Clark School Home UMD

March 2018


Physically based modeling for predictive simulation of a net-zero home

Alan Uy and Raymond Adomaitis

Image: Room-level thermal modeling, HVAC model elements, water, CO2 balances, etc. ==> ~10,000 lines of Python code.

This paper—presented as a poster—won the Best Poster Award in the Computing and Systems Technology Division at the 2017 AIChE Annual Meeting held in Minneapolis in October 2017. The research represented on the poster was undertaken in the context of the 2017 Solar Decathlon, in which the University of Maryland placed second overall and first among entries from the U.S.

Annual AIChE Meeting, Minneapolis, MN,  Oct. 29–Nov. 3, 2017.  A PDF of the poster can be downloaded here.


Effect of finger geometries on strain response of interdigitated capacitor-based soft strain sensors

Hee-Sup Shin and Sarah Bergbreiter

Figure 1 from the paper: a) Schema of basic sensor architecture having four fingers with its geometric variables. b) Image of the actual sensor sample with the close-up view of the sample’s corner with its dimensions. c) 3D profile of the sample showing comb electrode thickness.

Soft sensors are important for the control of soft robotic systems. This paper describes a comprehensive and analytical capacitance model to predict the strain response of a soft elastomeric capacitive strain sensor. The model is validated using both computational methods and experiments, and important design parameters to increase sensor sensitivity are identified.

Applied Physics Letters 112, 044101 (2018)  A PDF of the paper can be downloaded here.


Data privacy for a ρ-recoverable function

Ajaykrishnan Nageswaran and Prakash Narayan

A user’s data is represented by a finite-valued random variable. Given a function of the data, a querier is required to recover, with at least a prescribed probability, the value of the function based on a query response provided by the user. The user devises the query response, subject to the recoverability requirement, so as to maximize privacy of the data from the querier. Privacy is measured by the probability of error incurred by the querier in estimating the data from the query response. We analyze single and multiple independent query responses, with each response satisfying the recoverability requirement, that provide maximum privacy to the user. Achievability schemes with explicit randomization mechanisms for query responses are given and their privacy compared with converse upper bounds.

Computing Research Repository abs/1802.07851 (2018)  A PDF of the paper can be downloaded here.


A new unbiased stochastic derivative estimator for discontinuous sample performances with structural parameters

Yijie Peng, Michael C. Fu, Jian-Qiang Hu, Bernd Heidergott

This paper proposes a new unbiased stochastic derivative estimator in a framework that can handle discontinuous sample performances with structural parameters. This work extends the three most popular unbiased stochastic derivative estimators: (1) infinitesimal perturbation analysis (IPA), (2) the likelihood ratio (LR) method, and (3) the weak derivative method, to a setting where they did not previously apply. Examples in probability constraints, control charts, and financial derivatives demonstrate the broad applicability of the proposed framework. The new estimator preserves the single-run efficiency of the classic IPA-LR estimators in applications, which is substantiated by numerical experiments.

You can also read this interesting article about the research, which recently appeared on the Smith School’s website.

Operations Research, Articles in Advance 02 Feb 2018, DOI 10.1287 opre.2017.1674  A PDF of the paper can be downloaded here.


Small networks encode decision-making in primary auditory cortex

Nikolas A. Francis, Daniel E. Winkowski, Alireza Sheikhattar, Kevin Armengol, Behtash Babadi, Patrick O. Kanold

Image: The researchers used 2-Photon imaging in mice to discover how neurons become networked during active listening for sounds. The figure shows that when the mice correctly detected sounds, neurons formed small clusters with strong links.

In this paper, the researchers find that the responses of auditory cortex neurons in mice change when the mice become actively engaged in an auditory detection task, and that subsets of neurons process task-related information in different ways. The results show that activity in auditory cortex is not solely determined by the nature of the sensory signals, but also by internal brain activity related to attention and decision-making. Thus, the auditory cortex not only encodes the acoustic qualities of sound, but also the behavioral meaning of sound and the decisions we make based on what we hear.

Neuron 97, 885–897, Feb. 21, 2018.  A PDF of the paper can be downloaded here.


Localized three-dimensional functionalization of bionanoreceptors on high-density micropillar arrays via electrowetting

Sangwook Chu, Thomas E. Winkler, Adam D. Brown, James N. Culver and Reza Ghodssi

Image: (a) An overview of the custom-built system for the electrowetting-assisted 3-D biofabrication process. (b) Cross-sectional schematics describing the electrowetting-induced structural wettability transition and the introduction of TMV1cys into the deep microcavities.

Integrating biomaterials with 3-D microdevice components offers exciting opportunities for communities developing miniature bioelectronics with enhanced performance and advanced modes of operation. However, most biological materials are stable only in properly conditioned aqueous solutions, thus the water-repellent properties exhibited by densely arranged micro/nanostructures (widely known as the Cassie−Baxter state) represent a significant challenge to biomaterial integration. This work reports the characterization of the TMV1cys assembly on Au-coated Si-μPAs having different densities and identifies structural hydrophobicity as a key limiting factor for 3-D biofunctionalization. A major breakthrough of this study was achieved by adopting electrowetting principles, which circumvented structural hydrophobicity limitations to enable the controlled patterning of 3-D assembled bionanoreceptors on high-surface-area microstructures. Results from this study demonstrate the potential of electrowetting technologies to serve as robust platforms for the biofabrication of micro/nano/biointegrated 3-D devices and systems. These findings offer new possibilities for developing advanced 3-D components beneficial for a range of microdevice applications including microenergy storage/harvesting, biochemical sensing, microthermal management, water-repellent surfaces, etc. The simple principles, readily available system components, and minimal volumes of biological solutions featured by the custom-built fabrication system suit the major criteria for widespread application of this technology.

Langmuir 2018, 34, 1725−1732.  A PDF of the paper can be downloaded here.


Stochastic optimization models for transferring delay along flight trajectories to reduce fuel usage

James C. Jones, David J. Lovell and Michael O. Ball

Figure 1 from the paper: A typical “downwind” trajectory to absorb terminal area delay.

Typically, the precise landing times of en route aircraft are not set until each aircraft approaches the airspace adjacent to its destination airport. In times of congestion, it is not unusual for air traffic controllers to subject arriving aircraft to various maneuvers such as flying in circles or in zigzag patterns to create an orderly flow of aircraft onto an arrival runway. If the arrival time was established much earlier, that the delay could be realized by having the aircraft fly slower while still at a higher altitude, burning much less fuel than the low-altitude maneuvers. This paper proposes three integer programming models for assigning delay to aircraft approaching a single airport, well in advance of each aircraft’s entry into the terminal airspace.

Transportation Science, published online in Articles in Advance, DOI 10.1287 trsc.2016.0689  A PDF of the paper can be downloaded here.


Neural source dynamics of brain responses to continuous stimuli: Speech processing from acoustics to comprehension

Christian Brodbeck, Alessandro Presacco and Jonathan Simon

Figure 2 from the paper: Stimulus coding for kernel estimation. Illustration of the first four seconds of one of the two speech stimuli. The text at the top indicates the transcript; the next four lines show the raw acoustic waveform data and the three continuously coded predictor variables. The bottom illustrates the source localized MEG data from three arbitrary source dipoles from a representative participant, averaged across the three presentations of the stimulus. The analysis modeled the brain signal at each source dipole based on the three predictor variables using convolution with a kernel of 1 second length.

This paper shows that brain responses to continuous stimuli can be investigated in detail, for magnetoencephalography (MEG) data, by combining linear kernel estimation with minimum norm source localization. This suggests new avenues for analyzing neural processing of naturalistic stimuli, without the necessity of averaging over artificially short or truncated stimuli.

Neuroimage 172 (2018) 162–174  A PDF of the paper can be downloaded here.