Automated play group idea.
5 posters
Page 1 of 1
Automated play group idea.
I would like to head up a team whose responsibility is to create an automated play funcationality for the robots. We would have many goals to accomplish this semester that can mostly be broken up into two sections. Sensory, and intelligence.
Sensory:
1. Using the webcams on the robots and the sensory equipment in room 15, create a 2d projection of the robots position in the room.
2. Using the webcams on the robots and the sensory equipment in room 15, discover a method to determine the possible destination of the currently aimed dart.
3. (Optional) Using the webcams on the robots and the sensory equipment in room 15, determine if a dart has been fired from the opposing robot.
4. (Optional) Using the webcams on the robots and the sensory equipment in room 15, determine in a hit has occured either robot.
Intelligence:
1. Create a 3D model that we can use to test and develope the intelligence of our robots, using a 3D renderer like Oger3D.
2. Create attacking and defending strategies for the robots based on numerous scenarios.
3. Feed sensory data into our model frequently.
4. Use our model to determine the action of our robot.
5. (Optional - extremely difficult) If a dart has been fired from the opposing robot, determine if it would be possible to create a missile defense strategy, by either shooting the dart down or dodging.
I talked with Dr. Brooks about this last night and he seemd to like it, as long as it would be approved by the rest of the class. We would have to interface with atleast the robot group, the webcam group and the nerf dart group. I have even noticed other classmates who would be interested in this idea, if you are interested or oppose this idea please post your opinion in this thread.
Sensory:
1. Using the webcams on the robots and the sensory equipment in room 15, create a 2d projection of the robots position in the room.
2. Using the webcams on the robots and the sensory equipment in room 15, discover a method to determine the possible destination of the currently aimed dart.
3. (Optional) Using the webcams on the robots and the sensory equipment in room 15, determine if a dart has been fired from the opposing robot.
4. (Optional) Using the webcams on the robots and the sensory equipment in room 15, determine in a hit has occured either robot.
Intelligence:
1. Create a 3D model that we can use to test and develope the intelligence of our robots, using a 3D renderer like Oger3D.
2. Create attacking and defending strategies for the robots based on numerous scenarios.
3. Feed sensory data into our model frequently.
4. Use our model to determine the action of our robot.
5. (Optional - extremely difficult) If a dart has been fired from the opposing robot, determine if it would be possible to create a missile defense strategy, by either shooting the dart down or dodging.
I talked with Dr. Brooks about this last night and he seemd to like it, as long as it would be approved by the rest of the class. We would have to interface with atleast the robot group, the webcam group and the nerf dart group. I have even noticed other classmates who would be interested in this idea, if you are interested or oppose this idea please post your opinion in this thread.
making- Posts : 35
Join date : 2009-01-08
Re: Automated play group idea.
I would also like to work on the automated play functionality as well. However I think it is more important that we focus on the required on functionality while a few of us work on the automated play as a "side" project. I would be more than willing to do any image processing/computer vision software that needs to be done for automated play, I have had plenty of experience with this.
As far as creating a 3d model, I don't really see a need for this, the playing field only has two dimensions (x and y). So I think a 2d image that shows the location of the the two rovers would not only be simpler, but more practical for this project.
-Brian Peasley
As far as creating a 3d model, I don't really see a need for this, the playing field only has two dimensions (x and y). So I think a 2d image that shows the location of the the two rovers would not only be simpler, but more practical for this project.
-Brian Peasley
Peasley- Posts : 44
Join date : 2009-01-08
Re: Automated play group idea.
The main reason I believe we would need a 3d model is for testing and creation.
Creating a scenario such as this in Ogre3D (which is in C++ and open source) would take about 30 minutes if that, and we could use it to draw attack lines, show the range of attack of the tanks, and we wouldn't be confined to room 15 for the entire project. Working with the interface groups would make it extremely simple as we would only need basic move and fire functions for the tank and dart thrower.
I have done alot of work on game design and this is pretty much the exact thing, creating a tank with artifical intelligence. And as such there is alot of research and help on the subject as long as we keep it in the realm of game design.
Creating a scenario such as this in Ogre3D (which is in C++ and open source) would take about 30 minutes if that, and we could use it to draw attack lines, show the range of attack of the tanks, and we wouldn't be confined to room 15 for the entire project. Working with the interface groups would make it extremely simple as we would only need basic move and fire functions for the tank and dart thrower.
I have done alot of work on game design and this is pretty much the exact thing, creating a tank with artifical intelligence. And as such there is alot of research and help on the subject as long as we keep it in the realm of game design.
making- Posts : 35
Join date : 2009-01-08
Re: Automated play group idea.
Also it would be pretty simple and within our ability to map the tanks position on an image to the 2D space in a 3d model.
making- Posts : 35
Join date : 2009-01-08
Re: Automated play group idea.
I can set up another forum for this group if you want to do this. Would that be helpfull?
Re: Automated play group idea.
Definitely think that would be good idea to implement that feature in the project
ccherry- Posts : 20
Join date : 2009-01-08
Re: Automated play group idea.
Peasley wrote:I would also like to work on the automated play functionality as well. However I think it is more important that we focus on the required on functionality while a few of us work on the automated play as a "side" project. I would be more than willing to do any image processing/computer vision software that needs to be done for automated play, I have had plenty of experience with this.
As far as creating a 3d model, I don't really see a need for this, the playing field only has two dimensions (x and y). So I think a 2d image that shows the location of the the two rovers would not only be simpler, but more practical for this project.
-Brian Peasley
working on what NEEDS to be done would be best, but if people think that they want to do this and can get it working fast and neat then go ahead but I would suggest only a few peoplw (2-3) working on this or people doing this as a side project and take a section/group that does not require as much work, such as the group that does the robot movement interface to the laptop or the group that does the webpage.
-David Marshall
dsmarsh- Posts : 54
Join date : 2009-01-08
Age : 38
Re: Automated play group idea.
Yea I was thinking something along the same lines, focus on the requirements and those of us who are interested could take it upon ourselves to do this additional work.
-Brian Peasley
-Brian Peasley
Peasley- Posts : 44
Join date : 2009-01-08
Similar topics
» Just general idea from group leader meeting for us
» What is due next week?
» Group Meeting
» New Group Assignments
» Rover group needs.
» What is due next week?
» Group Meeting
» New Group Assignments
» Rover group needs.
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum
|
|