Progress for week 10 (2015)

Fra Robin

Gå til: navigasjon, søk

Innhold

André

Budget

  • Make presentation for Wednesday
  • Make Liveplot from beacon
  • Test Beacon with slipring. See if we can remove some of the errors
  • Design the top gripper of the robot
  • Write at least 3 pages

Accounting

  • Made the presentation
  • Liveplot, easier to see whats wrong
  • put the slipring on the tower
  • Made some tests with the slipring. A lot of code needed to be rewritten, so still something that's wrong
  • written 2 pages :(

Stian

Budget

  • Complete GA background chapter.
  • Write about algorithms used.
  • Do training set size test with a simpler algorithm and compare.
  • Continue robust objective testing.

Accounting

  • Made a presentation.
  • Ran the training set size tests with a simpler algorithm.
  • Created 8 small Google Compute Engine cloud computing machines to help with simulations. (using their free trial)
  • Created a helper batchfile for spreading each worker PC across multiple tests to ensure some fairness on runtime dependent objective types.
  • Learned not to leave RDP windows open as they time out and cancel the active simulation...
  • Made further notes in the background and results sections, but slow going.

Snorre

Budget

  • Test the results obtained when evolved using a transferability function
  • Make another transferabiliy function
  • Plan more experiments
  • Write about project and implementation

Accounting

  • Held a new master project presentation
  • Wrote pages about project
  • Prepared Motion Capture lab for experiments
  • Measured average height of robot in simulator (for transferability function)

Else-Line

Budget

  • Run more simulation tests
    • Changing environment during a run
  • Start testing on real robot
  • 30 pages

Accounting

  • Started implementation of changing environment, problem with reevaluation of parents.
  • Ran basic tests with crossover and simulated annealing
  • Fixed hardware problem with bbp-viewer, but robot still crashes in itself and freezes
  • 25-ish pages

Eivind

Budget

  • Model and print remaining parts
  • Test robot with "exact" position inputs
  • Test battery power systems

Accounting

  • Modelled some of the parts
  • Started testing beacon system with slipring
  • Battery systems work perfectly!
  • Researched and wrote about Hall Sensors and socket programming
  • Fixed a few bugs in the control program

Rune

Accounting

  • Implemented a reconstruction of prm based on observations
  • Fixed some bugs
  • Ran som tests


Bendik

Budget

Implement GOAP Rewrite parts of code in Java Researched learning

Accounting

  • Implemented GOAP framework
  • Written Eurobot implementation for GOAP framework
  • Researched learning strategies
  • Written 5 pages
Personlige verktøy