DocumentCode :
445870
Title :
Circuit implementation of multi-thresholded neuron (MTN) using BiCMOS technology
Author :
Zhu, Xiaolei ; Shen, Jizhong ; Chi, Baoyong ; Wang, ZhiHua
Author_Institution :
Dept. of Inf. Sci. & Electron. Eng., Zhejiang Univ., China
Volume :
1
fYear :
2005
fDate :
31 July-4 Aug. 2005
Firstpage :
627
Abstract :
By analysing the principle of multi-thresholded neurons (MTNs), a methodology is developed for designing multi-thresholded neuron circuits (MTNCs). First, two n-channel MOS transistors are employed to design the voltage-mode synapse circuit with high simplicity and linearity. Second, a BiCMOS technique based circuit named judgement-converting switch (JCS) is proposed. The JCS whose threshold is voltage-controlled current signal, provides an ability of converting the voltage signal to current signal. Based on this possibility, an approach is put forward to design multi-thresholded judgement function circuits (MTJFCs). Finally, single MTNC is designed for implementing XOR operation at switch level. Simulation results with PSPICE show that the designed circuits not only have the correct logic function but also small propagation delay.
Keywords :
BiCMOS integrated circuits; MOSFET; SPICE; electric potential; integrated circuit design; neural nets; BiCMOS technology; PSPICE; circuit implementation; judgement-converting switch; logic function; multithresholded judgement function circuit; multithresholded neuron circuit; n-channel MOS transistor; propagation delay; voltage-controlled current signal; voltage-mode synapse circuit; BiCMOS integrated circuits; Circuit simulation; Linearity; Logic functions; MOSFETs; Neurons; SPICE; Switches; Switching circuits; Threshold voltage;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks, 2005. IJCNN '05. Proceedings. 2005 IEEE International Joint Conference on
Print_ISBN :
0-7803-9048-2
Type :
conf
DOI :
10.1109/IJCNN.2005.1555904
Filename :
1555904
Link To Document :
بازگشت