Projects | Jongann Lee

Published papers

SIEVE: Overcoming Topological Obstruction in Equivariant Self-Supervised Learning

NeurIPS 2026 · Published paper. SE(3)-invariant point cloud embeddings learned through equivariant reconstruction and spherical projection masking.

Navigating in Uncertain Environments with Heterogeneous Visibility

IROS 2026 · Published paper. Planning under uncertain connectivity by balancing traversal cost against the value of observing blocked paths.

Other projects

Autonomous vehicle digital twin

A photorealistic Highbay digital twin combining Gaussian splatting, Gazebo vehicle dynamics, and obstacle-aware navigation.

MAESTRO: Agentic RAG for supply-chain optimization

A multi-expert chatbot combining document retrieval, spreadsheet analysis, and clear explanations of supply-chain optimization.

PIEJAM(Passive Inversion and Eversion Joint Ankle Mechanism)

PIEJAM is an inversion and eversion ankle joint mechanism that allows the user to perform ankle inversion and eversion while using an ankle exoskeleton.

Reinforcement learning for attitude-controller tuning

A reinforcement learning algorithm designed to tune the attitude controller gains for the geometric tracking controller for the quadrotor.

Collective pitch-tilting multirotor control

Geometric trajectory tracking, differential flatness, and Lyapunov stability analysis for an unconventional multirotor.

Unicycle Control using Disturbance Observer

Augmented an existing PD controller for a unicycle by adding a disturbance observer inner loop, improving the performance by reducing the overshoot. Implemented and tested the controller in Matlab and Simulink, confirming the performance improvement.

Autonomous Quadrotor System for Payload Delivery

Created an autonomous quadrotor capable of detecting and avoiding obstacles, automatically landing on a designated landing area, detecting a pre‑determined delivery point and delivering a payload to it.