Robotics Research on Human Intention

如今，机器人的应用正在扩展到服务、家务、家庭护理任务。有必要让这些机器人和人类安全共存。而人类友好沟通是开发此类机器人系统的重要线索之一。在这个项目中，我们专注于从视觉信息中识别人脸和手势。实际上，我们已经完成了以下六个研究项目。(1)人体手势识别——基于HMM和HMM的实时上下文手势识别——基于子空间方法和HMM的图像运动手势识别；(2)姿态估计——基于结构和运动模型的姿态估计；(3)人体跟踪和注意力控制——全向传感器机器人和双目视觉机器人协同发现和跟踪人；(4)人脸识别——整合人的身份和表情信息，实现鲁棒面部表情人脸识别人脸表情连续变化的识别程度(5)人机协作全向图像传感器自主制导机器人系统(6)HyperOmni vision微型全向图像传感器的研制人类友好型机器人YYOU的研制

Nowadays, robot applications are being extended to service, house keeping, home-care tasks. There is a necessity to have the safe coexistence of such robots and human-beings. And, human friendly communication is one of important cue for developing such robot system. In this project, we focused on face and gesture recognition from visual information. Actually, we have done following six research projects.(1) Human Gesture RecognitionReal-time context-based gesture recognition using HMM and automatonGesture recognition from image motion based on subspace method and HMM(2) Posture EstimationPosture estimation using structure and motion models(3) Human Tracking and Attention ControlFinding and tracking a person by cooperation between an omnidirectional sensor robot and binocular vision robot(4) Face RecognitionIntegrating person identity and expression information for robust facial expression recognitionRecognizing degree of continuous facial expression change(5) Human-Robot CooperationAutonomous guidance robot system with omnidirectional image sensor(6) HyperOmni VisionDevelopment of a tiny omnidirectional image sensorDevelopment of human friendly robot YYOU