Accurate modeling of low-cost SiGe:C-HBTs using adaptive neuro-fuzzy inference system

A low-cost BiCMOS SiGe:C-HBT is accurately modeled using adaptive neuro-fuzzy inference system (ANFIS) for the first time. The Volterra kernel-based approach can be suitable for this new kind of modeling. The model has been trained and tested with different sets of input/output data. Accuracy of the model is checked for all the DC and S parameters in a wide range of bias and frequencies. On the validation of the ANFIS model, the average error is found to be less than 4%. Especially in high-current and high-frequency regions, the ANFIS model is proved to be excellent unlike most of the physics-based equivalent circuit models that fail to track the actual device behavior.