EPIC Lab

EPIC Estimation · Perception · Intelligence · Computing

Reliable intelligence for the physical world.

Mission

From uncertainty to reliable action.

At the EPIC Lab, we study how robots estimate their motion from multiple sensors, detect when that estimate becomes unreliable, and relate the resulting uncertainty to measurable task performance in changing real-world environments.

Our work connects estimation, perception, computing, and action as one physical system. The goal is not only an accurate answer, but an answer a robot can obtain, trust, and use while it moves.

  1. 01

    Sense

    Collect complementary measurements from cameras, inertial sensors, LiDAR, GNSS, and the robot itself.

  2. 02

    Estimate

    Recover motion, calibration, and uncertainty from noisy, asynchronous observations.

  3. 03

    Understand

    Connect state estimates with geometry, maps, objects, and task-relevant structure.

  4. 04

    Act

    Use what the robot knows—and what it does not know—to plan, interact, and recover.

Featured work

Systems, not isolated components.

Representative collaborative work that shaped the technical questions EPIC now pursues—from estimator design to dense maps and sensing-compute co-design.

sqrtVINS visual-inertial tracking on the Aria Everyday Activities dataset

Reliable estimation · 2025

sqrtVINS

Square-root filtering for fast, numerically stable visual-inertial motion tracking.

IEEE Transactions on Robotics (TRO)

MINS multisensor navigation with LiDAR and camera observations

Multisensor navigation · 2025

MINS

A modular navigation system that fuses inertial, camera, wheel, GNSS, and LiDAR measurements.

Journal of Field Robotics (JFR)

miniVIO runtime composition and motion constraints

Efficient perception · 2026

miniVIO

A minimalist visual-inertial odometry system built around compact motion constraints.

IEEE/RSJ Int. Conf. Intelligent Robots and Systems (IROS)

ORBCam in-sensor feature processing pipeline

Sensing and computing · 2026

ORBCam

In-sensor ORB feature processing for ultra-low-power quantized visual-inertial odometry.

IEEE/RSJ Int. Conf. Intelligent Robots and Systems (IROS)

Plane-regularized VIO on the RPNG AR Table dataset

Geometry-aware perception · 2023

Plane-Regularized VIO

Monocular visual-inertial odometry that uses planar structure to constrain motion estimation.

IEEE Int. Conf. Robotics and Automation (ICRA)

NeRF-VINS localization and dense scene representation

Dense spatial representation · 2024

NeRF-VINS

Real-time visual-inertial navigation against a neural radiance field map.

IEEE Int. Conf. Robotics and Automation (ICRA)

View selected publications

People

The EPIC Lab team.

A growing group at The George Washington University working across robotics, estimation, perception, and intelligent physical systems.

Chuchu Chen

Principal Investigator

Chuchu Chen

Assistant Professor, Mechanical & Aerospace Engineering, The George Washington University.

Ph.D. Students

Yifu (Joseph) Tian

Ph.D. Student · Incoming Fall 2026

Yifu (Joseph) Tian

B.Eng. in Electrical and Computer Engineering, The Chinese University of Hong Kong.

Ruijie (Nerako) Li

Ph.D. Student · Incoming Fall 2026

Ruijie (Nerako) Li

M.S. in Computer Science, UC San Diego; dual B.Eng. degrees.

Master's Students

Zhentao Fan

Master's Student · Since 2026

Zhentao Fan

B.Eng., Guangdong University of Technology.

Undergraduate Researchers

Lucas Tovar-Gaytan

Undergraduate Researcher · Since 2026

Lucas Tovar-Gaytan

Mechanical & Aerospace Engineering, The George Washington University.

William Wang

Undergraduate Researcher · Since 2026

William Wang

Mechanical & Aerospace Engineering, The George Washington University.

Alumni

Matthew Chapin

Alumnus · Former Undergraduate Researcher

Matthew Chapin

Mechanical Engineering, The George Washington University.

Contact

Questions about EPIC Lab?

For questions about EPIC Lab research, publications, or the group, contact Chuchu Chen by email.

The George Washington University · Washington, DC