Sunday, February 27, 2005

long time no update

well, here's what's happened since last we met the crazy coder...

got all the programs on the car's computer.
debugged all the software required to work with gps
oh yea.. got the car driving around with the gps software!

next week... get obstacle avoidance with vision...
make gps navigation TIGHT

Friday, February 11, 2005

This week's interesting...

Monday
8 hours
  • Figured out how to read avi files from open cv
  • Meeting with Terri and Gus about aquiring computers
  • Drove around with Drew, capturing video for testing
  • Finished implementing reading avis with OpenCV
Tuesday
8 hours
  • Tested avi reading through OpenCV
  • Got avi files from Drew and ran them through tests
  • Met with Chris, Greg, Drew, and Rhubarb to go over code
  • Looked through ethernut code again
Wednesday
9 hours
  • Went over Overall Pathfinding code with Drew and Rhubarb
  • Aquired and began testing ethernut code
  • Walked a really really long time to get a straight through db9 cable to test with
  • Tested and fixed code
    • found incombatibilities with current libraries
Thursday
9 hours
  • Worked on ethernuts all day
  • Helped debug code for testing

Saturday, February 05, 2005

Overdue Update... lots of stuff this week

Monday
8.5 hours
  • Vision research (gabor kernels)
  • Worked on a Fake Sender (stub program for testing modules)
  • Began looking at Ethernut implementation
  • Vision implementation (c++)
  • note -
    • Sun's CC compiler is hard to deal with...
Tuesday
8.5 hours
  • Vision research
  • Simulator work
  • Ethernut implementation begun
  • Vision math completed
  • Got the video camera
  • implemented pixel voting for vision
Wednesday
11 hours
  • Vision implementation done
  • Bugs worked out in Hardware Interface
  • Ethernut implementation (compiler issues) discovered.
    • sent email to harold kipp
  • Ethernut implementation problem discovered
    • no such thing as a server socket
      • work around designed and planned to implement
  • Software meeting
    • reviews and integration schedule planned
Thursday
6 hours
  • Reimplementation due to protocol revision (from no server sockets on ethernuts)
    • Hardware Interface
    • Ethernuts
  • Ethernut compiles! ***** thanks Mr. Kipp for the advice
Friday
4 hours
  • Go over ethernut and Hardware Interface code
  • code review for Hardware Interface
Sunday
14
hours
  • Continued testing and debugging of Hardware Interface
  • Testing and integration with the Senior Design team (out at the barn)
  • Hang around barn doing random work

Sunday, January 30, 2005

Long week full of prototypes and deadlines

Last week analysis:
Good week in terms of writing test programs and testing various modules. One really bad day where we had problems getting the simulator up and running, but that's good now. Was also able to test the Hardware Interface with the NI boards that the senior design team had, discovering (and fixing) a few minor bugs in the process.

Last week's Events (tuesday - sunday ~45 hours):
  • Implemented test programs
    • Fake Sender
    • Test Boss
    • Stubs for rest of programs
  • Image Analysis
    • Researching Gabor filters
    • Thinking about how to make the process faster and more reliable
  • Simulator
    • Helped install new version of Gazebo
      • Took forever...started with red hat 8.0, after 12 hours of failure, switched over to Mandrake (~2 hours), where we were able to finish the install in about 1 hour.
  • Hardware Interface
    • Implementation finished, bugs worked out
    • Tried integrating with Senior Design Team's NI systems on Saturday.
      • Few bugs found (on listening sockets) and fixed.
  • Software Meeting
    • We know how people are implementing things, computational usage.
  • General Wiki documentation

Monday, January 24, 2005

Towards vanishing point detection

Tommorow's Goals:
  • Continue researching vanishing points, Gabor Filters
Today's Events:
  • 7:00 am - 12:00 pm
    • C-camera implementation
      • most of the vision code being converted to c++
  • 2:00 pm - 3:30 pm
    • Boss/World model meeting
      • How do we integrate the boss with the world model / database?
      • Extend world model for a limited time series
  • 3:30 pm - 3:45 pm
    • Quick net code for the simulator
  • 6:00 pm - 7:00 pm
    • Talk to chris, matt about GCART work
      • Need a new C coder for the ethernut
        • Action Item: Look over the ethernut code
      • Talked about the timeline
  • 8:00 pm - 9:00 pm
    • Fix the Unix Makefiles
  • 9:00 pm - 10:00 pm
    • Gabor filter research
      • application to edge angle detection
  • 11:00 pm - 1:30 am
    • Continued Gabor filter research

Friday, January 21, 2005

No one showed up for integration!

Tommorow's Goals:
  • Finish Camera implemenation in C++
  • Continued Utilities (formally Framwork) testing
Today's Events
  • 7:00 am - 9:00 am
    • Continued debugging of the Utility (framework) code
  • 9:00 am - 10:00 am
    • Metric implemented for vision system
  • 10:00 am - 11:00 am
    • review Hardware Interface code
      • review logic... over and over and over..
      • make it ready for integration
  • 11:00 am - 12:00 pm
    • changed docs
      • respecified what arrays will look like
    • stress test on the framework
      • memory usage test
        • located memory leak in SPacket
          • in parse() -- fixed
Rant:
  • Couldn't integrate the Hardware Interface.... meeting didn't happen.
  • Must get it integrated on monday.....

Thursday, January 20, 2005

Today's testing day

Tommorow's Goals:
  • Test Hardware Interface with the actual hardware
  • Stress test Framework ("Utility") communication
Today's Events:
  • 9:00 am - 10:00 am
    • Finish added Channel/Value information in the wiki
  • 10:00 am - 10:30 am
    • Updated this blog (for today and yesterday)
  • 10:30 am - 7:00 pm
    • Created stubs for
      • Vehicle Handler
      • Boss System
      • Overall Pathfinding
      • Stepper Calculator
      • Began full test for framework classes
  • 9:00 pm - 3:30 am
    • Testing/Bug finding/Bug fixing
      • Continued testing on framework classes via stubs
        • Fixed problems with communication between clients
        • Fixed problems with clients handling dropped clients incorrectly
        • Fixed memory leaks on dropped clients
        • Fixed memory leaks on reader requests
        • Fixed null pointer errors on writer thread deletions