Research
Embedded Files
Energy-Efficient Edge AI Architectures
Energy-Efficient Edge AI Architectures
This research focuses on developing energy-efficient edge AI architectures for resource-constrained IoT and embedded systems. I design scalable hardware-software solutions that enable real-time AI and image processing with low latency and reduced power consumption. My work includes developing near-memory CMOS VLSI architectures for rapid image edge detection using edge detection algorithms optimized through parallel processing and hardware-level optimizations and modeling memristive crossbar arrays. By integrating embedded systems, machine learning, and hardware-aware design, I aim to advance sustainable and high-performance intelligent edge computing for next-generation IoT and autonomous applications.
This research focuses on developing energy-efficient edge AI architectures for resource-constrained IoT and embedded systems. I design scalable hardware-software solutions that enable real-time AI and image processing with low latency and reduced power consumption. My work includes developing near-memory CMOS VLSI architectures for rapid image edge detection using edge detection algorithms optimized through parallel processing and hardware-level optimizations and modeling memristive crossbar arrays. By integrating embedded systems, machine learning, and hardware-aware design, I aim to advance sustainable and high-performance intelligent edge computing for next-generation IoT and autonomous applications.
Related Papers
R. Joshi, M. A. Zaman, and S. Katkoori, “Fast Sobel Edge Detection for IoT Edge Devices.” In SN Computer Science 3, no. 4 (2022): 302.
R. Joshi, M. A. Zaman, and S. Katkoori, “Novel Bit-Sliced Near-Memory Computing Based VLSI Architecture for Fast Sobel Edge Detection in IoT Edge Devices.” In 2020 IEEE International Symposium on Smart Electronic Systems (iSES)(Formerly iNiS), pp. 291-296. IEEE, 2020.
M. A. Zaman, R. Joshi, and S. Katkoori, “Early Design Space Exploration Framework for Memristive Crossbar Arrays.” In ACM Journal on Emerging Technologies in Computing Systems (JETC) 18, no. 2 (2022): 1-26.
M. A. Zaman, R. Joshi, and S. Katkoori, “High Level Modeling of Memristive Crossbar Arrays.” In 2020 IEEE Computer Society Annual Symposium on VLSI (ISVLSI), pp. 524-529. IEEE, 2020.
AI Hardware Co-Design and System Optimization
This research focuses on developing efficient hardware-aware machine learning models for resource-constrained IoT edge devices. I investigate optimization techniques that reduce latency, memory usage, and power consumption while preserving model accuracy. My work includes designing lightweight and compressed neural network architectures using hardware optimization strategies such as weight quantization, register resizing, weight sharing, and metaheuristic optimization techniques. I also validate these optimized AI models on FPGA and ASIC platforms to enable scalable, low-power, and high-performance edge intelligence for next-generation embedded and IoT systems.
Related Papers
R. Joshi, L. K. Kalyanam, and S. Katkoori, “Simulated Annealing Based Area Optimization of Multilayer Perceptron Hardware for IoT Edge Devices.” In IFIP International Internet of Things Conference, pp. 34-47. Cham: Springer Nature Switzerland, 2023.
R. Joshi, L. K. Kalyanam, and S. Katkoori, “Area Efficient VLSI ASIC Implementation of Multilayer Perceptrons.” In 2023 International VLSI Symposium on Technology, Systems and Applications (VLSI-TSA/VLSI-DAT), pp. 1-4. IEEE, 2023.
R. Joshi, L. K. Kalyanam, and S. Katkoori, “Simulated Annealing Based Integerization of Hidden Weights for Area-Efficient IoT Edge Intelligence.” In 2022 IEEE International Symposium on Smart Electronic Systems (iSES), pp. 427-432. IEEE, 2022.
L. K. Kalyanam, R. Joshi, and S. Katkoori, “Range Based Hardware Optimization of Multilayer Perceptrons with RELUs.” In 2022 IEEE International Symposium on Smart Electronic Systems (iSES), pp. 421-426. IEEE, 2022.
IoT and Intelligent Embedded Systems
This research focuses on enabling real-time, intelligent decision-making on resource-constrained edge devices. I investigate hardware-friendly machine learning approaches that optimize computational efficiency, memory usage, and energy consumption for distributed IoT environments. My work includes developing hybrid machine learning models that integrate dimensionality reduction, optimized hyperparameters, and lightweight classifiers to support efficient edge intelligence. By combining embedded systems, IoT technologies, and AI-driven optimization, I aim to design scalable and energy-efficient intelligent systems for next-generation smart and autonomous applications.
Related Papers
R. Joshi, R. S. Somesula, and S. Katkoori, “Edge-Driven Intelligence: A Hybrid Machine Learning Strategy for IoT Edge Nodes.” In IEEE International Conference on Cyber physical Systems, Power Electronics and Electric Vehicles (ICPEEV), 2023.
R. Joshi, R. S. Somesula, and S. Katkoori, “Empowering Resource-Constrained IoT Edge Devices: A Hybrid Approach for Edge Data Analysis.” In IFIP International Internet of Things Conference (IFIP-IoT), pp. 168-181. Cham: Springer Nature Switzerland, 2023.
S. Somesula, R. Joshi, and S. Katkoori, “On Feasibility of Decision Trees for Edge Intelligence in Highly Constrained Internet-of-Things (IoT).” In Proceedings of the Great Lakes Symposium on VLSI (GLSVLSI) 2023, pp. 217-218. 2023.
Accessible and Inclusive Computing Technologies
This research focuses on developing intelligent assistive systems that enhance accessibility, education, and digital inclusion for individuals with blindness and visual impairments. My work explores affordable refreshable Braille displays, tactile interfaces, and multimodal educational technologies that combine haptic feedback, audio interaction, and embedded intelligence to support accessible STEM learning. I investigate personalized and user-centered tactile display methods for Braille and graphics, while also designing educational STEM games integrated with refreshable Braille devices to improve engagement, comprehension, and learning outcomes. Through interdisciplinary research spanning embedded systems, AI, and human-centered design, I aim to create inclusive technologies that empower users and promote equitable access to education and digital experiences.
Related Papers
A. Han, R. Joshi, D. Tsivkovski, D. Ravel, M. Etezad and F. Cibrian “DotQuest: Accessible STEM Games to Support Blind and Visually Impaired Learners.” In International Society of the Learning Sciences Annual Meeting (ISLS), 2026.
R. Joshi, D. Tsivkovski, M. Etezad, and Franceli L. Cibrian “Designing Refreshable Braille STEM Games for Blind and Visually Impaired Children.” In International Conference on Ubiquitous Computing and Ambient Intelligence (UCAmI), pp. 247-259. Cham: Springer Nature Switzerland, 2025.
M. Etezad, R. Joshi, R. Alexander, and F. L. Cibrian, “3D-Printed Models for Optimizing Tactile Braille & Shape Display.” In IEEE Transactions on Haptics, 2024.
Page updated
Google Sites
Report abuse