Computational methods for active learning of dynamical models and model predictive control

Fields:

Model predictive control, system identification, numerical optimization, machine learning, design of experiments

Activity:

Carry out research activities within the scope of the ERC Advanced Grant “COMPACT” (COmputational Model Predictive and Adaptive Control Tools) on one of the following topics: active learning methods for designing system identification experiments coupled with robust MPC schemes for safe data collection; numerical optimization methods for offline nonlinear system identification and online model adaptation; preference-based MPC calibration algorithms based on human calibrators’ preferences, possibly in a game-theoretic setting; explicit and semi-explicit MPC algorithms for fast online controller evaluation and light MPC adaptation schemes. Co-supervision of laboratory activities to validate the research results on an experimental robotic platform.

Profile:

A highly motivated and talented individual to join our team as a Postdoctoral Researcher. The successful candidate should hold a PhD in one of the following fields: model predictive control, numerical optimization, system identification, or machine learning, with a strong background in computational methods and algorithm development. The candidate must have a proven track record of research excellence demonstrated by publications in top-tier journals and conferences, experience with numerical optimization techniques and/or system identification/machine learning approaches, very good programming skills in relevant languages for scientific computing (e.g., Python, Julia, MATLAB), and the ability to work independently and as part of a team.

Formal requirements:

• PhD in one of the following fields: model predictive control, numerical optimization, system identification, machine learning;
• Excellent knowledge of both written and spoken English.

Duration:

2 years, renewable

Gross amount:

40.530,00 €/year

SSD:

IINF-04/A – Systems and Control Engineering

IINF-04/A - Automatica

Project:

progetto “COmputational Model Predictive and Adaptive Control Tools” – COMPACT, GA n. 101141351, che è stato finanziato nell’ambito del bando Call for ERC Advanced Grant: ERC-2023 ADG rivolto alla ricerca di frontiera e valutato secondo il criterio dell’eccellenza