Regression Model-Based AMS Circuit Optimization Technique Utilizing Parameterized Operating Condition

Abstract

An analog and mixed-signal (AMS) circuit that draws on machine learning while using a regression model differs in terms of the design compared to more sophisticated circuit designs. Technology structures that are more advanced than conventional CMOS processes, specifically the fin field-effect transistor (FinFET) and silicon-on-insulator (SOI), have been proposed to provide the higher computation performance required to meet various design specifications. As a result, the latest research on AMS design optimization has enabled enormous resource savings in AMS design procedures but remains limited with regard to reflecting the intended operating conditions in the design parameters. Hereby, we propose what is termed a supervised learning artificial neural network (ANN) as a means by which to define an AMS regression model. This approach allows for rapid searches of complex design dimensions, including variations in performance metrics caused by process–voltage–temperature (PVT) changes. The method also reduces the considerable computation expense compared to that of simulation-program-with-integrated-circuit-emphasis (SPICE) simulations. Hence, the proposed AMS circuit design flow generates highly promising output to meet target requirements while showing an excellent ability to match the design target performance. To verify the potential and promise of our design flow, a successive approximation register analog-to-digital converter (SAR ADC) is designed with a 14 nm process design kit. In order to show the maximum single ADC performance (6-bit∼8-bit resolution and few GS/s conversion speed), we have set three different ADC performance targets. Under all SS/TT/FF corners, ±6.25% supply voltage variation, and temperature variation from −40 ∘C to 80 ∘C, the designed SAR ADC using our proposed AMS circuit optimization flow yields remarkable figure-of-merit energy efficiency (approximately 15 fJ/conversion step).

AMS Circuit Design Optimization Technique Based on ANN Regression Model With VAE Structure

“…To enhance the design optimization process, our proposed zoom-in algorithm in this study achieves greater precision and efficiency compared to previous research efforts. While previous studies have explored the use of regression models for analog circuit design automation [3], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], the original zoom-in algorithm presented in [3] had a limitation in the accuracy of the regression model during the initial step due to a smaller amount of data compared to single-step approaches. To address this limitation, our study focuses on improving accuracy by incorporating the variational autoencoder (VAE) structure and refining the verification step within the algorithm.…”

“…Therefore, considering the PVT variations, optimization with simulation results over a wide range can design a circuit with better performance. Similarly, studies including corner simulation, layout generator, and Monte Carlo simulation have been developed to consider PVT variations [10], [11], [12], [13], [14]. These simulations to check PVT variations have the disadvantage of requiring a significant computer processing time.…”

Optimized Design of a Self-Biased Amplifier for Seizure Detection Supplied by Piezoelectric Nanogenerator: Metaheuristic Algorithms versus ANN-Assisted Goal Attainment Method

“…Moreover, significant works has been performed in the field of AI-assisted optimal design of electronic circuits. A parameter optimization of a chaotic circuit by the use of Bayesian optimization and genetic algorithm has been reported by Acharya et al [ 33 ], and regression-model-based optimization of analog mixed signal circuits has been reported by Nam et al [ 34 ]. An overview of various AI applications for power electronics in design, control, and maintenance life-cycle phase [ 35 ] lists typical tasks (optimization, classification, regression, and data structure exploration) and methods (deterministic programming methods based on linear or quadratic programming; nondeterministic programming methods, such as metaheuristic ones) in over 500 publications [ 36 , 37 ].…”