Cognitive-Robotics
Projects
Open-Ended Learning Approaches for 3D Object Recognition
Active segmentation of cluttered scenes
Coupling between Perception and Manipulation
Description
Cognitive robots are expected to be more autonomous and efficiently work in human-centric environments. A cognitive robot should process very different types of information in varying time scales. Two different modes of processing, generally labeled as System1 and System2, are commonly accepted theories in cognitive psychology. The operations of System1 (i.e. perception and action) are typically fast, reactive, and intuitive. The operations of System2 (i.e. semantic) are slow, deliberative, and analytic.
For such robots, open-ended learning for object perception and grasping is a challenging task due to the high demand for accurate and real-time responses in dynamic environments. In this course, "open-ended" implies that the set of object categories to be learned is not known in advance, and the training instances are extracted from online experiences of a robot, and become gradually available over time, rather than being completely available at the beginning of the learning process. This way the robot adapts its perception and grasping skills over time to different environments. This year's theme of the course is "Simultaneous Multi-View Object Grasping and Recognition in Open-Ended Domains".
This course covers a diverse set of topics that focus on addressing the most critical aspects of building a cognitive robotic system. We recently wrote a survey paper about the state of lifelong learning in service robots (see https://link.springer.com/article/10.1007/s10846-021-01458-3). It covers all the topics of the cognitive robotics course in a concise and brief manner to help students in easy remembrance and quick revision.
The course is a combination of lectures, reading sessions, and lab sessions. The lectures discuss the fundamentals of topics required to develop a cognitive robotic system. During the reading sessions, students present and discuss recent contributions in the fields of object perception and manipulation. Webpage of the course: https://rugcognitiverobotics.github.io/
Learning outcomes
- Explain the main theories of open-ended learning and cognitive robotics.
- Explain meaning of different concepts often used in the field of 3D computer vision and Human Robot Interaction and their application in robotics.
- Exploit deep transfer learning algorithms for open-ended object category recognition.
- Implement and experiment deep learning architectures for object grasping.
- Create a tight coupling between object perception and manipulation and perform experiment using real Kinect data and a simulated Panda robotic arm.
- Design a cognitive robotic system capable of dealing with unseen object categories and performing manipulation tasks in open-ended environments.
- Use RACE cognitive robotic system, ROS, and Gazebo in practical projects.