shamil
|
098aed50f5
|
modifying readme 3
|
2021-04-10 12:46:14 +03:00 |
|
shamil
|
ded626374e
|
system matrices is generated in symbolic form as a function of standard rigid body parameters. Folder are slightly restructured
|
2020-03-06 15:55:03 +03:00 |
Shamil Mamedov
|
fe3cc01102
|
added file that check if optimized trajectory collides with table on which the robot is fixed. Moreover kinematics of UR is derived is analytical form, so it can be used in trajectory optimization in order to stay away from table, however is considerably slows down the optimization. Even on sever with 40 cpus patter search computes it for hours. Moreover, added files for identification of the pendubot.
|
2020-02-05 07:55:09 +03:00 |
|
Shamil Mamedov
|
8307ef7d99
|
identification of drive parameters fails even in simluation with some noise. We thoughts that weighted least sqaures can help, but for now it doesn't. It seems that we don't do it right.
|
2019-12-26 12:09:57 +03:00 |
|
Shamil Mamedov
|
6065d44dbd
|
New trajectory is found. The difference from previous ones is the fact that now the condition number of the whole whole regressor (with frinction and motor dynamics) is taken into accound
|
2019-12-23 08:55:18 +03:00 |
|
Shamil Mamedov
|
a5240b3574
|
Objectve function as well as constraint fucntion for trajectory optimization slightly changed. Now the structure of the robot is not an input argument. Moreover, objective function can work with QR decomposition based base parameters.
|
2019-12-19 13:31:18 +03:00 |
|
Shamil Mamedov
|
7e7f3512c4
|
experiment design is changed as to have two options for optimization algorithms - pattern search and genetic algorithm. Generation of full regressor and well as load regressor are put into seperate file
|
2019-12-19 09:48:18 +03:00 |
|
Shamil Mamedov
|
b382f14bc4
|
the last version of the project
|
2019-12-18 14:25:45 +03:00 |