Haoyang He

Hi! My name is 何昊洋 (hé hào yáng). I’m pursuing Masters in Robotic Systems Development (MRSD) in the Robotics Institute of Carnegie Mellon University School of Computer Science. My ongoing capstone project is centered around building a safe human-interactive autonomous wheelchair system for outdoor urban navigation, working with Cherie Ho and Sebastian Scherer at the AirLab.

Previously, I obtained my Integrated M.S./B.S. in Electrical and Computer Engineering here at CMU within 4 years, during which I interned at Apple and Bilibili.

My research interests include foundation models for robotics, multimodal embodied robot learning, and human-robot interaction, centering around enabling real-time human-interactive robotics applications with scene understanding and reasoning capabilities of multimodal foundation models through adaptation and distillation.

Recent Projects

RayFronts: Open-Set Semantic Ray Frontiers for Online Scene Understanding and Exploration

Omar Alama, Avigyan Bhattacharya, Haoyang He, Seungchan Kim, Yuheng Qiu, Wenshan Wang, Cherie Ho, Nikhil Keetha, and Sebastian Scherer

Submitted to IROS 2025, Apr 2025

Abs arXiv Video Website

Open-set semantic mapping is crucial for open-world robots. Current mapping approaches either are limited by the depth range or only map beyond-range entities in constrained settings, where overall they fail to combine within-range and beyond-range observations. Furthermore, these methods make a trade-off between fine-grained semantics and efficiency. We introduce RayFronts, a unified representation that enables both dense and beyond-range efficient semantic mapping. RayFronts encodes task-agnostic open-set semantics to both in-range voxels and beyond-range rays encoded at map boundaries, empowering the robot to reduce search volumes significantly and make informed decisions both within & beyond sensory range, while running at 8.84 Hz on an Orin AGX. Benchmarking the within-range semantics shows that RayFronts’s fine-grained image encoding provides 1.34x zero-shot 3D semantic segmentation performance while improving throughput by 16.5x. Traditionally, online mapping performance is entangled with other system components, complicating evaluation. We propose a planner-agnostic evaluation framework that captures the utility for online beyond-range search and exploration, and show RayFronts reduces search volume 2.2x more efficiently than the closest online baselines.