Team Twelve/Journal

From Maslab 2006

Jump to: navigation, search
Maslab teams
Team 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 15
Team Twelve's Journal · Paper

This is our Team, temp name BURRITO's, journal

1/9/06

The best group in the world [currently and temporarily known as BURRITO] finished sodering the orcboard and the orcpad componetns of Therobo Orca [our pegbot] and assembled the motors and wheels of it. Also a simple hello world java program was made and displayed to the orcpad.


1/10/06

  • Assembled the rest of the parts onto our board to complete the main structure of the robot
  • Mounted infrared sensor to the front of our bot
  • Created a Java class to encapsulate the motor functions of the robot
  • Successfully had the robot move forward before stopping at an object and notifying the botClient

1/11/06

  • For some reason our orcBoard died sometime between last we used it and today. We had to replace the orcBoard but everything seemed to be working fine when the new one was put in.
  • We glued sandpaper to the tires to reduce sliding, but we didnt get a chance to verify its success
  • Used the sample code from the MASLAB ImageTutorial to process the LogiCam's image to find red blob
  • Simply set a threshold for number of red pixels before detecting that a ball is seen to pass first checkpoint
  • This will definately be improved upon with a much more efficient algorithm
  • Heavy discussion about design and parallizing workload

1/12/06

  • Tried to adapt the code to find the red ball to find the green and black on the screen so we could read the barcode.
  • This attempt was unsuccessful because although we could detect green and black, the algorithm did not incorporate distinct blobs of color
  • We spent the rest of the time brainstorming on feature recognition
  • We didnt get a chance to complete the check point for today.

1/13/06

  • Had a meeting before lab and discuss overal design issues such as: Vision, Mechanics, and Navigation
  • Conceptually came up with a sound algorithm to search pixels
  • Basically checks each pixels color, and when it finds a pixel of a recognizable color, it performs a localized search to gather all of the pixels near it
  • This algorithm is then able to identify different blobs of color
  • This algorithm is incorporated in a beautifully designed class hierarchy which was continued to be developed today
  • As a consequence we were able to easily use the algorithm to identify and decode barcodes
  • We didn't, however, get the chance to work on the check point, which is going to be completed over the weekend

1/14/06

  • Finished setting up the SVN server on my home computer. Works great.
  • Still have to set up accounts for my other teammates
  • One of the team left the power supply cable in the lab, so will have to retrieve on Sunday @ noon. Prohibts direct testing.

1/15/06

  • Went back and got the power supply from the lab
  • Worked a bit the modularization of the components of the robot
  • Got the robot to rotate until it sees a ball by running vision on a separate thread
  • When the robot finally sees it, it moves toward the ball straighting itself using feedback from the camera

1/16/06 Holiday

1/17/06

  • Demonstrated to the rest of the team how to use the new svn server, hosted from my home server
  • Discovered that an LED on each encoder was broken by using a digital camera to detect infared (or lack thereof)
  • We also could test the encoders by using the orcPad. One of the menu items lets you look at the analog input from the encoders in real time
  • Replaced both LEDs and a phototransistor that was broken
  • Reattached the printed patterns to the wheels
  • Finished and uploaded the slides for the presentation tomorrow

1/18/06

  • Spent most of our lab time getting out encoders to work and building the robot chasis
  • It seems there was a manufacturing flaw in both of our encoder circuit boards, where one of the leads was touching a near by trace
  • Once identified, the problem was resolved by scraping the metal between these metal pieces away with a knife
  • The shop closed before we could finalize the chasis, so right now its held together with ducktape
  • We did end up able to catch a red ball


1/19/06

  • Finished the rest of the chasis for the robot. Still need to screw down eden, orcboard, orcpad and camera
  • Still need to mount two IR sensors
  • Incorporated encoders into code and was able to stabilize the robots movement by using feedback from the encoders
  • tweaking of the constants for the PID controllers still needs to be done
  • coded up software for an "ant mode" which wall follows, and every so often rotates to look for balls
  • still need to install ball catching mechanism
  • metal door for ball bay is created, needs to put in


1/20/06

  • Created a behavior systems to make it easier to implement new behaviors
  • Completed mounting all pieces on to the chasis of the robot
  • Created behaviors for retrieving balls, and wandering around the map


1/23/06

  • Began testing of behaviors
  • IR Sensors have yet to be mounted


1/24/06

  • Infrared sensors have been mounted
  • Camera mounted
  • Switched battery with a smaller battery
  • Ball catching mechanism complete
  • Successfully programmed the robot to lift the gate, move forward over a ball, and close the gate
  • The gate is controlled by a Servo, and I use the set method to set the angle of the gate to the desired position
  • Realized that the camera is not mounted at an angle, so it doesn't see anything directly in front of it

1/24/06

  • Successfully programmed the gyro sensor to make accurate turns
  • Used encoders to make the bot drive as straight as possible given these encoders
  • Successfully started the program from the orcPad

1/30/06

  • Conducted a series of testing in lab
  • The robot's wall following still needs some work
  • Currently the robot follows the wall, but the p-controller I wrote, so that it stays near the wall is a little buggy
  • I'm also having problems with my behavior system.
  • The problem with the behavior system, I suspect, has to do with threading.
  • I'm going to read up a little more on threading in Java to clear it up

1/31/06 - 1/1/06

  • Went into lab early to work out software threading issues
  • Was able to improve wall following by using a simple go-to-wall-and-turn-a-little-bit, as opposed to a pd-Controller to keep the distance to the wall the same while following
  • There are still multithreading problems that I have been trying to correct with a more directed approach with semaphores and mutexs to synchronize the threads
  • Added a interrupt in the wall following to every so often spin around for scanning of balls
  • The robot successfully found a ball, captured it, found a goal, and put it through the hole
Retrieved from "http://maslab.mit.edu/2006/wiki/Team_Twelve/Journal"