This talk will present our projects as examples
which demonstrate the use of a "art to science"
research methodology. Our team adopts a
creative-driven interdisciplinary approach in their
fundamental and applied research on wearable
technologies and its applications to effectively
enable knowledge transfer and realize market
applications. In short, these topics are centered on
understanding the customer experience, social
needs, and the past to shape the future. Aesthetic
research method (or Design thinking research) thus
provides a creative approach try to solve complex
problems in a user-centric manner. These research
works have contributed to discoveries, inventions
and developments which have been adopted by
some of the largest textile manufacturers and
implemented into their products, deriving concrete
financial gains. Yet, we believe there is still room for
improvement and we hope we can do better in the
future.

As indicated in a famous adage "A picture isworth a thousand words, images usually conveya lot of information. It is desirable to separateinput images into multiple layers via imageperception algorithms for people to understandthem easily. In this talk, I will present our imageperception works, which benefit outdoor visionsystems, multimedia, and healthcare. First, I willpresent our innovative methods to addressadverse weather image restoration for outdoorvision systems. Then, I will talk about our works onthe multimedia era, which separates input photosinto layers with textures, shadows, lanes, saliency.and so on. Lastly, l will also describe our Al-basedhealthcare algorithms.

Using robot arms, or a collection of them, to performvarious tasks is becoming increasingly popular inboth industry and our daily life. Recent advances inmachine learning provide us with an opportunity toemploy innovative learning to reach autonomouscontrol. Most existing neural network based robotcontrol methods are designed based on statisticalcriteria with reasonable average performance, butlacks amechanismtoensureworst-caseperformance. This talk will introduce our research onconstructing neural networks based on knowledge ofthe robot model and apply it to solve robot armmanipulation problems. This approach inherits theadvantage of neural networks in adaption andsatisfies guaranteed stability. It provides a tractablesolution using neural networks to address safetycritical tasks with certified performance

Many cities worldwide are embracing electric vehicle(EV) sharing as a flexible and sustainable means ofurban transit. However, it remains challenging forthe operators to charge the fleet because of limitedor costly access to charging facilities. We focus onanswering the core question-how to charge the fleetto make EV sharing viable and profitable. Our work ismotivated by the setback that struck San DiegoCalifornia, where car rental company car2go ceasedits EV-sharing operations. We integrate charginginfrastructure planning and vehicle repositioningoperations that were often considered separatelyMore interestingly, our modeling emphasizes theoperator-controlledchargingoperationsandcustomers' EV-picking behavior. With advancedoperations research models, we find that the viabilityof EV sharing can be enhanced by concentratinglimited charger resources at selected locations. Boththe high-level planning guidelines and operationalpolicies can be useful for practitioners.

Civil aviation is one of the drivers of global growth. It increasesat around 4.4% per annum and is poised to maintain thisgrowth for the next 20 decades. On the other hand, aviation isresponsible for 5% of total anthropogenic radiative forcing andis expected to increase substantially in the future. In the face ofthe continuing expansion of air traffic, mitigation of aviation'sclimate impact becomes challenging but imperative

Mining and analysis of massive population-based shape data can
result in knowledge of shape variability of the population. such
knowledge can lead to the construction of faithful
subject-specific 3D shape models from sparse measurements,
predict shape-specific functional performance
and
population-wide structural performance variation. Such an ability
brings about unprecedented capabilities and tantalizing
opportunities for mass customization, part-specific failure
prediction and just-in-time part maintenance, and patient-specific
biomedical intervention and treatment. This research proposes a
statistical atlas based approach that incorporates statistical shape
modeling in subject-specific shape reconstruction, finite element
(FE) modeling and analysis. The statistical atlas contains three
parts: the mean shape and the variation modes of the shape
population which span a linear shape space, the FE mesh of the
mean shape (template mesh), and the selected feature points and
sizing dimensions which are obtained by maximizing the total
variance they capture of the shape population. Given a subject
(e.g. a person), the corresponding dimensions are measured and
the3D shape model is synthesized. The template mesh can be
morphed to the subject shape to conduct subject-specific fe
analysis. The FE solution on the template mesh can also be
extrapolated to the subject shape through Taylor expansion. The
shape variances along the variation modes are obtained by the
principal component analysis. These variances tell the amount of
shape variabilities along the variation modes and are combined
with the Taylor expansion of the fE solution to obtain the
structural performance variation across the population. The 2D/3D
numerical examples demonstrate the efficiency and effectiveness
of the proposed approach

Despite the ambitious plans of vehicle manufacturers, vehicletechnologies usually take two to five decades to saturate theirpotential market,and currently, SAE level-2 automation (SAEJ3016 201401) is the state-of-the-art vehicle-automationtechnology. For this level, drivers are still required to monitor theroad and get ready to intervene in a timely manner whennecessary. Thus,visually monitoring the driving environmentand anticipating how a given situation may evolve (anticipatorydriving) may still bring benefits to driving safety. A good designneeds to support drivers' anticipation of both the environmentas well as the automated system that they are interacting withand consider drivers' states when transferring control. Throughdriving simulator experiments, my research aimed to answer thefollowing questions: (1) how automation affects anticipation indriving and (2) what types of feedback can be effective to supportdriver anticipation in automated vehicles. Results show thatdriving experience still matters to driving safety in automatedvehicles, with experienced drivers exhibited more anticipatorydriving behaviors. Distraction engagement may still need to beavoided in vehicles with the current level of driving automation(SAE Level-2), as distraction engagement can impede anticipatorydriving in automated vehicles. Further, this research indicates thatdrivers in automated vehicles should be trained with proper

Rail and urban transit systems often serve a large number of passengerson a daily basis and play an important role in supporting economicactivities. This presentation will discuss how to utilize modeling anddata analysis tools to improve our understanding and performanceof rail or urban transit systems. In particular, this presentation will firstdiscus data mining and modeling techniques for quantifying mobilitypatterns and demand prediction in rail/transit systems. Then, givencertain levels of demand information, pricing and service networkdesign problems in rail/transit systems will be presented. A fewexamples on game-theoretical models and network design models forpassenger-freight integrated transit/rail system will be illustrated. Inaddition, several empirical studies that examine the impacts of raildevelopment on accessibility, economy and equity will be discussedwhere the joint effect of rail development and megalopolis policy willalso be covered. In summary, this presentation illustrates with exampleshow modeling and analysis tools can be applied to rail and urban transitsystems for better policy and decision making.

Additive manufacturing (AM) enables seamless integration of product
design and manufacturing phases, and thus offers significant
advantages over conventional manufacturing. Despite the enormous
progress in recent AM technologies, certain intractable quality issues
persist. These product defects lead to considerable rework and scrap
rates, and thus pose significant impediments for sustainability of am.
Consequently, there is a vital need to advance online methods for
defect detection in AM processes. so that incipient process anomalies
can be identified and possibly prevented at an early stage of
manufacturing. With the above focus, this talk will introduce some of
the ongoing research related to real-time in situ process monitoring
for AM performed in the Sensing and Analytics of Smart Manufacturing
Laboratory at Virginia Tech. The topics cover: applications of machine
learning for real-time process monitoring in AM; 3D point cloud based
dimensional integrity assessment for AM; and real-time sensing based
process monitoring for AM cybersecurity.

The goal of my research is to understand the molecular and geneticregulatory circuits that control cell proliferation and differentiationduring development of multicellular organisms. Using Drosophila as amodel system, my lab focuses on the following two major researchtopics in the past decade: (1) investigating the molecular and cellularmechanisms that control stem cell biology in Drosophila ovary andgut; and (2) creating and optimizing novel genetic tools and resourcesto facilitate sophisticated analyses of Drosophila genome. Thesynergistic effects between two research projects allowed us to gaininsights into the molecular mechanisms that control stem-cell renewain ovary and intestine of Drosophila. I will briefly introduce my labachievements in the past five years.