We develop systematic approaches to effectively and efficiently augment artificial intelligence (AI)-guided predictive maintenance on critical assets (e.g. aviation) with existing maintenance processes carried out by humans. In this work, we investigate how AI solutions can be made more interpretable, computationally efficient, and robust by leveraging engineering- and physics-based digital twin models.

We develop theory and tools to extend the classical sense, learn/plan, and act functionalities present in traditional cyber-physical systems to social infrastructure so that community and human-centered objectives can be considered in infrastructure design and management in ways that go well beyond existing notions of traditional human-in-the-loop control.

We investigate how sensing and advanced technologies can be designed and used in responsible (i.e., privacy preserving) ways to measure human-human and human-infrastructure. Collected information is used to usher in a new era of cyber-physical-social systems aimed at creating a more equitable, innovative, and sustainable future. This work is deployed at full scale in partnership with communities to maximize the research’s impact.

Manufacturing processes have evolved in recent years to harness digital advances. In the new advanced manufacturing (AM) paradigm, human interaction is a first-class citizen, with AM making room for collaboration between humans, as well as between humans and machines. We develop cyber-physical-social digital twins for diagnosing human-induced root causes of advanced manufacturing faults, while providing real-time feedback and corrective actions to humans to improve their performance.

A key challenge along roadways and intersections is understanding the interactions between vulnerable road users for improved safety operations. We fuse two powerful sensing modalities—WiFi and cameras—on a standalone device to overcome longstanding limitations of the state-of-the-art computer vision solutions and improve pedestrian, cyclist, micromobility, and accessibility device user safety. This work enables real-time human-infrastructure communication to meet user-specific needs.

The sharing economy has revolutionized the way that people travel in urban areas to reach essential services (e.g., jobs, grocery stores, healthcare facilities). However, mobility infrastructure is not always fairly distributed across neighborhoods due to biases in the formulation of network design problems; these design problems often aim to maximize cost efficiency, which skews results in favor of larger population centers without regard to underserved areas. Building upon a routable and flexible multimodal network computational model developed by the team, this work addresses this spatial bias by optimizing the allocation of monetary mobility subsidies across communities, and optimizing mobility infrastructure investments that equitably provide access to essential goods and services.

We develop new scientific and engineering principles for the design and implementation of scalable and low-cost cyber-physical systems for health monitoring and management of railroad assets. The developed methodologies provide information that augment and inform existing decision-making practices carried out by humans in railroad maintenance processes.

We discover how the walkability of cities shapes the human experience (especially those with accessibility needs) and impacts accessibility. We develop smart city technologies and algorithms that can more accurately and widely measure, evaluate, and manage roadside and greenway infrastructure to support a safe and comfortable transit experience for a wide set of users.

Alaskan glacier mass loss during the last century outpaces any other glacierized region on Earth, and the future impacts on communities are unprecedented. We develop time-synchronizing wireless sensing networks capable of gathering in-situ measurements (e.g., timelapse images, differential GPS) in harsh glacial climates subject to stringent energy constraints. On-site measurements are used to assess biases, filter noise, and increase confidence in remote sensing products.

Preference-aware Human Spatial Behavior Modeling in Cyber-Physical-Human Systems [Accepted]
5th IFAC Workshop on Cyber-Physical & Human Systems, 2024
M Doctorarastoo, KA Flanigan et al.

Abstract: This study introduces a new approach for modeling preference-aware human spatial behavior in cyber-physical-human systems (CPHS) using Graph Neural Networks (GNN) and Reinforcement Learning (RL). Current models often overlook the causality and impact of factors influencing preferences. Our approach utilizes GNN for its advanced handling of graph-structured spatial data, capturing both physical and social environmental features and how they are perceived by humans. Integrated with RL, the model dynamically adapts to changes in the surrounding environment, improving adaptability and generalizability of simulations. We illustrate the approach in an educational conference room setting to compare student behavior simulation with and without preference inclusion. The results indicate that preference incorporation leads to significantly more realistic simulations, highlighting its potential to improve the design and control of CPHS and foster more adaptive human-system interactions.

Social Context-aware Dyadic Interaction Recognition in Cyber-physical-human Systems [Accepted]
5th IFAC Workshop on Cyber-Physical & Human Systems, 2024
C Lin and KA Flanigan

Abstract: Cyber-physical-social infrastructure systems—a subset of cyber-physical-human systems—extends traditional cyber-physical systems by integrating interactions between humans and infrastructure and control to meet human-centered objectives. This involves defining social benefits, measuring and interpreting human interactions, modeling interactions for prediction, linking outcomes to social benefits, and actuating the environment to achieve desired social outcomes. Within this feedback cycle, we propose a dataset that lays the foundation for contextualizing human interactions in a privacy-preserving way. We introduce a taxonomy for analyzing dyadic interactions, shedding light on the cultural and emotional dimensions of human behavior. We evaluate the performance of 7 human activity recognition algorithms.

State-of-the-art Review and Synthesis: A Requirement-based Roadmap for Standardized Predictive Maintenance Automation Using Digital Twin Technologies
Advanced Engineering Informatics, 2024
S Ma, KA Flanigan et al.
PDF

Abstract: Recent digital advances have popularized predictive maintenance (PMx), offering enhanced efficiency, automation, accuracy, cost savings, and independence in maintenance processes. Yet, PMx continues to face numerous limitations such as poor explainability, sample inefficiency of data-driven methods, complexity of physics-based methods, and limited generalizability and scalability of knowledge-based methods. This paper proposes leveraging Digital Twins (DTs) to address these challenges and enable automated PMx adoption on a larger scale. While DTs have the potential to be transformative, they have not yet reached the maturity needed to bridge these gaps in a standardized manner. Without a standard definition guiding this evolution, the transformation lacks a solid foundation for development. This paper provides a requirement-based roadmap to support standardized PMx automation using DT technologies. Our systematic approach comprises two primary stages. First, we methodically identify the Informational Requirements (IRs) and Functional Requirements (FRs) for PMx, which serve as a foundation from which any unified framework must emerge. Our approach to defining and using IRs and FRs as the backbone of any PMx DT is supported by the proven success of these requirements as blueprints in other areas, such as product development in the software industry. Second, we conduct a thorough literature review across various fields to assess how these IRs and FRs are currently being applied within DTs, enabling us to identify specific areas where further research is needed to support the progress and maturation of requirement-based PMx DTs.

Constructing a Routable Multimodal, Multi-cost, Time-dependent Network Model with All Emerging Mobility Options: Methodology and Case Studies
PDF
Transportation Research Part E, 2024
LK Graff, KA Flanigan et al.

Abstract: Cities aiming to improve their transportation networks are integrating emerging mobility options at a rapid pace. These modes provide commuters with greater flexibility to construct more convenient trips and reach a larger set of essential service destinations. A few open-source tools allow planners to conduct multimodal routing analysis in time-dependent networks, but they do not sufficiently capture the full set of travel mode combinations and disutility factors perceived by individual travelers. To this end, we introduce NOMAD: Network Optimization for Multimodal Accessibility Decision-making. NOMAD integrates the personal vehicle, transportation network company, carshare, public transit, personal bike, bikeshare, scooter, walking, and feeder micro-transit modes into a unified routable network model. A generalized travel cost function incorporates the following disutility factors: monetary cost, day-to-day mean travel time, (un)reliability as represented by day-to-day \nth{95} percentile travel time, crash risk, and physical discomfort. The proposed open-source tool can be used to create multimodal travel cost matrices, which may immediately serve as an input for accessibility analysis and other policy decisions related to emerging mobility options. This paper develops the network model that forms the basis of NOMAD and demonstrates four use cases in Pittsburgh, PA.

Read the Room: Inferring Social Context Through Dyadic Interaction Recognition in Cyber-physical-social Infrastructure Systems [Accepted]
ASCE International Conference on Computing in Civil Engineering, 2024
C Lin and KA Flanigan

Abstract: Cyber-physical-social infrastructure systems (CPSIS) account for human-centered interactions and benefits overlooked by traditional cyber-physical systems. This requires defining social benefits, measuring and interpreting human interactions with each other and with infrastructure in a privacy-preserving way, modeling these interactions for prediction, linking observed outcomes to social benefits, and operating and/or designing the physical environment to produce desired social out-comes. Within this feedback cycle, this paper specifically delves into recognizing dyadic human interactions using real-world data, which is the backbone to measuring and interpreting social behavior. This work addresses the existing need to enhance broader understanding of the deeper meanings and mutual responses inherent in human interactions. While RGB cameras have been informative for interaction recognition, privacy concerns arise. Depth sensors offer a privacy-conscious alternative. This study presents a taxonomy for analyzing dyadic interactions and com-pares five skeleton-based interaction recognition algorithms on a dataset of 12 dyadic interactions. Unlike monadic datasets, these interactions—categorized into communication types like emblems and affect displays—offer insights into the cultural and emotional aspects of human interactions.

GNN-based Predictive Modeling of Human Preferences in the Built Environment [Accepted]
ASCE International Conference on Computing in Civil Engineering, 2024
M Doctorarastoo, KA Flanigan, et al.

Abstract: This study introduces a novel approach to the predictive modeling of human spatial preferences in built environments leveraging Graph Neural Networks, which provide rich representation capabilities for graph-structured data, such as spatial environments. Existing models often struggle to capture the causality or the impact of the factors that influence preferences. To bridge this gap, we propose a methodology that captures the spatial, environmental, and social characteristics of the environment—structured in a graph format—to predict human spatial preferences for various activities. As a case study, the model is trained on a synthetic dataset generated to mimic real-world scenarios in a university conference room. It aims to predict the likelihood of spaces being selected for specific activities such as studying, eating, and socializing. The results demonstrate the model’s ability to incorporate multifaceted environmental and social cues into its predictions, offering insights into how preferences affect human spatial behavior.

Digital Twin Technologies in Predictive Maintenance: Enabling Transferability via Sim-to-Real and Real-to-Sim Transfer [Accepted]
ASCE International Conference on Computing in Civil Engineering, 2024
S Ma, KA Flanigan, et al.

Abstract: The advancement of the Internet of Things (IoT) and Artificial Intelligence has catalyzed the evolu-tion of Digital Twins (DTs) from conceptual ideas to more implementable realities. Yet, transition-ing from academia to industry is complex due to the absence of standardized frameworks. This paper builds upon the authors’ previously established functional and informational requirements sup-porting standardized DT development, focusing on a crucial aspect: transferability. While existing DT research primarily centers on asset transfer, the significance of “sim-to-real transfer” and “real-to-sim transfer”—transferring knowledge between simulations and real-world operations—is vital for comprehensive lifecycle management in DTs. A key challenge in this process is calibrating the “reality gap,” the discrepancy between simulated predictions and actual outcomes. Our research investigates the impact of integrating a single Reality Gap Analysis (RGA) module into an existing DT framework to effectively manage both sim-to-real and real-to-sim transfers. This integration is facilitated by data pipelines that connect the RGA module with the existing components of the DT framework, including the historical repository and the simulation model. A case study on a pedes-trian bridge at Carnegie Mellon University showcases the performance of different levels of inte-gration of our approach with an existing framework. With full implementation of an RGA module and a complete data pipeline, our approach is capable of bidirectional knowledge transfer between simulations and real-world operations without compromising efficiency.

Modeling Human Behavior in Cyber-Physical-Social Infrastructure Systems
The 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation
M Doctorarastoo, KA Flanigan et al.
PDF

Abstract: The concept of cyber-physical-social infrastructure systems (CPSIS) has been introduced to offer a paradigm shift necessary for including social and human-centered objectives in traditional cyber-physical systems (CPSs). In this paper we focus on the most challenging issue of developing CPSISs, which is modeling and predicting human behavior. Traditional approaches like rule-based models (e.g., agent-based models) and stochastic methods fall short in capturing the complex, multi-objective, and dynamic nature of human behavior. We propose the novel use of Hierarchical Imitation and Reinforcement Learning to combine the strengths of imitation learning and reinforcement learning for this purpose. Our framework offers a cost-effective and efficient approach to modeling human behavior, reducing reliance on extensive community surveys and expert opinions. The model imitates high-level human activities through imitation learning using real-world observed data and optimizes low-level actions via reinforcement learning, providing a dynamic and adaptable solution that generalizes to new, unseen environments. We demonstrate our framework through a conference room case study, illustrating its performance in modeling complex human behaviors and decision-making processes. This research advances understanding of human behavior modeling within a CPSIS framework and sets the stage to explore new avenues for social benefit.

Skeleton-based Human Action Recognition in a Thermal Comfort Context
The 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation
J Martins, KA Flanigan et al.
PDF

Abstract: Thermal comfort optimization is key to ensuring the well-being of building occupants and promoting intelligent energy use in the built environment. In the current space of thermal comfort analysis, many techniques are used ranging from more intrusive wearable sensors and qualitative occupant surveys to less intrusive infrared thermal monitoring and human action recognition (HAR). However, as these methods increase surveillance of many building occupants in often complex environments, privacy preservation and accurate analysis are essential for optimal thermal control. This paper focuses on uplifting the ability of skeleton-based HAR for use in thermal comfort-related action recognition as this method has been shown to have promising action recognition accuracy while maintaining user privacy. While we introduce this work in the context of thermal comfort, it is applicable to a wide range of areas that necessitate HAR in social, or human-centered, settings. We benchmark several fundamental deep learning models in the skeleton-based HAR space and compare their performance on a new dataset of thermal comfort-related actions.

Exploring the Potentials and Challenges of Cyber-Physical-Social Infrastructure Systems for Achieving Human-Centered Objectives
The 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation
M Doctorarastoo, KA Flanigan et al.
PDF

Abstract: For decades, cyber-physical systems (CPSs) have transformed the design and management of infrastructure systems, improving their performance through the combination of sensing, computing, and control. However, while the existing CPS paradigm supports infrastructure system actuation to achieve desired economic objectives (e.g., efficiency, safety), it fails to account for possible input-output relationships between physical system actuation and human-centered (or “social”) benefits and objectives. There are entirely unexplored social benefits to be derived from infrastructure systems that have yet to be scientifically understood and exploited. For instance, the impact of public space design and management on social interactions. To address this untapped potential, we first highlight the existing gaps and challenges to reimagining CPS theory capable of measuring, modeling, and actuating social objectives within physical infrastructure systems. We find that these challenges introduced by the integration of human-centered objectives necessitate an entirely new framework—as opposed to explicit extension of CPS—which we term cyber-physical-social infrastructure systems (CPSIS). Thus, we propose a high-level CPSIS framework to serve as a roadmap for advancements within this emerging field.

Street-Level Urban Gravity: A Quantum System Approach to Human-Centered Urban Space Design
The 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation
K. Kavee, K.A. Flanigan

Abstract: The fundamental questions of how the built environment influences why and where people meet have intrigued philosophers and researchers for centuries. This question has been amplified as we have witnessed the evolution of streets from vibrant public spaces to traffic conduits following the advent of the automobile and highway engineering practices. More than ever, infrastructure such as public open spaces and busy transportation networks can either have the capacity to support communal gatherings and provide societal benefits to society or polarize people and communities. However, there exists little ability to quantitatively capture the relationship between infrastructure and the intricacies of walkability and street-level activities. This paper sets the stage for discussions on how the design of the built environment can influence the use of public spaces (e.g., walkability), with profound implications for engineers, urban planners, and designers seeking to create human-centered infrastructure. The history of key themes such as social interactions supported by public open spaces, walkability, and the relationship between infrastructure and well-being are introduced in this paper, emphasizing their interplay within communities. The paper illustrates how traditional traffic…

Human Trajectory Estimation Using Analog Privacy-Preserving Urban Sensing Technologies
Sensors and Smart Technologies for Civil, Mechanical, and Aerospace Systems, 2023
C Lin, KA Flanigan
PDF

Abstract: In recent years, there has been considerable interest globally in “smart cities,” which aim to improve the performance of urban systems and the experiences of citizens. However, growing interest in smart cities has given rise to many underlying challenges. Society is at a critical juncture where the decisions made to integrate technologies into daily life can either help create an equitable future, or will heighten the inequitable distribution of resources, knowledge, and power in society and infringe on privacy. Nowhere is this more notable than in civil infrastructure systems (e.g., transportation, social infrastructure, the grid, buildings), which are the foundation of society, provide basic public services to communities, and play a critical role in the distribution and usage of energy, goods, and mobility resources. Underpinning the management of many of these civil infrastructure systems is the spatio-temporal tracking of humans and the measurement of human-infrastructure interaction. Already, we have witnessed countless engagements where camera-based sensing systems are designed and deployed to track humans in public spaces. While camera-based solutions often promise to anonymize data by processing video footage using automated data processing tools, many communities are resistant to trust camera-based monitoring due to infringements…

Anomaly Identification Algorithms for Indirect Structural Health Monitoring Using a Laboratory-Scale Railroad Track System
Health Monitoring of Structural and Biological Systems XVII, 2023
G Montero, J Yin, KA Flanigan et al.
PDF

Abstract: Presently, railroad monitoring strategies focus on preventative maintenance by detecting wheel anomalies using wayside detection methods (e.g., wheel-impact load detection), and direct detection of track anomalies using onboard systems (e.g., track geometry vehicles). Both approaches are periodic, manual, and do not support real-time track damage detection. Recent research has focused on detecting damage from acceleration signals obtained onboard moving vehicles and identifying anomalies from derived structural dynamic properties. Though promising due to inherent scalability and cost efficiency, its main goal is to detect damage on the supporting infrastructure and has never before been tested for detecting rail crack damage. Among other reasons, a robust anomaly detection algorithm is missing to allow the industry to embrace an automated and more cost-effective monitoring technique. In this work, we leverage a lab-scale track and moving vehicle actuation system that is scaled with the assistance of industry experts, and comprises a vehicle instrumented with two onboard vertical accelerometers. Cracked rails are simulated by introducing discontinuities (longitudinally and transversely). Several types of feature extraction and dimensionality reduction techniques are employed to evaluate their ability to separate damaged…

Scientific Evaluation of Ultrasonic Sensors for Use in Applications of Water and Hydraulic Monitoring
Sensors and Smart Technologies for Civil, Mechanical, and Aerospace Systems, 2023
Y Zhang, KA Flanigan et al.
PDF

Abstract: As easy-to-deploy, off-the-shelf sensors decrease in cost and increase in accessibility through their integration with well-documented, entry-level electronics platforms, low-quality sensors are increasingly being inappropriately used to characterize physical and natural processes under varying environmental and operational conditions. This is notably occurring across water and hydraulic system applications, which necessitates measuring water levels using ultrasonic sensors. To lay a roadmap for future water and hydraulic system monitoring research and implementation, there is a need to develop a well-informed mapping between sensors and application areas for which the use of the sensors is deemed appropriate. In this work, we identify commonly used ultrasonic sensors, develop systematic experimental setups to simulate their use in common application areas, and evaluate their accuracy under varying conditions and parameters. From the results of these experiments, we present a suggested mapping between sensors and application areas/conditions for which the use of the sensors is appropriate, in addition to limitations placed on each sensor-application pairing are identified.

Open-Source Hardware and Software for a Laboratory-Scale Track and Moving Vehicle Actuation System Used for Indirect Broken Rail Detection
Sensors and Smart Technologies for Civil, Mechanical, and Aerospace Systems, 2023
J Yin, G Montero, KA Flanigan et al.
PDF

Abstract: There is an urgent need to better understand vehicle-rail interaction dynamics to pave the way for more consistent and automated rail crack detection methodologies, as opposed to relying on periodic and manual detection via track circuits or dedicated track geometry cars. Designing an open-source hardware framework for a lab-scale rail testbed would open the doors to further data collection and analysis needed to understand the dynamic response of cracked rails. We present a framework and the corresponding open-source hardware and software (published to GitHub) for developing a laboratory-scale motorized railroad testbed, with a vehicle that is modularly tuned to the dynamics of an in-service rail car.

Distributed MPC-ILC Thermal Control Design for Large-Scale Multi-zone Building HVAC System
ACM Energy Informatics Review, 2023
W Liang, S Ma, E Cochran, KA Flanigan
PDF

Abstract: With building heating, ventilation, and air conditioning (HVAC) systems accounting for 50% the energy consumption in the building sector in the United States, there is a need to develop and implement optimal control strategies for building HVAC systems that reduce energy consumption while achieving thermal comfort for users. In the author’s previous work, an integrated model predictive control (MPC) and iterative learning control (ILC) design approach was presented that took advantage of both controllers. It did not rely on model accuracy compared to conventional MPC and reduced the learning curve compared to conventional ILC. Albeit the previous numerical results showed fast convergence in most of the VAV subsystems, the gap between room air temperature and its set point remains noticeable in several zones. This paper further proposes an approach to extend the centralized MPC-ILC controller to take into account the distributed factor and the spatial distribution of the thermal zones of the VAV system. The improved control strategy allows all VAVs to interact with each other and contribute collectively to the overall convergence of the whole system. The proposed controller is implemented on a thirty-two zone VAV reheat system and compared with different controllers, including our previous MPC-ILC design. The outcomes show that the proposed controller results in faster and closer convergence among all zones even when the number of subsystems is large.

Measuring Time-Dependent Accessibility with Emerging Mobility Options: A Generic Multi-Modal Network Modeling Framework
TRB Annual Meeting, 2023
L Graff, KA Flanigan et al.

Abstract: As cities aim to improve their holistic transportation networks, emerging mobility options are being integrated at a rapid pace. These modes provide commuters with greater flexibility to construct more convenient trips and reach a larger set of essential service destinations. However, a way to quantify their respective impacts on accessibility across time and space has not yet been introduced in a large-scale network that allows for general cross-modal trips. Moreover, most classical metrics of accessibility in single-mode networks have considered the single trip cost of travel time while also assuming a homogeneous population. To address this challenge of measuring time-dependent accessibility in a multimodal transportation network associated with a diverse set of travel costs, this paper develops a multimodal network modeling framework that accounts for five major factors across all travel modes: day-to-day average travel time, price, reliability represented by day-to-day travel time variability, safety risks, and discomfort. The generalized travel cost of the least-cost path in the multimodal network serves as a metric of accessibility, where the full set of travel modes includes personal vehicle, transportation network companies, car share, public transit, personal bike, bike share, scooter, and walking. The network design was tested with four examples, which showed how shared mobility…

Optimal Event-Based Policy for Remote Parameter Estimation in Wireless Sensing Architectures Under Resource Constraints
IEEE Transactions on Wireless Communications, 2022
KA Flanigan et al.
PDF

Abstract: Energy is a resource bottleneck in wireless sensing networks (WSNs) relying on energy harvesting for their operations. This is pronounced in WSNs whose data is used for remote parameter estimation because only a subset of the measured information can be transmitted to the estimator. While much attention has been separately paid to communication schemes for energy-aware data transmission in WSNs under resource constraints and controlled parameter estimation, there has yet to emerge a censoring policy that minimizes the variance of a measured process’ estimated component parameters subject to realistic constraints imposed by the WSN. Consequently, this paper presents the derivation of an optimal event-based policy governing data collection and transmission that accounts for energy and data buffer sizes, stochastic models of harvested energy and event arrivals, value of information of measured data, and temporal death. The policy is optimal in the sense that it maximizes the information rate of transmitted data, thereby producing the best possible estimates of the process parameters using the modified maximum likelihood estimation given the system constraints. Experimental and simulation-based results reflect these objectives and illustrate that the framework is robust against significant uncertainty in the initial parameter estimates.

Rapid-to-Deploy Wireless Water Pressure Sensors for the Assessment of Water Distribution Systems
IEEE International Smart Cities Conference
KA Admassu, KA Flanigan et al., 2022
PDF

Abstract: Drinking water systems require efficient operations to ensure high-quality water is delivered to customers with an adequate level of pressure. Municipalities adopt methods to analyze system operations to ensure they are meeting operational objectives. For example, pressure and flow sensors are commonly deployed but such sensors are spatially sparse in the distribution system. Hydraulic models are also employed to simulate the performance of drinking water systems, with operational sensor data used to refine the models for greater accuracy. However, there remains the need for a data collection method that can be easily deployed to collect measurements that densely map pressures across the drinking water system. This paper describes the design and validation of a low-cost, rapid-to-deploy wireless water pressure sensor designed to monitor water pressures in a municipal drinking water system. The sensors attached to a household hose spigot, register their GPS location, and communicate average measured pressures using a cellular modem. The system was validated using the water distribution system in Benton Harbor, Michigan to map areas suspected of low operational pressures. The study proved the ease and reliability of the monitoring system while providing reliable pressure data that was used to model the performance of the monitored system.

Integrated MPC-ILC control design for thermal control of a large-scale multi-zone building HVAC system
Proceedings of the 9th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, 2022
W Liang, S Ma, E Cochran, KA Flanigan
PDF

Abstract: With building heating, ventilation, and air conditioning (HVAC) systems accounting for 20% of the energy consumption in the United States, there is a need to develop and implement optimal control strategies for building HVAC systems that reduce energy consumption while achieving thermal comfort for users. Given its ability to handle model discrepancies and repetitive disturbances, iterative learning control (ILC) has been introduced as a feasible approach for controlling building HVAC systems. However, the majority of previous research on ILC design for building HVAC systems has been conducted either using estimated thermal parameters or building energy simulation engines. In addition, there has been limited progress toward the application of ILC on large-scale HVAC systems because the effectiveness of ILC can be intractable where a large number of zones need to be controlled. This paper proposes an integrated model predictive (MPC) and ILC control strategy for temperature regulation of large-scale variable air volume (VAV) systems based on real-world data. The use of an integrated MPC-ILC approach achieves faster convergence and traceability of room temperature setpoint, which facilitates its implementation in more complex systems, such as large-scale VAV systems. Numerical results are provided that illustrate the model identification, model validation, implementation of the proposed integrated MPC-ILC control strategy, and resulting fast convergence property of a thirty-two zone VAV reheat system with extensive measurements taken in an office building.

Functional Requirements Enabling Levels of Predictive Maintenance Automation and Autonomy
Proceedings of the IEEE 2nd International Conference on Digital Twins and Parallel Intelligence, 2022
KA Flanigan et al.

Abstract: Artificial Intelligence (AI) supporting Digital Twins (DTs) has undoubtedly changed the ways predictive maintenance (PMx) is carried out on assets by enabling processes to be increasingly automated. However, without a standard definition for such evolution, this transformation lacks a solid foundation upon which to base its development. Other fields, namely, autonomous vehicles (AVs), use standardized levels of automation to outline coherent, agreed-upon criteria for AI-driven developments supporting autonomy that minimize barriers to interdisciplinary collaboration. In this work, we draw inspiration from the autonomy levels present in AV industry and propose levels of PMx DT automation. These levels define a clear path forward for AI-driven PMx DT developments. Motivated by our understanding that standardized processes for deploying AI-driven DTs (not only for PMx) in practice must have stakeholder buy-in that requires scalability, transferability, and integration into existing processes, we explore the functional requirements that facilitate systematic approaches at each of the proposed automation and autonomy levels.

Enhancing Undergraduate Students' Sensing and Data-Informed Decision-Making Through a Smart Cities Project
ASEE Annual Conference and Exposition, 2022
J Moore, C Lin, KA Flanigan

Abstract: Smart cities promise the ability to use data to inform city planning, resource allocation, and so much more. To do so, they require capturing, processing, and interpreting data. Considerable design work is required to ensure that data captured within smart cities can actually be used to inform decisions. For smart cities and the sensed infrastructure they comprise to be as widely adopted, as current interest suggests they will be, future engineers will need to be familiar with both the design and data aspects of smart cities. Today’s engineering students will be those future engineers. Our junior-level Civil and Environmental Engineering (CEE) project course has typically included a project involving sensing and data analysis. This year, for the first time, we deployed a project that used smart cities as the context for a project requiring full-scale design, sensing, data analysis, and decision-making amid uncertainty. Importantly, while many smart cities technologies are privacy invasive, our project was done using technology that is not privacy invasive. We assessed whether the project achieved the content and skill-oriented objectives by surveying students quantitatively and qualitatively.

Levels of Predictive Maintenance Automation
Association for the Advancement of Artificial Intelligence Fall Symposium, 2022
S Ma, KA Flanigan et al.

Abstract: Artificial Intelligence (AI) is radically changing the way predictive maintenance (PMx) is performed on assets, allowing for more parts of the process to be automated. However, without a standard measure defining such evolution, the discussion around this topic is lacking a solid basis to assess and plan its development. In other fields, notably in autonomous vehicles (AVs), standardized definitions of the levels of automation have been used to chart a coherent and agreed-upon path forward, as well as to reduce barriers to interdisciplinary collaboration. Drawing inspiration from AVs, in this paper we propose levels of PMx automation, and use them to define a roadmap for AI-driven PMx developments.

Deriving the Climatic Mass Balance Gradients of Alaskan Glaciers Through the Integration of Field Measurements and Remote Sensing
American Geophysical Union Conference, 2022
A Wells, D Rounce, L Sass, C Florentine, C Lin, KA Flanigan

Abstract: Alaskan glacier mass loss during the 21st century outpaces any other glacierized region on Earth. However, the range of glacier types and climates in Alaska limits the capacity of regionally-constrained models to accurately project the fate of diverse, individual glaciers across the region. New comprehensive remote sensing datasets for Alaska glaciers provide unprecedented opportunities to quantify the spatiotemporal variation in mass loss at a glacier scale, but confidence in these products is limited due to a paucity of in-situ observations. Our approach leverages in-situ field measurements on Gulkana Glacier in the Alaska Range to assess biases, filter through noise, and gain confidence in remote sensing products, namely surface velocity, ice thickness, and elevation change estimates. We combine these remote sensing data and apply mass conservation principles to resolve the climatic mass balance (i.e., the sum of surface and internal accumulation and ablation). Using time-lapse cameras, monitored ablation stakes, and low-frequency ground-penetrating radar, we aim to produce contemporaneous velocity, elevation change, ice thickness, and surface mass balance measurements around the equilibrium-line altitude on Gulkana Glacier across a range of spatial and temporal scales. We calculate the climatic mass balance gradient via a flux-gate approach and use in-situ measurements…

Smart and Equitable Parks: Quantifying Returns on Investments Based on Probabilistic Mobility-Dependent Correlates of Park Usage Using Cyber-Physical System Technologies
Mobility21, 2022
KA Flanigan et al.

Abstract: Parks are integral to the success of any vibrant city and have long been touted as engines of economic growth that also improve public health, clean the air, manage stormwater, and enable patrons to commune with nature while enjoying a rich set of social experiences within their community. Today, 165 parks are maintained in Pittsburgh ranging from small neighborhood parks to large greenways. Unfortunately, the financial constraints of the city have challenged its ability to maintain its parks; Pittsburgh parks are underinvested in comparison to both regional and aspirational peers. A key challenge for local governments is to develop and maintain parks and other public goods in ways that equitably distribute benefits to health, well-being, livability, accessibility to essential services, and the economy. This is critical because in areas where essential services are unevenly distributed across a community, parks and greenways often lead to a bifurcation: they either serve as barriers that result in social polarization, or serve as enabling public facilities that connect citizens in under-resourced areas to their wider communities and services; the polarizing or unifying nature of parks is heavily dependent on the configuration and health of surrounding mobility services. The overarching goal of this work is to explore urban park use and correlates of use (measured by time-dependent accessibility) in order to bring to light ways…

Objective Resilience Monitoring for Railroad Systems
Objective Resilience Manual of Practice, ASCE Press, 2022
KA Flanigan et al.

Abstract: This chapter explores different issues related to the resilience monitoring of railroad systems. Profiles of some of the most important hazards considered by railroad managers and owners are described, with specific focus on the various stressors that pose a risk to asset and network operations. Available data sources that can be used to objectively quantify the components of resilience are described, along with the NIAC's four resilience components, in the context of the railroad sector. Several case studies that illustrate the quantification of resilience in various railroad applications are presented. The Principle of Resilience Monitoring and associated lemmas are applied to project a general basis for performing comprehensive, and ultimately rewarding, resilience monitoring practices as part of a resilience management program for civil infrastructure systems. A set of recommended practices is included.

Quantitatively Linking Long-Term Monitoring Data to Condition Ratings Through a Reliability-Based Framework
Structural Health Monitoring, 2021
KA Flanigan et al.

Abstract: The holy grail of structural health monitoring is the quantitative linkage between data and decisions. While structural health monitoring has shown continued growth over the past several decades, there is a persistent chasm between structural health monitoring and the ability of structure owners to make asset management decisions based on structural health monitoring data. This is in part due to the historical structural health monitoring paradigm cast as a problem of estimating structural state and detecting damage by monitoring changes in structural properties (namely, reduced stiffness). For most operational structures, deterioration does not necessarily correspond to changes in structural properties with structures operating in their elastic regimes even when deteriorated. For structures like bridges, upkeep decisions are based on federally mandated condition ratings assigned during visual inspection. Since condition ratings are widely accepted in practice, the authors propose that condition ratings serve as lower limit states (i.e. limit states below yielding) with long-term monitoring data used to quantify these lower limit states in terms of the reliability index. This article presents a method to quantify the reliability index values corresponding to the lower limit states described by existing condition ratings. Once the reliability index thresholds are established, the data-driven reliability index of the in-service asset can be monitored continuously and explicitly mapped to a condition rating at any time. As an illustrative example, the proposed framework for tracking structural performance is implemented…

Optimal Stochastic Data Collection and Transmission Policy for Self-Sustaining Structural Health Monitoring Sensing Architectures
Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems, 2021
KA Flanigan et al.

Abstract: Data acquisition methods for structural health monitoring (SHM) historically rely on schedule-based or transmit-all data collection strategies. For sensing systems that are self-sustaining (e.g. those relying on harvested energy), these data collection methods are unable to explicitly account for the availability of energy and the value of structural response data. As a result, sensing systems often fail to capture and transmit key data to end users and require an excessive amount of time to characterize statistical parameters of response data. As structural monitoring data is increasingly incorporated into decision-making processes for asset management, there is a need for an automated data collection and transmission strategy that facilitates the characterization of the statistical parameters of structural response data with minimum variance so that bridge managers can increase the frequency with which they track structural condition without compromising accuracy. This paper presents a stochastic data collection and transmission policy that minimizes the variance of estimated component parameters of a measured process subject to constraints imposed by a sensing node’s energy and data buffer sizes, stochastic models of the incoming energy and event arrivals, the value of data, and temporal death. This work then extends the optimal data collection and transmission policy to a proposed SHM application in which a standard steel pin-and-hanger assembly is monitored to track…

Probabilistic Fatigue Assessment of Monitored Railroad Bridge Components Using Long-Term Response Data in a Reliability Framework
Structural Health Monitoring, 2020
KA Flanigan et al.

Abstract: Fatigue is a primary concern for railroad bridge owners because railroad bridges typically have high live load to dead load ratios and high stress cycle frequencies. However, existing inspection and post-inspection analysis methods are unable to accurately consider the full influence of bridge behavior on the fatigue life of bridge components. Reliability-based fatigue analysis methods have emerged to account for uncertainties in analysis parameters such as environmental and mechanical properties. While existing literature proposes probabilistic fatigue assessment of bridge components, this body of work relies on train parameter estimates, finite element model simulations, or controlled loading tests to augment monitoring data. This article presents a probabilistic fatigue assessment of monitored railroad bridge components using only continuous, long-term response data in a purely data-driven reliability framework that is compatible with existing inspection methods. As an illustrative example, this work quantifies the safety profile of a fracture-critical assembly comprising of six parallel eyebars on the Harahan Bridge (Memphis, TN). The monitored eyebars are susceptible to accelerated fatigue damage because changes in the boundary conditions cause some eyebars to carry a greater proportion of the total assembly load than assumed during design and analysis; existing manual inspection practices aim to maintain an equal loading distribution across the eyebars. Consequently, the limit state function…