Automatic behavioral model generator for mixed-signal circuits based on structure recognition and auto-calibration

Building the behavioral model for each circuit block is an efficient approach for mixed-signal system verification. If an automatic model generator is available to extract the required behavioral model from the given circuit netlist, it is useful for designers to reduce the extra efforts. Instead of modeling the relationship between circuit inputs and outputs directly, this paper proposes a divide and conquer approach to alleviate the difficulty on building behavioral models. By dividing the circuit into several small building blocks, it is much easier to model the behavior of each block by the black-box approach Therefore, the model construction efforts can be greatly reduced without losing the generality for different circuits. As shown in the experimental results, the proposed approach does generate the corresponding behavioral models automatically with good accuracy.