SURF 2011: Receding Horizon Temporal Logic Planning Toolbox

From MurrayWiki
Jump to: navigation, search

2011 SURF Project Description

  • Mentor: Richard Murray
  • Co-Mentor: Ufuk Topcu

TuLiP (Receding Horizon Temporal Logic Planning Toolbox) is a collection of Python-based code for automatic synthesis of correct-by-construction embedded control software as discussed in [1, 2]. It combines ideas from control theory (reachability analysis, receding horizon control) and computer science (finite transition systems, linear temporal logic) to automatically synthesize a planner-controller hierarchy that guarantees to satisfy given linear temporal logic based specifications. TuLiP can be used in a wide range of applications such as autonomous driving, vehicle management systems in avionics and smart camera networks. Additional information and related references can be found at: http://www.cds.caltech.edu/tulip/. However it is still in development stage and there is room for quite a bit of advancements. The goals of this project are:

  • To develop a visualization environment for simulating the results for application areas such as robot navigation and autonomous driving. It would be useful to have an interactive tool where the user can dynamically change the environment to see the response of the designed controller. One possibility for visualization is to integrate a tool like Player/Stage (available at http://playerstage.sourceforge.net/).
  • To develop a user interface to guide the designer in defining short horizon specifications. In particular, one should define a partial order among the sets of states based on their “closeness” to the goal state. Visualizing the finite state machine that models the system could help in finding such a partial order.
  • To improve the discretization procedure used for abstraction. We would like to explore different possible representations of the regions in the partition (e.g. using zonotopes instead of polytopes).
  • To explore/identify other possible extensions to TuLiP, its applications and underlying theory.

Required Skills: This project requires programming experience. The programming language underlying TuLiP is Python. The student is expected to know Python or to have enough programming experience to learn it in a short time.

References

  1. T. Wongpiromsarn, U. Topcu, N. Ozay, H. Xu, and R. M. Murray, TuLiP: a software toolbox for receding horizon temporal logic planning, International Conference on Hybrid Systems: Computation and Control, 2011 (to appear).
  2. T. Wongpiromsarn, U. Topcu, and R. M. Murray, Receding Horizon Temporal Logic Planning, IEEE Transactions on Automatic Control, 2010 (submitted).