In Edge AI applications, the integration of MEMS-based sensors into System-on-Chip (SoC) designs presents a promising avenue for enhancing efficiency and performance. This research addresses the pressing need for optimized SoC designs tailored to the unique requirements of Edge AI, aiming to overcome existing challenges in area utilization. The current landscape lacks a comprehensive solution that seamlessly integrates MEMS sensors and employs advanced optimization techniques for SoC area efficiency. The research begins by delving into the intricacies of Edge AI applications and the pivotal role played by MEMS-based sensors. It identifies a critical gap in existing SoC designs, where the full potential of MEMS technology remains underutilized due to suboptimal area allocation. The overarching problem addressed in this study is the lack of a systematic approach to optimize SoC area for Edge AI applications, hindering the realization of compact and efficient devices. To bridge this gap, a novel methodology is proposed, leveraging the power of Deep Evolutionary Algorithms (DEA) for SoC area optimization. The DEAs are tailored to adapt and evolve the architecture based on the specific requirements of Edge AI tasks, ensuring an optimal allocation of resources. The methodology integrates seamlessly with MEMS-based sensors, ensuring a symbiotic relationship between hardware and sensor technologies. Results from extensive simulations and benchmarks demonstrate the efficacy of the proposed methodology, showcasing significant improvements in SoC area utilization for Edge AI applications. The optimized designs exhibit enhanced performance metrics, validating the effectiveness of the Deep Evolutionary Algorithm in tailoring SoC architectures to the unique demands of Edge AI.
M. Ramya Devi1, I. Jasmine Selvakumari Jeya2, G. Sakthi3, B. Senthilnathan4
Hindusthan College of Engineering and Technology, India1, Vellore Institute of Technology, Bhopal, India2, Galgotia University, India3, Jansons Institute of Technology, India4
Edge AI, MEMS-based sensors, System-on-Chip (SoC), Deep Evolutionary Algorithm (DEA), Area Optimization
| January | February | March | April | May | June | July | August | September | October | November | December |
| 16 | 2 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Published By :
ICTACT
Published In :
ICTACT Journal on Microelectronics
( Volume: 9 , Issue: 3 , Pages: 1646 - 1651 )
Date of Publication :
October 2023
Page Views :
260
Full Text Views :
41
|