Robotics: from material to cognitive - WPMGROB9

Goal(s)

• Acquisition of fundamental knowledge on perception and control architectures in robotics, with associated learning methods
• Acquisition of basic skills in robotic system modeling (hardware and control architecture) and control (e.g., servoing)

Content(s)

After a historical retrospective of robotics, partially explaining its current definitions and its inclusion in various disciplines, the evolution of robotics challenges will be presented, from the emergence of the field to the present day. The course will illustrate the diversity of existing approaches and solutions, depending on the objectives and application contexts.
In the first major part of this course, J-C. Quinton will address the issue of coupling between sensors and actuators, first from a hardware design perspective and then from a control perspective. To enable the transition between hardware and control, concepts from mechanics and kinematics will be introduced. Several robotic control architectures (classical and more recent) will be presented, as well as methods spanning the disciplinary spectrum of robotics (e.g., from automation or AI).
In the second part of the course, M. Lefort will address the learning aspects involved in the majority of robotic architectures that must interact with less controlled environments. The main classes of machine learning and their applications to robotics will be introduced, along with methods dedicated to interactive systems (e.g., learning by demonstration or imitation). Various methods for abstracting sensorimotor interactions and the robotic control component will be presented, integrating the spatial, temporal, and multimodal aspects of signals. The limitations of current approaches for cognitive and developmental robotics will be discussed.
Depending on the needs of the participants and the time available, both parts will be structured and illustrated by simulated case studies on a computer platform. These will allow participants to manipulate the concepts developed in the course and solve small robotics challenges (in teams, in the form of mini-projects).

Test

Written exam

Additional Information