Fundamental Research on Intelligent Sensors

我们的研究目标是研究智能传感器的新技术，包括用于生物医学和机器人应用的超声波传感器，用于机器人的光学传感器和用于生物医学诊断的正电子发射传感器。本年度，根据日美双方的讨论，我们主要致力于开发将传感器的周边预处理器所提供的感觉信息进行融合的新技术。这些新技术包括将组织定征和临床诊断用的相移参数和结构变化参数等超声波非线性参数进行融合，正电子发射断层扫描数据与磁共振图像在分层神经网络中的融合以及多通道诱发电位波形的融合在活体神经束定位中的应用，通过对上述应用的实验，我们证明了分层智能处理器，就像生物神经系统一样，从外周到中枢，极大地提高了传感系统的性能，特别是用于诊断用途。

The objective of our research is to investigate the new technologies for intelligent sensors including ultrasonic sensors for biomedical and robotic applications, optical sensors for robotics and positron-emission sensors for biomedical diagnosis. This fiscal year, based on the discussion among the Japanese and US project participants, we mostly concerned with the development of new technologies for fusing sensory information given by peripheral pre-processors attached to each transducer.These technologies include fusion of ultrasonic nonlinear parameters such as the phase-shift parameter and structure-change parameter for tissue characterization and clinical diagnosis, fusion of positron emission tomographic data with magnetic resonance images in a layred neural network and the one to fuse multi-channel evoked potential waveform to localize the neuro-bundle position in the living body.Through the experiments regarding the above applications, we showed that the hierarchically arranged intelligent processors, from the peripheral to the central, just like the biological nervous systems, extremely improve the performance of the sensing system especially for diagnostic use.