A data-driven control framework for constrained nonlinear systems with application to unmanned ground vehicles

The objective of this project is the development of a novel data-driven control framework for nonlinear dynamical systems. The proposed solution will be validated and tested on the autonomous ground vehicles available at Concordia University. To achieve the desired goal, the research intern, with the supervision and help of the host and home supervisor, will take a three-step approach. In the first step, the intern will jointly develop a reinforcement learning algorithm and behavioural approach to obtain a data-driven outer and convex differential inclusion of the dynamics of the nonlinear system. Then, in the second part, it will use the developed data-driven model to design a data-driven approach inspired by an existing model-based counterpart. Finally, in the last step, the proposed control architecture will be tested on autonomous ground vehicles and contrasted with other model-based and data-driven alternative schemes.

The conducted research is anticipated to lead to a prestigious peer-reviewed conference/journal publication co-authored by all the participants of this project (student, home supervisor, host supervisor). In general, the intended research and collaboration are believed to contribute to increasing and improving the research and innovation capacities of both host and home institutions in the field of data-driven control.

Faculty Supervisor:

Walter Lucia

Student:

Partner:

University of Calabria

Discipline:

Engineering

Sector:

Education

University:

Concordia University

Program: