| Title | : | Probabilistic Model-Checking and System-theoretic Process Analysis Methods: A Case for Autonomy Management in Networked Systems |
| Speaker | : | Kaliappa Ravindran (Grove School of Engineering) |
| Details | : | Wed, 5 Feb, 2025 11:00 AM @ SSB 334 |
| Abstract: | : | The focus of my talk is on "probabilistic model-checking" (PMC) as a flexible design paradigm for fault-adaptive and resilient network systems. My talk dwells on two sub-parts: 1. State-machine programming style that underlies the PMC approach, where a system designer explicitly assigns "state-transition probabilities" to depict the likely occurrence of various "events" and "actions" incident on the system at run-time; 2. A case-study for the PMC approach: namely, a terrain surveillance system composed of multiple drones with on-board cameras where the images used for object detection often exhibit "data fuzziness", thereby requiring a probabilistic treatise of system behavior. For part-1, I shall use a well-known model-checking language, PROMELA (developed by G. Holzmann at AT&T Bell-Labs in the 90's). A software overlay for PROMELA developed at CUNY for programmers to explicitly assign state-transition probabilities will be explained. Other PMC languages are: PRISM from University of Oxford and STORM from University of Aachen. The part-2 in my talk is about process anlysis and mining methods for intelligent decision-making that are layered on top of a PMC-based system modeling platform. My talk will cover some key aspects of a cyber-fence system being developed at CUNY for metro-area surveillance by a team of drones. The system employs a confusion-matrix of probabilities to characterize the accuracy of drone-mounted sensing system for object detection. How the PMC methods improve the design & development phase of such complex systems will be highlighted. The improved reusability and reconfigurability of system models & algorithms, as offered by PMC-based design approaches, is a major advantage in today's world of complex cyber-physical systems. |
