«Strengthening the Nation through Diversity, Innovation & Leadership in STEM San Antonio,Texas · October 3-6, 2013 Get Connected! Connect with the ...»
Previously we presented TOA, a network data visualization framework using open source interactive visualization libraries based in NetFlows. In this framework we apply visualization analytics techniques to aid system administrators to monitor complex network infrastructures with big amounts of data, and the possibility to identify anomalies in the network. In this work, we present the new graphic user interface (GUI) and the administrator’s interface for the network data visualization framework. The interface manages the system data for the generation of the dictionary that allows the dynamic data mining of the network data, the generation of the graphs, and the GUI menus. Some of the features of this administrative interface are encryption of passwords, assignment of unique session identification, a session timeout that logs the user out of the system if there is inactivity, multiple session notifications that allow the user to close sessions in other machines, and administration levels to create an administrative hierarchy and change the password. We look forward to keep improving and adding features to the interfaces, not only to make them pleasing to the eye but to maintain them securely.
COMPARING AND EXTENDING APPROACHES THAT ACCESS SCIENTIFIC DATABASES THROUGH
SEMANTIC WEB ONTOLOGIESJoshue Cardoza1, Natalia Villanueva-Rosales2.
University of Texas at Brownsville, Brownsville, TX, 2University of Texas at El Paso, El Paso, TX.
1 The increasing amount of scientific data produced and managed using heterogeneous formats and schemas has created the need to automate the access, search, and retrieval of this data. Semantic web technologies provide a promising solution for this challenge through the creation of ontologies that further describe scientific data. The goal of this research is to evaluate current approaches that integrate legacy scientific databases with ontologies. One of the challenges when creating ontologies from legacy databases is that the data is typically migrated into a triple store or an ontology management system, which creates an additional data source to maintain and update. We are investigating approaches that allow us to access these databases with ontologies without migrating them into other systems. We are currently comparing the scalability and expressivity of D2RQ (http://d2rq.org/) and Virtuoso (http://virtuoso.openlinksw.com) which implement the RDB to RDF Mapping Language (R2RML). To compare these approaches, we are using 2 test scenarios: 1) sensor-generated data for ecological research generated by the Systems Ecology Lab, and 2) data from the Virtual Geocaching project developed at the Cyber-ShARE Center, both located at the University of Texas at El Paso. These projects require the storage and integration of large amounts of heterogeneous data that will be integrated via ontologies as part of our future work. We anticipate that efforts that bridge legacy databases and ontologies will facilitate the adoption of semantic web technologies for the handling and processing of large amounts of scientific data.
THE FEASIBILITY OF USING JAVA REAL-TIME AS THE EXECUTION CODE FOR AN AUTONOMOUS
UNDERWATER VEHICLESamori Price, Antonia Boadi.
California State University, Dominguez Hills, Carson, CA.
Autonomous underwater vehicles (AUVs) are very useful devices often used in detection and exploration. AUVs must be able to search, survey, document, and possibly clear dangerous areas that may be too hazardous for humans to explore. They must complete these missions without any human interaction unless the administrator chooses to take
162 UNDERGRADUATE POSTER ABSTRACTS
control. During the course of our research, we created a model for an AUV and simulated it in an obstacle course.
The model was created using the AUV Workbench simulation software from the Naval Post-Graduate School. With the data gathered from the model and simulation, we were able to use a joystick to simulate controlling the AUV.
AN INTERACTIVE ASSISTIVE APPLICATION FOR PEOPLE WITH COGNITIVE IMPAIRMENTSAlex Santos1, Jing Wang2, Harshal Mahajan2, Dan Ding2.
University of Puerto Rico, Rio Piedras Campus, San Juan, PR, 2University of Pittsburgh, Pittsburgh, PA.
1 People with cognitive impairments (e.g., traumatic brain injury, Alzheimer’s disease, etc.) usually suffer from various cognitive deficits with memory, attention, planning, and executive functions. These cognitive deficits can limit individuals from completing activities of daily living (ADLs) independently. Among ADLs, cooking tasks have been identified as essential skills for living independently, which can play an important part in an individual’s overall health implications, accomplishment in social roles, self-esteem, and sense of control. The objective of this study is to develop interactive surfaces in the kitchen that can give intuitive prompts for people with cognitive impairments when doing multi-step cooking tasks. Our touch-based interactive application is developed based on the WorldKit technology which pairs Kinect’s depth camera with a projector to create interactive surfaces. In this application, an interactive kitchen island that shows image lists of required ingredients have been created and different types of prompts (text, audio, and visual) have been enabled according to detected user response activities. Also, we integrate a MySQL database of the kitchen inventory into this assistive application, making it easy to adjust different recipes and kitchen storage. The main parts of the application development have been completed, and further user assessment will be needed. This study presents an innovative method to assist people with cognitive impairments. Through further user assessment of this interactive application, we may able to improve the functions of this application, identify the efficiency of interactive prompts for people with cognitive impairments, and increase individuals’ independence levels.
A WEB SERVICE TO VISUALIZE AND MONITOR NETFLOW DATA USING THE WEB IMPLEMENTATION OF THE
SPINNING CUBE OF POTENTIAL DOOMJhensen Grullón Sanabria Sr. Eric Santos, Jose Ortiz.
University of Puerto Rico, Rio Piedras Campus, San Juan, PR.
The monitoring and analysis of computer networks to detect abnormal network behavior is a complex task for network administrators, which is assisted by the use of visualization analytics. In this work, we present a web-based implementation of the three dimensional cube known as the Spinning Cube of Potential Doom using WebGL and the Three.js library to provide an easy-to-use and access network visualization interface. In the first version of the cube, we worked with NetFlow data stored in an application server. Our current implementation works as a web service to allow the visualization of Netflow data uploaded by users on the internet without the need to install and configure any additional software in their servers. This application allows system administrators to visualize distinct network events such as network and port scanning. Also, a control panel is provided to filter data by network classes. The computing needed for the filtering and parsing of the file is done on the server side; the web browser is only used as the graphic user interface. For future work, we will implement the feature where the user can upload and hold multiple files on the server and select which one to display.
A CONSTRAINT-BASED MODEL TO GENERATE T-WISE TEST SUITESLuis Gutierrez, Martine Ceberio, Carlos Nieto, Francisco Zapata.
University of Texas at El Paso, El Paso, TX.
Having reliable software is a difficult task to achieve. Today, many people rely on software to accomplish everyday life activities. Moreover, software failures can cause significant costs to companies. Software testing is used to find errors before releasing software. Ideally, testing all possible scenarios will result in high confidence that the software is reliable. However, testing all possible scenarios is unrealistic for large systems. Therefore, it is important to determine
163 UNDERGRADUATE POSTER ABSTRACTS
a cost-effective and reliable testing approach. In order to accomplish this, common causes of software failures have to be identified. Empirical studies show that most software errors are caused by the interaction between 2 to 6 input parameters. This strategy, called combinatorial testing, leads to generate test suites that include such interactions.
This strategy is cost-effective since it significantly reduces the number of test cases generated, when compared to generating all possible test cases, and ensures high levels of reliability. Some of the algorithms developed to generate such test suites include greedy algorithms and genetic algorithms. However, there is no algorithm that can ensure the generation of the smallest set of test cases. The main goal of this project is to generate a t-wise test suite, where t is the number of interactions to be included in the test suite, with the least amount of test cases. To address this problem, we present a constraint-based model that can generate t-wise test suites in an optimal manner. Preliminary results of this model are presented.
THOUGHT-FEELING-EXPRESSION INTERFACES FOR COMMUNICATION AND CONTROLKhoa Nguyen, Hong Lin, Lucino Flores, Steve Leon.
University of Houston-Downtown, Houston, TX.
Interaction between machines and humans is limited to conscious interaction only. Non-conscious communication such as intuition and perception are reserved for humans. However, Emotiv has changed this by developing a tool by which humans and machines may communicate through non-conscious means. Emotiv has developed the Epoc headset which is a high resolution, multi-channel, wireless neuroheadset. Furthermore, the Epoc uses 14 sensors and 2 references to tune into electric signals produced by the brain to detect a user’s thoughts, feelings, and expressions in real time. The objective of this research is to create an application program which is controlled completely by thoughts, feelings, and expressions to perform pre-designed tasks. By doing so, the application will give people the experience of controlling and influencing their virtual environments with their minds. The results of this project will have a wide range of applications in controlling devices with thoughts, feelings, and expressions in real time.
RIZZY THE ROBOTAntonio Gomez, John Fernandez.
Texas A&M University-Corpus Christi, Corpus Christi, TX.
Over the fall semester of 2012, our team built a 4-wheeled robot named Rizzy from the ground up. The design was entirely our own, as well as all the welding, wiring, and other construction that was necessary. We programmed an Arduino microcontroller to control the motors and to respond to input commands to allow us to steer and accelerate the machine using an Xbox controller. A Microsoft Kinect was mounted to the front of Rizzy to act as machine eyes.
RGB along with depth information is transmitted from the Kinect to provide a video display either onto a directly connected laptop or across a wireless network to a remote terminal. In this way, we can drive Rizzy directly with a wired controller or from a remote location. We also programmed a heads-up display to output relevant information to the video display, such as the camera angle. This semester we have many plans to expand its capabilities.
Ideally, Rizzy the robot will become a tool for computer science students at Texas A&M University-Corpus Christi to experience writing software and watching it come to life through a robot.
NAVIGATION SYSTEM FOR 3D MODELSMelissa Greenlee, Ongard Sirisaengtaksin.
University of Houston-Downtown, Houston, TX.
The purpose of this project is to create a program using Microsoft XNA that will allow users to upload any 3D model and navigate through the model. Model navigation will include collision detection to make movement experience more realistic. This program can be used for familiarization with facilities prior to visitation or as a training implement for personnel in case of emergency situations. In addition, this program would be useful in previewing a facility model and determining functionality and aesthetics before construction. The main problem is to develop a collision detection algorithm to detect model objects. A camera object was created for navigation and collision detection. Another problem is placing the camera so it is not placed in the model where it will detect a collision on loading of the model.
In order for the developed collision detection algorithm to function properly, the 3D models uploaded must be drawn in a specific way. Another issue encountered was how to ensure that models would load without the camera object initially detecting a collision. Due to this issue, another algorithm was created that uses model vertex information
164 UNDERGRADUATE POSTER ABSTRACTS
to determine initial load position of the camera. With the algorithms for collision detection and camera loading, the program created was successful for user navigation and collision detection. Several different models were tested and it appears that any bugs in the program have been resolved. Further testing with more models will bring to light any Computer Science other program bugs that may need to be addressed.
STRATEGIES FOR PERSUASION IN COMPUTER-MEDIATED DIALOGUESBrandon DeBord, Mohamed Munir Mohamed Riyaj, Yehuda Gutstein, Rachel Adler.
Northeastern Illinois University, Chicago, IL.