During the spring semester I have been taking a Molecular Biology course to fulfill a requirement for my biochemistry major. In the latter half of the semester we conducted unguided research on characterization of genes from the rice, Oryza sativa, genome. The research itself is far from novel or interesting, but I figured I would post the lab report for the project as an example of the biochemical work I do since the computer science half of me controls most of this site. The report can be found here.
I have finally completed the final revision of the paper for my senior project on genetic algorithms for pathfinding with robots, which will be turned in tomorrow. My partner and I have also printed out the poster for the project for one of our two presentations we must give. The first presentation is just a poster presentation where we will be in a large room with other seniors presenting their own projects. Later that day we will have to give a 30 minute presentation and demonstration for faculty. I will soon update the pathfinding GA page to reflect all of the research that went into the project and where it currently stands. For now you can view the paper and poster here and here.
The pathfinding genetic algorithm I have been working on during my senior year is entering the final week. My paper and source code are due to my advisor by a week from today. After that I have until the beginning of May to prepare a poster and presentation for the students and faculty. Currently the OPEN-ROBOT we have been recently using can view and map the world around it using the two IR sensors mounted to its front. My genetic algorithm successfully generates nearly optimal paths which are sent as movement commands to the robot. My algorithm in conjunction with the vision algorithm my partner developed allows the robot to slowly but surely pathfind around objects to a specified goal. As of right now the goal is hard-coded in and there is only one robot (so no predator-prey dynamics). When I have time (and when it is finished) I will update the main project page with the full details of the project along with my paper. We are planning to continue development on this project after graduation and are looking to get published in an undergraduate research journal and hopefully have the school fund us to attend a robotics or AI conference in the coming year.
The OPEN-ROBOT has arrived and we have assembled and began testing its various features. Some work has been done integrating the robot’s functionality with my AI and driver code. Hopefully we can soon have the robot turning and scanning the area with the IR sensors and then sending the data to my AI code for path planning. It seems as though some of the wiring was done incorrectly with the IR sensors and wheel encoders, so fixing those are first priority. Once that has been done, actually getting the robot to sweep its field of view and return values we can work with should be fairly simple. The major hurdle will be figuring out the “resolution” of our robot’s world view and how it is represented in code. Getting a feel for what distance corresponds to the IR sensor values should provide major insight. Here is a short video clip showing some of the robot’s current abilities:
After much deliberation, we decided to stop building our current robot and instead buy a kit which basically contains all of the parts and functionality we had been tyring to attain with our own robot. This decision was made after we realized it would take us most of the semester just to finish building the robot with all of the features we wanted, leaving little or no time for properly integrating and testing my AI algorithm. The kit we will be working with is called OPEN-ROBOT and can be found here.
This kit comes with some open source software written in C# for communicating with the robot. Because of the availability of the C# code, I went ahead and ported my AI code written in C and my front-end code written in Python to C#. Currently all of the code external to the robot has been properly combined and now all we are waiting for is the robot to arrive so that we may begin testing.
I have been working on my senior project in robotics over the winter break, mostly trying to get my AI code to run on the robot. My partner and I realized that our chip just did not have enough memory or power to run my algorithm (we bought a really cheap model to save money). We decided to run my pathfinding algorithm directly on a computer and send the appropriate data back and forth to the robot. My partner is currently working on setting up a board with a Xbee wireless module that will be plugged into a computer which will be used to communicate with the other robots (which will have wireless modules of their own).
Shortly after submitting Bubble Blow to IGF 2009, we were contacted by GameCareerGuide asking myself and the other three people who worked on Bubble Blow for an interview. We happily obliged and after a short wait they have finally put our interview in the features section on the front page much to our delight. The article can be found here.
I have finally finished the DNA sequence finding program for the Biology department at my school. As I was finishing it, I realized the program was much more powerful than I had previously thought. The program was originally designed to search for transcription factor binding sites in DNA but in actuality it is a graphical front-end to regular expressions.
I have finally “finished” the projects page of the site where you can find a collection of my more interesting projects. Some of the projects are available to download from their respective pages.
With my old site up for renewal, I decided to switch hosts and overall website format. ….so please bear with me as I update this site and get everything back up to speed.
The OPEN-ROBOT has arrived and we have assembled and began testing its various features. Some work has been done integrating the robot’s functionality with my AI and driver code. Hopefully we can soon have the robot turning and scanning the area with the IR sensors and then sending the data to my AI code for path planning. It seems as though some of the wiring was done incorrectly with the IR sensors and wheel encoders, so fixing those are first priority. Once that has been done, actually getting the robot to sweep its field of view and return values we can work with should be fairly simple. The major hurdle will be figuring out the “resolution” of our robot’s world view and how it is represented in code. Getting a feel for what distance corresponds to the IR sensor values should provide major insight. Here is a short video clip showing some of the robot’s current abilities: