MS in Robotics Student, University of Minnesota - Twin Cities
BS in Robotics and Mechatronics, Nazarbayev University

Projects

Tactile Surface Exploration with UR5 GitHub

May 2026 – Present | Tactile Robotics Laboratory | Astana, Kazakhstan

Tools & Technologies: UR5, URScript, Python, trimesh, NumPy, SciPy (ICP), pandas, Matplotlib, Docker (URSim), ATI Nano17 force sensor, NI-DAQ

Built an end-to-end pipeline for automated tactile exploration of a silicone cone with a UR5 collaborative arm, from CAD-derived surface geometry to executed touch sequences with synchronized contact-force recording on both simulated and physical robots.
Sampled 3,000 surface points with normals from the cone STL using trimesh, then calibrated to the robot base frame with an axis-constrained, translation-only ICP fit (cone axis pinned vertical to avoid a spurious tilt from apex-clustered touch points), reaching ~2.2 mm RMS alignment error.
Generated full-surface and 15-strip vertical touch-pose patterns with height-scaled orientation tilt (5°–15°) for tool clearance, then executed them via a custom Python UR5 IK solver (needed because the controller's single-seed IK fails to converge for poses spread around the cone), streaming URScript approach-press-retract sequences over a raw TCP socket.
Synchronized ATI Nano17 force readings with the UR5's real-time TCP pose stream via NI-DAQ, auto-detecting each press and logging its peak force and pose for offline analysis.

March 2026 | ME 5286 – Robotics, University of Minnesota | Minneapolis, MN

Tools & Technologies: UR5, RoboDK, Python, Cartesian & joint-space motion planning, pneumatic clamping

Programmed a UR5 collaborative robot via the RoboDK Python API to autonomously assemble a three-part flashlight (head, battery, endcap) without human intervention.
Defined 15 Cartesian targets for pick-and-place, insertion, and fastening operations; used joint-space motion for fast transits and Cartesian linear motion near contact points to prevent collisions.
Implemented a multi-stroke pre-threading routine followed by torque-controlled final tightening, coordinated with a pneumatic chuck and optical sensor for stable part fixturing.
Completed the full assembly sequence in 1 minute 39 seconds, within the 110-second requirement.

Fall 2025 | CSCI 5551 / EE 5271 Final Project, University of Minnesota | Minneapolis, MN

Tools & Technologies: TurtleBot3 Burger, ROS2 Humble, LiDAR, Raspberry Pi camera, Gazebo, RViz, Python

Developed a LiDAR-based obstacle avoidance pipeline on TurtleBot3 Burger using two ROS2 nodes: obstacle_detector (monitors the front LiDAR sector and publishes Boolean obstacle topics) and obstacle_avoidance (overrides velocity commands to turn and back away when an obstacle is within 30 cm).
Integrated a Raspberry Pi camera onto TurtleBot3 Burger, a non-standard hardware addition, calibrated its intrinsic parameters using a ROS2 checkerboard calibration package, and added a static TF frame to correctly represent the camera pose relative to the robot base link.
Implemented ArUco marker tracking and following: the aruco_follower node fuses camera-derived heading to the marker with LiDAR range measurements to drive the robot to a target distance while stopping when the marker leaves the camera's field of view.
Validated both algorithms first in Gazebo simulation, then transferred them to the real robot with LiDAR noise filtering adjustments for reliable real-world performance.

September 2023 – April 2025 | Nazarbayev University & CEMRR | Astana, Kazakhstan

Tools & Technologies: CAD, 3D printing, four-bar linkage design, cable-driven parallel mechanisms, mechanical prototyping

Designed and built a 5-DOF hybrid robotic shoulder exoskeleton, a 2-DOF rotational four-bar linkage combined with a 3-DOF cable-driven parallel mechanism, enabling full-range shoulder rotations while preserving alignment with the human glenohumeral joint center of rotation.
Developed 3D-printed miniature prototype models to validate human-robot mechanism coupling before scaling up to the full device.
Completed a fully assembled, wearable exoskeleton targeting intensive task-specific exercises for stroke survivors, bridging the gap between mechanical design and clinical rehabilitation requirements.
Work originated as a BS graduation project at Nazarbayev University and was extended into a fully functional device at the Center of Excellence in Medical Robotics & Research (CEMRR).

January 2023 – September 2023 | Nazarbayev University | Astana, Kazakhstan

Tools & Technologies: SolidWorks, MATLAB, four-bar linkage kinematics, cable-driven parallel mechanisms, 3D printing

Designed a 5-DOF miniature shoulder coupling mechanism combining a 2-DOF crank-rocker four-bar linkage for the shoulder girdle with a 3-DOF cable-driven parallel mechanism for the glenohumeral joint, modeling the shoulder's three rotational and two translational degrees of freedom.
Derived forward and inverse kinematics for both the four-bar linkage and the cable-driven mechanism, then validated the equations against a SolidWorks CAD model in MATLAB, matching cable lengths and joint angles to within 1–2 mm and a few degrees.
Designed a pretension mechanism to eliminate cable slacking and derailment, and built an adjustable 3D-printed prototype with reconfigurable cable connection points for testing different coupling strategies.
This foundational prototype later informed the full-scale 5-DOF exoskeleton developed at CEMRR.