This week I learned about the basics of motion planning. I was introduced to the concept of configuration space, or c-space. Furthermore, I learned about the concepts of obstacle space and free space, and how these affect motion planning. Furthermore, I am now aware that motion planning becomes increasingly difficult when a robot has many degrees of freedom. I have learned about the various approaches used to address the problem of motion planning, from the early complete method, to the probabilistic road-mapping approach. I was particularly intrigued by the various variations of the PRM method, like OBPRM, MAPRM, and Toggle PRM.
This week I got a chance to begin familiarizing myself with the Group Behavior group's code-base. Playing around with the code has allowed me to see the way road-mapping techniques are implemented.
Next week, I am going to learn more about the code base, and do more experimentation with the environments. I am also going to familiarize myself with the field of motion planning by reading more papers. Furthermore, I will utilize the slides from Dr. Amato's algorithms course to learn more about that subject. The future looks bright at Parasol!
This week I achieved two things: greater familiarity with the field by reading papers, and greater familiarity with the code base. I read three papers: "OBPRM: An Obstacle-Based PRM for 3D Workspaces", "Better Group Behaviors in Complex Environments with Global Roadmaps", and "Reactivity and Deliberation: A Survey on Multi-Robot Systems". These helped me understand the motion planning field, the solutions that have been implemented, and the problems that persist. Secondarily, I achieved greater familiarity with the code base by going tracing through files and asking questions when necessary. Tracing through code was challenging initially, because I was unaware of certain convention, however by unitizing my resources I've surpassed many hurdles.
Next week, I will be going through more of the code base, and beginning the primary phases for developing the marker class. Albeit the fact that I'm reading less papers, I feel that I am still learning a great deal.
This week I learned a great deal working through, and experimenting, with the PMPL code base. The latter was interesting because it was very structured, in my opinion. Experimenting with the PMPL code-base and corresponding GUIs helped me to see how the techniques of motion planning look when implemented. In addition, I was privileged to attend two PhD defenses by students from Parasol. Nathan and Timmie's presentation of their work on on STAPL was illuminating, It was also inspiring to foresee what it will look like when I defend. Next week, I will be completing my first mini-project, an implementation of a vision sensor for agents, which I began this weekend.
This week I made progress in the implementation of my agent sensor, and learned to work proactively in the lab. Firstly, I completed several of the assigned iteration for the mini-project. In addition, I tested the class I wrote in the preexisting code-base and environments. The testing I conducted helped reveal several flaws in my implementation, which led to revisions.
My mini-project focuses on creating a sensor that detects agents within a particular view radius. The project was divided into several smaller iterations. The first part was the creation of the vision sensor, with a function to return a vector of pointers to all the agents. This fits into the larger picture, because it gives me experience with the Group Behaviors sensor suite, which I will be extending to include marker detection.
I tested my code by loading my sensor in other files in the code-base. Then, I worked to remove errors that occurred when compiling. Secondarily, I tested my sensor by initializing it in one of the preexisting environments. This was to make sure that the sensor was functional in the environment. The latter revealed that the sensor was indeed functional. I am currently finishing the second iteration - sorting of the agents being returned.
Secondly, this week I learned to be proactive in the lab. The physical absence of my graduate mentor in the lab meant that I had to learn to ask questions, and communicate via email. This experience was instrumental in helping understand that independence is necessary, to some extent. Furthermore, I learned to utilize the resources available at Parasol, i.e. the other graduate students (Jory), who were very helpful.
Next week, I will continuing with my project, and finishing the current implementation. Furthermore, I will begin preparing my progress report.
Finally, next weekend I'be going to Austin with some friends for the July 4th celebration. Hopefully, we'll also get to visit Houston!
My friends and I are definitely going to Austin and Houston this weekend! We're going rent a car, plug-in the GPS, and explore. This is on of the best parts of the REU expierence: the ability to have mini-adventures. It's going to be a great July 4th celebration!
This week I was able to complete the first three iterations of my mini-project. I successfully implemented the agent sensor class I was working, and tested its functionality in simulation. Testing revealed several flaws in my implementation that I had to correct before moving ahead. After doing that, I was able to verify that my sensor was functionally properly.
I have started implementing what we discussed at Thursday's meeting - the creation of a triangulation function, that will allow an agent to localize based on information about two other agents. I have been researching how triangulation is accomplished, in a mathematical sense. I am currently re-familiarizing myself with the geometric {and trigometric) concepts that I'll need to utilize in solving this problem.
Next week I will be completing the triangulation function, and finishing my progress report.
Last weekend was great! My friends and I went to Houston on Saturday, and then to Austin on Sunday, for the July 4th celebration. Unfortunately, the fireworks shows were cancelled due to a fireworks ban in Texas. Nevertheless, we had a great checking out vintage shops and eating BBQ!
This week I made a great deal of progress in completing the triangulation program that was discussed last week. At this point, the program functions as follows. It takes in two points, each represented by a x and y location, and their distances from the current robot. Using this information, the program calculates: the distance between those two objects, and the current position of the robot. I am very close completing the implementation of the program. I will be running several tests on Monday of next week,to make sure that the implementation is sound.
Secondarily, I completed the progress report required by the REU. I will completing the progress report required by DREU before Monday.
I am slightly concerned about the preparation of my final poster and paper. I feel, to some extent, that I have not completed enough work this summer.
My summer is slowly approaching its end. Research has been awesome here at TAMU. I am already beginning to reminisce!
This week I made further progress in the completion of my agent sensor program, particularly its triangulation function. Based on preliminary testing, the triangulation function works well, but runs into trouble in certain scenarios. When attempting to triangulate the position of a robot using two given landmarks, the program miscalculates the position. This problem might be a result of the fact that there is no method currently implemented in my program to check the robot's facing direction, and determine the position of landmarks: to its left or right. I am currently working on adding this functionality to my program. Nevertheless, I am unsure as to whether or not this is exactly where the problem lies.
This weekend, and next week I will be working on the following: completing the agent sensor; writing my research paper; completing my website (for DREU).
This was my last week at Parasol and it was truly bittersweet. I was excited to be completing my work, but I am also sad to have to leave the great friends and mentors I've made here. I went out with several friends to Houston and Austin - on the same day! We ate really good food and did some fun activities.
This week I completed my portion of my research, particularly a localization sensor that uses principles from triangulation. I had alot of debugging and testing to do, but everything finally works. My partner Thaddeus will be taking over and finishing the implementation of the marker-based system. I am looking to seeing the work he completes.
If you're reading, you are awesome!!! I encourage you to participate in a summer research program. Please contact me if you want any more information. Peace!