Duration: September 2024-January 2025
Collaborators: Prof. Debasish Chatterjee, Prof Abhir De, Souvik Das, Prashik Arun Patil
Tools/Method: Python, algorithmic game theory, multiple shooting
Description: Developed a game-theoretic framework to defend linear time-invariant control systems against stealthy adversarial attacks. Designed a stabilising controller that anticipates worst-case attacks and implemented it using an indirect multiple shooting algorithm
Duration: April 2025-Present
Collaborators: Prof. Debasish Chatterjee, Souvik Das
Tools/Method: Julia, algorithmic game theory, multiple shooting, Actuator placement
Description: Extend the PRE-TADV controller to homogeneous multi-agent systems, focusing on optimal actuator placement that minimises LQR cost while resisting worst-case adversarial attacks.
Course: Embedded Control and Robotics (SC 649), IIT Bombay
Tools: ROS, Gazebo, ViCON Motion Capture, Python
Description: Implemented an Extended Kalman Filter to estimate the 2D pose of a Turtlebot3 using trilateration from known landmarks. Simulated in Gazebo and validated with ViCON motion capture. Analysed trajectory tracking accuracy using MSE across varying initial conditions.
Course: Mathematical Structures (SC 639), IIT Bombay
Description: Modelled a binary integer program to maximise menu profit under cost, time, and demand constraints. Focused on formulation and analysis using LP relaxation and the branch-and-bound approach.
Course: Active Vibrational Control (SC 664), IIT Bombay
Tools: MATLAB-Simulink
Description: Studied and simulated a direct approach to Adaptive Feedforward Cancellation (AFC) of unknown-frequency harmonic disturbances using Amplitude-Frequency-Phase (AFP) estimation. Implemented adaptive update laws for disturbance parameter tracking and validated performance under noise and varying frequencies.