Workshop #2: Control Using Logic and Switching

Workshop #6: Networked Autonomous and Semi-autonomous Vehicles

ORGANIZERS:
James C. Spall, Applied Physics Laboratory, Johns Hopkins University (USA)
I.-Jeng Wang, Applied Physics Laboratory, Johns Hopkins University (USA)

TITLE:
Simulation and Monte Carlo Methods

DESCRIPTION:
Computer simulation and related Monte Carlo methods are widely used in engineering and scientific work. Simulation provides a powerful tool for the analysis of real-world systems when the system is not amenable to traditional analytical approaches. In fact, recent advances in hardware, software, and user interfaces have made simulation a "first line" method of attack for a growing number of problems. Areas where simulation-based approaches have emerged as indispensable include decision aiding, prototype development for large-scale control systems, performance analysis of communication networks, control and scheduling of complex manufacturing processes, and computer-based personnel training. Large stochastic simulations arising in these and other areas present great challenges in the design of the simulation models and in the analysis of output data. The objective of this workshop is to introduce concepts and statistical techniques that are critical to constructing and analyzing effective simulations, and to discuss certain applications for simulation and Monte Carlo methods with a focus on problems of interest to the control systems community.

The target audience for the workshop includes practitioners, researchers, and students with an interest in simulation. While certain aspects of the underlying theory will be discussed, the course will not dwell on technical details. Rather, the instructors will attempt to convey the ideas that are of greatest practical significance, with pointers to the appropriate literature for the technical details associated with these ideas. The material will be presented at a level that should be accessible to almost all CDC attendees. In particular, it is expected that participants have had undergraduate courses in probability and statistics, multivariate calculus, and linear algebra, although we will review critical aspects of this material as needed. Some practical experience with statistical analysis and simulation will be helpful. Some MATLAB-based exercises will be discussed for participants to carry out after the workshop.

SCHEDULE:
08:30-08:45 Course overview and introduction to instructors (JCS and IJW)
08:45-09:15 Brief review of relevant probability and statistical methods (JCS)
09:15-09:45 Computer-based (pseudo) random number generation (IJW)
09:45-10:15 Stochastic timed discrete-event systems model and discrete-event simulations (IJW)
10:15-10:30 Break
10:30-11:00 Simulation validation and output analysis (IJW)
11:00-11:30 Variance reduction techniques for simulation output (IJW)
11:30-12:00 Simulation-based optimization (use of simulations for optimizing system performance) with gradient-based methods such as IPA, LR, etc. (IJW)
12:00-01:00 Lunch
01:00-01:30 Continuation of simulation-based optimization with gradient-based methods (IJW)
01:30-02:30 Simulation-based optimization with non-gradient methods (finite-difference, SPSA, etc.); common random numbers in optimization (JCS)
02:30-03:00 Issues in model building (bias-variance tradeoff, model selection, etc.) (JCS)
03:00-03:15 Break
03:15-03:45 Comparisons of systems via simulation; selection from discrete number of options (JCS)
03:45-04:15 Computer resampling methods (Markov chain Monte Carlo, etc.) (JCS)
04:15-04:45 Input selection and optimal experimental design (JCS)
04:45-05:00 Wrap-up and open discussion (JCS and IJW).

ORGANIZERS:
Joao P. Hespahna, University of Southern California (USA)
Daniel Liberzon, University of Illinois, Urbana-Champaign (USA)
A. Stephen Morse, Yale University (USA)

TITLE:
Control Using Logic and Switching

DESCRIPTION:
The overall objective of this tutorial workshop is to overview a variety of methods for synthesizing and analyzing logic-based switching control systems. By a logic-based switching controller is meant a controller whose subsystems include not only familiar dynamical components {integrators, summers, gains, etc.} but logic-driven elements as well. The overall models of systems composed of such logics together with the processes they are intended to control, are concrete examples of what might be called "hybrid dynamical systems." An important category of such systems are those consisting of a continuous-time process to be controlled, a family of fixed-gain or variable-gain candidate controllers, and an "event-driven switching logic" called a supervisor whose job is to determine in real time which controller should be applied to the process. Examples of supervisory control systems include re-configurable systems, fault correction systems, and certain types of parameter-adaptive systems. Major reasons for introducing logic and switching are to deal with communication, actuator and sensor constraints, with model uncertainty, with unforeseen events or to avoid performing difficult tasks e.g., precise equipment calibration which might otherwise be necessary were one to consider only conventional controls. The aim of this workshop is to provide an overview of algorithms with these capabilities, as well as to discuss various techniques for analyzing the types of switched systems which result.

SCHEDULE:
08:00-08:45 Switched Control System: (Liberzon)
08:45-10:15 Stability of Switched Systems: (Liberzon)
10:30-12:00 Framework for Self-Adjusting Control: (Morse)
12:00-01:00 Lunch
01:00-02:15 Achieving Detectability: (Morse)
02:15-03:45 Supervisory Control of Families of Linear Controllers: (Hespanha)
03:45-05:15 Supervisory Control of Families of Nonlinear Controllers: (Hespanha)

ORGANIZERS:
Joćo Sousa, Raja Sengupta, University of California-Berkeley (USA)
Anouck Girard, Karl Hedrick, Shankar Sastry, Stavros Tripakis, Tunē Simsek, Pravin Varaiya, University of California-Berkeley (USA)
Datta Godbole, Honeywell (USA)
Sergio Yovine, VERIMAG-CNRS (France)

TITLE:
Networked Autonomous and Semi-autonomous Vehicles

DESCRIPTION:
The objective of the workshop is to present an in-depth description of technologies and tools for designing and implementing communication and control systems for networked semi-autonomous and autonomous vehicles. The tutorial is organized into four parts: systems, technologies, tools and a challenge problem.
(i) Systems. Description of beginning-to-end design and implementation of sea, air and land applications with special emphasis on architectures and problem decomposition, and on layers and theories. The selected applications are the Automated Highway System (AHS), Mobile Offshore Base (MOB), Air Traffic Management System (ATM), Unmanned Combat Air Vehicles (UCAV) and Autonomous Underwater Vehicles (AUV).
(ii) Technologies. Discussion of enabling control and communication technologies for networked autonomous and semi-autonomous vehicles with examples from the five applications and from the challenge problem. The technologies are hybrid control design and verification, protocol design and verification, networking, and enterprise control. The exposition is organized around problems and solutions, with a final segment on related topics and state-of-the-art.
(iii) Tools. Presentation of computational tools for formal modeling, simulation, and implementation. The exposition addresses the main issues in formal modeling, simulation, and implementation with examples from the AHS and MOB systems, and gives an overview of the state-of-the-art in this field.
(iv) Challenge problem. The challenge problem calls upon the attendees to exercise the techniques and tools that have been presented, and illustrates the advantages of an integrated approach for design and implementation.

SCHEDULE:
08.00-08.15 Overview of the workshop (Joćo Sousa)

Systems
08.15-08.45 Automated Highway System (Karl Hedrick)
08.45-09.15 UCAV and ATM (Shankar Sastry)
09.15-09.45 MOB and AUVs (Joćo Sousa)

Case study I
09:45-10:15 Presentation of the challenge problem (Raja Sengupta)
10.15-10.30 Coffee break

Tools I
10.30-11.30 Formal modeling and implementation (Anouck Girard)

Enabling technologies
11.30-12.30 Hybrid control synthesis and verification (Joćo Sousa)
12.30-01.30 Lunch
01.30-02.30 Protocol design and verification (Sergio Yovine, Stavros Tripakis)
02.30-03.30 Networking (Raja Sengupta, Pravin Varaiya)
03.30-03.45 Coffee break
03.45-04.15 Enterprise control (Raja Sengupta, Wenbin Wei)

Tools II
04.15-05.00 Simulation (Tunc Simsek)

Case study II
05.00-05.45 Case study (Raja Sengupta)
05.45-06.00 Conclusions (Joćo Sousa)