The Energy-Efficient and Distortion Blocking Leaky Integrate-and-Fire Neuron Model

Neuromorphic technologies have the potential to revolutionize cognitive computing by accelerating biologically inspired learning algorithms. These algorithms and architectures rely on basic processing and memory units called neurons and synapses. In neuromorphic circuit design, there is a continuous quest to create energy-efficient and distortion-free spiking neurons for integration into neuromorphic system architectures. In this context, this paper introduces a novel approach to prevent the distortion encountered in the membrane voltage of the LIF neuron circuit. By implementing a low-power Schmitt trigger-based spike generator in the LIF neuron design, the circuit effectively mitigates distortions in the membrane voltage. The presenting leaky integrate-and-fire neuron has been realized by 12 transistors and two capacitors, which offer a significant energy efficiency advancement in contrast to their current equivalents. In addition, the LIF neuron proposal showcases optimal energy and power savings, like 8.56fJ energy consumption per spike, and the model consumes an average power of 1.62nW. Furthermore, it offers the flexibility to adjust firing frequency within the biological frequency range of 1Hz-84KHz using a control voltage V
C
. These findings highlight the potential of creating energy-efficient and distortion-free neuromorphic systems.