Qatar Foundation Annual Research Forum Volume 2012 Issue 1

In order to analyze and deduce valuable information from big image data, we have developed a framework for distributed image processing in Hadoop MapReduce. A vast amount of scientific data is now represented in the form of images from sources like medical tomography. Applying algorithms on these images has been continually limited by the processing capacity of a single machine. MapReduce created by Google presents a potential solution. MapReduce efficiently parallelizes computation by distributing tasks and data across multiple machines. Hadoop, an open source implementation of MapReduce, is gaining a widespread popularity due to features such as scalability and fault tolerance. Hadoop is primarily used with text-based input data. Its ability to process image data and its performance behavior with image processing have not been fully explored. We propose a framework that efficiently enables image processing on Hadoop and characterizes its behavior using a state-of-the-art image processing algorithm, Edge Detection. Existing approaches in distributed image processing suffer from two main problems: (1) input images need to be converted to a custom file format and (2) image processing algorithms require adherence to a specific API that might impose some restrictions on applying some algorithms to Hadoop. Our framework avoids these problems by: (1) bundling all small images into one large file that can be seamlessly parsed by Hadoop and (2) relaxing any restriction by allowing a direct porting of any image processing algorithm to Hadoop. A R educe-less job is then launched where the code for processing images and a mechanism to write the images back individually to HDFS are included in Mappers. We have tested the framework using Edge Detection on a dataset of 3760 biomedical images. Besides, we characterized Edge Detection along several dimensions, such as degree of parallelism and network traffic patterns. We observed that varying the number of map tasks has a significant impact on Hadoop's performance. The best performance was obtained when the number of map tasks equals the number of available slots as long as the application resource demand is satisfied. Compared to the default Hadoop configuration, a speedup of 2.1X was achieved.

An empirical evaluation of constraint diagrams (CD) as a program specification language is conducted by comparing it to natural language (NL) with computer science students included two experiments, one on the interpretation of CD to evaluate the comprehension of notational system, and the other on the construction of program specifications. The first experiment took the form of a web-based competition in which 33 participants were given instructions and training either on CD or on equivalent NL specification expressions. After each example, they responded to three multiple-choice questions requiring the interpretation of expressions in their particular notation. Although the CD group spent more time on the training and had less confidence, they obtained comparable interpretation scores as the NL group and took less time to answer the questions, although they had no prior experience of CD notation. In the second experiment, which focused on the construction of CD, 20 participants were given instructions and training either on CD or on equivalent NL specification expressions. After each example, they responded to three questions requiring the construction of expressions in their particular notation. We built a mini-editor to allow the construction of the two notations, which automatically logged their interactions. Although the CD group supplied more accurate answers, they spent more time in answering the questions. The NL group supplied answers that were partially correct but with some missing information. Moreover, the CD group had spent more time in training, but their returns to the training examples were fewer than the NL group.

Wireless technology such as GSM/GPRS and Wi-Fi has taken on an important role in our daily life. The noticeable improvement of these technologies has led to advancement in home automation, where people can enjoy comfort and secure living places. This project proposes a ZigBee-based cost effective solution to control and monitor home appliances remotely and to enable home security against intrusion in the absence of a homeowner using the wireless sensor network kit from national instruments. A wireless sensor network, connected to a GSM modem, can send and receive alarming and controlling SMSs and remotely control the system using an internet web page. Also a security camera is embedded within the system to send alarm SMS and snap an image in case of a visitor or intruder. The system is implemented and evaluated among different sample applications spanning home security, climate control, hazard alarming, and appliance control with remote access and control features to demonstrate the feasibility and effectiveness of the proposed system. Moreover a small-scale live demo model of the smart house is constructed to proof the concept

This work explores the differences in direction-giving strategies between two groups, native Arabic and native English speakers. This study will help influence design decisions for multi-lingual, cross-cultural human robot interaction. There are clear cultural influences on modes of communication. Previous research studies found that direction-giving techniques and strategies vary between different cultural groups. Burhanudeen compared Japanese and English native speakers and found that locator remarks are more frequently used by Japanese natives, while the use of directives is more common with English natives. In this work, we examine the discourse for navigation instructions of members of two target groups, Arabic native speakers and English native speakers. We address the following questions: How do languages and strategies used for providing directions vary between these two groups? What are the differences and what are the similarities? Are there any possible gender-related differences in giving directions? We recorded 56 participants giving oral direction instructions for three specific locations at the Carnegie Mellon Qatar campus, and 33 participants giving oral direction instructions for three different locations at the Student Center. We transcribed the audio recordings and annotated our transcriptions. We categorized the spatial perspectives used by participants into route perspective which involves the use of left/right turns and landmarks, and survey perspective, which involves the use of units (time and distance) and cardinals (north, south, east and west). Our analysis also included number of pauses, repetitions, error corrections, number of words and intermediate details. Our results showed that the way-finding strategy favored by English natives and Arab natives is the landmark-based navigation strategy. However, English natives had a higher frequency of using cardinals, pauses and intermediate information while Arab natives used units of distance, left/right turns and error corrections more frequently than English natives. Male participants from both groups are more likely to rely on survey perspective than female participants. Based on these results, we conclude that culture, language, and gender influence a speakers discourse and strategy for giving directions.

Due to the requirement of high data rates and broad utilization of wireless technologies (e.g., 3G, 4G and beyond), the radio frequency (RF) spectrum has become a very limited resource for modern wireless communication systems. Studies indicate that the spectrum is being under-utilized. As a promising solution, cognitive radio (CR) is an encouraging candidate to achieve more efficient RF spectrum utilization. The previous studies motivate us to utilize the dynamic channel fading model (hyper-fading) to perform a unified analysis for CR multiple-access channel (CR-MAC) networks. Since the nature of the CR networks is multiuser communication, deliberating CR-MAC is more pertinent than point-to-point communication systems. The objective is to maximize the capacity of CR-MAC network over hyper-fading channels under both secondary user's (SU's) transmit power (TP) and interference temperature (IT) constraints. Multiple SUs transmit to the secondary base station under the TP and IT constraints. In order to perform a general analysis, a theoretical dynamic fading model termed hyper-fading model, which is suitable to the dynamic nature of cognitive radio channel, is considered. The optimal power allocation method (water-filling) is employed to maximize the capacity of CR-MAC for hyper-fading channel with TP and IT constraints. Throughout the results, the capacity of the hyper-fading channels are compared with that of other channel fading models such as Rayleigh, Nakagami-2, and with an AWGN channel. Furthermore, the impact of the number of SUs on capacity is investigated. Numerical results along with relevant discussions for capacity measure under AWGN, and Rayleigh, Nakagami-2 and hyper-fading channel models are provided to compare the behavior of CR-MAC in these environments. The results reveal that in the case of very strict IT constraint the water-filling method gives good capacity improvements. This benefit is lost when the IT constraint is relaxed. Through comparison of hyper-fading with other fading environments, it is found that the hyper-fading channel fills the gap in the capacity profiles obtained from the other channel fading types. Further, a study of such a CR-MAC system that undergoes a hyper-fading model can provide unified and comprehensive insights on performance analysis of the CR networks.

Background & Objectives: Advances in the radio frequency identification (RFID) technology have made it ubiquitously present. Due to the recently emerging applications, RFIDs have been used in a plethora of different scenarios. Oil and gas industry is no exception; developments in the RFID technology enabled it to be used in oil and gas quality and pipeline infrastructure monitoring through low frequency RFID tags. The presence of moisture in liquefied petroleum gas (LPG) / liquefied natural gas (LNG) pipes can create two main problems. It can degrade the heating ability of the fuel and it can also react with the refrigerant of the liquefied gas to create hydrates, which further lower the quality of the gas. A wireless monitoring system using low frequency RFIDs as moisture sensors can prevent such hazards by detecting the received energy of the transmitted signal to determine the presence of moisture. Method: A simple energy detector concept was utilized to exploit the well-known behavior of the RFID's signal reception and energy absorption with varying environments. A decision for the presence of moisture was made using the threshold values of the energy detector. Results: The experimental results show that the energy detector approach detects the presence of moisture in the oil and gas system. As another remark, when we decrease the distance between the receiver and transmitter antennas of RFID sensors, one can easily notice the increase on the energy detection performance. Different power levels, modulation schemes and frequency ranges have been analyzed to better understand the energy detection output of RFID system. An optimized solution can be looked at for using the best frequency and input power to maximize the distance between the two antennas. Conclusion: In this study, one can see that the RFID system can be used to send information regarding the presence of moisture for quality monitoring of liquefied gas pipes in Qatar. By using this information, we can detect and make decisions on the basis of the energy detection output from the RFID antenna system. This study can be used in a wireless cyber-physical moisture detection system targeted at Qatar's needs.

Background & Objectives: Spectrum sensing is a feature of cognitive radio (CR) systems which is proposed to improve spectral utilization of wireless signals. One of the sensing methods is the non-coherent energy detector, and even though it is computationally more effective than coherent methods, it has critical drawbacks, e.g. requirement of certain signal-to-noise ratio and number of samples. Moreover, type of digital modulation employed also affects the performance of the energy detectors. Therefore, the performance of the energy detectors is investigated for phase shift keying (PSK) and quadrature amplitude modulated (QAM) signals. Such an analysis is essential, because once the performance of the different modulated signals are understood; next generation wireless networks (NGWNs) and CRs can be designed in such a way that the arduous and expensive planning stage is omitted. This way, a higher data rate can be achieved by using the proper modulation type and/or order for indoor CRs and NGWNs. Methods: Instead of false alarm and missed detection analysis, probability mass function vs. energy detection statistics are introduced to better understand the effect of modulation type, order and wireless channel. A measurement setup is developed to consider line-of-sight and non-line-of-sight conditions. Signals are constructed, transmitted and recorded based on the signal model provided. All experiments took place in the Wireless Research Laboratory of the department of ECEN at Texas A&M University at Qatar. Results: The results show that performance of the energy detector changes drastically with the digital modulation scheme employed at the transmitter side. Another interesting point is the impact of the used energy detector samples (N). As expected, with the increase in N, the impact of central limit theorem (CLT) can be felt in a clearer way as well. Conclusions: By using the experimental results, the design and deployment of future cellular mobile radio systems such as femtocells in Qatar could be optimized. This is crucial, since most communications such as voice/internet/text traffic occur inside buildings especially for rapidly changing network topographies as is the case with the city of Doha. This way, data rate and speed can be optimized for the indoor environments.

Information resources over the World Wide Web are increasingly growing in size and creating immense demand for relevant and context-sensitive information. Thus, information retrieval and knowledge management systems need to exhibit high capacity for the mass of information available and most importantly should be able to handle the constant and rapid changes in information in order to produce results in a reasonable time. In this presentation we would like to highlight the research work done for building a new system for incremental information management relative to a domain of knowledge. The new system (named IPS) utilizes new conceptual methods developed using new formal concept analysis constructs called pseudo concepts for managing incremental information organization and structuring in a dynamic environment. The research work in hand focuses on managing changes in an information store relevant to a specific domain of knowledge attempted through addition and deletion of information. The incremental methods developed in this work should support scalability in change-prone information stores and be capable of producing updates to end users in an efficient time. We will also discuss practical aspects related to the macro and micro information organization built using the new system. These include handling incremental corpus organization for a specific domain, performing context-sensitive text summarization of news articles and news articles features extraction. In addition, initial evaluation results will be also discussed showing the improvement in execution time and time complexity while maintaining reasonable comparable quality of incremental structures produced.

There are still specific areas where our knowledge of carbon storage is in need of improvement, particularly in carbonate reservoirs, since currently we extrapolate data from limited sources and the predictive modelling technologies employed have a level of uncertainty that needs to be addressed. We will highlight our efforts through the Qatar Carbonates and Carbon Storage Research Centre (a $70 million, 10 year research programme with currently 20 PhD students and 10 postdoctoral researchers along with 14 faculty members) to investigate the underlying science and engineering concerning carbonate reservoir characterisation, rock-fluid-CO₂ interactions and multiphase flow experiments under reservoir conditions linked to complimentary simulation and modelling advances, including the rapidly developing field of digital rocks. This has involved developing unique HTHP experimental rigs and pioneering new modelling techniques, enhancing the toolbox available to engineers and geoscientists to select suitable reservoirs and optimally design CO₂ storage processes. These capabilities extend over molecular-pore-core-field scales. We have four focused research laboratories (Qatar Stable Isotope Lab; Qatar Thermophysical Property Lab; Qatar Complex Fluids Lab; Qatar CCS Multiscale Imaging Lab) and will discuss the highlights of major research findings to date in the context of carbon storage in the Middle East.

Maximizing recovery in oil and gas fields relies on geological models that realistically portray the spatial complexity, composition, and properties of reservoir units. Present day arid climate coastal systems, like the coastline of Qatar provide analogues for depositional and diagenetic processes that control reservoir quality in ancient reservoirs. Many major reservoirs in Qatar are formed under conditions that are remarkably similar to those shaping the coastlines of today. Among the major controls on coastal sedimentation patterns are: 1) wind, wave and tidal energy, 2) coastline orientation, 3) relative sea level, 4) depositional relief and 5) sediment sources. Strong NW prevailing winds (shamal winds) drive shallow marine circulation patterns, creating four very distinct coastal profiles: windward, leeward, oblique, and protected. In addition, winds supply quartz sand to the leeward coast, as the dune fields of Khor Al-Adaid are blown into the sea. Elsewhere, carbonate sands are formed by wave breakdown of skeletal material in the shallow marine environment. These sands are washed ashore to form beaches. The grain size, composition, and dimensions of coastal sands vary due to wave energy. Coastal deposits are equally affected by high frequency oscillations in sea level. Approximately 8,000 years ago, the sea level was about 3 meters higher than it is currently and the Qatari coastline was up to 15 km inland. Most coastal deposits and sabkhas are relicts of this ancient highstand in sea level. Punctuated sea level drops to present day levels have led to the formation of seaward-stepping spit systems. Understanding these coastal and near coastal areas, the processes that form them, and developing geologic models based on this understanding, is a focus of the Qatar Center for Coastal Research (QCCR) within ExxonMobil Research Qatar. The observed spatial complexity and heterogeneity of modern coastal systems are important aspects to be considered for conditioning three-dimensional geological models. The studied modern outcrops along the Qatar coastline are particularly useful as analogs for conditioning subsurface data sets in geologic (static) and reservoir (dynamic) models.