2023 Semester 2 Study Reflection

My Goals

So at the beginning of the semester I wrote this post about starting this blog for uni and the goals I had for the semester. The gist was that I felt I was spending too much time taking notes and I wanted to try writing blog posts instead as I thought this would help me to learn the content better by having to summarise and explain it.

So, did I follow through with these goals? Well partly… I don’t think I took any notes in classes this semester but writing technical blog posts was more time consuming than I imagined and it was unrealistic to write one for each class each week. Most of the courses I took this semester were quite time consuming with assessment items and some amount of team work in each which meant I made the decision to focus on completing assessment and going to classes rather than writing more blog posts.

I finished my last exam almost two weeks ago now and so more recently I have been writing more blog posts on some of the content from some of the classes (mostly things relating to control engineering and robotics). I aim to continue writing some of these technical posts about the things I have learned this semester over the coming weeks.

Results

So how did my strategy of taking no notes in class go? Well firstly - as a good engineer - I think it’s important to note some of the limitations of this little experiment. Namely, this is just my opinion and experience so it really just serves as an anecdote rather than any hard evidence.

To cut to the chase I got all 7s (High Distinctions) for the four subjects I did (Control Engineering 2, Robotics and Automation, Advanced Dynamics and Vibrations and Professional Practice and the Business Environment).

And to be honest, it didn’t feel like choosing NOT to take any notes impacted my learning very much compared to other semesters. It is of course hard to be objective about these things since I don’t have a perfect control for this experiment. However, I do think I was able to engage in class better and understand concepts as they were presented a bit more easily since I could think about them rather than trying to write them down.

One other new thing I tried this semester was using the recommended textbook excerpts as my primary resource for content for my Control Engineering 2 course. In the past, I have not bothered much with textbooks and mostly just learned content from the lecture slides. I have to say though, the textbook excerpts that were provided were very helpful. They covered the content thoroughly and in many cases to a greater depth than the slides. Yes, sometimes it could be a bit difficult getting stuck on a part and not being able to ask questions but usually continuing on and coming back to it later was a good strategy.

Short Course Reviews

Advanced Dynamics and Vibrations

This course was a classic mechanical course broken into three sections.

The first part delved into multi-degree of freedom linear systems with a focus on modal analysis. The assignments focused on a hypothetical consulting problem for a wind turbine where passive methods were used to control vibrations in the system. This means designing dimensions for the blades or gearbox e.t.c to minimise vibrations during operation.

The second part moved into continuous systems focusing on strings and beams. Again, modal analysis was used to analyse torsional, longitudinal and transverse vibrations. Furthermore, some approximate techniques (Dunkerley’s and Rayleigh’s methods) were used to model the effect of point masses. Finally, there was some brief dicussion on Finite Element Analysis (FEA) which uses software to model continuous systems by breaking them down into lots of small parts making them MDOFs. We used ANSYS software for part of the consulting assignment to model the effect of tapering turbine blades.

Finally, there was a brief discussion about instruments used to practically collect vibration data such as accelerometers and how to analyse this data. The effects of noise and discretisation of data on the spectrums produced using FFTs are some examples of this.

Overall, I didn’t mind this course. There are parts such as the signal analysis at the end which I thought were slightly more interesting but the ANSYS modelling felt a bit tedious and uninteresting.

Robotics and Automation

This was my favourite subject this semester as I predicted at the beginning. I have heard in the past this course was more heavily focused on inverse kinematics for robotics however this year there was a new course coordinator and she seems to have made the course much more broad.

The first part of the course mostly focused on stochastic models in robotics for things like movement and sensor models, mapping, localisation and SLAM. Next some path planning was introduced with some classical graph traversal techniques including Dijkstra’s Algorithm and A* as well as techniques for creating these graphs such as PRMs and RRTs. Next, we looked at some forward and inverse kinematics using screw coordinates and then the homogeneous transforms associated with cameras. Finally, there was a brief overview of how Reinforcement Learning could be used for some more difficult tasks in robotics. RL is something I am very keen to learn more about so this was cool to see.

The assessment was based on programming some turtlebots using ROS2 (in Python) to run SLAM. We had to design an algorithm to fully explore a static map (that we did not know ahead of time) and localise some ArUco markers around the map.

Control Engineering 2

This course was also pretty interesting and introduced more control techniques.

The course actually began with nonlinear control using Lyapunov theory which I think was kind of a strange choice but nevertheless I thought the concepts were quite interesting and this was probably my overall favourite part of the course. Next, linear state space models were explored including many of the concepts I have been writing blog posts about recently such as state feedback and controllabliity, reference tracking, LQR/I, observability and Kalman Filters.

Overall, I thought this course was pretty good but I think some more depth or even other techniques for control such as RL would have been cool to see.

Professional Practice and the Business Environment

Although I didn’t mind the idea and content of this course it just seemed like it was missing a lot of depth. The main focus was on business management with assessment based around a startup idea in groups. The course began with an overview of how to come up with a value proposition and then continued with different analyses that could be conducted for the proposal. First the Discounted Cashflow Analysis was used to help get an idea for the value of the business cashflows. Next there were some details about what it means to be a professional engineer and how to get chartered after graduating. Finally, there were some details about business sustainability.

Most of the content was very easy and lacked depth. I really wish there was a bit more to it because I think I could have really liked this subject if that were the case.

What Next?

Well in the next couple of months I have a few things I will be focusing on.

Firstly, I will hopefully continue to add some posts about the things I learned in the last semester of uni to help consolidate my knowledge, have some personal notes to look back on and to hopefully help anyone else who might stumble upon them. It would be cool to look into adding some of these posts to a more public blogging website so that could be worth a look too.

I also want to lean in more to software engineering and particularly AI since this is what I am most interested in at the moment. For now, I have been focusing more on some fundamentals and trying to complete the daily leetcode challenges mostly in C++ to keep up my skills in that language. Aside from that, I plan on doing the Advent of Code challenges this year starting on the 1st of December in C++ or maybe Python or even TypeScript. Again, I just want to keep my skills up in these languages and this seems like a good fun way to do so similar to leetcode.

I’m also hoping to finish off an MVP for an app for a tutoring business I co-founded. It’s quite small at the moment but I am hoping that getting this app done that helps automate some of the admin will mean we can start to scale up soon.

After that I will be starting my placement at Hatch in mid January where I will be commencing a six month project for my Master’s thesis. The project will be based on computer vision which is great because as I said before I am keen to learn more about AI and deep learning.

In the longer term, I will have to start thinking about what I want to do after I finish hopefully at the end of next year. The safest route most people take is to just apply for a bunch of grad positions and work that for at least a couple of years and I will definitely be on the lookout for applications soon but there are a couple of other options I am also seriously considering.

Firstly, I do think that I will be interested in further study and probably a PhD at some point and so I may start applying for those opportunities as well. I’d really like to research something in AI, IoT or biomedical engineering that could benefit disadvantaged people. For example, I think there are some really exciting developments happening in AI right now related to logical reasoning and chain-of-thought and I would love to look at how these could be applied to education systems. From tutoring I have learned that helping students to learn is all about guiding them by prompting them with the right question which relies on gauging what they do an don’t understand. In other words, its important for you and the student to understand what steps in they struggle with in problem solving and focusing on that. If AI was capable of doing this I think this support could become affordable for so many more students and improve educations worldwide. Regardless, I want to start looking for papers in these areas and see where might be good to apply.

Another option I am considering is looking at a startup. This is the most scary because I feel like there are so many things I don’t know both technically and in terms of business and management. It also seems obvious that my initial income would be much worse and if things do not go well I am not sure how convincing employers of my skills would go. On the other hand, it seems like being young is the ideal time to take these kinds of risks because I don’t have much responsibility to other people that I need to worry about. Regardless, at some point in my life this is something I would like to do since I believe it is one of the most impactful things I could do.