Mark P. Hansen, Ph.D. - US grants

ABSTRACT
0120778
U of Calif - Los Angeles

The research focus of the Center for Embedded Networked Sensing (CENS) will be the fundamental science and engineering research needed to create scalable, robust, adaptive, sensor/actuator networks. The vision of densely distributed, networked sensing and actuation requires advances in many areas of information technology. Moreover, there is a critical interplay between the technology and the applications and physical context in which it is embedded. By conducting research in the context of specific and high-impact scientific applications, CENS will enable new scientific discovery through high resolution, in situ monitoring and actuation. At the same time, CENS will explore the fundamental principles and technologies needed to apply embedded networked sensing to a wide range of applications.

The Center will focus initially on fundamental technology and on four experimental application drivers: habitat monitoring for bio-complexity studies, spatially-dense seismic sensing and structure response, monitoring and modeling contaminant flows, and detection and identification of marine microorganisms. To support this scope, CENS will combine the expertise of faculty from diverse engineering disciplines with the expertise of biological, environmental and earth scientists. During the lifetime of the Center, additional opportunities for applying the technology to natural and engineered systems will be pursued.

The CENS educational focus will be twofold: new hands-on experimental capabilities for grades 7-12 science curriculum through access to real-world, real-time, sensor-network interrogation, along with materials for teacher-training, and undergraduate research opportunities in cutting-edge technologies (e.g., wireless systems, MEMS, embedded software) and scientific applications (e.g., bio-complexity, seismic and environmental monitoring), with emphasis on under-represented minority students.

CENS will benefit from and contribute to a large number of related activities on its participating campuses, and in the larger research and education community, including: UCLA's California Nanosystems Institute, Institute for Pure and Applied Mathematics, Nanoelectronics Research Facility; USC's Information Sciences Institute, Wrigley Institute for Environmental Studies; UC Reserve systems; Cal State and GLOBE Teacher training programs; INEEL, JPL government laboratories; DARPA, and NSF-related research activities. Many of the constituent technologies will have near- and long-term commercial relevance.

Wireless sensor networks with their ability of in situ and dense spatiotemporal sampling of physical phenomena have rapidly emerged as a valuable new class of instrument for the basic sciences. Like any instrument, however, sensor networks suffer from impairments that adversely impact the integrity of the sensed information about the physical phenomena. The causes of reduced integrity in sensor networks are many, and include various internal and external uncertainties in the sensing, processing, and communication elements of the system. Examples include calibration errors, faults, sensing channel degradation and obstructions, bio-fouling etc. This project is comprehensively addressing the integrity problem by studying its causes, understanding fundamental limits, developing algorithms and system support, and validating the approach in real-life. The two key elements of the overall approach are autonomous detection of integrity compromise, and resilient estimation and aggregation. Underlying these two are novel statistical and signal-processing techniques that exploit learned models of the physical phenomenon, the measurement process, and the faults which result in corrupted or missed data. In addition, the project is investigating approaches for guiding remediation of the causes of integrity problems, and for integrity-driven deployment of sensor network to achieve desired resilience. The research would result in a thorough understanding of integrity failure in sensor networks, and result in a toolkit of design tools and run-time software for ensuring high-integrity operation. These would be validated via terrestrial, under-soil, and aquatic ecological observation application, and also disseminated for broader use via the project web site at http://nesl.ee.ucla.edu/projects/integrity.

This project is investigating the impact on the network architecture of a new class of applications involving embedded sensing technology as it moves from scientific, engineering, defense, and industrial contexts to the wider personal, social and urban contexts. These applications draw on sensed information about people, objects, and physical spaces to enable new kinds of social exchange and offer new and unexpected views of our communities. They require new algorithms and software mechanisms because unlike scientific applications of distributed sensing, a single system is widely distributed, intermittently connected, and privately administered; and unlike traditional Internet applications the physical inputs are critical to the behavior. The project is developing principles and abstractions that are vital for the future Internet to incorporate such applications, to identify the network fabric architecture options, and to develop the key components and services to realize them. These include network services for context verification and resolution control to enable privacy-aware and verifiable sensing, and application services for naming, dissemination, and aggregation of sensed data. In addition, the project is developing concrete instances of personal, social and urban embedded networked sensing applications to act as design drivers for the broader community of researchers architecting the new Internet core, as well tools to assist in application authoring.
Broader Impact: Ultimately, this research will yield fundamental understanding of software architectures, networking models and data processing techniques that are needed to support a citizenry actively participating in collection, sharing, and interpretation of physical sensor data in the public sphere at multiple scales.

The mobile phone network is emerging as the largest sensor network on the planet. Mobile phone users, however, are generally unaware of the dual uses of this network, in which their communication devices are also information gathering devices. In participatory urban sensing, everyday mobile devices become a platform for coordinated investigation of the environment and human activity. But transforming phones into data collection instruments raises both technical and ethical challenges. The PI believes that researchers should utilize this network of sensors with the consent and active participation of users. Facilitating responsible, socially trusted, and participatory ethics for data collection and analysis with urban sensing systems remains an open problem, and is the challenge undertaken in this research and education project. In this project, the PI and her team will formalize and qualitatively assess an important test case in participatory ethics: a participatory approach to managing privacy in urban sensing applications. They will create both an immersion curriculum (using a hands-on laboratory approach) and an interdisciplinary seminar-style curriculum to teach participatory ethics for urban sensing to diverse STEM undergraduate and graduate students, and will evaluate these curricula and synthesize classroom findings into best practices which will then be disseminated for education in participatory urban sensing ethics to urban sensing, ubiquitous computing, and broader technology education communities through white papers, guest lectures, video presentations and discussions, and an active website.

Broader Impacts: This work will benefit many areas of mobile and ubiquitous technology research. The multidisciplinarity and rapid pace of system development, the social diversity of users, and the diversity of urban sensing applications are exhilirating yet pose significant challenges to developing a participatory ethics framework. Much as traditional human subjects research guidelines apply to a broad and diverse research scope, the PI believes that similarly powerful principles can be specified for human sensing research. The pedagogical tools developed in the education phase of this project will train a diverse group of STEM students to align technological advances with human practices and ethics. Involving students in discussions and practical implementation of participatory ethics will integrate considerations of values into their research and design practice. Students and researchers trained in participatory urban sensing ethics will design systems that reflect participatory ethics and balance technical and human values. Students will also bring the ethical thinking and value commitments formed during their education to careers in academia, technology industries, and policy arenas. As a critical test case in participatory ethics for urban sensing, formalizing and refining participatory privacy regulation will contribute to fields struggling with meaningful privacy design, including mobile and ubiquitous computing, social networking, and web community systems design.

MOBILIZE: Mobilizing for Innovative Computer Science Teaching and Learning is a Targeted Math and Science Partnership between the University of California Los Angeles (UCLA) as the lead and the Los Angeles Unified School District (LAUSD) as the Core Partner school district. The Computer Science Teachers Association (CSTA) and the Association for Computing Machinery, Inc. are Supporting Partners. The Partnership promotes computational thinking, with an overarching goal of fostering inventiveness and innovation among students and teachers through increasing the computer science instructional capacity of high schools, especially in a large urban school district. This project brings together the Center for Embedded Networked Sensing (CENS), an NSF Science and Technology Center (STC) in the Henry Samueli School of Engineering and Applied Sciences, and Center X in the Graduate School of Education & Information Studies at UCLA. MOBILIZE deploys challenging and engaging hands-on computer science projects and curricula using new participatory sensing technologies in high school mathematics and science courses. MOBILIZE prepares a large number of high school teachers to use inquiry teaching methods with the intent that technologies that are ubiquitous with students today, such as mobile phones, peak the interests of students and motivate them in ways that ultimately increases both students' achievement and their identities as "doers" of science, while enhancing the computer science knowledge and pedagogical acumen of teachers.

MOBILIZE:
--creates exciting, challenging, multi-disciplinary, real-life based, cutting-edge, hands-on inquiry projects, tools, and materials for teaching computer science concepts in high school computer science AND in standards-based mathematics and science classes. Piloting occurs locally with dissemination nationally, with the long term goal of increasing student engagement and achievement in computer science, mathematics, and the other sciences. High school teachers work with STEM and education faculty to develop new computer science materials that build on the CENS Participatory Sensing systems, which involve students in observing and analyzing environmental and social processes where they live, work, and play;
--develops an innovative model of professional development for current and future high school teachers around the implementation of these projects, to include multi-disciplinary teams of teachers organized into learning communities with STEM and Education faculty, coaching, and a new pre-service computer science methods course, in order to create a cadre of teachers with expertise in both computer science content and pedagogy; and
--informs state and national policy changes (including changing the academic status of computer science from vocational to academic and establishing a computer science teaching credential) necessary to improve the quality of high school computer science instruction.

The research hypotheses explored focus on: students as computer scientists, whether in mathematics, science or computer science courses; teachers as computer scientists, whether teaching mathematics, science or computer science; and impact on pre-service teachers in terms of their likelihood to utilize inquiry-based instructional strategies during their early years of teaching.

Cognitive diagnosis models have received increasing attention within educational and psychological measurement. The popularity of these models largely may be due to their perceived ability to provide useful information concerning both examinees and test items. However, the validity of such information may be undermined when diagnostic models are misspecified. This project focuses on one aspect of model misspecification: violations of the local item independence assumption. It examines potential causes and consequences of such dependence, with particular attention to those causes unrelated to the attributes a diagnostic test is intended to measure. The project proposes and evaluates a hierarchical diagnosis model as an alternative to traditional diagnosis models in which nuisance dependence is ignored. This model maintains the desirable properties of existing models while allowing for greater complexity in the underlying response process. Importantly, the model may be estimated efficiently, even for models with a large number of nuisance latent variables, using an analytical dimension reduction technique described by Gibbons and Hedeker (1992).

There is growing interest in extracting model-based diagnostic information from assessments in order to provide more useful feedback to stakeholders. Up to this point, however, the question of whether traditional cognitive diagnosis models fit real test data has been somewhat neglected. This project examines the issue of model fit and presents a model that may better account for certain causes of misfit than the traditional diagnosis models. To the extent that the proposed framework better accounts for the structure of real test data, its application will contribute to improved test development and lead to more valid model-based diagnostic inferences, such as the classification of test takers according to cognitive attributes or skills. This, in turn, is expected to enhance decision making, as better diagnostic information may allow for more effective (or better targeted) delivery of instructional strategies or clinical interventions. The results from this project will benefit the psychometric practice in any social and behavior science discipline that involves testing and measurement. As a Doctoral Dissertation Research Improvement award, support is provided to enable a promising student to establish a strong, independent research career.