Technological innovations

Asset management is a time consuming and error prone process. Information Technology (IT) personnel typically perform this task manually by visually inspecting assets to identify misplaced assets. If this process is automated and provided to IT personnel it would prove very useful in keeping track of assets in a server rack. A mobile based solution is proposed to automate this process. The asset management application on the tablet captures images of assets and searches an annotated database to identify the asset. We evaluate the matching performance and speed of asset matching using three different image feature descriptors. Methods to reduce feature extraction and matching complexity were developed. Performance and accuracy tradeoffs were studied, domain specific problems were identified, and optimizations for mobile platforms were made. The results show that the proposed methods reduce complexity of asset matching by 67% when compared to the matching process using unmodified image feature descriptors.

This thesis research was funded by the Southeast National Marine Renewable Energy Center (SNMREC) at Florida Atlantic University. Its objective is the development of Quality of Service (QoS) mechanisms for the wireless communications architecture used by the Prognosis and Health Monitoring (PHM) subsystem. There are numerous technical challenges that the PHM Communications Subsystem tries to solve. Due to ocean platform mobility from waves, currents, and other environmental factors, signal quality can vary significantly. As a result, the wireless link between the electric generator platform and shore systems will have variable quality in terms of data rate, delay, and availability. In addition, the data traffic that flows from generator sensors and PHM applications to the shore systems consists of numerous types of messages that have different QoS demands (e.g. delay) and priority that depends on the message type, user ID, sensor location, and application-dependent parameters. The PHM Communications subsystem must handle effectively high priority messages, such as alarms, alerts, and remote control commands from shore systems. It also performs QoS in the application layer, so it can read the contents of every message to prioritize them. In order to perform QoS in the application layer the PHM subsystem relies on Java Servlet multithreaded technology and different queuing techniques to control message transmission order. Furthermore, it compresses all traffic that comes from the ocean-based electric generator/turbine platform to reduce the load on the wireless link. The PHM Communications subsystem consists of three components: the wireless link, the Link Manager, and the Web Services Network Proxy. We present experimental results for the Web Services Network Proxy and demonstrate the effectiveness of XML data compression and semantic-based message scheduling over a link with variable capacity.

Possibly the largest problem when working in bioinformatics is the large amount of data to sift through to find useful information. This thesis shows that the use of feature selection (a method of removing irrelevant and redundant information from the dataset) is a useful and even necessary technique to use in these large datasets. This thesis also presents a new method in comparing classes to each other through the use of their features. It also provides a thorough analysis of the use of various feature selection techniques and classifier in different scenarios from bioinformatics. Overall, this thesis shows the importance of the use of feature selection in bioinformatics.

With the increasing demands of rising medical costs in combination with a boom in elderly patients in need of quality patient care medical practices are being stressed. Patient to nurse ratios are increasing and government spending in the medical domain is at an all-time high threatening the futures of government medical programs such as Medicare and Medicaid. In this thesis we propose a framework for the monitoring of a patient's vital statistics in a home-based setting using a mobile smart device. We believe that in taking advantage of the wireless sensor technology which is readily available today we can provide a solution that is both economically and socially viable offering a solid quality of healthcare in a comfortable and familiar environment. Our framework exposes both 802.11 and Bluetooth wireless protocol transmitting medical sensor devices using an Android platform device as a monitoring hub.

One of the greatest challenges to data mining is erroneous or noisy data. Several studies have noted the weak performance of classification models trained from low quality data. This dissertation shows that low quality data can also impact the effectiveness of feature selection, and considers the effect of class noise on various feature ranking techniques. It presents a novel approach to feature ranking based on ensemble learning and assesses these ensemble feature selection techniques in terms of their robustness to class noise. It presents a noise-based stability analysis that measures the degree of agreement between a feature ranking techniques output on a clean dataset versus its outputs on the same dataset but corrupted with different combinations of noise level and noise distribution. It then considers classification performances from models built with a subset of the original features obtained after applying feature ranking techniques on noisy data. It proposes the focused ensemble feature ranking as a noise-tolerant approach to feature selection and compares focused ensembles with general ensembles in terms of the ability of the selected features to withstand the impact of class noise when used to build classification models. Finally, it explores three approaches for addressing the combined problem of high dimensionality and class imbalance. Collectively, this research shows the importance of considering class noise when performing feature selection.

This work presents the development of the Statistical Location-Assisted Broadcast (SLAB) protocol, a multi-hop wireless broadcast protocol designed for vehicular ad-hoc networking (VANET). Vehicular networking is an important emerging application of wireless communications. Data dissemination applications using VANET promote the ability for vehicles to share information with each other and the wide-area network with the goal of improving navigation, fuel consumption, public safety, and entertainment. A critical component of these data dissemination schemes is the multi-hop wireless broadcast protocol. Multi-hop broadcast protocols for these schemes must reliably deliver broadcast packets to vehicles in a geographically bounded region while consuming as little wireless bandwidth as possible. This work contains substantial research results related to development of multi-hop broadcast protocols for VANET, culminating in the design of SLAB. Many preliminary research and development efforts have been required to arrive at SLAB. First, a high-level wireless broadcast simulation tool called WiBDAT is developed. Next, a manual optimization procedure is proposed to create efficient threshold functions for statistical broadcast protocols. This procedure is then employed to design the Distribution-Adaptive Distance with Channel Quality (DADCQ) broadcast protocol, a preliminary cousin of SLAB. DADCQ is highly adaptive to node density, node spatial distribution pattern, and wireless channel quality in realistic VANET scenarios. However, the manual design process used to create DADCQ has a few deficiencies. In response to these problems, an automated design procedure is created that uses a black-box global optimization algorithm to search for efficient threshold functions that are evaluated using WiBDAT. SLAB is finally designed using this procedure.

This thesis consists of the development of a web based wireless sensor network (WSN) monitoring system using smartphones. Typical WSNs consist of networks of wireless sensor nodes dispersed over predetermined areas to acquire, process, and transmit data from these locations. Often it is the case that the WSNs are located in areas too hazardous or inaccessible to humans. We focused on the need for access to this sensed data remotely and present our reference architecture to solve this problem. We developed this architecture for web-based wireless sensor network monitoring and have implemented a prototype that uses Crossbow Mica sensors and Android smartphones for bridging the wireless sensor network with the web services for data storage and retrieval. Our application has the ability to retrieve sensed data directly from a wireless senor network composed of Mica sensors and from a smartphones onboard sensors. The data is displayed on the phone's screen, and then, via Internet connection, they are forwarded to a remote database for manipulation and storage. The attributes sensed and stored by our application are temperature, light, acceleration, GPS position, and geographical direction. Authorized personnel are able to retrieve and observe this data both textually and graphically from any browser with Internet connectivity or through a native Android application. Web-based wireless sensor network architectures using smartphones provides a scalable and expandable solution with applicability in many areas, such as healthcare, environmental monitoring, infrastructure health monitoring, border security, and others.

A Wireless Sensor Network (WSN) is composed of low-cost electronic devices with sensing, data storage and transmitting capabilities, powered by batteries. There are extensive studies in the field of WSN investigating different algorithms and protocols for data collection. A data collector can be static or mobile. Using a mobile data collector can extend network lifetime and can be used to collect sensor data in hardly accessible locations, partitioned networks, and delay-tolerant networks. The implementation of the mobile data collector in our study consists of combining two different platforms: the Crossbow sensor hardware and the NXT Legos. We developed an application for data collection and sensor querying support. Another important contribution is designing a semi-autonomous robot control. This hardware prototype implementation shows the benefits of using a mobile data collector in WSN. It also serves as a reference in developing future applications for mobile WSNs.

Sensors are shaping many activities in our society with an endless array of potential applications in military, civilian, and medical application. They support different real world applications ranging from common household appliances to complex systems. Technological advancement has enabled sensors to be used in medical applications, wherein they are deployed to monitor patients and assist disabled patients. Sensors have been invaluable in saving lives, be it a soldier's life in a remote battlefield or a civilian's life in a disaster area or natural calamities. In every application the sensors are deployed in a pre-defined manner to perform a specific function. Understanding the basic structure of a sensor node is essential as this would be helpful in using the sensors in devices and environments that have not been explored. In this research, patterns are used to present a more abstract view of the structure and architecture of sensor nodes and wireless sensor networks. This would help an application designer to choose from different types of sensor nodes and sensor network architectures for applications such as robotic landmine detection or remote patient monitoring systems. Moreover, it would also help the network designer to reuse, combine or modify the architectures to suit more complex needs. More importantly, they can be integrated with complete IT applications. One of the important applications of wireless sensor networks in the medical field is a remote patient monitoring system. In this work, patterns were developed to describe the architecture of patient monitoring system.

A Wireless Sensor Network (WSN) is composed of a large number of sensor nodes that are densely deployed in an area. One of the main issues addressed in WSNs research is energy efficiency due to sensors' limited energy resources. WSNs are deployed to monitor and control the physical environment, and to transmit the collected data to one or more sinks using multi-hop communication. Energy efficiency protocols represent a key mechanism in WSNs. This dissertation proposes several methods used to prolong WSNs lifetime focusing on designing energy efficient communication protocols. A critical issue for data gathering in WSNs is the formation of energy holes near the sinks where sensor nodes participate more in relaying data on behalf of other sensors. The solution proposed in this dissertation is to use mobile sinks that change their location to overcome the formation of energy holes. First, a study of the improvement in network lifetime when sinks move along the perimeter of a hexagonal tiling is conveyed. Second, a design of a distributed and localized algorithm used by sinks to decide their next move is proposed. Two extensions of the distributed algorithm, coverage and time-delivery requirement, are also addressed. Sensor scheduling mechanisms are used to increase network lifetime by sending redundant sensor nodes to sleep. In this dissertation a localized connected dominating set based approach is used to optimize network lifetime of a composite event detection application. A set of active nodes form a connected set that monitor the environment and send data to sinks. After some time, a new active nodes set is chosen. Thus, network lifetime is prolonged by alternating the active sensors. QoS is another main issue encountered in WSNs because of the dynamically changing network topology.