An end-to-end tutorial on modern approaches to NLSI
Register NowNonlinear System Identification (NLSI) is essential in engineering, robotics, biomedical systems, finance, and process industries. Modern systems require data-driven techniques such as NARMAX, Deep NARX, Gaussian Processes, and hybrid modeling approaches to handle complexity, uncertainty, and real-world variations.
This workshop introduces practical and theoretical tools for identifying nonlinear dynamical systems. Participants will gain hands-on experience through MATLAB demos and real case studies covering biomedical signals, financial forecasting, process control, robotics, and hardware systems.
| Topic | Key Elements | Duration |
|---|---|---|
| Overview of NLSI | Challenges, classification, modeling approaches | 30 min |
| Parametric Methods | Volterra/Wiener/Hammerstein, NARMAX, Deep NARX | 60 min |
| Case Studies — Parametric | ECG modeling, financial volatility | 90 min |
| Gaussian Processes (GPR) | Kernels, noise modeling, GP-NFIR | 90 min |
| Case Studies — GPR | Two-tank system, Thermal Control Lab | 90 min |
| Transfer Learning | Sim-to-Real, cross-environment adaptation | 60 min |
Graduate students, researchers, and industry professionals working in control, signal processing, AI/ML, applied mathematics, and dynamic systems.
Prerequisites: Basic linear system identification, control theory, MATLAB.
Professor, IIT Tirupati. Expert in system identification, causality, multiscale modeling, optimization, and data analytics.
Assistant Professor, Amrita School of AI. Works in data-driven modeling, large-scale dynamical systems, and modern control.
Researcher, IIT Madras. Focus on Gaussian Process modeling, system identification, and bridging theory and real-world systems.
