Introduction

The sot-dynamic package is a bridge between the stack of tasks framework and the dynamicsJRLJapan library. It provides an inverse dynamic model of the robot through dynamic-graph entities. More precisely it wraps the newton euler algorithm implemented by the dynamicsJRLJapan library to make it accessible for the stack of tasks controllers (in the Stack of Tasks Framework as defined in Mansard2007.)

This package depends on the following packages:

• dynamicsJRLJapan
• sot-core
• dynamic-graph
• dynamic-graph-python

See the JRL umi3218's page on github for instructions on how to download and install these packages at https://github.com/jrl-umi3218.

Python bindings

As most packages based on the dynamic-graph framework (see https://github.com/jrl-umi3218/dynamic-graph), the functionnality is exposed through entities. Hence python sub-modules of dynamic_graph are generated. See sphinx documentation for more details.

The following entities are created by this package:
(all entites are placed in the namespace sot::)

• sot::ZmprefFromCom
• sot::ForceCompensation
• sot::IntegratorForceExact
• sot::MassApparent
• sot::IntegratorForceRk4
• sot::IntegratorForce
• sot::AngleEstimator
• sot::WaistAttitudeFromSensor
• sot::Dynamic - provides the inverse dynamics computations for of a humanoid robot

See each entity's documentation page for more information (when available).

References

"Task sequencing for sensor-based control", N. Mansard, F. Chaumette, IEEE Trans. on Robotics, 23(1):60-72, February 2007