Difference between revisions of "EECI 2020: Computer Session: TuLiP"
(→Lecture Materials) 
(→Lecture Materials) 

Line 9:  Line 9:  
* amination script: [home/tichakorn/Shared/eeci/animate.py animate.py]  * amination script: [home/tichakorn/Shared/eeci/animate.py animate.py]  
* [http://www.cds.caltech.edu/~murray/courses/eecisp2020/tulip_examples.zip Example TuLiP files] (zip file):  * [http://www.cds.caltech.edu/~murray/courses/eecisp2020/tulip_examples.zip Example TuLiP files] (zip file):  
−  ** 6 cell robot, discrete state space: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/robot_simple_discrete.  +  ** 6 cell robot, discrete state space: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/robot_simple_discrete.ipynb robot_simple_discrete.ipynb] 
−  ** 6 cell robot, with dynamics: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/robot_simple_continuous.  +  ** 6 cell robot, with dynamics: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/robot_simple_continuous.ipynb robot_simple_continuous.ipynb] 
−  ** 3x3 exercise: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/exercise_3x3.  +  ** 3x3 exercise: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/exercise_3x3.ipynb exercise_3x3.ipynb] 
−  ** Left turn exercise: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/exercise_leftturn.py exercise_leftturn.  +  ** Left turn exercise: [http://www.cds.caltech.edu/~murray/courses/eecisp2020/exercise_leftturn.py exercise_leftturn.ipynb] 
== Further Reading ==  == Further Reading == 
Revision as of 09:42, 12 March 2020
Prev: Reactive Synthesis  Course home  Next: Minimum Violation Planning 
This lecture provides an overview of TuLiP, a Pythonbased software toolbox for the synthesis of embedded control software that is provably correct with respect to a GR[1] specifications. TuLiP combines routines for (1) finite state abstraction of control systems, (2) digital design synthesis from GR[1] specifications, and (3) receding horizon planning. The underlying digital design synthesis routine treats the environment as adversary; hence, the resulting controller is guaranteed to be correct for any admissible environment profile. TuLiP applies the receding horizon framework, allowing the synthesis problem to be broken into a set of smaller problems, and consequently alleviating the computational complexity of the synthesis procedure, while preserving the correctness guarantee.
A brief overview of TuLiP will be followed by handson exercises using the toolbox.
Lecture Materials
 Lecture slides: TuLiP
 amination script: [home/tichakorn/Shared/eeci/animate.py animate.py]
 Example TuLiP files (zip file):
 6 cell robot, discrete state space: robot_simple_discrete.ipynb
 6 cell robot, with dynamics: robot_simple_continuous.ipynb
 3x3 exercise: exercise_3x3.ipynb
 Left turn exercise: exercise_leftturn.ipynb
Further Reading

TuLiP: A Software Toolbox for Receding Horizon Temporal Logic Planning, T. Wongpiromsarn, U. Topcu, N. Ozay, H. Xu and R. M. Murray, Hybrid Systems: Computation and Control, 2011.
Additional Information

JTLV Project Home Site JTLV provides the framework for the underlying digital design synthesis routine used in TuLiP.