Dexterous skills in humans are arguably one of the most challenging abilities for robots to learn. These require incessant coupling between action and tactile perception maneuvered by a central core such as the brain. Specifically, the implementation of tactile feedback in a robotic system has only begun recently along with the advancement in tactile sensor technology, which now calls for novel methods to realize human-like touch covering from sensors to perception. This tactile intelligence is the key to the development of robots that are able to grasp and manipulate objects in various circumstances. Development of robot tactile intelligence would also deepen our understanding of tactile information processing in the human nervous system.
To make a robot touch and feel like a human, it is natural to translate the principles of the human nervous system into robot intelligence. However, relatively little is known about neural mechanisms of tactile sensation and perception compared to visual or auditory modalities. Hence, more engineering-oriented approaches employing cutting-edge machine learning algorithms have been adopted to implement tactile intelligence in robots. They can provide practical solutions to the problems of object manipulation by a robot but may benefit much more from the understanding of biological bases of human tactile perceptual capabilities. For instance, it is desirable to create artificial neural networks that mimic somatosensory nervous systems and integrate them with advanced machine learning techniques in order to have robots gain human-like dexterous skills.
This research proposal aims to investigate the implementation of tactile intelligence in robots inspired by neural mechanisms of tactile information processing in order to improve object grasping and manipulation with tactile feedback. It also aims to collect current learning methods for robots to exploit tactile information from the sensors. Theoretical work as well as cognitive models to explain emergence of tactile perception from biological systems are welcomed. This proposal also emphasizes human studies on tactile perception that can provide insights on robot learning for object manipulation tasks.
This research topic is interdisciplinary with a goal to integrate state-of-the-art researches from diverse domains. Topics include, but not limited to:
- Learning mechanisms through touch
- Human tactile perception mechanisms translated into robots
- Biologically inspired tactile information processing of tactile sensor signals
- Methods to transform raw tactile sensor signals into perceptual information
- Artificial tactile intelligence for object manipulation and recognition
- Robot learning algorithms to gain tactile intelligence
- Neural mechanisms for closed-loop of hand action and tactile perception
- Intelligent tactile sensor technology
Dexterous skills in humans are arguably one of the most challenging abilities for robots to learn. These require incessant coupling between action and tactile perception maneuvered by a central core such as the brain. Specifically, the implementation of tactile feedback in a robotic system has only begun recently along with the advancement in tactile sensor technology, which now calls for novel methods to realize human-like touch covering from sensors to perception. This tactile intelligence is the key to the development of robots that are able to grasp and manipulate objects in various circumstances. Development of robot tactile intelligence would also deepen our understanding of tactile information processing in the human nervous system.
To make a robot touch and feel like a human, it is natural to translate the principles of the human nervous system into robot intelligence. However, relatively little is known about neural mechanisms of tactile sensation and perception compared to visual or auditory modalities. Hence, more engineering-oriented approaches employing cutting-edge machine learning algorithms have been adopted to implement tactile intelligence in robots. They can provide practical solutions to the problems of object manipulation by a robot but may benefit much more from the understanding of biological bases of human tactile perceptual capabilities. For instance, it is desirable to create artificial neural networks that mimic somatosensory nervous systems and integrate them with advanced machine learning techniques in order to have robots gain human-like dexterous skills.
This research proposal aims to investigate the implementation of tactile intelligence in robots inspired by neural mechanisms of tactile information processing in order to improve object grasping and manipulation with tactile feedback. It also aims to collect current learning methods for robots to exploit tactile information from the sensors. Theoretical work as well as cognitive models to explain emergence of tactile perception from biological systems are welcomed. This proposal also emphasizes human studies on tactile perception that can provide insights on robot learning for object manipulation tasks.
This research topic is interdisciplinary with a goal to integrate state-of-the-art researches from diverse domains. Topics include, but not limited to:
- Learning mechanisms through touch
- Human tactile perception mechanisms translated into robots
- Biologically inspired tactile information processing of tactile sensor signals
- Methods to transform raw tactile sensor signals into perceptual information
- Artificial tactile intelligence for object manipulation and recognition
- Robot learning algorithms to gain tactile intelligence
- Neural mechanisms for closed-loop of hand action and tactile perception
- Intelligent tactile sensor technology
