ABOUT THE COURSE: Cyber-physical systems (CPS), which consist of physical systems tightly integrated and/or controlled by software, are ubiquitous in many safety critical domains, including automotive, avionics, railways, healthcare, atomic energy, power, and industrial automation. The principles of design and implementation of cyber-physical systems are remarkably different from that of other embedded systems because of the tight integration of real valued and dense time real time systems with software based discrete automated control. The objective of this course is to develop an exposition of the challenges in implementing a cyber-physical system from a computational perspective, but based equally on the principles of automated control. The course aims to expose the student to real world problems in this domain and provide a walk through the design and validation problems for such systems. With the advent of Al techniques, their increased use in CPS is also a promising growth vertical along with the necessity of safety assurance. In this course we also touch upon concepts of Neural Network based decision making for Continuous Systems while guaranteeing safety and stability using control theoretic constraint solving.INTENDED AUDIENCE: UG/PG students of CSE/EE/ECEPREREQUISITES:1. Basic Programming Knowledge2. Engineering MathematicsINDUSTRY SUPPORT: "Tier 1 Automotive companies:Robert Bosch Engineering, OEM Automotive companies:TML, BMW, Daimler, Mahindra etc, Govt Labs like DRDO, HAL"
Foundations of Cyber Physical Systems
Indian Institute of Technology, Kharagpur and NPTEL via Swayam
Overview
Syllabus
Week 1:
i. Cyber— Physical Systems (CPS) in the real world: Industry 4.0, Automotive, Building Automation, Medical CPS
ii. Low power compute platforms for CPSWeek 2:Real time sensing and communication for CPS
i. Sensors, Actuators
ii. CAN protocol in automotive systemsWeek 3:
i. Real time task scheduling for CPS
ii. Worst Case Execution Time, Res ponse time analysis of CPS softwareWeek 4:
i. Dynamical System modeling for CPS
ii. Different notions of stabilityWeek 5:
i. Controller Desig n (using pole placement)
ii. Delay aware Controller DesignWeek 6:
Stability and Control Performance in presence of Platform uncertaintiesWeek 7:
i. Lyapunov Stability
ii. Barrier FunctionsWeek 8:
Quadratic Program based Controller Design ensuring Safety and StabilityWeek 9:
Neural Network (NN) Based Controllers in CPSWeek 10:
Safety of NN enabled CPS: switching between NN and conventional controllersWeek 11:
State Estimation using Kalman Filter and other techniquesWeek 12:
i. False Data Injection (FDI) Attack detection in CPS
ii. Attack Mitigation in C
i. Cyber— Physical Systems (CPS) in the real world: Industry 4.0, Automotive, Building Automation, Medical CPS
ii. Low power compute platforms for CPSWeek 2:Real time sensing and communication for CPS
i. Sensors, Actuators
ii. CAN protocol in automotive systemsWeek 3:
i. Real time task scheduling for CPS
ii. Worst Case Execution Time, Res ponse time analysis of CPS softwareWeek 4:
i. Dynamical System modeling for CPS
ii. Different notions of stabilityWeek 5:
i. Controller Desig n (using pole placement)
ii. Delay aware Controller DesignWeek 6:
Stability and Control Performance in presence of Platform uncertaintiesWeek 7:
i. Lyapunov Stability
ii. Barrier FunctionsWeek 8:
Quadratic Program based Controller Design ensuring Safety and StabilityWeek 9:
Neural Network (NN) Based Controllers in CPSWeek 10:
Safety of NN enabled CPS: switching between NN and conventional controllersWeek 11:
State Estimation using Kalman Filter and other techniquesWeek 12:
i. False Data Injection (FDI) Attack detection in CPS
ii. Attack Mitigation in C
Taught by
Prof. Soumyajit Dey
