjects from different fields. Angel Borja wrote a three articles on writing your article and I would like to share these with you:

In addition, I found the UW-Madisons Writer’s Handbook and especially the page about Transitional Words and Phrases useful.

Finally, without further ado, the three articles I have written.

Frequency-domain analysis of real-time and networked control systems with stochastic delays and data drops

Abstract—We present a novel frequency-domain analysis framework for a closed-loop model capturing a wide range of real-time and networked control systems with stochastic delays and packet drops. Our results allow for inferring the mean and variance of the output response to deterministic inputs, based on a new frequency response plot. We illustrate the usefulness of our results in the context of real-time control systems with input-to-output delays resulting from the use of a shared processor.

Keywords—Frequency-domain analysis, networked control systems, real-time systems, stochastic systems, data losses, delays

pdf

The impact of deadline misses on the control performance of high-end motion control systems

Abstract—In high-end motion control systems the real-time computational platform must execute tasks from multiple control loops operating at high sampling rates. In recent years traditional special-purpose platforms have been replaced by general purpose multi-processor platforms, which introduce significant fluctuations in execution times. While considering worst-case execution times would severely reduce the sampling rates, accepting deadline misses and assuring that the control system still meets the desired specifications is challenging. In this paper, we provide a framework to model and assert the impact of deadline misses in a real-time control loop. We consider stochastic models for deadline misses and characterize the mean and the variance of closed-loop output variables based on a time-domain analysis. We illustrate the usefulness of our framework in the control of a benchmark motion control experimental setup and in the control of a wafer stage in a lithographic machine.

Keywords—Deadline misses, data losses, packet drops, performance analysis, stochastic analysis, industrial case study, hybrid systems, cyber-physical systems, real-time systems

pdf

Smarter balancing between performance and costs in control systems

Abstract—In order to ensure the performance of control systems, the digital platforms on which they are executed are often over-dimensioned, which are unnecessary costly. State-of-the-art model based techniques make it possible to create cheaper control systems, without performance degradation.

Keywords—Deadline misses, data losses, packet drops, performance analysis, stochastic analysis, industrial case study, hybrid systems, cyber-physical systems, real-time systems

pdf

Tags: TU Eindhoven, Work

GEWIS GEMOLD 2014

From 28th of May to the 1st of June GEMOLD14 takes place. GEMOLD14 is the GEWIS version of the popular Dutch television program wie is de mol?, literally translated who is the mole?. The program is about 10 candidates who have to complete several assignments with which they can earn money. However, one of the candidates is the mole, it is his solely purpose to make sure that assignments fail and the team of candidates earn as less money as possible. At the end of each episode there is an execution, the candidates have to answer a number of questions about the identity of the mole. The candidate who answers the least amount of questions correct has to leave. Candidates can earn jokers, which turns a falsely answered question into a correct one, or free-cards with which they do not have to do the answer the questions at all and guarantees continuation to the next episode. At the end of the program there are only three candidates left, the mole and two candidates. The winner is the candidate, who is not the mole of course, who answers the most correct questions about the identity of the mole.

On the 23rd of February I got an e-mail from the GEMOLD14 committee with the message I was selected to participate as a candidate. Together with the e-mail I received a questionnaire which I immediately filled in. You can download it here, however, it is in Dutch. I also had to write a small introduction about myself for the website, which you can view here, also in Dutch. After that it was just waiting for the 28th of May.

On the 28th of May all 14 candidates gathered at GEWIS in the Metaforum at the university. Before I went to GEWIS I made a visit to the bathroom where I wrote mole on mirror with soap just as a tease for the other male candidates who might had to go to the bathroom after I went. Later we were delegated to come to balcony of the Metaforum. After a group picture with all the candidates GEMOLD14 was finally about to start. Tim the host of GEMOLD14 showed up and welcomed us. He explained some organizational matters and how jokers and free cards looked like. We handed in our mobile phones, wallets and everything else that we were not allowed to have in our possession during the weekend. Afterwards, he told us to prepare ourselves for execution! With only the questions from questionnaire someone was already going to leave the group. Everyone was in shock, while I only had an eye for the joker and free card which Tim showed us as an example and left on a small table. With an easy pace I walked up to the table and grabbed them while everyone was gasping and watching me. Unfortunately however, one of the other organizers came up to me and told me to give them back. As polite as I am, I gave them back. Very stupid afterwards since they really lay in the open and the organizers were only to blame themselves to leave them so in the open. In the meantime the monitor was placed. Ruben, the host who did the executions, called the first name: Rik — who I think is the mole by the way — got a green screen. He was followed by Celine, Mark, Jos, Arno and Levi who all got a green screen. Then it was Wout’s turn, he was one of the finalist of GEMOLD13. Unfortunately for him he got a red screen and we were left with 13 candidates.

Wout was executed and the remaining candidates and myself gathered between the Metaforum and the Main building. Tim showed up. He asked for two volunteers who were lazy, Jos and Annebel presented themselves after which they were escorted away from the rest. Tim asked for someone who still had to recover from the shock, Mike presented himself and he was also escorted. At the end Tim needed two more volunteers who would enjoy a little walk, Ingmar and Leroy presented themselves. The rest had to fetch our bicycles and were escorted by two people from the organisation, Merel and Serge. For clearance here a list of the people:

Two volunteers who were lazy: Jos and Annebel.
One volunteer who still had to recover from the shock: Mike
Two volunteers who enjoyed a walk: Ingmar and Leroy
The rest who were gonna travel by bicycle: Wouter (me), Jorik, Celine, Wim, Levi, Rik (mole!), Patricia and Arno.

We drove to the Opwettense Watermill. Tim was waiting for us, after we parked our bicycles Tim explained us the assignment. Two people would arrive at the Watermill in approximate 10 minutes. From there on we had to guide them within one hour to the camp location were we would spend the night. We were to act as a traffic signpost and were not allowed to communicate by any other means than that. We were given three maps in which not all roads were clearly visible. On the left side the street directions were given in plain text. After the instructions of Tim we gathered in a circle around one of the maps, discussing which positions were best to post someone as a signpost. In this discussing Levi took the lead. Celine also got hold of one of the maps, I asked her if I could have a look at it. She gave me the map, but not before saying that she wanted it back afterwards, of course I never gave it back. At the first point were we had to leave someone behind I told the group we had to leave Celine as a signpost, I took the lead in this discussion since I did not suspect her. Celine was left at the first street crossing and the rest bicycled further. During the road there was some discussion whether or not someone had to be left at a street junction however of course our logic kicked in and we decided to post no-one as long as the directions was straight ahead. Jorik was left next at a roundabout. Then Patricia was left at a crossing. After that there was some discussion if we took the correct turn. I quickly ended that discussion because from reading the map and street names I knew we were correct. It was very funny and suspicious to see that no one seemed to be bothered to actually look at the street names and the ones on the map. The discussions were more like — when you look at the scale of the map and the distance from this crossing to this crossing, this had to be the correct crossing. At the next crossing I told the group which was left, that I want to act as a signpost and to be left behind. Levi did not like that and wanted to leave someone else behind — I guess he did not trust me. Rik (mole!) however told Levi however that we could keep discussing but that it was useless and he should accept the situation as it was. I was left behind as a signpost and the rest continued. However at the end of the road they stopped and were discussing. Later I learned that there was confusion about the map since there should have been a straight clear road ahead but there was not. However during their discussion Ingmar and Leroy arrived at my location, quickly I acted as a signpost and pointed them to the correct direction. Ingmar and Leroy were gaining upon the rest of the bicyclists, since they were still discussing. Suddenly they saw them and drove of, however, leaving Wim behind as a signpost. From this point on I could only guess and learn from others what happened further on. Suddenly however Mike arrived, I was kind of surprised and without thinking I pointed him in the correct direction, very foolish of me so it appeared afterwards.

While I was waiting to be picked up by one of the organizers I started to write in my mole booklet. After a while the group reappeared and we drove back to the starting point to pick up everyone after which we drove straight to the camp location. At the camp location we had to step off from our bicycles and one by one we had to go to the gates from the encampment. Jorik went first, he however came back and told us that he did not get in. He told us that we were getting a yes or no question about if we wanted to get into the encampment for -500 mole money. In the meantime Arno did went to the gate and got in. At first I thought he told yes. Then Wim and Celine got in as well. I got to the gate and got the question which Jorik told me I would get, I clearly said no and got in. From what I learned afterwards Jorik told not clearly no but said “euh no“. Inside the main building of the encampment I learned that while there was some confusion on the road, after I was left behind, we succeeded with the assignment. When every candidate was inside Tim greeted us. He told us that Ingmar and Leroy were inside the encampment in time with which we earned 500 mole money. However, Mike also made it also into the encampment with which he earned a free-card. This did however cost -1000 mole money. So actually we did fail with the assignment, since we got -500 mole money in the end. The mole had succeeded and we went to sleep.

The next morning all candidates gathered in a small room. Tim joined and explained the lie and cheat assignment. The moles of the three previous editions of GEMOLD would join us they would act as a jury. Tim would then ask each candidate two questions, one had to be answered truthfully the other with a lie. The jury would then try to guess which answer was a lie. Each candidate could earn mole money when the jury did not manage to find out which answer was a lie. When a previously candidate managed to fool the jury the next candidate could choose to call. The mole money earned by the previous candidate would then be safe. If he continued and the jury would guess correctly which answer was a lie then the previously earned mole money would be gone. The amount of mole money earned would raise by:

10 mole
30 mole
100 mole
200 mole
500 mole
1000 mole
1500 mole

First Ingmar and Patrice who both did not manage to fool the jury. Then other candidates told me I had to go, I resisted a bit but gave in to the pressure. Tim asked me two questions, if I had a part-time job and what my favorite book was. Since I did not quiet understand the game at that moment, I was wondering if the question which were asked stood in the questionnaire, I decided to answer the first question falsely and the second truthfully. The jury then guessed wrong. The next candidate did not decide to call and unfortunately he did not fool the jury. The 10 mole earned by me were gone. After my question, I gave the assignment some more thought and I believed that we could have lied always regardless what type of question we were asked because the questions did not actually come from the questionnaire. Hence it could never be checked if the candidate was actually lying towards the jury. After I got executed, I looked it up and indeed both questions which I was asked were not in the questionnaire. As such I think I was correct during the game and we could all just have lied towards the jury. Unfortunately everyone was to truthfully towards the jury and we only managed to obtain 100 mole. We could have earned a lot of mole money very easily. This time the mole was not clever, we were dumb.

After the next assignment, I unfortunately got executed. I wrote this blog almost immediately after I was executed. However, I did not describe the events during the last assignment, I did plan to finish it at a later stage. However, as you can see that never happened. Because the episodes are being aired now, I decided not describe these events anymore from my point of view.

Mole booklet

Tags: Funny, Games, TU Eindhoven

Testimonial Master’s program TU/e Systems & Control

In November 2013 I was asked by the PR of the Mechanical Engineering department of the TU/e to write a story about myself and how I perceived the master Systems & Control. It would be used as a testimonial on the TU/e website. In the end only a part of my full story was incorporated, below is the full-text which I submitted.

TUe - Testimonial — Testimonial on the TU/e website.

Hello, my name is Wouter Geelen. Currently I am following the Master Systems & Control at the Eindhoven University of Technology. First I will tell you something about why I decided for the Master System & Control, then something about how I experience the Master track and what currently keeps my life occupied.

When I was finished with my Bachelor in Mechatronics (HBO) it was clear to me that I wanted to continue studying. Choosing at which University I wanted to do my Master was a really easy choice for myself for 2 main reasons. Firstly, since I was raised in a small town in the province of Limburg and did not yet wanted to drift off to far from my roots and secondly I did two years of my Bachelor in Germany near Düsseldorf and at that time I still went regularly to Germany to visit friends. Therefor I decided that the University in Eindhoven was best the best choice. However, choosing the master was a really hard and difficult decision. After a Bachelor in Mechatronics there are several possible Masters which you can choose from. In my case there were basically two options: i) Embedded Systems or ii) Systems & Control.

During my graduation project for my Bachelor I created a double inverted pendulum (video) which would be a typical project for students from Systems & Control, but with the knowledge at that time it was a difficult problem for myself. However, I successfully completed the project. After a lot of talking with my former teachers and a graduate from Systems & Control, who helped me with my graduation project, I decided to start with the Master Embedded Systems and that I would choose several Systems & Control courses in my free electives. When you come from a University of Applied Sciences you will have to do a pre Master first, before you can actually start with the “real” Master. I did the pre Master of Embedded Systems, it had a lot of fun and interesting courses. After I finished the pre Master and had started with the real Master it started nagging me that the control courses which I wanted to do for my electives where really hard because of a knowledge gap. Furthermore, the master Embedded Systems also focused really on energy-efficient chip design which I really did not find interesting at all. I am the type of person which just wants to pick a chip from the shelve after which I work my magic on it. For those two reasons I decided to switch to Systems & Control. Luckily most of the courses of the pre Master of Systems & Control where the same as the pre Master of Embedded Systems so I finished it quickly. Furthermore, after asking the examination board of Systems & Control for permission, I could take 3 courses from Embedded Systems with me to the Systems & Control. So luckily that detour of half a year Embedded Systems was not a complete waste.

The core program of Systems & Control covers really all the basics about control engineering and the whole cycle of model based design. There are very theoretical courses but also very practical courses, which, I think is perfect for a University of Technology. You can then namely apply what you learnt in your theoretical courses in your practical courses. You also have a lot of freedom in choosing your electives: i) there are a lot of elective courses from which you can choose from and ii) if there is a course which you really like but is unfortunately not on the list then, if you have good arguments, you can ask the examination board for permission. Which I did as you have read earlier.

For a lot of courses you have to work together with another student or a group of students. Which is nice since most students in System & Control all have a different background, as well in study, as in culture. As such groups get really mixed up, which is good for learning how to work with a very diverse group of people. Also all the Professors excel in their field of expertise, they really challenge you during college. Since the TU/e is a small University you really get to know them and can almost always ask them anything if you don’t understand something. So the distance between you as a student and the Professor is really small.

Inside the master Systems & Control there are several specialization groups. At first I did choose for the Hybrid Systems group. However , during the summer they did merge with the Control Systems Technology group, which made them one of the strongest groups at the TU/e and one of the strongest in control theory worldwide. The reason why I did choose for this group is very simple. During my graduation presentation of the double inverted pendulum I was asked if I could compute the best possible transition between the two controllers. The system namely consisted out of two controllers one for bringing the pendulum in upwards position and one for stabilizing it. At that time I could not answer that question yet, but now and thanks to the help of Hybrid Systems, which combines continuous and discrete systems, I can answer that question! Which makes me feel kind of proud of myself.

Nowadays I am busy with a literature survey about several Matlab toolboxes which help in analyzing and simulation of systems in which wireless networks are involved. Wireless networks are namely increasingly more being used for control systems. To give an example where one could have used a wireless network inside a control system is my double inverted pendulum, namely at the encoder where the two pendulums meet. In that case the encoder at the connection of the two pendulums is not wired anymore. Wireless networks however bring all kind of nasty behavior with them which influences the stability of the system. Furthermore, wireless networks combined with control systems are still hard to analyze mathematically. Therefore there are still a lot of open questions to be answered around this subject.

Tags: TU Eindhoven

TU/E LATEX HOUSE STYLE TEMPLATE

Last month I blogged about a template I made for my thesis at the Eindhoven University of Technology. It so happens that the TU/e actually has a $\latex$ house style template for reports. It can not be found threw any public channels, such as the corporate identity webpage of the TU/e. You will have to connect to the network drive \\winstorage\MikTex\MiKTeX29\. Which is accessible from the TU/e network or threw a VPN connection. An example of the package output can be downloaded here. If you want to have the source please contact me.

For now I will just stick with my own template. The TU/e template namely generates some warnings and errors. Furthermore, it is also limited in its capabilities. Of course I can expand it, however, I do not feel the need to do that at this time.

Tags: LaTeX, TU Eindhoven

TU/e LaTeX Template

On the second of June, my birthday actually, I will start with my Master thesis at ASML Holding NV. ASML produces machines for the production of integrated circuits. Using their machines other companies, such as Intel, Apple, Samsung and many more, can create, for instance, CPUs, SoCs, DRAM or flash memory. But this blog post is not about my graduation at ASML actually. It is about my $\LaTeX$ template which I created for my thesis.

The university where I study, the TU/e, does unfortunately not provide a nice $\LaTeX$ template for their students to use. Joos Buijs actually created a kind of template à la TU/e style. However, as geeky as I am, I decided to create my own template. An example of how it looks can be downloaded here, the complete source can be downloaded here. It might be possible that you find a bug or you have a question about the template, if so, please contact me. To give you a heads-up, most likely you will not have the beautiful font MinionPro installed. If so, and you want to compile the source, you will need to remove line 590 (\figureversion{lining,proportional}%) from the class file (thesis.cls) and select one of the other fonts in the main $\LaTeX$ file (thesis.tex).

Most features of the class will speak for themselves some might need a little explanation. The \thSupervisors command for instance can take an arbitrary number of supervisors. The input has to have the format {name} / {Affiliation}, and elements need to be separated by a comma.

\thSupervisors{%
{F.A.J. Rompa} / {Eindhoven University of Technology}, {D.S. Addiks} / {ASML Holding NV}, 
{J. Appelhof} / {Embedded Systems Institute}, {S.R.E. Wessel} / {Van der Lande Industries}}

The class also provides a nomenclature. Symbols are to be put in the file ./frontmatter/symbols.tex. In the file the same format is used as the \thSupervisors command accepts with an extra element, namely, {symbol} / {meaning} / {unit}. The unit can be expressed using the siunitx package. Any acronyms can be inserted in the ./frontmatter/acronyms.tex file which has the format {acronym} / {meaning}. Furthermore the class can insert a custom title page. The title page can either be a .pdf or .tex file which is set as \thTitlePageOne{pdf}{./frontmatter/titlepage.pdf}.

To create this class I used a lot of internet resources, especially the Q&A site tex.stackexchange, which is the number one spot on the internet for all $\LaTeX$ related questions. Here follows a list of all the internet resources which I consulted.

The template is ready for use but I will most likely give it an update in the coming months. I tend to make it more robust with use of the etoolbox package instead of the ifthen package.

Tags: LaTeX, TU Eindhoven