Doctoral Project: Human-inspired interactive perception for robotic manipulation
Part of research project: Capabilities and consequences of recursive, hierarchical information processing in visual systems
Description of the research project
The human perceptual system works unbelievably well. But we do not understand the principles of information processing that enable such versatile, robust, and accurate perceptual processes. In this project, we combine the study of the human perceptual system with our knowledge of how to construct technical perceptual systems, combining robotics and computer vision. Through this combination, we can ask novel questions and pursue innovative technical approaches. Our goal is both to better understand the human perceptual system and to build robotic perceptual systems with previously unseen abilities.
Description of the doctoral project
Research in interactive perception does not distinguish between action and perception but regards them as two sides of the same coin, similar to what we know about biological perception systems. Online perceptual approaches in interactive perception are based on interconnected recursive estimation processes, each informed by prior knowledge about a specific aspect of the perception task. In this project, you will–in collaboration with psychologists that study human perception–develop novel algorithms in interactive perception. This includes theoretical algorithm development as well as practical experiments on a mobile manipulation platform.
Project start date: October 1, 2019
- Strong interest in innovative, interdisciplinary research
- MS degree in computer science or similar field
- Research experience in robotics, computer vision, machine learning, and/or control
- Excellent software engineering and programming skills in C++
- Excellent English writing and communication skills