Skip to content

Compute walking trajectory from a desired 2D velocity vector

Notifications You must be signed in to change notification settings

laas/nmpc_walkgen

Repository files navigation

Gait Pattern Generator for Humanoids

Author

Manuel Kudruss

Address

Im Neuenheimer Feld 368 69120 Heidelberg Germany

Contact

manuel.kudruss@iwr.uni-heidelberg.de

Authors

Manuel Kudruss; Maximilien Naveau ...

organization

Interdisciplinary Center for Scientific Computing

date

$Date: 2014-08-17 15:12:43 +0000 (Thu, 17 Aug 2014)$

status

This is a "work in progress"

Dedication

Roboticists that want their humanoid to walk

abstract

General

This is a collection of different approaches to gait pattern generation for humanoid robots. A paper is available here ... giving an overview of some of the algorithms in the module. The algorithms itself are documented with docstrings. Usage examples are given in the test cases.

Third Party Software

The algorithms in this module use different third party libraries to solve the resulting optimization problems for the different pattern generator types.

qpOASES

Download the QP solver qpOASES from here (here). Academic license is free.

Follow the instructions to install the pythong interface and add qpOASES to your PYTHONPATH in your bash.rc.

gurobi

For the QCQP version of the pattern generator we use gurobi as solver (here). Academic license is free.

Because their python interface is crappy. Please Download a real one from ...

Follow the instructions on how to install the python interface for gurobi.

SNOPT7

For the nonlinear version of the pattern generator the SQP implementation SNOPT7 is used. Get it (here). A proper license is needed to use this software.

The python interface is available through ....

Follow the instructions on how to install the python interface for SNOPT7.

About

Compute walking trajectory from a desired 2D velocity vector

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages