Progress for week 6 (2019)
From Robin
(Difference between revisions)
(→Malin Aandahl) |
|||
Line 11: | Line 11: | ||
* Lifetime learning in Revolve | * Lifetime learning in Revolve | ||
** Where to be implemented | ** Where to be implemented | ||
+ | === Acconting=== | ||
+ | * SurfSara/virtual machines are working and doing experiments | ||
+ | * Testing new fitness function, looks OK. Currently checking variables to the size-parameter. | ||
== Wonho Lee == | == Wonho Lee == |
Revision as of 12:25, 8 February 2019
Contents |
Anders Rønningstad
Budget
- Start writing introduction chapter
- Find a good way to finish the updated drone.
Malin Aandahl
Budget
- Do more experiment on testing different discount on sizes and compare
- Decide how to deal with position
- Lifetime learning in Revolve
- Where to be implemented
Acconting
- SurfSara/virtual machines are working and doing experiments
- Testing new fitness function, looks OK. Currently checking variables to the size-parameter.
Wonho Lee
Budget
- Writing
- Continue writing on NEAT
- Draft table of contents for essay
- Reading
- more input on NN-based CPGs - capable of dynamic gait/locomotion transition?
Accounting
- Writing
- Started on CPG - bullet points
- Draft on table of contents
# Introduction # Adaptive Behavior ## CPG # NeuroEvolution ## ANN ### CTRNN ## EA ### NEAT ### HyperNEAT
- Reading
- started on Neural bases of goal-directed locomotion in vertebrates—An overview - CPG analysis from neurology perspective too much field specific terms/concepts, may get some insights on evolving primitive CPG to complex ( lamprey to mammalian )
- Central pattern generator and feedforward neural network-based self-adaptive gait control for a crab-like robot locomoting on complex terrain under two reflex mechanisms - system of diff.eq based cpgs with coupled NN to map the signal to joint angles
David Kolden
Budget
- Write implementation chapter
Accounting
- Wrote everything related to pose estimation