Mobile & Real-Time Electrochemical Sensing Platform
Project - Georgia Institute of Technology, August 2024 - December 2024
Mobile healthcare sensing is a critical tool for real-time health monitoring and is immensely helpful for users with
chronic health conditions. This assists the users to perform essential tests like glucose level monitoring at home
that eliminates the need for frequent lab visits. There are many off-the-shelf devices in the market that can perform
testing at home and have a good form factor that allows users to carry it easily. Most of these devices use the concept
of electrochemical sensing to perform the tests and are reliable. The downside of these devices is that they are confined
to a particular use case.
A compressed version of potentiostat is built from scratch that enables real-time health monitoring.
This version of potentiostat is not confined to a specific use-case but rather can be used in other fields like environmental
testing, food quality assessment and more. This miniature version facilitates in vivo testing and the compact design lower costs,
improves accessibility and integrate seamlessly into wearable or handheld devices.
The hardware solution has a software counterpart that collects the information processed from potentiostat through a
Bluetooth Low Energy (BLE) module. The information is received by a mobile and it allows real-time display of charts and helps
with quick analysis of the data. The application is built using a flutter framework and currently integrated with an ESP32 microcontroller
as the current hardware built does not support a dedicated Bluetooth transmission. This will be made available in the next iteration.
CMP Memory System: Advanced Cache & DRAM Design for Multicore Processors
Project - Georgia Institute of Technology, October 2024 - December 2024
Designed and implemented a hierarchical cache system, including private L1 instruction and data caches, a shared
L2 cache, and DRAM-based main memory with configurable policies like LRU and Random replacement.
Extended the simulator to support multicore processors, incorporating techniques like Static Way Partitioning
(SWP) and exploring advanced dynamic partitioning strategies for efficient cache resource management.
Evaluated the system using metrics like cache miss rates, row buffer hit rates, and weighted speedup across different
workloads and memory configurations, ensuring scalability and high performance.
Out-of-Order Pipeline Simulation
Project - Georgia Institute of Technology, September 2024 - October 2024
Designed and implemented a seven-stage out-of-order pipeline, integrating key components like Register Alias Table
(RAT) and Reorder Buffer (ROB) to support dynamic instruction scheduling and precise state management.
Developed and tested both in-order and out-of-order scheduling policies, optimizing instruction issue and execution
to enhance performance metrics such as cycles per instruction (CPI).
Extended the pipeline to a 2-wide superscalar design, enabling concurrent instruction execution while maintaining
correctness and efficiency across varied latency scenarios.
Superscalar Pipeline and Branch Prediction Simulator
Project - Georgia Institute of Technology, August 2024 - September 2024
Designed and implemented a seven-stage out-of-order pipeline, integrating key components like Register Alias Table
(RAT) and Reorder Buffer (ROB) to support dynamic instruction scheduling and precise state management.
Developed and tested both in-order and out-of-order scheduling policies, optimizing instruction issue and execution
to enhance performance metrics such as cycles per instruction (CPI).
Extended the pipeline to a 2-wide superscalar design, enabling concurrent instruction execution while maintaining
correctness and efficiency across varied latency scenarios.
Health Camp Management System (HCMS)
Project - Oracle, November 2022 - December 2023
The health condition of people living in urban and rural slums is dire due to the absence of nearby hospitals,
forcing them to rely solely on health camps. Patient data collected at these camps is currently paper-based and
consolidated manually in Excel sheets, hindering the extraction of insights such as disease outbreaks or resource
requirements for future camps. Despite support from NGOs, volunteers struggle with data consolidation and tracking
of medical equipment, severely impacting their ability to effectively serve the population. The lack of rural
statistics exacerbates the gap between the country's health infrastructure and the needs of rural communities.
To address these challenges, a simplified application will be developed for efficient collection and management of
patient data. This will enable doctors to treat individuals based on their comprehensive health history gathered during
health camps. Utilizing data insights and predictive analytics, the application aims to optimize the scheduling and
execution of health camps, as well as streamline the delivery and monitoring of medical supplies to targeted regions.
Key features of the application will include user-friendly interfaces tailored for easy adoption by field volunteers,
ensuring daily operational usability. Data security will be paramount, achieved by eliminating intermediaries from
transaction processes.
Multi Agent System & IoT based Plug and Play implementation in Smart grid
Project - Sri Sivasubramaniya Nadar College of Engineering, 2019 - 2023
Dynamic energy management using smart grid, implementing plug and play feature with Multi Agent System(MAS) and Internet of Things(IOT).
Energy Management System Using Multi Agent Systems with IoT and Machine Learning
IEEE Publication - 2023 International Conference on Circuit Power and Computing Technologies (ICCPCT), September 2023
Combining the implementation of smart grid using multi agent system and machine learning algorithms for short term load prediction together for providing an efficient energy management system to sustain high energy demands in industries and households.
Blockchain based energy transaction in microgrid
IEEE Publication - 2022 IEEE 19th India Council International Conference (INDICON), November 2022
This project aims to provide a potential solution of how blockchain facilitates cost reduction when used in micro grids. The energy is transferred from producer to consumer and the additional service charges are eliminated when using the blockchain method. This idea uses a bidding mechanism in the grid so that the producer who gives energy at a minimum cost is automatically mapped to the consumer. The smart contract is included in blockchain shared by the participants and keeps immutable transaction records.
Application of Machine Learning Algorithms for Short term Load Prediction of Smart grid
IEEE Publication - 2020 International Conference on Smart Electronics and Communication (ICOSEC), October 2020
Short term load prediction in power grid using machine learning algorithms like Decision Tree Regression, Random Forest, Neural Networks and Radial Basis Function.