.:: Previous Week ::: Next Week ::.
Week 3:
~ Monday, June 11 ~
Yay! A new work week :) I am very excited about starting on my project – it looks like we will finally be doing some real hands-on research.
I tried writing C++ code to capture data from the wiimotes. In particular, I was wondering whether we can record the data from each controller into a separate file. However, this morning when I was working on solving this problem I got stuck :-( For some reason VS was giving me a run-time error, something to the extent “A problem has occurred; we don't know what is wrong; all the info you've worked with is lost; hope you know how to fix it!” I was also unable to debug the code because “debugging info is not defined or does not match. Binary was not defined for debug info.” And then, about half-way into solving this problem I realized that the mocap lab doesn't even have VS 2005 installed on their computers. I really didn't want to go through all the trouble of installing it today. I decided to switch to GlovePie.
I quickly modified my GlovePie code to work for 5 wiimotes. I was also wondering how I should organize the output: the GlovePie writes all the information into a single output.txt file, which meant I better know in which order I am recording the information, because there is going to be a lot of data, all in one file. It was a tough decision to make because I didn't know exactly how we will be extracting and using this data.
 |
| Khristine: "Do I look like a mummy?" Hmmm, wii-mified mummy - that's trendy :) |
Thanks to Kristen I was spared from wearing a tight spandex suit required for motion capture. She volunteered to help and be my "test subject." The girls in the lab helped her put all the markers on and then I figured out how to place and secure the wiimotes on her. We had some problems with the cameras' orientation and I had a tough time with GlovePie, which refused to record the data. However, in less than an hour all the issues were resolved and we were rolling along. The part that took the most time was when I had to open the output.txt file after each take and copy the last test's info into a different file. However, I very quickly found an easy way of operating the mouse so that I could select the data a lot faster!
Things to consider for the next time:
~ Tuesday, June 12 ~
I've been thinking about the problem with my code that I was having yesterday, and today, when I turned on my PC I knew exactly where to look for the culprit. The reason it was complaining was because I added the fifth wiimote but didn't change the declaration (why couldn't the compiler just tell me that?).
Other than that the day was not very busy. I was working on my journal for the website and the travel request for the SIGGRAPH conference. At the end of the day we had our usual Tuesday night meeting: Jim was presenting an amazing paper on Simple Biped Locomotion Control. It was a very interesting talk but it was even more intriguing seeing how many different aspects of the paper the Graphics people could find to criticize. I liked how meticulously they went through the animations that showed the results of the project and found faults and bogus details. It is so great to be around professionals and learn from the best. I am very happy to be a part of this group!
~ Wednesday, June 13 ~
 |
| Bluetooth interface by Toshiba: supports up to 5 wiimotes, is very simple and intuitive. |
Today I also had a meeting with Jimmy. We went over [one of Jessica's papers]. We've discussed our plans for the next few days and I outlined the steps that need to be taken. One of the major things that both of us have to work on is understanding Principal Component Analysis (PCA).
Just in case you are curious, here's a simple definition of PCA taken from Umetrics :
Principal Component Analysis ~ A transformation where the data set receives a new coordinate system, in which new axes follow the direction of greatest variance in the data set.
~ Thursday, June 14 ~
Today I worked on my C++ code for the wiimotes. One of the goals that I set at yesterday's meeting with Jimmy was to have each recording of data time-stamped, so that we can check how well the controllers are synchronized with the mo-cap cameras. I found a timer class that I can use for this purpose.
However, before I could test the timer class I need to have an application file that handles the recording procedure. I am planning on giving the user the ability to start / pause / stop the recording using the wiimote's buttons. In between working on my code and getting my website to run on the CMU server, I attended an interesting presentation by Prof. Auke Jan Ijspeert from Switzerland on locomotion control in animals and robots: the robot's motion was modeled after a salamander!
~ Friday, June 15 ~
Today I got an amazing opportunity to talk to Jernej, who is graduating with his PhD this summer, and Alla, a postdoc at CMU, who will be teaching in the fall at the University of Pittsburgh. I am so pleased that I got to discuss issues with them concerning grad schools, teaching careers and job applications, industry vs. academia and computer science in general. I got a lot of great ideas and the best kind of information I could ever get: an insight from people who have "been there and done that."
After lunch I had another meeting with Jimmy. We discussed our progress at the moment: none of us seemed to have any issues with PCA, so we ended up discussing the problem I had with my code. It looked like once the button was pressed its state wouldn't change even though the user released it. I found the portion of code that dealt with parsing the buttons' signal. Afterwards, we tracked the various variables until we found where in the code the states of the buttons were updated. We've agreed that I will look into it and try to fix the API. The rest of the day I spent learning about input channels and how they work, about packets that are sent over these channels and about C-like syntax and the meaning of hexadecimal values in bit masks.
I am ready for the weekend!
~ Saturday, June 16 ~
 |
| The view of downtown from the top of Mt. Washington |
It was a wonderful day! I met up with Kristine and Manfred and together with a couple of Manfred's friends we headed to Mt. Washington :) The weather was very nice and the view form the top of the hill was gorgeous. Riding the inclines was also fun: I loved watching downtown unfold in front of me as we rode up.
It was great that after the trip to the mountain we also got a chance to visit one of the most famous eating places in Pittsburgh: the Primanti Brothers restaurant. Their claim to fame is that you get an entire meal in a sandwich: french fries and coleslaw are served inside the can't-open-my-mouth-wide-enough-to-fit-it sandwich!! It is very yummy, especially if you put Red Devil hot sauce on it (the sauce is not too spicy but it adds a nice flavor).
When Manfred and I got back to Oakland, he taught me how to play frisbee. I had a great time but, ouch, does it hurt when the frisbee gets your nail or the tip of your finger!
.:: Previous Week ::: Next Week ::.