Social RoboticsTU Graz works on making Furhat give natural conversational feedback
Speech scientists want the robot head to say 'mhm' at the right moments.
At TU Graz, Barbara Schuppler and student Michael Paierl are researching how a Furhat social robot can take turns more naturally, hold attentive gaze and give small verbal signals of listening. The plan is to analyse recordings of human conversations to model the acoustic cues that tell a listener when to respond. The team noted that the robot's topics are still limited to keep delays down and that speech recognition struggles with some Austrian dialects. Such backchannelling is a key gap between today's talking robots and fluid human conversation.
Source: TU Graz
This archive entry summarises reporting published elsewhere; follow the source link for the full story. The summary was drafted with AI from that reporting.