Study on the Modeling of a Neuromorphic Transistor Circuit System

When the traditional von Neumann architecture calculates large and complex problems, the data transmission rate and energy consumption between CPU and memory become a huge bottleneck, and the application of neuromorphic computing can effectively solve this problem. In this work, because the neuromorphic computing circuit is composed of neurons and synapses, we mainly achieved the basic functions of neurons and synapses through circuits, and further complete the circuit design and simulation of neuromorphic system. Besides, in the design of the neuron circuit, we innovatively proposed the monostable trigger and pulse emission circuit, which can emit the required pulse waveform, simulate the integrated emission function of biological neurons, and solve the problem of controlling the time length of the pulse emission waveform. Besides, In neural synaptic circuits, we produce approximately linear changes in synaptic weights that enable the function of learning, computation, and forgetting. In summary, using the circuit designed in this paper can improve the learning rate and improve the work efficiency.