Modeling of Voltage-Controlled Oscillators Including I/O Behavior Using Augmented Neural Networks

Abstract

This paper proposes augmented neural networks (AugNNs) for modeling the behavior of steady-state oscillators in time-domain. Multi-output AugNNs with the corresponding gradient scheme and training methodology are proposed for the modeling of multi-phase oscillators. Using the proposed AugNNs, a novel technique is presented for the modeling of voltage-controlled oscillators (VCOs) including I/O behavior. In the proposed AugNNs, a periodic unit is introduced to capture the relation between the instantaneous frequency and the control signal, and to provide the phase information of the oscillation, which is used to predict the oscillatory output waveforms using feed forward neural networks (FFNN). For the modeling of VCOs including output buffers, an AugNN-based model is proposed, where recurrent neural networks (RNNs) are used to capture the nonlinear dynamic current-voltage relation at the output port. The simulated data of transistor-level oscillator circuit are used to train the proposed model. As a black-box approach, the proposed model protects intellectual property (IP) and can be implemented in Verilog-A. The examples using transistor-level oscillator circuits demonstrate the effectiveness of the proposed model for time-domain analysis.INDEX TERMS Behavioral modeling, voltage-controlled oscillator (VCO), neural network, output buffer, Verilog-A.

Advanced Space Mapping and Artificial Neural Network-Based Techniques for Microwave Device Modeling

“…ANN has also been an important vehicle for parametric modeling of EM behavior in radio frequency and microwave area [44][45][46], [49], [50], [109][110][111][112]. ANN benets from its strong learning and generalization capabilities and it has been used for a wide variety of microwave applications [56][57][58][59][60][61][62][63][64], [66][67][68][69][70][71][72][73][74][75][76][77], [80]. The universal approximation theorem [78] of ANNs provides the theoretical foundation that if sucient data are used in training, the good accuracy of ANN models can be achieved within the training region.…”

“…Articial neural networks (ANNs) have been recognized as important vehicles for device modeling in RF and microwave area [56][57][58][59][60][61][62][63][64], [66][67][68][69][70][71][72][73][74][75][76][77], [80]. As important datadriven models in the passive and active RF and microwave device modeling area, ANN models benet from their strong learning and generalization capabilities [78].…”

“…ANNs have been used for modeling a wide variety of microwave components [56][57][58][59][60][61][62][63][64], [66][67][68][69][70][71][72][73][74][75][76][77], [80], including microstrip lters [47], [68], antennas [57], vias [58], high-speed interconnects [59], [60], transmission line components [61], [94], coplanar waveguide (CPW) components [62], waveguide lters [63], [64], embedded passives [54], [67], voltage-controlled oscillators [69], mixers [71], bends [80], spiral conductors [66], power ampliers [34], [35], eld-eect transistors (FETs) [16], [36], high-electronmobility transistors (HEMTs) [6], [37][38][39], etc.…”

Advanced Parametric Modeling and Yield Optimization Approaches for Microwave Passive Components

“…antennas [52], vias [53], power amplifiers [5], high-speed interconnects [54], [55], high electron mobility transistor (HEMT) devices [8], transmission line components [56], [57], coplanar waveguide (CPW) components [58], waveguide filters [9], [34], [59], embedded passives [60], [61], voltage-controlled oscillators [7], mixers [62], bends [63], and spiral conductors [64] etc.…”

“…7 shows the yield of the waveguide FSI filter before and after optimization using the proposed approach. The optimal yield solution found by the proposed approach is [9.3682 4.9542 3.7585] T (mm).…”