论文部分内容阅读
A circuit system of on|chip BP(Back|Propagation) learning neural network with programmable neurons has been designed,which comprises a feedforward network,a n error back|propagation network and a weight updating circuit.It has the merit s of simplicity,programmability,speediness,low power|consumption and high densi ty.A novel neuron circuit with programmable parameters has been proposed.It gene rates not only the sigmoidal function but also its derivative.HSPICE simulations are done to a neuron circuit with level 47 transistor models as a standard 1 2 μm CMOS process.The results show that both functions are matched with their res pective ideal functions very well.The non|linear partition problem is used to v erify the operation of the network.The simulation result shows the superior perf ormance of this BP neural network with on|chip learning.
A circuit system of on chip BP (Back | Propagation) learning neural network with programmable neurons has been designed, which comprises a feedforward network, an error back | propagation network and a weight updating circuit. It has the merit s of simplicity, programmability, speediness, low power | consumption and high densi. A novel neuron circuit with programmable parameters has been proposed. Not genetically not the sigmoidal function but also its derivative. HSPICE simulations are done to a neuron circuit with level 47 transistor models as a standard 1 2 μm CMOS process. The results show that both functions are matched with their res pective ideal functions very well. non | linear partition problem is used to v erify the operation of the network.The simulation result shows the superior perf ormance of this BP neural network with on chip learning.