Deterministic State Estimation

Stage II, Course 7 · 4 credits · 14 weeks · Inferring hidden state from observations

Course Purpose

Students learn to infer hidden state from limited observations. Core question: "Given what we can observe, what can we conclude about the internal state of a system?" This is central to diagnosis, control, and understanding systems where not everything is visible.

By end: Students can determine whether a system is observable, reconstruct hidden state from observations, apply optimal estimation methods, and understand limits of inference.

Learning Outcomes

1. Understand observability formally — when state inference is possible
2. Reconstruct state from observations — methods and guarantees
3. Analyze detectability and reconstructibility — partial observability
4. Apply optimal estimation methods — Luenberger observer, Kalman-like approaches
5. Understand estimation accuracy — sensitivity to noise and model error
6. Design for observability — choosing what to measure
7. Recognize when inference fails — unobservable modes, limits

Core Concepts

observability, state-reconstruction, hidden-variables, measurement, noise, optimal-estimation, observer, Luenberger, Kalman, observability-matrix, reconstructibility

Course Structure (14 weeks)

← Back to Stage II