Віктор

Viktor Hotsii​
​
Lviv • [відкрити контакти](див. вище в блоці «контактна інформація») • [відкрити контакти](див. вище в блоці «контактна інформація»)

Education
Lviv Polytechnic University​ 2022-2026
Major: AI & Robotics
Relevant Coursework: Control Theory, ML, Deep Learning, Rigid Body Dynamics, Linear Algebra, Calculus, Theory Of
Probability,

Experience

DroneProject Project Repository

The DroneProject is an ongoing robotics class project focused on creating a drone that can automatically find a landing
spot and land safely (PX4, MAVSDK, UE5, AirSim).
●​ Circle Tracking: The drone detects a moving circle on the ground and adjusts its position to stay above it.

RL GYM Reacher PPO Solution Project Repository
The PPO Reacher Solution demonstrates the application of Proximal Policy Optimization (PPO) to solve the Reacher-v4
problem in the Gymnasium environment
●​ Continuous Control: The PPO agent successfully learns to control a 2D arm to reach random targets.
●​ Stability: After 10,000 episodes, the agent exhibits smooth and stable behavior in reaching targets
●​ Target Precision: The agent not only reaches the target but also stays near it after reaching, maintaining its position
with high accuracy.

IK-Lagrangian-Solver Project Repository
TThe IK-Lagrangian-Solver uses a Lagrangian-based optimizer to solve inverse kinematics for a 7‑DOF arm while minimizing
user‑defined joint rotation costs.
●​ Cost‑Aware IK: Minimizes weighted joint deviations to find the least “expensive” configuration.
●​ Constrained/Unconstrained Modes: Optionally enforces joint limits via a barrier method and clipping, or ignores limits
entirely.
●​Gradient Descent Optimization: Iteratively updates joint angles and Lagrange multipliers until end‑effector error falls
below tolerance.

Inverted Pendulum Control Simulation Project Repository​
Simulates the control of an inverted pendulum on a cart using discrete-time Linear Quadratic Regulator (LQR) and Kalman
Filtering techniques.

●​ LQR Control Design: Implements conservative and aggressive LQR strategies for stabilizing the pendulum in upright
position.
●​ Kalman Filter Estimation: Estimates full system state from noisy measurements using a discrete-time Kalman Filter.
●​ Linearized Dynamics Model: Models the cart-pendulum system using linearized equations of motion for control
synthesis.
●​ Simulation & Visualization: Visualizes system response, estimation error, and controller eigenvalues; includes
optional animation comparing true vs estimated motion.
 Skills & Interests

Technical: C++, Python, ROS2, PX4, MAVSDK, Unreal Engine, OpenCV, Github, Linux
Language: Fluent Ukrainian and English
Interests: Robotics, Computer Graphics, LLM Agents