Progress for week 12 (2015)

Fra Robin

(Forskjeller mellom versjoner)
Gå til: navigasjon, søk
 
(13 mellomrevisjoner ikke vist.)
Linje 17: Linje 17:
=== Accounting ===
=== Accounting ===
 +
* 8 pages
 +
* Looked at results from objective type testing
 +
** Found critical bug in the current crowding distance calculation
 +
* Did experiments on a softer non-dominating sorting for non-deterministic objectives (runtime)
 +
* Got cameras for a possible stereo implementation.
== Rune ==
== Rune ==
Linje 25: Linje 30:
=== Accounting ===
=== Accounting ===
 +
* Refreshing statistics and gotten familiar with R
 +
* Done more simulations with boring results
 +
* Filled in missing background parts
== Snorre ==
== Snorre ==
=== Budget ===
=== Budget ===
 +
* Write
 +
* Make a new transferability function
 +
* Simulate with this function
=== Accounting ===
=== Accounting ===
 +
* Wrote about walking methods
 +
* Planned new experiments in simulator (on uneven terrain)
 +
 +
== Else-Line ==
 +
=== Budget ===
 +
* Continue testing on real robot
 +
** Baldwinian solution
 +
** With learning?
 +
* Change implementation of obstacle scenario
 +
* 36 pages
 +
 +
=== Accounting ===
 +
* Started testing on real robot
 +
* Running experiments with changing environments
 +
* Testing only local search
 +
* 35 pages
 +
 +
 +
 +
== Eivind ==
 +
=== Budget ===
 +
* Test with the new AI
 +
* Print remaining parts
 +
* Create parts for holding stack of "stands"
 +
 +
=== Accounting ===
 +
* Spent most of the week researching and writing.
 +
* Lift/gripper tests.
 +
* Modelled parts in Solidworks
 +
 +
== Bendik ==
 +
=== Accounting ===
 +
Implemented simulation tools for driving/lift
 +
Improved simulation visualizer
 +
Tested planner on simulator
 +
Researched strategies for avoiding the enemy

Nåværende revisjon fra 24. mar 2015 kl. 10:01

Innhold

André

Budget

  • Test with different transition matrices, and covariancematrices for the kalmanfilter
  • Find error for tower, and setup a noise matrice
  • Write 3 pages....
  • Make a plot over distance - voltage for the distance sensor, so it's easy to get the distance from it..

Accounting

Stian

Budget

  • Write 10 pages
  • Look at results from objective type testing.
  • Do more runs of objective type testing or move to the next experiment depending on results
  • Meeting with Johannes.

Accounting

  • 8 pages
  • Looked at results from objective type testing
    • Found critical bug in the current crowding distance calculation
  • Did experiments on a softer non-dominating sorting for non-deterministic objectives (runtime)
  • Got cameras for a possible stereo implementation.

Rune

Budget

  • Keep on testing
  • Improve/adjust parameters for learning
  • Write (35 ~ ~~pages total)

Accounting

  • Refreshing statistics and gotten familiar with R
  • Done more simulations with boring results
  • Filled in missing background parts

Snorre

Budget

  • Write
  • Make a new transferability function
  • Simulate with this function

Accounting

  • Wrote about walking methods
  • Planned new experiments in simulator (on uneven terrain)

Else-Line

Budget

  • Continue testing on real robot
    • Baldwinian solution
    • With learning?
  • Change implementation of obstacle scenario
  • 36 pages

Accounting

  • Started testing on real robot
  • Running experiments with changing environments
  • Testing only local search
  • 35 pages


Eivind

Budget

  • Test with the new AI
  • Print remaining parts
  • Create parts for holding stack of "stands"

Accounting

  • Spent most of the week researching and writing.
  • Lift/gripper tests.
  • Modelled parts in Solidworks

Bendik

Accounting

Implemented simulation tools for driving/lift Improved simulation visualizer Tested planner on simulator Researched strategies for avoiding the enemy

Personlige verktøy