Skills and Projects

Top Skills

Programming Languages

  • Python – Expert
  • C++ – Advanced
  • MATLAB and Simulink – Advanced
  • ML Libraries (TensorFlow, Scikit-Learn)
  • Data Manipulation Libraries (Pandas, NumPy)
  • Distributed Computing (Apache Spark)
  • OpenCL – intermediate

Core Machine Learning Skills

  • Supervised Learning (e.g., regression, classification)
  • Unsupervised Learning (e.g., clustering, dimensionality reduction)
  • Representation Learning
  • Deep Learning (e.g., neural networks, CNNs, RNNs)
  • Fairness in Machine Learning
  • Generative Models (e.g., GANs, VAEs)
  • Information Theory – Probability and Statistics

Hardware and controllers

ARM processors, BeagleBone, Raspberry Pi, dlp light-crafter, Arduino. Intel Realsense 3D cameras.

Other skills

ROS, Gazebo, OpenCV, GIT version control, Linux Command, Intel RealSense SDK 2.0.

Projects

WoundSprayIQ Intelligent Spray Device – [Project Contributor]

  • Developed and integrated sensing components for real-time wound analysis, including visual odometry for speed estimation using dense images from Intel RealSense cameras.
  • Engineered a customized projection system with BeagleBone and DLP mini projector to enhance precision in treatment application.
  • Designed and modeled the spray-painting mechanism to ensure accurate dosing and distribution of wound care substances.
    [click here for more details].

ST A* Based Path Planning for Multi-agent With Speed Profiles Assignment – [Main Developer]

  • Designed and implemented the ST A path-planning algorithm* tailored for mobile robots in warehouse environments, allowing dynamic speed profiles with acceleration (torque) as an input parameter rather than a fixed speed.
  • Developed a simulation environment based on the Robotarium testbed, enabling robust testing for systems with hundreds of robots.
  • Engineered a parallelized, multi-threaded GPU implementation using OpenCL to enhance computational efficiency and support large-scale robotic coordination.
Figure 1: Simulation of ST A* Path Planning Algorithm: Ten robots navigate from square starting points to circular goal locations of matching colors. The algorithm enables robots to coordinate by stopping and waiting, ensuring collision-free paths.

Development of Adaptive Control Algorithms for Multi-Agent Systems Formation – [Master’s Thesis Project]

  • Adaptive Consensus Control for Linear Multi-Agent Systems with Unknown Sinusoidal Disturbances: Developed a distributed adaptive control algorithm to achieve consensus among agents with unknown sinusoidal disturbances, using LaSalle’s invariance principle to prove convergence.
  • Formation Control with Disturbance Compensation: Designed control laws for stable formation control of agents under unknown disturbances, considering both relative displacement sensing and global position awareness for specific agents.
  • Formation Control for Manipulators with Unknown Parameters: Formulated a distance-based formation control algorithm for manipulator end-effectors with unknown parameters, incorporating an extended observer for robust control against unknown dynamics.
Figure 2: Consensus control of the first state variable for four agents with unknown sinusoidal disturbances. [Click the link for more details].
Figure 3: Comparison of formation control for six mobile robots with distance compensation (top) and without distance compensation (bottom). [Click the link for more details]
Figure 4: Rigidity-based formation control of four 2R manipulators with unknown parameters, aiming to form a square shape. [Click the link for more details]

Development of a Control Algorithm for Self-Balancing One-Wheel Skateboard – [Bachelor’s Degree Graduation Project]

  • Developed a control algorithm and mathematical model for a self-balancing skateboard akin to an inverted pendulum.
  • Designed a custom Brushless DC motor driver for torque control and implemented stabilization algorithms for angle and velocity. [Click the link for details about the motor driver]
  • Validated the control model through real-world testing and simulations, incorporating user behavior as a variable torque input.
Figure 5: Key components of the self-balancing one-wheel skateboard, including the BLDC motor, 3-phase H bridge, and system controller.
Figure 6: Testing the self-balancing one-wheel skateboard developed as part of the graduation project.


Additional projects

Explore my GitHub repositories for more projects covering topics such as control systems [project 1], finite-time stability [project 2, project 3], multi-agent systems [project 4], sensorless control [project 5], switched systems [project 6], and path planning using OpenCL [project 7]. These projects demonstrate a broad range of expertise in control systems, robotics, and high-performance computing.