My Job... in summary

Well... since I didn't have much to write about the weekend I have decided to give people a run down on what I'm doing for work. Next weekend I'm going to Lucerne and Zurich in Switzerland, so will probably have a bit more to write about.

I have had many people ask "are you actually working or just writing a travel guide?" and also several requests to write something about what I'm doing. So... here goes... I'm pretty sure most of you will think "geek" and will go to sleep during this :-P

First a bit of a summary of what I'm doing here. I'm doing a PhD in Computer Vision: getting computers to see the world as humans do. This is a massive field, so my primary focus is; free-space detection in construction scenes for driver assistance systems using binocular stereo vision. (In English: using 2 cameras to look at the road, in road-works situations, and tell the assistance system where the car is allowed to go)

I am working at a DaimlerChrysler (Mercedes Benz for those of you not familiar) in a research department. The business model is very different here, basically they are doing the work that their suppliers should be doing. So if they generate anything beneficial, it is given freely to their suppliers. The company only gets a 6 months first to market advantage, then the other car companies can just use the same products.

The first few weeks here I spent setting my my laptop and learning the framework of the software that I am going to be using / writing. Basically learning how everything fits together and how the analysis works with the camera system. There is 20 years of research / development that has gone into it, so I was only scratching the surface. I hadn't touched coding in over 2 years, so it was a drop in the deep end and getting my head around things.

I also learnt how advanced they are in research. The research is about 10 years ahead of serial production. So the systems that are coming out now which you may be familiar with (auto-parking, ABS, anti-slip systems, collision avoidance) were started 10 years ago. They have already got cars that can drive themselves in certain situations, e.g. following another vehicle and grid-locked traffic. Of course, our main focus is the algorithms for the systems, I.e. designing "how a system works" as opposed to "designing a working system". So kind of the opposite facet of what I was doing in my Engineering Degree.

The next few weeks I spent writing manuals on what I had learnt. How to set up a laptop from scratch, and also overviews on the whole system aimed specifically at Camera Calibration (taking the left camera and mapping it onto the same coordinate plane as the right camera). It was during this time that my supervisor (Prof. Reinhard Klette) needed some assistance for setting up the project back in NZ. We are applying for funding so we can set up a "basic" car in NZ to record and analyse data. The aim is we will have a small research facility in Auckland that is compatible with the research here at DaimlerChrysler.
I was very comfortable with this part of it as, it was similar to the organisational stuff I had to do at COGITA. A nice way to ease me into my stay here.

The last couple of weeks I have actually started getting my hands dirty back in programming. Remembering how everything works with computer software and trying to read other peoples code. I got a lot of assistance from Clemens (guy at work that knows everything) to get me up to speed. I have started producing some real results now, working in my research area. The theory is starting to settle in, and because it is all on-screen visually, you can see your work immediately which is nice.

The rest of my time here will be continuing my research and increasing my theoretical knowledge about computer vision. Also assisting the Masters Student back in NZ when he is having trouble with the software. I may be going to a conference or two, so I can network and get new ideas on what types of algorithms / methods I can apply to my area of research.

In my final weeks I will be concentrating on learning how the car computers communicate to each other, and how to get the information out of the car. The aim is to get enough knowledge so we can set up the car back in NZ. The aim of this task is I will be co-supervising a Masters students who will do the car set-up with me assisting when they get stuck.

When I get back we will hopefully have some more PhD and Masters students, who I will be working with and co-supervising, so we can expand the project.

Hopefully that brief summary of what I am doing didn't go over your heads too much... and I won't blame you if you went to sleep and/or skipped half of it. I know I'm a geek... and better start accepting the fact ;-)

Bis später (see ya later)