Included here are some of the class projects that I have conducted over my graduate studies at the University of Minnesota Twin Cities.

1. Robust Control (taught by Dale Enns)

Longitudinal Plane Missile Stabilization

Overview: A robust controller was designed for self-stabilization and disturbance rejection of a missile system with 6 degrees of freedom. The missile system is simulated using Matlab Simulink. The performance of the designed controller is analyzed in terms of nominal stability, nominal performance, robust stability, and robust performance.

2. Sensing/Estimation in Robotics (taught by Stergios Roumeliotis)

Simultaneous Localization and Mapping (SLAM) Navigation Robot

Overview: A 2D SLAM-based navigation system is developed for a Pioneer P3-DX robot that is equipped with a 2D SICK laser scanner. Hough Transform is used to extract line and landmark features from the environment. The extracted landmarks are then used for SLAM through the usage of Extended Kalman Filter and Structural Compass.

3. Advanced Control System Design (taught by Perry Li)

Control Methods for the "Papi Rubber" Game

Overview: Three different controller approaches are used to solve the "Papi Rubber" Game. First, two open-loop controllers based on least norm control and input shaping respectively are developed and used to solve the "Papi Rubber" Game. Next, a closed-loop controller is designed using a pole placement approach, where a least-square estimator is designed to estimate the system's states. The "Papi Rubber" game is simulated using Matlab Simulink with Gaussian process noise and Gaussian measurement noise.

4. Convex Optimization in Control (taught by Pete Seiler)

Robust Filtering for Systems with Parameter Uncertainties

Overview: The robustness performance of a traditional Kalman Filter is analyzed and compared to a robust filter designed using a semi-definite programming approach. A 1D cart displacement problem subjected to mass uncertainties, air drag, and rolling resistance is used as a case study.

5. Navigation and Guidance System (taught by Demoz Gebre-Egizabher)

Integration of Inertial Navigation System (INS) with LiDAR Based Localization

Overview: Sensor fusion between Inertial Navigation System (INS) and Light Detection and Ranging (LiDAR) scanner. The INS is integrated with a LiDAR-based localization using an Extended Kalman Filter. The proposed localization method was applied to real driving data taken from the KITTI dataset and can achieve more accurate state estimates of the vehicle with an estimation of the INS's error model.

6. Advanced Control System Design with Applications to Smart Vehicles (taught by Rajesh Rajamani)

Path Planning with Collision Avoidance for Highway Operations with Obstacle Tracking

Overview: A path planning algorithm with collision avoidance is developed using a modified rapidly exploring random trees (RRT*) and Linear Quadratic Regulator (LQR). The developed LQR-RRT* algorithm uses LQR to generate an optimum path connecting the proposed vertices from the RRT* algorithm and to evaluate the cost-to-go of these paths. The LQR-RRT* algorithm is evaluated on two simulation scenarios and was able to plan a path to safely overtake other moving vehicles in the scene using a lane-change maneuver.

7. Artificial Intelligence II (taught by Paul R. Schrater)

Autonomous Wildfire Surveillance using Reinforcement Learning

Overview: A reinforcement learning agent is trained using Q-learning to control a fixed-wing aircraft for autonomous wildfire surveillance.