Objective
- How to formulate your domain specific problem as a CPS
- How to apply some suitable AI techniques to solve a CPS problem
Modus Operandi
- Ask doubts: 5%
- Five minutes teaching (Pick any small topic that is taught in the class before your schedule and teach it for five minutes.): 10%
- Attendance: 5%
- Course project: 10%
- Mid-sem: 30%
- End-sem: 40%
Syllabus
The course is dynamic and the content may include explanations of the questions asked by students as well as some LinkedIn or X discussions related to AI. The following syllabus will be followed to give story to mathematics and programming.
In other words, more focus will be given to fundamentals and underlying rigorous mathematics instead of easy applications of CPS.
Module 1: Introduction
- What is AI?
- What is CPS?
- Key features of CPS
- Modeling, sensing, planning and acting for a physical system
Module 2: Acting
- Case study
- Reinforcement learning in Gymnasium
- Deep Reinforcement Learning in HuggingFace
- MPC (if time permits)
Module 3: Planning
- Case study
- \( A^\ast \)
- RRT
- PRM
Module 4: Sensing
- Case study
- Overview of sensors
- Kalman filter and sensor fusion
- CNN
- YOLO
Books and references
Slides with topicwise references will be provided. A single book with all topics does not exist.