# Progress for week 49 (2014)

(Forskjeller mellom versjoner)
Gå til: navigasjon, søk
 Versjonen fra 3. des 2014 kl. 09:53 (vis kilde) (Ny side: == Student template (copy this for your entry) == === Budget === * Todo 1 * Todo 2 === Accounting === * Done 1 * Done 2)← Eldre redigering Nåværende revisjon fra 13. jan 2015 kl. 09:56 (vis kilde) (→Accounting) (16 mellomrevisjoner ikke vist.) Linje 7: Linje 7: * Done 1 * Done 1 * Done 2 * Done 2 + + == Else-Line == + === Budget === + * Run more tests + ** different number of iterations over longer time + ** with crossover? + ** on different environment? With obstacles? + * Start planning how thesis should look + + === Accounting === + * Ran tests + ** Lamarckian 1,2,5,10,20 iterations + ** Baldwinian 25,32,40,50 iterations + ** Baldwinian 1,2,5,10,20 iterations with crossover + ** Baldwinian and Lamarckian 1,2,5,10,20 iterations with low friction + * Read some of the related work listed in the "Learning to walk" video + * Checked resulting walking patterns is bbp viewer + * Listed a few headlines for thesis + + == Snorre == + === Budget === + * Hopefully fix the problem with non-deterministic results. + * Do more experiments, and plot them. + * Start to write the thesis. + + === Accounting === + * Fixed the problem with non-deterministic results + * Planned the thesis: Wrote some headlines, looked at the UiO Master template + * Planned the experiments in more detail + + == Stian == + === Budget === + * Testing with the new evaluation method. + * Update the other algorithms to use the new code. + * More algorithms. + + === Accounting === + * Plenty of simulation + * Looked at how well parameters were explored by NSGA2. + * Problem with individual uniqueness over time. + * Had a look at theoretical resolution vs range for a given camera setup. + * Tested another new evaluation with minimization of max error. Returns more robust results (less variation within dataset). + * Started experimenting with different prefiltering techniques. + + == Eivind == + === Budget === + + === Accounting === + * 3d-modelled and printed lift mechanism + * Rewrote parts of the position-control program for smoother and more accuracte control. + * Worked on implementation of IR beacon-system

• Todo 1
• Todo 2

• Done 1
• Done 2

## Else-Line

### Budget

• Run more tests
• different number of iterations over longer time
• with crossover?
• on different environment? With obstacles?
• Start planning how thesis should look

### Accounting

• Ran tests
• Lamarckian 1,2,5,10,20 iterations
• Baldwinian 25,32,40,50 iterations
• Baldwinian 1,2,5,10,20 iterations with crossover
• Baldwinian and Lamarckian 1,2,5,10,20 iterations with low friction
• Read some of the related work listed in the "Learning to walk" video
• Checked resulting walking patterns is bbp viewer
• Listed a few headlines for thesis

## Snorre

### Budget

• Hopefully fix the problem with non-deterministic results.
• Do more experiments, and plot them.
• Start to write the thesis.

### Accounting

• Fixed the problem with non-deterministic results
• Planned the thesis: Wrote some headlines, looked at the UiO Master template
• Planned the experiments in more detail

## Stian

### Budget

• Testing with the new evaluation method.
• Update the other algorithms to use the new code.
• More algorithms.

### Accounting

• Plenty of simulation
• Looked at how well parameters were explored by NSGA2.
• Problem with individual uniqueness over time.
• Had a look at theoretical resolution vs range for a given camera setup.
• Tested another new evaluation with minimization of max error. Returns more robust results (less variation within dataset).
• Started experimenting with different prefiltering techniques.

## Eivind

### Accounting

• 3d-modelled and printed lift mechanism
• Rewrote parts of the position-control program for smoother and more accuracte control.
• Worked on implementation of IR beacon-system