Clark School Home UMD

July 2016

Automated Planning and Acting (book)

Malik Ghallab, Dana Nau and Paulo Traverso

Autonomous AI systems need complex computational techniques for planning and performing actions. Planning and acting require significant deliberation because an intelligent system must coordinate and integrate these activities in order to act effectively in the real world. This book presents a comprehensive paradigm of planning and acting using the most recent and advanced automated-planning techniques. It explains the computational deliberation capabilities that allow an actor, whether physical or virtual, to reason about its actions, choose them, organize them purposefully, and act deliberately to achieve an objective. Useful for students, practitioners, and researchers, this book covers state-of-the-art planning techniques, acting techniques, and their integration which will allow readers to design intelligent systems that are able to act effectively in the real world.

The book covers a wide range of different techniques, their pros and cons, and how they can be used effectively in different real-world applications. It presents state-of-the-art techniques for planning, and how to integrate these techniques with acting to design intelligent systems.

Cambridge University Press, August 2016. View more information about the book here.


Risk-based path planning optimization methods for UAVs over inhabited areas

Eliot Rudnick-Cohen, Jeffrey W. Herrmann, and Shapour Azarm

Examples of the solutions generated by the network approach with the 40x16 grid (solid line) and the non-network approach with 20 waypoints (dashed line) for the College Park case.

Operating unmanned aerial vehicles (UAVs) over inhabited areas requires mitigating the risk to persons on the ground. Because the risk depends upon the flight path, UAV operators need approaches that can find low-risk flight paths between the mission's start and finish points. Because the flight paths with the lowest risk could be excessively long and indirect, UAV operators are concerned about the tradeoff between risk and flight time. This paper presents a risk assessment technique and bi-objective optimization methods to find low-risk and time (flight path) solutions and computational experiments to evaluate the relative performance of the methods (their computation time and solution quality). The methods were a network optimization approach that constructed a graph for the problem and used that to generate initial solutions that were then improved by a local approach and a greedy approach and a fourth method that did not use the network solutions. The approaches that improved the solutions generated by the network optimization step performed better than the optimization approach that did not use the network solutions.

The paper presents a novel approach for finding safe, efficient unmanned aerial systems paths.

Journal of Computing and Information Science in Engineering, JCISE-15-1185  A PDF of the article can be downloaded here.


Robotic Fish: Flow-relative control behaviors using distributed flow sensing

F. Zhang, F. D. Lagor, H. Lei, X. Tan, and D. A. Paley

Speed control experimental testbed, front view.

Over millions of years of evolution, fish have developed a flow-sensing system to detect the surrounding fluid motion, which consists of hundreds of receptor organs distributed on – and under – the skin. Flow sensing serves an important role in swimming behaviors such as rheotaxis (orientation into or against the flow direction), station holding, predation, and schooling. Advanced underwater vehicles that are biologically inspired attract scientific attention because of their potential for energy efficiency and maneuverability. A flow-sensing capability enables robotic fish to navigate in unknown, murky, and cluttered environments. To demonstrate bio-inspired flow sensing and control using distributed pressure and velocity sensors, a rigid airfoil-shaped robotic fish and a flexible, self-propelled robotic fish have been developed at the University of Maryland. The robots are capable of rheotaxis, station holding, and speed control using a recursive Bayesian algorithm to assimilate measurements of the flow. A closed-loop control strategy that comprises feedback and feedforward designs has been validated in experiments.

This work includes contributions by Xiaobo Tan (EE Ph.D. 2002), a professor of electrical and computer engineering and mechanical engineering at Michigan State University.

ASME Dynamic Systems and Control Magazine, March 2016  A PDF of the article can be downloaded here.


The forecasting and impact of the loss of critical human skills necessary for supporting legacy systems

Peter Sandborn and Varun J. Prabhakar

Annual cost to support a legacy system with and without the inclusion of skills obsolescence (H = 0). For this figure, w = 0 (no cost of money), therefore, the results are actual dollars.

The loss of critical human skills that are either non-replenishable or take very long periods of time to reconstitute, impacts the support of legacy systems ranging from infrastructure, military and aerospace to IT. Many legacy systems must be supported for long periods of time because they are prohibitively expensive to replace. Loss of critical human skills is a problem for legacy system support organizations as they try to understand and mitigate the effects of an aging workforce with highly specialized, low-demand skill sets. The existing research focuses on workers that have skills that are obsolete and therefore need to be retrained to remain employable; alternatively this paper addresses the system support impacts due to the lack of workers with the required skill set. This paper develops a model for forecasting the loss of critical human skills and the impact of that loss on the future cost of system support. The model can be used to support business cases for system replacement. A detailed case study of a legacy control system from the chemical manufacturing industry is provided and managerial insights associated with the support of the system drawn.

This is the first paper that addresses the loss of the irreplaceable human skills necessary to support critical legacy systems.

IEEE Transactions on Engineering Management, August 2015  A PDF of the article can be downloaded here.


Measured changes in limbal strain during simulated sleep in face down position using an instrumented contact lens in healthy adults and adults with glaucoma

Alison Flatau, PhD; Francisco Solano, MD; Sana Idrees, MD; Joan L. Jefferys, ScM; Peter Volpe, MS; Christopher Damion; Harry A. Quigley, MD

Images shown to study subjects to instruct them on the positions to assume during the study corresponding to: (a) the lateral decubitus position; and (b) the face down position.

Does limbal strain differ between control eyes and glaucoma eyes when in simulated face down sleep positions with one side of the face against a pillow? This case-control study showed a significant increase in limbal strain in patients with glaucoma, not in controls, especially those with past visual field worsening. Mean face down strain increase among participants with glaucoma was equivalent to that expected from an increase in intraocular pressure of 2.5±1.1 mmHg. Face contact with a pillow during simulated sleep produced a sustained strain increase in eyes with glaucoma, particularly those with past progressive visual field loss.

The paper received a favorable review in the International Glaucoma Review, known for providing quarterly post-publication assessments of important papers: “This is an original study on a clinically-relevant topic and may result in practical applications.”

JAMA Opthamology,  doi: 10.1001/jamaophthalmol.2015.5667, January 21, 2016,  A PDF of the article can be downloaded here.


A surface acoustic wave biofilm sensor integrated with a treatment method based on the bioelectric effect

Young Wook Kim, Mariana T. Meyer, Andrew Berkovich, Sowmya Subramanian, Agis A. Iliadis, William E. Bently, Reza Ghodssi

Schematic of the integrated microsystem of the SAW sensor and electrodes for induction of the bioelectric effect. A microfluidic chamber is placed over the system to introduce bacterial growth media and antibiotics.

This paper presents a real-time microsystem for detection of biofilm growth as well as its removal through integrated treatment. Detection of biofilms is achieved using a surface acoustic wave (SAW) sensor that monitors the change in total biomass by measuring the resonant frequency of the system. Biofilm treatment is based on the bioelectric effect (BE), a combination of low-dose antibiotics with application of both alternating and direct current signals. The detection limit of the SAW system is approximately 166 pg, corresponding to a bacterial population on the order of hundreds of bacteria. The system is used to observe an 80% reduction of total biomass when treated by the BE as compared to traditional antibiotics. Through system integration of the BE with the SAW sensor, simultaneous biofilm detection and treatment is achieved. The system consumes 194 W of power, with the sensor and treatment consuming 100 W and 94 W, respectively. These results suggest that the designed system can be used for treatment of diverse biofilms at low antibiotic concentrations. The system integration of sensor and treatment into a single platform can be used for managing biofilm associated problems as well as for fundamental chemical and biological research, such as new drug discovery studies.

Sensors and Actuators A: Physical, 238 (2016), 140-149  A PDF of the article can be downloaded here.


Encoding of natural sounds by variance of the cortical local field potential

Nai Ding, Jonathan Z Simon, Shihab A. Shamma, Stephen V David

Correlation between the shape of STRFs measures the similarity of tuning across neural signals. Delta-, theta-, and alpha-variance STRFs are highly correlated, and the higher frequency bands (gamma, high gamma, MUA) also show a cluster of similarity to each other.

Neural encoding of sensory stimuli is typically studied by averaging neural signals across repetitions of the same stimulus. However, recent work has suggested that the variance of neural activity across repeated trials can also depend on sensory inputs. Here, we characterize how inter-trial variance of the local field potential (LFP) in primary auditory cortex (A1) of awake ferrets is affected by continuous natural sound stimuli. We find that natural sounds often suppress the inter-trial variance of low-frequency LFP (< 16 Hz). However, the amount of the variance reduction is not significantly correlated with the amplitude of the mean response at the same recording site. Moreover, the variance changes occur with longer latency than the mean response. Although the dynamics of the mean response and inter-trial variance differ, spectro-temporal receptive field (STRF) analysis reveals that changes in LFP variance have similar frequency tuning to multi-unit activity at the same recording site, suggesting a local origin for changes in LFP variance. In summary, the spectral tuning of LFP inter-trial variance and the absence of a correlation with the amplitude of the mean evoked LFP suggest substantial heterogeneity in the interaction between spontaneous and stimulus-driven activity across local neural populations in auditory cortex.

Journal of Neurophysiology, 24 February 2016 Vol. no. , DOI: 10.1152/jn.00652.2015. A PDF of the article can be downloaded here.