About

Hae Young Noh is a professor in the Department of Civil and Environmental Engineering at Stanford University. Her research introduced the innovative concept of “structures as sensors,” which enables physical structures such as buildings and vehicle frames to be user- and environment-aware by sensing their own structural responses, particularly vibrations. Instead of relying on additional dedicated sensors like cameras or motion detectors, her work leverages the natural vibrations caused by human activities and environmental conditions to infer information about human behaviors, environmental states, and system performance. This approach represents a paradigm shift in how structures are viewed and interact with their surroundings, transforming them from passive objects into active sensing platforms. Her work addresses the traditional view of structures as passive and unchanging, which are monitored using dense sensor networks often complicated by noise from occupants and environmental factors. Noh’s methodology utilizes this “noise” as a valuable source of information, simplifying hardware requirements and enabling long-term, practical deployment. Her research involves high-rate dynamic sensing and multi-source inferencing to analyze structural responses and extract meaningful data about users and environments. Her ultimate goal is to develop structural systems that can serve as general sensing platforms, enhancing sustainability and quality of life through smarter, more interactive infrastructure. Noh received her PhD and MS degrees in Civil and Environmental Engineering from Stanford University, along with a second MS in Electrical Engineering, and her BS in Mechanical and Aerospace Engineering from Cornell University.

Research topics

• Computer Science
• Computer Security
• Engineering
• Artificial Intelligence
• Real-time computing
• Construction engineering
• Systems engineering
• Geology
• Seismology
• Forensic engineering
• Environmental science
• Transport engineering
• Knowledge management
• Embedded system

Selected publications

• EmotionVibe: Human Emotion Recognition Through Footstep-Induced Floor Vibrations

  IEEE Transactions on Affective Computing · 2026-01-01

  articleSenior author

  Emotion recognition is critical for various applications, including the early detection of mental health disorders and emotion-based smart home systems. Previous studies utilized various sensing methods for emotion recognition, such as wearable sensors, cameras, and microphones. However, these methods are often intrusive or raise significant privacy concerns, which may reduce user acceptance for continuous, long-term deployment. This paper introduces a non-intrusive and privacy-friendly personalized emotion recognition system, EmotionVibe, which leverages footstep-induced floor vibrations for emotion recognition. The main idea of EmotionVibe is that individuals' emotional states influence their gait patterns, subsequently affecting the floor vibrations induced by their footsteps. However, there are two main research challenges: 1) the complex and indirect relationship between human emotions and footstep-induced floor vibrations and 2) the large between-person variations within the relationship between emotions and gait patterns. To address these challenges, we first empirically characterize this complex relationship and develop an emotion-sensitive feature set including gait-related and vibration-related features from footstep-induced floor vibrations. Furthermore, we personalize the emotion recognition system for each user by calculating gait similarities between the target person (i.e., the person whose emotions we aim to recognize) and those in the training dataset and assigning greater weights to training people with similar gait patterns in the loss function. We evaluated our system in human walking experiments with 20 participants, summing up to 37,001 footstep samples. EmotionVibe achieved the mean absolute error (MAE) of 1.11 and 1.07 for valence (unpleasant to pleasant) and arousal (calm to excited) score estimations, respectively, reflecting 19.0% and 25.7% error reduction compared to the baseline method (using only gait-related features without personalization).

  PublisherDOI

• Mechanics-informed machine learning: A gray-box framework for pore-pressure denoising

  Engineering Geology · 2026-03-21

  article
  PublisherDOI

• Non-contact Mass Estimation of Static Objects on Kirchhoff–Love Plates via Active Vibration Sensing

  SSRN Electronic Journal · 2026-01-01

  preprintOpen accessSenior author
  PublisherDOI

• VibraFarrow: Pig Farrowing Time Prediction Using Ambient Floor Vibrations

  2025-11-11

  articleOpen access

  Farrowing, the onset of parturition in mother pigs (i.e., sows), is a high-risk period for both the sow and her newborn piglets. Early and accurate prediction of farrowing time, along with monitoring indicators such as vital signs and pre-farrowing behaviors, enables timely assistance and can lead to lower stillbirth rates. However, existing methods have limitations: camera-based systems require constant lighting that disrupts pigs' circadian rhythms, while wearable sensors can cause discomfort to pigs and are prone to be damaged.

  PublisherOA PDFDOI

• Poster Abstract: VibraFarrow - Pig Farrowing Time Prediction Using Ambient Floor Vibrations

  2025-11-11

  article

  Farrowing, the onset of parturition in sows, is critical for both sows and piglets. Accurate prediction of farrowing time enables timely intervention to reduce stillbirths, but existing methods have limitations: cameras require constant lighting that disrupts pigs' circadian rhythms, while wearable sensors cause discomfort to animals. We present VibraFarrow, a non-intrusive system that senses floor vibrations in pig pens to predict whether a sow will farrow within 20 hours. The main challenge is the highly uncertain timing, duration, types, and patterns of pre-farrowing behaviors, which complicate feature extraction and prediction. To address this, VibraFarrow introduces Hierarchical Adaptive Window Selection (HAWS) to extract features across multiple time scales, applies unsupervised clustering to capture implicit behavior indicators, and fuses these with expert-defined features such as heart and respiration rates. We deployed VibraFarrow on a commercial farm for seven months, monitoring 18 farrowing events and 384 hours of vibration data. The system achieved a weighted F1-score of 0.735, outperforming baselines by up to 20% and demonstrating floor vibration sensing as an effective method for farrowing prediction and management.

  PublisherDOI

• Bridge Monitoring Using Existing Telecom Fiber-Optic Networks

  Research Square · 2025-09-16

  preprintOpen accessSenior author
  PublisherOA PDFDOI

• ViLA: Leveraging General-Purpose Audio for Training Vibration-Based Stadium Crowd Monitoring Models

  2025-11-11

  articleOpen access

  Crowd monitoring in sports stadiums is important to enhance public safety and improve audience experience. Existing approaches mainly rely on manual observation, cameras, and microphones, which can be disruptive and often raise privacy issues. Recently, floor vibration sensing has emerged as a less disruptive and more non-intrusive method for crowd monitoring in sports stadiums. However, because vibration-based crowd monitoring is newly developed, open-source datasets are lacking, making it challenging to develop data-driven models.

  PublisherOA PDFDOI

• Personalized Emotion Detection from Floor Vibrations Induced by Footsteps

  ArXiv.org · 2025-03-06

  preprintOpen accessSenior author

  Emotion recognition is critical for various applications such as early detection of mental health disorders and emotion based smart home systems. Previous studies used various sensing methods for emotion recognition, such as wearable sensors, cameras, and microphones. However, these methods have limitations in long term domestic, including intrusiveness and privacy concerns. To overcome these limitations, this paper introduces a nonintrusive and privacy friendly personalized emotion recognition system, EmotionVibe, which leverages footstep induced floor vibrations for emotion recognition. The main idea of EmotionVibe is that individuals' emotional states influence their gait patterns, subsequently affecting the floor vibrations induced by their footsteps. However, there are two main research challenges: 1) the complex and indirect relationship between human emotions and footstep induced floor vibrations and 2) the large between person variations within the relationship between emotions and gait patterns. To address these challenges, we first empirically characterize this complex relationship and develop an emotion sensitive feature set including gait related and vibration related features from footstep induced floor vibrations. Furthermore, we personalize the emotion recognition system for each user by calculating gait similarities between the target person (i.e., the person whose emotions we aim to recognize) and those in the training dataset and assigning greater weights to training people with similar gait patterns in the loss function. We evaluated our system in a real-world walking experiment with 20 participants, summing up to 37,001 footstep samples. EmotionVibe achieved the mean absolute error (MAE) of 1.11 and 1.07 for valence and arousal score estimations, respectively, reflecting 19.0% and 25.7% error reduction compared to the baseline method.

  PublisherOA PDFDOI

• Vehicle Speed Invariant Drive-By Bridge Damage Detection

  2025-09-09

  articleOpen accessSenior author

  Given the central role of bridges in transportation networks, continuous monitoring of these structures is crucial to detect damage, ensure long-term serviceability, and prevent catastrophic failures. Traditional inspection methods, however, are costly, laborintensive, and often subjective. Sensor-based approaches, such as strain gauges or accelerometers installed directly on bridges, require significant installation and maintenance efforts, limiting their scalability. An emerging alternative leverages vehicle–bridge interaction: by analyzing dynamic responses recorded by in-vehicle sensors, bridge conditions can be assessed without installing dedicated instrumentation on the structure. Vehicle accelerations reflect bridge-induced vibrations during crossings and indirectly encode dynamic properties of the bridge. However, variability in vehicle speed poses a significant challenge, as it affects the vibration signatures captured by accelerometers while remaining independent of bridge health. This confounding factor hinders the extraction of reliable, damage-sensitive features. To overcome this challenge, we propose a vehicle speed invariant drive-by bridge damage detection model that employs adversarial learning to extract features from vehicleinduced vibrations that are sensitive to structural damage while invariant to vehicle speed. The model integrates long short-term memory (LSTM) layers with a Gradient Reversal Layer (GRL). The LSTM layers capture temporal patterns in the frequency components, enabling the extraction of features that reflect subtle temporal variations across the full vibration spectrum. The GRL imposes speed invariance by adversarially optimizing the feature representation to maximize accuracy in damage classification while minimizing performance in speed prediction. We evaluate our method through a lab-scale experimental vehicle–bridge system with the vehicle running at varying speeds. Our model performs 1.4× better than baseline methods for bridge damage detection, achieving an accuracy of 91.01% and an F1-score of 93.29%.

  PublisherOA PDFDOI

• P-189 Oocyte SOS: can nicotinamide mononucleotide save the egg? – a systematic review and human oocyte transcriptomic analysis

  Human Reproduction · 2025-06-01

  reviewOpen access1st authorCorresponding

  Abstract Study question How does nicotinamide adenine dinucleotide (NAD+) metabolism influence oocyte quality under various reproductive conditions, particularly in relation to nicotinamide mononucleotide (NMN) supplementation? Summary answer NAD+ metabolism influences oocyte quality by improving mitochondrial function, reducing oxidative stress, and maintaining meiotic progression. NMN supplementation enhances these processes in preclinical models. What is known already Oocyte quality refers to the ability of an egg to mature, fertilise, and support early embryo development. Its decline is associated with ageing, oxidative stress, and metabolic dysfunction. NAD+ metabolism plays a crucial role in cellular energy production, DNA repair, and redox balance. NMN, a key precursor of NAD+, has been shown to increase NAD+ levels. While studies suggest that NMN supplementation enhances oocyte quality, specific mechanisms and outcomes remain underexplored. Study design, size, duration A systematic review was conducted following PRISMA guidelines. An initial search across three databases (MEDLINE, EMBASE, Scopus) yielded 486 results. After screening, seven high-quality studies were included in the final synthesis. These studies, published between 2015 and 2024, met inclusion criteria of peer-reviewed original research and included NMN in experimental interventions. The NHLBI–NIH tool was used to evaluate study quality, with included papers scoring 8.5 to 10. Participants/materials, setting, methods Preclinical studies were conducted in China (six studies) and Australia (one study). Five studies used mouse models, while two papers investigated porcine and bovine oocytes. Experimental designs involved NMN supplementation via in vitro (oocyte culture) and in vivo (oral/intraperitoneal) delivery. Control groups assessed the impact on parameters under stress conditions, including ageing, high-fat diets, and environmental toxins. Single-cell transcriptomic data from forty-six human oocytes was used to analyse differentially expressed genes. Main results and the role of chance NAD+ metabolism was shown to be essential for oocyte quality by supporting mitochondrial function, reducing oxidative stress, and maintaining meiotic progression. NMN supplementation consistently improved mitochondrial function by upregulating biogenesis genes (e.g. PGC1A, NRF1), enhancing redox balance through antioxidant enzymes (e.g. SOD1, CAT), and stabilising mitochondrial dynamics (e.g. MFN2, DRP1). NMN reduced oxidative stress markers such as reactive oxygen species, supported meiotic spindle integrity, and corrected chromosomal misalignments. Key NAD+-dependent enzymes, including sirtuins, were identified as regulators of redox balance, apoptosis, and mitochondrial health. The review highlighted variability in NMN dosing, delivery methods, and timing. Functional improvements were evidenced by enhanced oocyte maturation rates and spindle morphology. Human oocyte transcriptomic data showed differential expression in genes highlighted in the systematic review. Limitations, reasons for caution The systematic review findings are based on animal studies, which may not fully reflect human oocyte biology. Additionally, variations in NMN supplementation protocols could influence outcomes. Standardisation in future research is necessary to ensure consistency and translatability to clinical settings. Wider implications of the findings These findings underscore the potential of targeting NAD+ metabolism and using NMN supplementation to improve oocyte quality, offering promising avenues for mitigating age-related fertility decline. This research provides a foundation for developing NAD+-based strategies in assisted reproductive technologies to enhance reproductive outcomes. Trial registration number No

  PublisherOA PDFDOI

Recent grants

• CAREER: Structures as Sensors: Elder Activity Level Monitoring through Structural Vibrations

  NSF · $437k · 2020–2024

• CAREER: Structures as Sensors: Elder Activity Level Monitoring through Structural Vibrations

  NSF · $516k · 2017–2020

Frequent coauthors

• Pei Zhang

  University of Michigan–Ann Arbor

  92 shared

• Shijia Pan

  University of California, Merced

  70 shared

• Jonathon Fagert

  Baldwin–Wallace College

  53 shared

• Yiwen Dong

  53 shared

• Mostafa Mirshekari

  Stanford University

  51 shared

• Jingxiao Liu

  Stanford University

  35 shared

• Susu Xu

  Johns Hopkins University

  30 shared

• Mario Bergés

  Carnegie Mellon University

  23 shared

Labs

• Structures as Sensors LabPI

  1-2 sentence research focus

Education

• Ph.D., Civil Engineering

  Stanford University

  1997

• M.S., Civil Engineering

  University of California, Berkeley

  1992

• B.S., Civil Engineering

  University of California, Berkeley

  1990