My Projects

Efficient Pose Estimation and Sensor feedback Optimization for Enhancement of Robotic Ultrasound 

Innovation Hub for Foundation of Cobotics Technology (IHFC) - Programme in Autonomous Robotics (PAR) Laboratory

Indian Institute of Technology (IIT), Delhi

Guide: Dr S K Saha, Dr. Deepak Raina 

•Developed a cutting-edge Bayesian Optimization algorithm for optimizing probe orientation in Robotic Ultrasound procedures, utilizing end-effector reaction force data as the objective function while incorporating sampling noise regularization for probe normalization to enhance acoustic coupling. 

•Collaborated with AIIMS hospital, deploying a sophisticated 7-Degree of Freedom (DOF) robotic manipulator arm and achieved an outstanding 98 percent accuracy across diverse body types

3D Tracking for Robotic Endoscopic Surgery Using Image Segmentation Maps and Heuristics Analysis for Geometric Cues in Intervals                                                                                                                              

Embedded Systems Lab | Autonomous Robotics and Multi Robot Systems Lab

Indian Institute of Technology (IIT), Bombay

Guide: Dr Leena Vachhani, Dr. Shubhangi Nema

•Developed a pioneering algorithm using Computer Vision and Heuristics to track raw 2D motion data into precise 3D motion with semantic segmentation for feature extraction followed by interval-based dimensional transformation.

•Led research using Davinci Surgical robotic video datasets as well as real-time motion capture using VICON to create real-time simulations aligned with ground truth data, achieving a remarkable 94.99 percent improvement over industry-standard methods

AeroRL: Empowering UAVs with Reinforcement Learning Control 

This project explores the application of Proximal Policy Optimization (PPO) in Reinforcement Learning (RL) for drone and manipulator control. RL enables autonomous agents to adaptively learn control policies through interactions with their environment. PPO, a robust algorithm for policy optimization, ensures stability and controlled learning, making it suitable for scenarios where safety is crucial. The mathematical foundation of PPO involves nuanced policy optimization, balancing policy improvement and stability. 

PathBot: Revolutionizing Robot Path Planning with Particle Filter Techniques 


Created a particle filter method for effective overhead terrain mapping and path planning for mobile robots with RRT algorithm to verify

DeepRacer RL Racecar

Amazon DeepRacer Student League

Utilizing state-of-the-art deep reinforcement learning techniques, our project aims to train an autonomous race-car to autonomously navigate complex racetracks by employing Proximal Policy Optimization (PPO) while adhering to the midline and track boundaries to maximize rewards

AI-StockAnalytica: An Intelligent Stock Prediction System 

Prediction using LSTMs and Deep  RL                                    



Large scale Segmentation with Clustering and PCA

Dr. Ryan Ahmed

McMaster University | Coursera

Designed an Unsupervised ML technique - Clustering combined with Principal Component Analysis for a Bank to create innovative marketing programs for various segments of customers

CFD-Driven Fluid Flow Analysis for Pump Design Optimization

Conducted with SimScale

Analysed the flow profile in a pump that powers hydraulics in automobiles and robots, for optimizing design characteristics and improving overall performance

FEM Project

Fluid Flow Analysis for a Racecar Spoiler in a Continuous Air Flow Simulation



Library Book Management System

Using SQL and Python, I built an app for a library to manage its books and enter info for new ones, I also made a face recognition system with Cpp for admins to sign in.

Intelligent CO2 Emission Levels Predicter

Using real-world data for predicting the CO2 levels in any city, for identifying the factors causing it and ways to reduce carbon footprints.

And many more in the works........