Alexis DUBURCQ

Results 92 comments of Alexis DUBURCQ

You should also limit yourself to invertible transmissions,. Which are n-to-n motor to joint mappings. And rename the abstract case accordingly. Non-invertible transmission must be modelled through constraints instead.

It is necessary to access the motors directly at robot level for several reasons, including the calling the *All methods of motors (such as resetAll ), which cannot be done...

Cassie could be the first robot using an advanced transmission in Jiminy. Currently it is done by solving a optimization problem, which is likely to be more costly. Could be...

`rotorInertia` should be a config option of the motors, and the transmission inertia should be accessed by the robot through the transmissions. This is necessary for advanced transmission. Ideally, once...

Just added `flake8` for Python PEP8 conformance.

Added `pylint` and `mypy` for 'gym_jiminy/common' and 'gym_jiminy/toolbox'.

Part of the code is now partially tested though `gym_jiminy` unit tests, but it is clearly not enough, considering that nearly 40% of the whole code base is pure Python.

Also, add a wiki page about the physics, especially about the contact model.

Now that it is clearly impossible to have a system named `""` if there is multiple systems, you could probably simplify the logics in `Simulator`. By the way, you should...

The viewer segfault causes some tests to fail for reference: https://github.com/panda3d/panda3d/issues/1372