Spatial Algebra Development Framework for Rigid Body Dynamics
| Title | Spatial Algebra Development Framework for Rigid Body Dynamics |
|---|---|
| Summary | Design and implement a C or C++ library that seamlessly integrates spatial vector algebra with the Stanford_WBC robot control framework. |
| Keywords | |
| TimeFrame | |
| References | |
| Prerequisites | |
| Author | |
| Supervisor | Roland Philippsen |
| Level | Master |
| Status | Draft |
Project Description
Advanced control approaches for robotic manipulators, such as Whole-Body Operational Space Control, rely on models of the physical implementation in order to achieve sufficient agility [1]. For instance, accelerating a certain part of the mechanism can create undesired dynamic couplings in other parts, and a model which accurately predicts these terms can be used to compensate for them. As these models need to be integrated with controllers, which in turn have to be compatible with hard real-time constraints, their computational efficiency is very important. To some extent, thanks to control feedback, accuracy can be sacrificed for speed, but it is important to strike the right balance.
In current robotic implementations, rigid-body dynamics models rely on traditional geometric descriptions of the involved motions and forces. This has been shown to lead to cumbersome equations and inefficient code, but a very promising alternative exists and has been worked out in detail [2]: spatial vector algebra.
The objective of this MS project at the Halmstad Intelligent Systems Lab is to understand spatial vector algebra, and then design and implement a C or C++ library that seamlessly integrates it with the whole-body control codebase. Evaluations will be done in simulation. Time and resources permitting, it may be possible to also perform experiments on a real robot platform at the Human Centered Robotics Lab, one of our partners.
References
- [1] R. Philippsen, L. Sentis, O. Khatib. An Open Source Extensible Software Package to Create Whole-Body Compliant Skills in Personal Mobile Manipulators. Proc IROS, 2011.
- [2] R. Featherstone. Rigid Body Dynamics Algorithms. Springer, New York, 2008. ISBN 0387743146.
