Social robots require excellent communication and interaction skills to be able to operate effectively in dynamic social environments. This includes the ability to capture, understand and respond to complex social content given by humans, such as verbal (e.g. semantics, pauses) and non-verbal (e.g. visual cues, body language) information. Therefore, social robots must be able to adjust to individuals on a case by case basis, as well as being able to identify and provide appropriate responses.
The aim of this project is to develop algorithms that can improve the verbal/non-verbal interactivity of a social robot in a one-to-one conversational setting.
Research activities include either the creation/refinement of techniques for improving verbal exchanges or vision applications for identifying and responding to different types of information provided in a conversational setting. This project is related to the Humanoid Robotics project using the Pepper Robot from Softbank Robotics.
The project outcomes include being able to demonstrate an enhanced interaction in a conversational setting over previous methods, including being able to function with different people taking part in the same one-to-one interaction with the robot.
Skills and experience
Excellent programming skills in Python or C++ and basic knowledge of working with a robotic system.
Basic knowledge of AI, machine learning, natural language processing or image processing would be desirable.
Contact the supervisor for more information.