My Journal

Week 10- Conferences

Objective: Work on the Technical Design Document. Debug current work. Attend two conferences. This week I will be attending the SIGGRAPH Sandbox Symposium in San Diego, California, followed immediately by the STARS Celebration Conference in Charlotte, North Carolina.

What was learned: Lots... Plenty of applicable research.

Problems encountered: Airplane luggage troubles...

Week 9- Presenting my work

Objective: Create a presentation to show off my work for the summer. Write my paper. Get as much of the game done as possible.

What I've learned: Nothing. this week I'm regurgitating recently acquired knowledge.

What problems encountered: The more I work, the more problems come up from being unfamiliar with the system and not being able to find documentation. this is the reoccuring theme for this summer. If I had to do this over again, I would find a group on campus that is working (even if on unrelated things) on Microsoft Robotics Studio. Just working in an environment where many people are doing the pioneering work would have taken a lot of th pressure off of me. Oh well. Live and Learn.

Future work: Write a technical design document for future students that choose to work on this project.

Week 8- Reprioritizing...

Objective: So, because of major setbacks in the last couple of weeks. I have to seriously reconsider my priorities for this game. This week, I need to stop working on issues associated with not-working- assets and start getting the interface working. The robot needs to be controlled in ways specific to this project and player should be able to receive and accept missions.

What was learned: Documentation is important. It seems like new information is being added everyday, days after I figure out a solution to a particular project. It's important that I document my pioneering efforts for this project, as it looks like I will not be able to get a working game by the end of the summer.

What problems encountered: I'm running out of time.

Future work: Finish game(yeah right). Start working on final paper.

Week 7- Asset Generation continued

Objective: This game is starting to look more and more impossible. At this point I'm focusing on finishing designing level and starting on the GUI.

What was learned: I should really start looking to design smaller games. I'm still having trouble with the terrain, and I also have another problem with the terrain on top of the current problem. The image that I'm using to paint the terrain is high enough resolution, and in order to fix that I have to make the image insanely large. So large in fact that it takes all of my computers physical memory (2 gigabytes) in order to open the file in an image editing program. For obvious reasons, this isn't going to cut it.

Problems encountered: Listed above.

Future work: Finish the game.

Week 6- Asset Generation

Objective: So the design is nearly complete and now its time to start creating levels. This week is to be spent creating the assets that will be used in MARS2020.

What was learned: I've started with the MARS level, specifically with creating the terrain. The problem, however, is that I can't seem to make a terrain real enough to be a convincing Mars terrain. Ultimately, we want a terrain that maps directly with an area of Mars, so that students wll be able to see the direct translation of what they would be doing in this situation. The problem, however, is that I can't seem to translate overhead images of Mars into terrain data.

Problems encountered: Terrain generation is harder than I thought.

Future work: Finish he terrain and start developing other assets.

Week 5- Search and Research

Objective: Find related works from other researchers. My work is starting to resemble a black box which is not good. I’m hoping to find others that have attempted to do something similar with teaching robotics.

What was learned: There have indeed been many attempts at teaching engineering. Unfortunately, their aren't many game attempts that have reported positive results (as far as learning is concerned). Bad design and unclear objectives seem to be a common theme among games intended to teach engineering or computer science related concepts. Of the projects that seemed the most successfull they usually werent abstracted into a game form.

Problems encountered: There isn't enough time to do read these papers. But they're important to the design of this game. Game Design Document was not completed.

Future work: Start implementation, wrap up the GDD.

Week 4- Game idea confirmed

Objective: New game idea has been confirmed, but I still feel like there are a lot of unanswered questions. I decided that before I continue development, I should spend the week reading up on related works in the programming/engineering field. I also started evaluating some of my pasts works on the topic of learning in an immersive virtual environment.

What was learned: Caught up on a lot of reading concerning these topics and from there began writing a concept document for MARS2020. MARS2020 will be designed to tackle some of the past failures of mine and other related projects. I don't want the problem of bad design this go round (a problem a lot of people seem to have). I'm taking extra steps to ensure that time dedicated to a certain aspect of the game is proportional to the amount of influence it has on my overall goal.

What problems were encountered: It's taking too long to write the GDD, everytime I think I'm done, another aspect of the game comes up. I don't know if I even want to attempt a Technical Design Document. Maybe I can take the cheap way out and take a code and go approach? Heck, I don't even know if anyone else will see this project when I'm done with it... Also, Georgia Tech doesn't allow me to look through their online resources for technical papers. I'm beginning to like this school less and less. >.<

Future work:

Finish GDD. Start development. Make an immersive game in 5 weeks.

Week 3 - Change of plans

Objective: Pitch a Game Idea to Dr. Howard and Sekou that is doable within the remaining time. Dr. Howard decided that she would like my research here to focus more on teaching robotics to a younger audience.

What was learned: After harrassing Sekou about my new project until I knew exactly what Dr. Howard wanted, I did a lot of research on the applications of robotics, more specifically future applications. I decided that making a game based on a realistic future would be a good way to get the next generation interested in robotics and get them psyched about learning in game. Eventually, I was able to come up with a game idea that was doable, fun, and with varying levels of difficulty that I'm fairly confident that I'll be able to pull off in the remaining seven weeks.

What problems were encounted: I'm never sure that I'm doing the right thing... *sigh. Oh yeah, and I had to scrap the beautiful world that was sooo close to being finished, oh well... Next go round should be a peice of cake. Plus I'm sure that the Search and Rescue environment can be used for something. I'll hold on to it.

Future work: Game plan has been finished, presented, and green lighted. The remainder of this week will be spent creating a user manual, GDD, and TDD. Plus I'll be researching whether or not anyone has attempted to make a game with Microsoft Robotics Studio (although I think I already know the answer to that question)...

 

Week 2 - Microsoft Simulator 101

Objective: Get down the foundation for the urban environment that I will be using for the Urban Search and Rescue simulation environment.

What was learned: I learned a little bit about the Wavefront object format. Apparently a material file is needed in addition to the actual mesh file itself. None of the tutorials made mention of this, it was yet another important criterion for creating a world that was assumed or omitted. I intend to make mention of this on my blog. Also, there were differences between my desktop and laptop that caused my previous build to crash when I ported it to my desktop. I’d like to play some more with moving my build from one platform to another and see what causes it to act up the way it does.

What problems were encountered: Many hours were wasted trying to make the mesh textures work. The solution is actually pretty simple, though I’ve yet to figure out how to do it through my 3d modeling application.

Another half of a day was spent trying to port the application to my desktop. I’m still not sure what I did to solve my problem but as of right now it works.

The metro modeler was a little more complicated to use than I originally thought, but it didn’t take too long to adapt to the environment. I’d say the only result is that some of the buildings that I am using in my world look a little funky. I figure I’ll fix those later.

Future work: I still intend to port a robot into my environment but I’m not entirely sure what that robot is supposed to do once he’s in the world. I’m still awaiting briefing on what these robots will do once they are ported into my environment. As of now, all I know that the robot will do is walk around.

I also intend to create a very basic GDD for the robotics world. A Technical Design Document doesn’t seem necessary because most of my work as of now seems to be more scripting than programming.

Week 1- Learning the Ropes

Objective: Familiarize myself with the project and the tools that I’ll be using for the completion of this project.

What was learned: My objective is to provide a virtual environment for robot simulations. For the summer, I will be working directly with a type of Urban Search and Rescue project that was started prior to my arrival. After the summer, standard robots and associated protocols should be able to work within the environment that I’ve set up.

This week I started by familiarizing myself with Microsoft Robotics Studio. I’ve gone through various tutorials provided by Microsoft and feel like I should be fairly comfortable working in this environment in a short amount of time.

I’ve also become familiar with the work being done in the lab that I’m working in and why my work is necessary. It’s a very costly endeavor to engineer robots. Working out some of the design issues and testing in a simulated environment should cut cost significantly if done correctly. So far, I’ve seen some of the simulations that they’ve set up in 2 dimensional environments, it’s pretty cool buts I think it limits the amount you can accomplish within the simulation. Microsoft has provided a free solution for homebrew robot enthusiast and researchers. This should make it easier to create more realistic simulation solutions for our robots.

What problems were encountered?
Robotics studio seems like a very solid solution. So far, I haven’t found any real limitations, but a lot of the work going on behind the scenes is so far removed from the code that I’m working with, that it’s very difficult to control the things that I’d like to control. Also, because of the newness of the studio, documentation is limited.

Because of my limited knowledge on what’s going on behind the scenes, I feel like I’m wasting a lot of valuable time guessing on what’s happening behind the simulation engine. For example, it seems that a lot of the processes are occurring within different threads, even though when I write them, I’m writing them sequentially. Why didn’t they tell me that somewhere? Was I supposed to assume that’s what’s going on? Because I did, after about four hours of guessing….

Future work
I’ve started designing the world I would like to create in the simulator and looking at the different way’s to bring this world together. If the environment is to be Search and Rescue, I have to look at methods to make the world more dynamic, which is something they don’t show you how to do (and I’m not sure if they want you to do).