Structured learning in fuzzy spiking neural networks for human state estimation

In this paper, we focus on the monitoring of sleep using an optical oscillosensor and a pneumatic sensor for the health care of the elderly people. The system composed of these sensors applies thresholds for the estimation of human behaviors. We can use membership functions to extract the feature of sensor data, and spiking neural networks to estimate the human state in the bed. However, it is difficult to design the membership function in advance because of environmental condition and personal difference. Therefore, we propose a structured learning in fuzzy spiking neural networks to enable optimization of the membership functions in the learning process. We discuss the effectiveness of the proposed method through comparative experiments.