Suyash Kushwaha
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Suyash Kushwaha has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Rohit Sharma
Indian Institute of Technology Ropar
5 shared publications - Sourajeet Roy
Department of Electronics and Communication Engineering, Indian Institute of Technology Roorkee, Roorkee, India
4 shared publications - Avirup Dasgupta
Department of Electronics and Communication Engineering, Indian Institute of Technology Roorkee, Roorkee, India
1 shared publication Chintu Bhaskara Rao
Department of Electrical Engineering, Indian Institute of Technology Ropar, Rupnagar, India
1 shared publication- Devarshi Mrinal Das
Indian Institute of Technology Ropar
1 shared publication - Domenico Spina
ELEC Department, Vrije Universiteit Brussel, Brussel, Belgium
1 shared publication
Latest Publications
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Components, Packaging and Manufacturing Technology
- Vol. 16 (3) , 638-641
Machine Learning-Based Performance Optimization of Copper–Graphene Heterointerconnects
This work introduces a model-based optimization approach for on-chip copper-graphene interconnects. The aim of this study is to optimize the signal integrity (SI) characteristics of these nano interconnects. Employing existing optimization techniques to improve the SI performance of advanced technology node copper-graphene interconnects remains challenging due to prohibitive computational costs as these interconnects require precise characterization through computationally exhaustive full-wave simulations. Our current work proposes a flexible and efficient modeling and optimization framework that improves SI compared to electromagnetic (EM) simulations using advanced, data-intensive optimization algorithms such as Genetic Algorithm (GA) and Pattern Search. We use the Gamultiobjective and Paretosearch algorithms to demonstrate the capabilities of the proposed framework and verify the accuracy of results in both single- and multiobjective scenarios with the help of two numerical examples.
- 28 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Components, Packaging and Manufacturing Technology
- Vol. 16 (3) , 638-641
Machine Learning-Based Performance Optimization of Copper–Graphene Heterointerconnects
This work introduces a model-based optimization approach for on-chip copper-graphene interconnects. The aim of this study is to optimize the signal integrity (SI) characteristics of these nano interconnects. Employing existing optimization techniques to improve the SI performance of advanced technology node copper-graphene interconnects remains challenging due to prohibitive computational costs as these interconnects require precise characterization through computationally exhaustive full-wave simulations. Our current work proposes a flexible and efficient modeling and optimization framework that improves SI compared to electromagnetic (EM) simulations using advanced, data-intensive optimization algorithms such as Genetic Algorithm (GA) and Pattern Search. We use the Gamultiobjective and Paretosearch algorithms to demonstrate the capabilities of the proposed framework and verify the accuracy of results in both single- and multiobjective scenarios with the help of two numerical examples.
- 1 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Multiscale and Multiphysics Computational Techniques
- Vol. 10, 379-387
Performance Enhanced Copper-Graphene Hetero Interconnect Structures in Crossbar Arrays for Neuromorphic Computing
In this paper, novel copper graphene heterogeneous interconnect structures are proposed which retain the ease of fabrication while having far better electrical performance when compared to the conventional copper interconnects. In the nanoscale regime, signal integrity (SI) of the copper interconnects degrades significantly. To address the signal integrity issues, these heterogeneous interconnects are developed at 7nm technology nodes which are further used to make the crossbar arrays for neuromorphic computing. The proposed copper graphene heterogeneous interconnects were designed by stacking the layers of copper and multilayer graphene nanoribbons (MLGNRs) one over the other and a detailed signal integrity analysis is done based on the quantities like the per unit length Resistance, Insertion Loss (IL), Return Loss (RL), eye diagrams, surface charge density and volume current density. The results shows that the proposed interconnects outperformed the copper interconnects based on each and every SI quantity. Finally, in the application example, the best performing heterogeneous interconnects are used to create larger crossbar arrays with sizes 64x64, 128x128. Further, the key performance matrices such as the delay time, the rise time and the fall time are analyzed and compared with the conventional crossbars made from the copper interconnects. The results in application example proved that the heterogeneous interconnects performs better than the copper interconnects for neuromorphic computing.
- 1 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Multiscale and Multiphysics Computational Techniques
- Vol. 10, 379-387
Performance Enhanced Copper-Graphene Hetero Interconnect Structures in Crossbar Arrays for Neuromorphic Computing
In this paper, novel copper graphene heterogeneous interconnect structures are proposed which retain the ease of fabrication while having far better electrical performance when compared to the conventional copper interconnects. In the nanoscale regime, signal integrity (SI) of the copper interconnects degrades significantly. To address the signal integrity issues, these heterogeneous interconnects are developed at 7nm technology nodes which are further used to make the crossbar arrays for neuromorphic computing. The proposed copper graphene heterogeneous interconnects were designed by stacking the layers of copper and multilayer graphene nanoribbons (MLGNRs) one over the other and a detailed signal integrity analysis is done based on the quantities like the per unit length Resistance, Insertion Loss (IL), Return Loss (RL), eye diagrams, surface charge density and volume current density. The results shows that the proposed interconnects outperformed the copper interconnects based on each and every SI quantity. Finally, in the application example, the best performing heterogeneous interconnects are used to create larger crossbar arrays with sizes 64x64, 128x128. Further, the key performance matrices such as the delay time, the rise time and the fall time are analyzed and compared with the conventional crossbars made from the copper interconnects. The results in application example proved that the heterogeneous interconnects performs better than the copper interconnects for neuromorphic computing.
- 22 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 14 (3) , 563-574
Design and Analysis of 3D Integrated Folded Ferro-Capacitive Crossbar Array (FC²A) for Brain-Inspired Computing System
This paper presents a novel 3D folded capacitive synaptic crossbar array designed for in-memory computing architectures. In this architecture, the bitline is folded over the wordline to enhance the synaptic density. The proposed folded capacitive crossbar array ( FC 2 A ) architecture decreases the wordline interconnect length and physical crossbar area by 50%. Thus, it helps to reduce the crossbar-associated parasitics and optimize space utilization. The proposed folded capacitive synaptic crossbar is used for designing a brain-inspired computing system (BiCoS) to recognize different patterns using CMOS technology. The BiCoS systems are prone to various reliability issues caused by the crossbar’s parasitics. Hence, the 3D folded capacitive crossbar’s Q3D model is developed to investigate the crossbar-associated parasitics and its effect on the proposed system is analyzed. The impact of crossbar parasitics is investigated for two cases: Firstly, how the three different spiking patterns (regular spiking, fast-spiking, and chattering) of the Izhikevich neuron change for the different crossbar sizes. Secondly, the impact is analyzed on the pattern recognition rate, which gets reduced to 70%. Addressing these challenges is critical to ensure the correct and robust working of the proposed system. Therefore, we propose a solution to effectively overcome and resolve these adverse effects. The energy consumed to recognize each pattern is calculated, and the average energy needed is 0.25 nJ , which is significantly less when compared to the other state-of-the-art works. The circuit is implemented using 65nm standard CMOS technology.
- 22 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 14 (3) , 563-574
Design and Analysis of 3D Integrated Folded Ferro-Capacitive Crossbar Array (FC²A) for Brain-Inspired Computing System
This paper presents a novel 3D folded capacitive synaptic crossbar array designed for in-memory computing architectures. In this architecture, the bitline is folded over the wordline to enhance the synaptic density. The proposed folded capacitive crossbar array ( FC 2 A ) architecture decreases the wordline interconnect length and physical crossbar area by 50%. Thus, it helps to reduce the crossbar-associated parasitics and optimize space utilization. The proposed folded capacitive synaptic crossbar is used for designing a brain-inspired computing system (BiCoS) to recognize different patterns using CMOS technology. The BiCoS systems are prone to various reliability issues caused by the crossbar’s parasitics. Hence, the 3D folded capacitive crossbar’s Q3D model is developed to investigate the crossbar-associated parasitics and its effect on the proposed system is analyzed. The impact of crossbar parasitics is investigated for two cases: Firstly, how the three different spiking patterns (regular spiking, fast-spiking, and chattering) of the Izhikevich neuron change for the different crossbar sizes. Secondly, the impact is analyzed on the pattern recognition rate, which gets reduced to 70%. Addressing these challenges is critical to ensure the correct and robust working of the proposed system. Therefore, we propose a solution to effectively overcome and resolve these adverse effects. The energy consumed to recognize each pattern is calculated, and the average energy needed is 0.25 nJ , which is significantly less when compared to the other state-of-the-art works. The circuit is implemented using 65nm standard CMOS technology.
- 12 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
Space Mapped Neuromodeling for Fast & Accurate Signal Integrity Analysis of Rough On-chip Copper Interconnects
In this paper, an accurate modeling of on-chip copper interconnects with surface roughness is performed considering the parametric variability. This modeling is highly accurate as per-unit-length parameters of the on-chip rough copper interconnects are extracted via full wave EM solver. Further, a space-mapped artificial neural network (ANN) is developed for accurate prediction of eye height and eye width from the geometrical and material parameters of the rough copper interconnects. The novel space-mapping ANN developed in this work is more efficient in terms of accuracy and requires fewer training samples when compared to conventional ANNs.
- 12 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
Space Mapped Neuromodeling for Fast & Accurate Signal Integrity Analysis of Rough On-chip Copper Interconnects
In this paper, an accurate modeling of on-chip copper interconnects with surface roughness is performed considering the parametric variability. This modeling is highly accurate as per-unit-length parameters of the on-chip rough copper interconnects are extracted via full wave EM solver. Further, a space-mapped artificial neural network (ANN) is developed for accurate prediction of eye height and eye width from the geometrical and material parameters of the rough copper interconnects. The novel space-mapping ANN developed in this work is more efficient in terms of accuracy and requires fewer training samples when compared to conventional ANNs.
- Open Access
- 20 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 131191-131204
Fast Multi-ANN Composite Models for Repeater Optimization in Presence of Parametric Uncertainty for on-Chip Hybrid Copper-Graphene Interconnects
In this paper, composite models are developed to predict the statistics of the optimal number and size of repeaters required to minimize the power delay product (PDP) of on-chip hybrid copper-graphene interconnects when subject to parametric uncertainty. Specifically, two distinct artificial neural network (ANN) based composite models are developed in this paper. Each composite model is made up of three individual ANNs that are interconnected. Depending on the way in which the ANNs are interconnected, the total number of full-wave electromagnetic (EM) simulations and SPICE circuit simulations required for training the composite models are reduced. Overall, the composite models enable the use of analytic expressions instead of expensive and repeated full-wave EM and SPICE simulations to solve the repeater optimization problem within a Monte Carlo framework for efficient statistical analysis.
- 20 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 131191-131204
Open AccessFast Multi-ANN Composite Models for Repeater Optimization in Presence of Parametric Uncertainty for on-Chip Hybrid Copper-Graphene Interconnects
In this paper, composite models are developed to predict the statistics of the optimal number and size of repeaters required to minimize the power delay product (PDP) of on-chip hybrid copper-graphene interconnects when subject to parametric uncertainty. Specifically, two distinct artificial neural network (ANN) based composite models are developed in this paper. Each composite model is made up of three individual ANNs that are interconnected. Depending on the way in which the ANNs are interconnected, the total number of full-wave electromagnetic (EM) simulations and SPICE circuit simulations required for training the composite models are reduced. Overall, the composite models enable the use of analytic expressions instead of expensive and repeated full-wave EM and SPICE simulations to solve the repeater optimization problem within a Monte Carlo framework for efficient statistical analysis.