Progress for week 11 (2015)
From Robin
(Difference between revisions)
(→Eivind) |
|||
(8 intermediate revisions not shown) | |||
Line 1: | Line 1: | ||
- | == | + | == André == |
=== Budget === | === Budget === | ||
- | * | + | * Finish writing the code of the tower, and figure out whats wrong at the moment |
- | * | + | * Look at the protocol for the positioning system (zeroMQ) |
+ | * Make the griptower | ||
+ | * Write 3 pages | ||
=== Accounting === | === Accounting === | ||
- | * | + | * Fixed the tower code. |
- | * | + | * Installed ZeroMQ |
+ | * Implemented the EKF in c++ | ||
+ | * Tested the IR distance sensor | ||
+ | ** The data sheet is totally wrong | ||
+ | * "Learned" tikz | ||
== Stian == | == Stian == | ||
Line 14: | Line 20: | ||
=== Accounting === | === Accounting === | ||
+ | * Wrote 10 pages. | ||
+ | * Created a parallel version of the ELAS algorithm for a possible later real-time implementation. | ||
+ | * More objective type testing. | ||
== Else-Line == | == Else-Line == | ||
Line 22: | Line 31: | ||
=== Accounting === | === Accounting === | ||
+ | * Started testing on changing environments | ||
+ | * Ran tests on static environment with obstacles, not exactly a great success | ||
+ | * Started testing on real robot, larger reality gap on lamarckian | ||
+ | * 30 pages | ||
== Rune == | == Rune == | ||
Line 31: | Line 44: | ||
=== Accounting === | === Accounting === | ||
- | + | * Updated the scenario generator and created two scenarios | |
- | + | * Written 5~ pages (25 pages total) | |
+ | * Done a bit of testing | ||
== Snorre == | == Snorre == | ||
Line 41: | Line 55: | ||
=== Accounting === | === Accounting === | ||
+ | * Wrote about project, implementation and results | ||
+ | * Did experiments in the MoCap lab | ||
+ | * Started planning new transf. func. | ||
== Eivind == | == Eivind == | ||
Line 49: | Line 66: | ||
=== Accounting === | === Accounting === | ||
+ | * Modelled new lift-mechanism for Dynamixel servo | ||
+ | * Modelled structural parts | ||
+ | * Restructured thesis | ||
+ | * Wrote 7 pages | ||
== Bendik == | == Bendik == | ||
=== Budget === | === Budget === | ||
- | Research and implement Q-learning | + | * Research and implement Q-learning |
- | Modify GOAP framework with feature vectors for Q-learning | + | * Modify GOAP framework with feature vectors for Q-learning |
- | Test "QOAP" | + | * Test "QOAP" |
- | Implement motor controller communication in java | + | * Implement motor controller communication in java |
Current revision as of 10:04, 17 March 2015
Contents |
André
Budget
- Finish writing the code of the tower, and figure out whats wrong at the moment
- Look at the protocol for the positioning system (zeroMQ)
- Make the griptower
- Write 3 pages
Accounting
- Fixed the tower code.
- Installed ZeroMQ
- Implemented the EKF in c++
- Tested the IR distance sensor
- The data sheet is totally wrong
- "Learned" tikz
Stian
Budget
- Write another 10 pages.
- Continue nsga2 objective type testing.
Accounting
- Wrote 10 pages.
- Created a parallel version of the ELAS algorithm for a possible later real-time implementation.
- More objective type testing.
Else-Line
Budget
- Finish implementing changing environment in simulation and start testing
- Start testing on real robot
- 35 pages
Accounting
- Started testing on changing environments
- Ran tests on static environment with obstacles, not exactly a great success
- Started testing on real robot, larger reality gap on lamarckian
- 30 pages
Rune
Budget
- Test
- Create more scenarios
- Adjust/improve learning component (parameters)
- Write (30 pages total)
Accounting
- Updated the scenario generator and created two scenarios
- Written 5~ pages (25 pages total)
- Done a bit of testing
Snorre
Budget
- Write more about project and implementation
- Do experiments and tests in the Motion Capture lab
- Then make the new transferability function
Accounting
- Wrote about project, implementation and results
- Did experiments in the MoCap lab
- Started planning new transf. func.
Eivind
Budget
- Model remainding parts and get them printed
- Test robot with new AI
- Help improve the position system
Accounting
- Modelled new lift-mechanism for Dynamixel servo
- Modelled structural parts
- Restructured thesis
- Wrote 7 pages
Bendik
Budget
- Research and implement Q-learning
- Modify GOAP framework with feature vectors for Q-learning
- Test "QOAP"
- Implement motor controller communication in java