Nano-Carbon Neuromorphic Circuits for on-body AI -CoNoCo

On-body AI refers to the use of wearable sensors for personalized and secure healthcare, enabling the intelligent performance of complex tasks with low power consumption, small size, and lightweight devices.

To achieve this, the neuromorphic design of analog circuits that emulate the nervous system provides a promising solution.

However, artificial neuron networks developed worldwide have mostly been composed of excitatory neurons, unlike the nervous system which also contains inhibitory neurons.

Furthermore, artificial neurons based on organic and polymer films have limitations in delivering bio-plausible spiking performance.

Therefore, this project aims to develop artificial excitatory and inhibitory neuron circuits using carbon nanotubes and graphene as key materials in the circuit design.

The goal is to achieve bio-plausible spiking performance with a spike width of 1-2 ms and a spiking rate range of 0-500 Hz.

Additionally, we plan to demonstrate the superiority of a network comprising both artificial excitatory and inhibitory neurons over a conventional network of only artificial excitatory neurons in terms of energy and computation efficiency.

Our approach is based on the belief that the evolved nervous system represents the most efficient engine for cognitive computing.

Thus, a network comprising both artificial excitatory and inhibitory neurons with bio-plausible spiking performance is essential to achieve the functions and efficacies of the brain.