Research

A full, always-current list is on my Google Scholar profile. Bold marks my name; * marks equal contribution.

Selected Projects

Pro-squared-Assist: a continuous, step-aware proactive assistant guiding a user through a coffee-brewing task through smart glasses.

Pro2Assist: Continuous Step-Aware Proactive Assistance with Multimodal Egocentric Perception for Long-Horizon Procedural Tasks

L. Xu, B. Yang, S. Jiang, K. Liu, Kaiyuan Hou, Y. Fan, H. Chen, Z. Yan, X. Jiang

IMWUT 2026

An egocentric assistant that continuously tracks where a user is inside a long procedural task and decides when to speak, not just what to say. It combines motion-based perception with multi-scale temporal context and expert knowledge to deliver step-aware guidance without interrupting the user mid-action.

Moth: a moth drawn to a light-based landing station, the palm-sized drone platform, and the staged precision-landing sequence.

Moth: A Low-Cost IR-Based Approach Towards Autonomous Precision Drone Landing

Y. Liu, M. Zhao, Kaiyuan Hou, J. Xia, C. Carver, S. Xia, X. Zhou, X. Jiang

ICRA 2026

Precision landing for small drones without GPS, cameras, or heavy compute. Borrowing from how moths navigate toward light, the drone homes in on an infrared "landing station" using a handful of cheap IR sensors, making autonomous docking practical on palm-sized, power-constrained platforms.

FlexiFly: Interfacing the Physical World with Foundation Models Empowered by Reconfigurable Drone Systems

M. Zhao*, J. Xia*, Kaiyuan Hou*, Y. Liu, S. Xia, X. Jiang

SenSys 2025 · related demo: MobiCom 2024 Best Demo Runner-Up

A drone platform that lets a foundation model act in the physical world. Given a natural-language task, the system plans which sensing and actuation modules it needs, reconfigures the airframe accordingly, and executes — closing the loop between language-level reasoning and physical capability.

Anemoi: a micro-drone flying a path through a room, with the reconstructed 3D airflow field drawn as arrows, and an inset of the sensorless quadrotor platform.

Anemoi: A Low-Cost Sensorless Indoor Drone System for Automatic Mapping of 3D Airflow Fields

S. Xia, M. Zhao, C. Adhivarahan, Kaiyuan Hou, Y. Chen, J. Nie, E. Wu, K. Dantu, X. Jiang

MobiCom 2023

Mapping indoor airflow normally means a grid of anemometers or a wind tunnel. Anemoi instead infers the 3D flow field from a micro-drone's own flight behavior — no airflow sensor on board at all — turning a cheap quadrotor into a mobile probe that surveys a room automatically.

ARSteth: an AR-guided intelligent stethoscope for at-home cardiac self-screening.

ARSteth: Enabling Home Self-Screening with AR-Assisted Intelligent Stethoscopes

Kaiyuan Hou, S. Xia, E. Bejerano, J. Wu, X. Jiang

IPSN 2023 · related: SenSys 2022

Cardiac auscultation requires placing a stethoscope on precise anatomical landmarks — the reason it stays in the clinic. ARSteth uses augmented reality to guide an untrained person to the right spots on their own chest, and analyzes the captured heart sounds on-device for at-home screening.

The in-situ multi-person fever screening system deployed in a building entrance.

A Low-Cost In-Situ System for Continuous Multi-Person Fever Screening

Kaiyuan Hou, Y. Liu, P. Wei, C. Yang, H. Kang, S. Xia, T. Spada, A. Rundle, X. Jiang

IPSN 2022

A low-cost thermal system that screens people continuously as they walk past, rather than one at a time at a checkpoint. Deployed in a real building, it has run for over three years and screened more than 40,000 individuals.

Publications

  1. L. Xu, B. Yang, S. Jiang, K. Liu, K. Hou, Y. Fan, H. Chen, Z. Yan, X. Jiang. Pro2Assist: Continuous Step-Aware Proactive Assistance with Multimodal Egocentric Perception for Long-Horizon Procedural Tasks. IMWUT, 2026.
  2. K. Hou, X. Di, X. Jiang. T-Gated: A Scene-Aware Framework for Occlusion-Resilient Passenger Analytics in Bus Interiors. BuildSys, 2026.
  3. Y. Liu, M. Zhao, K. Hou, J. Xia, C. Carver, S. Xia, X. Zhou, X. Jiang. Moth: A Low-cost IR-based Approach Towards Autonomous Precision Drone Landing. ICRA, 2026.
  4. M. Zhao*, J. Xia*, K. Hou*, Y. Liu, S. Xia, X. Jiang. FlexiFly: Interfacing the Physical World with Foundation Models Empowered by Reconfigurable Drone Systems. SenSys, 2025. [paper]
  5. L. Xu*, K. Hou*, X. Jiang. Exploring the Capabilities of LLMs for IMU-based Fine-grained Human Activity Understanding. FMSys, 2025.
  6. Y. Sui, Y. Zhang, Y. Liu, M. Zhao, K. Hou, J. Nie, X. Jiang, S. Xia. DomAIn: Towards Programless Smart Homes. HumanSys, 2025. Best Paper Award
  7. K. Hou*, M. Zhao*, L. Xu, Y. Fan, X. Jiang. TDBench: Benchmarking Vision-Language Models in Understanding Top-Down Images. arXiv:2504.03748, 2025. [arXiv]
  8. Y. Guo, K. Hou, H. Fu, H. Chen, Z. Yan, G. Xing, X. Jiang. Vivar: A Generative AR System for Intuitive Multi-Modal Sensor Data Presentation. arXiv:2412.13509, 2024. [arXiv]
  9. K. Hou, Y. Guo, H. Fu, H. Chen, Z. Yan, G. Xing, X. Jiang. Improving On-Device LLMs' Sensory Understanding with Embedding Interpolations. MobiCom, 2024.
  10. M. Zhao*, K. Hou*, J. Xia, S. Xia, X. Jiang. EmbodiedRDA: Connecting Foundation Models with the Physical World using Reconfigurable Drone Agents. MobiCom, 2024. Best Demo Runner-Up
  11. S. Xia, M. Zhao, C. Adhivarahan, K. Hou, Y. Chen, J. Nie, E. Wu, K. Dantu, X. Jiang. Anemoi: A Low-cost Sensorless Indoor Drone System for Automatic Mapping of 3D Airflow Fields. MobiCom, 2023. [paper]
  12. K. Hou, S. Xia, E. Bejerano, J. Wu, X. Jiang. ARSteth: Enabling Home Self-Screening with AR-Assisted Intelligent Stethoscopes. IPSN, 2023. [paper]
  13. M. Zhao, S. Xia, J. Nie, K. Hou, A. Dhupar, X. Jiang. LegoSENSE: An Open and Modular Sensing Platform for Rapidly-Deployable IoT Applications. IoTDI, 2023. [paper]
  14. K. Hou, S. Xia, J. Wu, M. Zhao, E. Bejerano, X. Jiang. AI Stethoscope for Home Self-Diagnosis with AR Guidance. SenSys, 2022.
  15. M. Zhao, Y. Liu, A. Dhupar, K. Hou, S. Xia, X. Jiang. A Modular and Reconfigurable Sensing and Actuation Platform for Smarter Environments and Drones. MobiSys, 2022.
  16. K. Hou, Y. Liu, P. Wei, C. Yang, H. Kang, S. Xia, T. Spada, A. Rundle, X. Jiang. A Low-Cost In-Situ System for Continuous Multi-Person Fever Screening. IPSN, 2022. [paper]