Photo of Helen Oleynikova

Helen Oleynikova!

Senior Robotics Researcher

Projects & Demos

Instead of reading what I do, watch some videos! A semi-random selection of project videos and demos.

HERON: Autonomous Road Maintenance

Final pilot demonstration for the EU project HERON, integrating autonomous surveying, planning, and execution for road maintenance tasks.

nvblox: Online 3D Mapping for Navigation

Demonstration of nvblox running in real time on a warehouse mobile robot, generating obstacle-aware 3D maps for online navigation and safe motion planning in cluttered industrial environments.

CUROBO: GPU-Accelerated Manipulation

CUROBO is a massively parallel, GPU-accelerated manipulation planning library for high-speed trajectory generation. It integrates nvblox for fast 3D scene reconstruction and collision checking, enabling real-time manipulation in dynamic environments.

Mixed Reality Robotics: Bring me a Beer

Work from the Microsoft Mixed Reality & AI Lab on human–robot interaction using mixed reality. A HoloLens and a mobile robot are co-localized so you can interact with the robot in MR and send a TurtleBot to bring you a beer in the real world.

COIN-LIO: Robust LiDAR-Inertial Odometry in Degenerate Environments

COIN-LIO fuses LiDAR intensity, geometry, and IMU data to deliver odometry that remains reliable even in tunnels, flat fields, and other geometrically degenerate environments. The demo showcases stable localization under conditions where classic LiDAR-inertial methods fail.

Voxgraph: Globally-Consistent Volumetric Mapping

Voxgraph builds large-scale, dense 3D maps via overlapping signed-distance-function submaps, aligned via pose-graph optimization — enabling globally consistent volumetric mapping on CPU, even for resource-constrained robots. The demo showcases real-time construction of a globally consistent volumetric map.

(PhD Thesis) Full Open-Source, On-Board System for Drone Mapping and Planning

A high-level summary of my PhD research on volumetric reconstruction, local planning, and safe autonomous navigation for micro-aerial vehicles. Covers voxel-based mapping (Voxblox), local and global path planning, and real-world deployment of dense mapping and planning pipelines in GPS-denied, cluttered environments.

Visual–Inertial Navigation & Thermal Submapping for Search-and-Rescue

A prototype system combining visual–inertial odometry, dense 3D mapping, and thermal submapping for realistic search-and-rescue. The robot autonomously explores tight, cluttered spaces, reconstructs the environment in real time, and highlights heat signatures of potential victims using an integrated thermal camera.