Simple Applications of WiBIC with Asynchronous Pulse Code Multiple Access
Jiaying LIN, Nagasawa RYUJI, Kien NGUYEN, Hiroyuki TORIKAI, Mikio HASEGAWA, Won-Joo HWANG, Hiroo SEKIYA
IEICE Technical Report, vol.122, no.453, pp.85-90, Mar., 2023. [pdf document]

<Abstract>

In this study, we propose to combine Spiking Neural Network and IoT network to construct a distributed information processing system, the WiBIC system. From the idea of WiBIC, SNN should be wireless. The CSMA/CA protocol based on IEEE802.15.4e adopted by most IoT communication terminals will generate random waiting time before transmission, thus losing the time information of the spike generation moment, which is not suitable for SNN. Therefore, the Asynchronous Pulse Code Multiple Access(APCMA) protocol is used to construct the WiBIC system. The APCMA protocol can retain the information about the spike generation time and make the SNN operate stably. This wireless SNN is used to construct networks and applied to two learning problems: ”Pavlovian conditioning” and ”XOR function”. Experimental results show that the WiBIC system has learning functions.