Visuo-Haptic Sensor Fusion System with Active Sensing

人的手是一个良好的传感系统，其中作为传感器的功能和作为执行器的功能很好地融合在一起。在人眼中，作为移动眼睛的致动器的功能深深地影响了作为传感器的功能。这种获取信息的主动运动对实现不同感官之间的融合具有重要作用。主动感知是工程中获取信息的主动运动，而视觉信息和触觉信息为目标识别提供了不同性质的信息。考虑到它们之间的相互关系进行处理，可以实现灵活可靠的识别和动作，作为一个视觉-触觉信息处理或视觉-运动信息处理的传感器信息处理系统。本研究建立了一个传感器融合试验台系统，以实现工程中的视觉-触觉-运动融合系统，并从理论和实验上阐明了其信息处理功能。实际上，我们设计了一个由视觉传感器和触觉传感器组成的完整的传感系统，并在制作这些元素及其基础实验的基础上进行了目标识别实验。结果表明，主动感知在传感器融合系统中具有重要意义。在这些基础实验的基础上，我们提出了新的传感器融合识别算法，并通过对整个系统的评估，形成了传感器融合的系统化理论。

A human hand is a good sensing system in which a function as a sensor and a function as an actuator are fused well. In human eyes a function as an actuator for moving eyes affects a function as a sensor deeply. Such active motion for obtaining information has an important role to realize fusion between different senses. Active sensing is such active motion for obtaining information in engineering.On the other hand, visual information and haptic information provide information with different property for object recognition. To do processing considering mutual relation between them, we can realize a flexible and reliable recognition and actions as a sensor information processing system with visuo-haptic information processing or visuo-motor information processing.In this research we make a testbed system for sensor fusion to realize such visuo-haptic-motor fused system in engineerign, and cleared its information processing function theoretically and experimentally. In fact we design a whole sensing system with vision sensor and haptic sensor, and we did an object recognition experiment based on making the elements and their fundamental experiments. As a result it is cleared that active sensing is important in sensor fusion system. Based on such basic experiments we developed new recognition algorithms for sensor fusion and systemized theory for sensor fusion by evaluating a whole system.