What This Document Is
This document presents a focused exploration extending the capabilities of HyTech, a symbolic model checker used in the verification of hybrid automata. It delves into the limitations of existing methods when dealing with complex system dynamics and proposes approaches to overcome these challenges. The material is based on a presentation delivered at the University of California, Berkeley, as part of the Introduction to Embedded Systems course (ELENG C249A). It represents joint work with leading researchers in the field.
Why This Document Matters
This resource is valuable for students and professionals working with embedded systems, control theory, and formal verification. It’s particularly relevant for those seeking to understand the practical challenges of applying symbolic model checking techniques to real-world hybrid systems. Individuals interested in expanding the applicability of verification tools beyond restrictive assumptions will find this a useful study aid. It’s best utilized when you’re already familiar with the fundamentals of hybrid automata and symbolic model checking and are looking for advanced concepts.
Topics Covered
* Hybrid Automata fundamentals and representation
* Symbolic Model Checking techniques and requirements
* The architecture and capabilities of the HyTech model checker
* Limitations of polyhedral automata in complex systems
* Interval numerics as a method for extending dynamic analysis
* Algorithms for exploring state spaces in hybrid systems
* Techniques for improving the accuracy and efficiency of verification
* Case studies illustrating the application of these concepts
What This Document Provides
* A detailed overview of the core concepts behind extending HyTech’s functionality.
* An examination of the trade-offs between accuracy and computational cost in hybrid system verification.
* A discussion of methods for handling dynamics that fall outside the scope of traditional polyhedral automata.
* Insights into the implementation of a prototype system designed to address the shortcomings of existing tools.
* A comparative analysis of different approaches to bounding and approximating system behavior.
