Existing robotic platforms in the market have serious limitations for natural face-to-face communication. An alternative approach that could overcome many of these problems is to use state-of-the-art character animation technologies to project lifelike 3-D models onto a robotic mask that can display the 3-D models? natural speech and facial expressions. Currently, there are no such animation-based human robotic agents available in the market that can be acquired, nor can such capabilities be assembled from plug-and-play mechatronic components without significant design work and customization. In this proposal the team has developed a robotic head that projects an animation on a translucent custom designed facial mask to mimic emotive facial gestures and visual speech.
The proposed product (ExpressionBot) is an extremely low-cost and portable robotic head using animation-projection and a custom designed facial mask that can display rich emotive facial gestures and visual speech generated by an avatar agent. The agent can speak, produce and mirror facial expressions synchronized with agent?s visual and prosodic speech. The ExpressionBot can be used as science tutor in classrooms, companionbots in nursing homes, museums tour guide, hotels receptionists, and tourist information kiosks. The ExpressionBot1 comprises a rich Physical Display Portal complete with software libraries for robotic character instantiation including facial expression generation and visual speech generation, human facial recognition, human speech recognition and natural language processing, and world environment recognition.