I. Introduction
Neuromorphic sensory systems would offer a promising approach for achieving high energy-efficient perception based on biological strategies [1], [2], [3]. Artificial sensory neurons are regarded as one building blocks for neuromorphic sensory system, which have been demonstrated with sophisticated functions similar to their biological counterparts [4], [5], [6]. An emerging trend is dedicated to pursue a high biological plausibility by introducing the spike-encoding scheme [7]. Such spike-encoding scheme is believed with considerable superiorities such as energy-efficiency, fault-tolerance, and noise-robustness [8], [9], [10]. In the peripheral nervous system, different forms of external information are detected by various types of receptors and are converted into spike trains through the soma of a sensory neurons (Fig. 1a) [11]. In this case, a sensory neuron can be analogous to of the integration of a sensor that is able to detect the external stimulation and an oscillating unit that is able to generate spikes based on the properties of stimulation.
(a)Schematic of sensory neurons encoding external stimulus into spike trains. (b) Structure schematic and (c) optical photographs of the oscillating device.