Advisor: Prof. Alireza Ramezani
Duration: May 2025 - May 2026
- Proposed a hierarchical motion planning and control framework for agile trajectory tracking on M4, a morphing aerial robot (6 kg) capable of transitioning between ground and aerial configurations, developing an NMPC trajectory optimizer and whole-body controller via inverse-dynamics-based modeling of multi-joint kinematics, accelerated with ACADOS and ALTRO for real-world hardware deployment.
- Demonstrated high-speed sharp turns of up to 120° in MATLAB and Simscape, showing that actively reconfiguring the robot's appendages via posture manipulation and thrust vectoring reduces peak tracking error by over 60% versus fixed-geometry drone baselines at entry speeds up to 10.5 m/s.
- Benchmarked and performance-optimized a QP solver-based real-time control loop with solve times below 0.6 ms, validating stable hover under randomly changing joint configurations.
- Performed hardware-in-the-loop integration and motor servo control on a VOXL2 + ESC embedded system via frequency-domain system identification, estimating system parameters (gain K, time constant τ, delay L) using Nonlinear Least Squares and Bode analysis, and built a thrust-to-PWM pipeline modeling nonlinearity, motor lag, and delay.
Project Duration: January 2023 - March 2024
- Worked as a research intern at the Multi-Robot Autonomy Lab at IISER Bhopal under the guidance of Dr. P. B. Sujit and Dr. Manoj Kumar Tripathi.
- Designed a trajectory optimization and motion planning algorithm for UAVs using nonlinear MPC with CasADi and IPOPT, enforcing obstacle-avoidance constraints and wind-aware path replanning in dynamic, cluttered environments.
- Trained a data-free physics-informed neural network (DeepXDE, TensorFlow backend) to solve steady-state RANS equations for millisecond-level wind-field inference, enabling 100% collision-free navigation under randomized wind disturbances while reducing control effort by 9% versus a full CFD-driven planner.
Project Duration: June 2022 - July 2022
- Developed Prota: The ROS Bot as part of the e-Yantra Summer Internship, a low-cost open-source educational autonomous ground vehicle designed to teach ROS, SLAM, and navigation from the ground up.
- Performed hardware bring-up and multi-sensor integration of LiDAR, IMU, wheel encoders, depth camera, and proximity sensors on a Raspberry Pi, developing ROS drivers for data acquisition and performing sensor calibration, fusion, and time synchronization for reliable odometry and localization in a GPS-denied environment.
- Executed real-time SLAM using particle-filter-based localization (GMapping, Hector SLAM) and the ROS move_base navigation stack with AMCL and costmap-based obstacle avoidance, validated in Gazebo/RViz simulation and then on physical hardware.