NOVEL TECHNOLOGY INDEPENDENT NEURAL-NETWORK APPROACH ON DEVICE MODELING INTERFACE

A novel, fast and accurate neural network tool is proposed for efficie nt technology independent realisation of the interface between device modelling and circuit simulation. Enhanced back-propagation neural net work based algorithms are applied to the problem of modelling various device characteristics. These algorithms include the modified back-pro pagation algorithm, the conjugate gradient algorithm and the Levenberg -Marquardt algorithm. Also, the radial basis function neural network i s tested in the device modelling problem. Simulations show fast conver gence or learning rate and an excellent fit of recalled characteristic s to the measured device data. The algorithm utilised is robust and ca pable of presenting the entire device characteristics unaltered even w ith largely reduced amount of the teaching material. The good monotoni city of the neural network generated device data facilitates the usage of the method in circuit simulation purposes. Possible further applic ations of implementing circuit level macromodels with this technique a re discussed.