NASA Tensegrity Robotics Toolkit

Read here about the NASA Tensegrity Robotics Toolkit (NTRT), open source software enabling the modeling, simulation, and control of tensegrity robots.

Overview

The NASA Tensegrity Robotics Toolkit (NTRT) is a collection of C++ and MATLAB software modules for the modeling, simulation, and control of Tensegrity Robots. Tensegrity Robots are a biologically inspired approach to building robots based on the tension networks of tensegrity structures, which have no rigid connections between elements. The NTRT was created to enable: the rapid co-exploration of structures and controls in a physics based simulation environment; the development of tensegrity robotics algorithms such as structural analysis, kinematics, and motion planning; and the validation of the algorithms and controls on hardware prototypes of the tensegrity robots.

NTRT is an evolving collection of loosely connected open-source modules designed by the NASA Ames Intelligent Robotics Group. Modules of the NTRT will be released one at a time as they reach maturity. The first module to be released is the NTRT Simulator Core, which is the core library for doing physics based simulations of Tensegrity Robots, and will be described in further detail below.

Stable releases of the software modules of NTRT will be provided below, and the source code for all NTRT modules is available on Github.

Videos

Tutorial introduction to NTRT

Sample creation from NTRT: Snake Walk

Bullet Physics Engine

Bullet Physics Engine is a Collision Detection and Rigid Body Dynamics Library. The Library is Open Source and free for commercial use, under the zlib license. Bullet is used in the NTRT.

NTRTsim

The NTRT Simulator is a tensegrity-specific simulator built to run ontop of the Bullet Physics Engine, version 2.82. The NTRTsim includes a set of builder tools for specifying rods and strings as a set of points in Cartesian coordinates. Structures built out of these rods and strings can be specified as a tree of substructures, and can be rotated and moved, which greatly simplifies the task of creating new tensegrity structures. The NTRTsim also includes libraries for controllers such as Central Pattern Generators and a machine learning framework, which allows users to specify their own learning algorithms. For strings, instead of the default Bullet softbodies, which are not physically accurate, we used a two point linear string model using Hooke's law forces with a linear damping term. Finally, terrains can be created, and the performance of the controller can be tested as the tensegrity robot moves through the simulated world.

Links and References

NASA Tensegrity Robotics Toolkit (NTRT) official site: http://ti.arc.nasa.gov/tech/asr/intelligent-robotics/tensegrity/ntrt/

Bullet Engine: http://bulletphysics.org/wordpress/?p=413