George J. Pappas - US grants

Embedded systems, such as controllers in automotive, medical, and avionic systems, consist of a collection of interacting software modules reacting to and controlling an analog environment. Engineering disciplines such as control theory focus on continuous dynamics, and offer foundations for designing robust control laws for ensuring optimal performance of dynamical systems. Computing disciplines such as software engineering focus on discrete programs, and offer structured ways of implementing complex control and analysis tools for validating distributed software. For networked embedded devices with multiple modes of operation, the combination of the complexity in both discrete and continuous aspects leads to fundamental problems that are not yet well understood, and this makes the programming of reliable embedded systems a particularly challenging task. A systematic approach to designing embedded devices requires combining tools from control theory and modern software engineering, and the emerging theory of hybrid systems---systems with tightly integrated discrete and continuous dynamics, has the potential to provide the foundation. Despite the great appeal of hybrid systems as a model, the applicability of the state-of-the-art analysis and design techniques for hybrid systems has been limited to examples of small size due to complexity. This ITR research aims to develop foundations and tools for automatic abstraction and hierarchical decomposition as a means of simplification and scalability.

To facilitate high-level design of embedded software, modeling concepts such as hierarchy, modularity, reuse, compositionality, and object-orientation, are explored to develop a theory of hierarchical hybrid systems with an accompanying a compositional calculus of refinement. This will be the basis for behavioral interfaces and descriptions of components at different levels of abstractions. For rigorously specifying and evaluating design alternatives and correctness requirements, automated techniques such as model checking are very effective. To apply these techniques for formal analysis of hybrid systems, this research is developing automated schemes for constructing abstractions of hybrid models. The technical directions being pursued include model checking algorithms that exploit hierarchy, algorithms for extracting finite-state approximations using predicate abstraction, counter-example guided refinement of abstractions, property-preserving bisimulation-based reductions of continuous differential equations, and assume-guarantee reasoning. The results of this research are being integrated in software tools for modeling and analysis of hybrid systems. The benefits of the techniques for developing embedded systems with higher assurance for safety and reliability are evaluated in an experimental testbed of multiple, autonomous, mobile robots.

Engineering - Mechanical (56)

The project is modeled on successful existing implementations of undergraduate robotics efforts at MIT and Swarthmore, but with a specific emphasis on the freshman and sophomore experience. The investigators are purchasing equipment to develop and implement the Laboratory for Undergraduate Robotics Education (LURE), permitting them to develop new, technologically advanced laboratory space for undergraduate education. The laboratory allows them to change to a mode of teaching that provides analysis, design, and manufacturing skills in a robotics setting. The equipment requested also permits them to inject engineering content with a hands-on laboratory component into the curriculum at an early stage (freshman year). This provides some perspective and motivation to beginning students, who currently receive the impression that engineering consists only of theoretical physics and mathematics. In the evaluation study, they are investigating how the differences in background preparation and training of incoming students affect development for high-tech courses related to robotics. They are developing and disseminating robotic-related curricular materials for use both in interdisciplinary college-level education, as well as K-12 outreach programs.

EIA-0130797 -Harvey Rubin-University of Pennsylvania-Modeling and Analysis of Biological and Information Networks

The overall goals of our research are to: 1) create enabling technologies and experimental systems that are necessary to understand and predict the integrated functions of two bacterial sensing and regulatory networks--porim osmo-regulation in E.coli and oxygen sensing regulation of DNA synthesis in Mycobacterium tuberculosis; 2) model and abstract principles of organization, design control and coordination of biological systems. We believe that a better understanding of networked, hybrid models in biology will provide deeper insights into networked, embedded systems. No systematic approach to designing and developing such hybrid systems exists today.

Our research on the porin osmo-regulatory system in E. coli will investigate crosstalk between the porin osmo-regulatory system and other signaling systems. We suggest that the ability of the sensing element of the system, EnvZ, to act as both a kinase and phosphatase is crucial for the control of information flow and to minimize crosstalk. We will extend our models in a related series of experiments on the PhoQ/PhoP two component systems, which responds to changes in the extracellular magnesium concentration. Since the levels of the histidine kinase PhoQ and response regulator PhoP are modulated by the concentration of phosphorylated PhoP, we will be able to establish the effect of this feedback and its influence on robustness of the overall system behavior.

Proposal Title: PECASE: Hierarchical Abstractions of Hybrid Systems
Institution: University of Pennsylvania

Recent years have witnessed the availability of extremely powerful and inexpensive computers, the explosion of communication networks, and revolutions in sensor and actuator technology. These accelerating advances have found their way inside many physical systems such as cars, aircraft, chemical processes, power networks, micro-mechanical systems, and manufacturing systems, resulting in the so-called embedded control systems or embedded software systems which merge physical processes with information systems. These application domains are important not only from an engineering perspective but also because of their prevalence in everyday life and the well-being of the economy.
Designing efficient, yet robust, large-scale embedded systems is a significant systems challenge. It is of immediate importance to develop the research foundations and educational programs to enable next-generation engineers to optimally utilize the opportunities offered by information technology. The difficulty with embedded systems arises from the fact that the software design and the control design is highly decoupled. As a result, physical constraints, such as differential equations, are not taken explicitly into account in the software design process. Consequently, fundamentally novel approaches to the design of embedded systems are needed as well as the development of models and tools that address the analysis and design of the integrated system with its many different physical, functional and logical aspects.
The research discipline of hybrid systems provides a theoretical foundation for the modeling, analysis, and design of embedded systems. Hybrid systems naturally combine discrete-event and continuous-time systems in a manner that can capture software logic, physical dynamics, and communication protocols, in a unified modeling framework. Hybrid systems have been used as mathematical models for automated highway systems, air traffic management systems, embedded automotive controllers, manufacturing systems, chemical processes, and, more recently, biomolecular networks. The wide applicability of hybrid systems has inspired a great deal of research from both control theory and theoretical computer science.
Despite the great success of hybrid systems as a model, the applicability of state-of-the-art analysis and design techniques for hybrid systems has been limited to examples of small size due to complexity. One of the fundamental approaches for reducing complexity involves the use of hierarchical decomposition. One of the main challenges in hierarchical systems is the extraction of a hierarchy of models at various levels of abstraction which are compatible with the functionality and objectives of each layer. Hierarchies have been instrumental in separately managing the complexity of both control and software designs. However, for hybrid systems which re-integrate software and control, the right notion of hierarchy is a great challenge, and is the critical obstacle for scaling our models, theories, algorithms, and tools to large scale, embedded systems.
Next-generation, large-scale embedded systems have motivated new control as well as software paradigms. As a result, there are numerous opportunities for fundamental and significant contributions from both a theoretical and an applied perspective. In this exciting intellectual landscape, the research and educational agenda of the proposed research focuses on developing the theoretical foundations for the hierarchical decomposition of hybrid systems at various levels of abstraction. The long term goal of the research agenda will address the fundamental problem of given a class of hybrid models, and a class of properties that must be preserved, extract modeling abstractions that preserve the properties of interest. Achieving this goal will consist of first developing robust notions of bi-simulation for purely continuous systems, and then unifying the continuous and discrete notions in a manner that is consistent with the dynamics of hybrid systems.

Merging discrete and continuous systems creates serious educational issues in order to combine the traditionally separate threads of discrete and continuous mathematics. These issues must be addressed at all educational levels. At the graduate level, creating new courses on hybrid systems are needed, but with focus on complexity reduction methods in both control theory and computer science. However, given the one-sided backgrounds of most undergraduate students, what is needed at the undergraduate level, is a cross-departmental, embedded systems course that will highlight both the discrete and continuous nature of embedded systems in various application domains. This will allow the earlier uniformization of educational backgrounds, expose students to related application domains in other departments, as well as demonstrate the cross-disciplinary power of the models, methods, and tools.

This project was originally funded as a CAREER award, and was converted to a Presidential Early Career Award for Engineers and Scientists (PECASE) award in May 2004.

This project, a collaborative with 03-24864 Papanikolopoulos at University of Minnesota, and 03-25017 Joel Burdick at California Institute of Technology, addresses research issues key to an important application of robot teams and information technology (emergency response in hazardous environments for various tasks). The research sets 6 goals:
Development of new algorithms that enable collaborative sensing.
Development of distributed localization/mapping methods that leverage capabilities of the heterogeneous robots.
In-depth study of communication issues with emphasis on transparent integration of ultra wideband communication methodologies.
Development of methods for team coordination and dynamic distribution of tasks to robots.
Creation of algorithms for the presentation of sensory information to users.
Experimental validation of the scalability of the aforementioned algorithms and techniques.
The PIs use the Scout and MegaScout robotic platforms designed at the University of Minnesota along with other testbeds at CalTech and U Penn to conduct the research. The project integrates the algorithms with first responder teams, emphasizing realistic scenarios; mentors students from underrepresented groups in order to retain them in CS/EE programs; conducts outreach activities through demonstrations at local schools and youth groups; conducts workshops that emphasize cross-disciplinary interaction; creates web resources; innovates classroom uses of multi-robot teams; and includes parts of the research in design projects for seniors. The project also includes international collaboration with groups at NTUA (Greece) and the University Louis Pasteur-Strasbourg I (France).

LGORITHMIC SYNTHESIS OF EMBEDDED CONTROLLERS

In recent years, microprocessors have invaded the physical world, resulting in
sophisticated but complicated networked embedded systems that defy theoretical
understanding, resulting in inadequate design tools for the modern embedded
system engineer. Embedded systems require the development of new theories,
methods, and tools providing the correct understanding of these systems and
accelerated analysis and design methods in order to ensure safety
but also substantially decrease their design time. Embedded systems require very novel, very challenging specifications that have
to deal with synchronization, sequencing, and temporal ordering of different
tasks. Mathematically formulating such desired specifications cannot be
achieved using traditional mathematical formulations in control theory. On the
other hand, computer aided verification has popularized the use of several
temporal logics to describe complex specifications. However, the emphasis has
been on verification of these properties for purely discrete systems, and not
on synthesis for systems with a continuous component.

In this research, a novel approach for automatically synthesizing hybrid
controllers is pursued in order to satisfy specifications that are expressed
in temporal logics. In particular, methodologies are developed to extract
finite abstractions of linear control systems, that will be used to design
controllers meeting the desired temporal logic specifications. Contrary to
other approaches, this project considers specification dependent abstractions
for continuous control systems as opposed to continuous dynamical systems. The
proposed framework will provide algorithms and tools for the computation of
discrete controllers, which by refinement will lead to embedded, hybrid
controllers for the original system while providing performance and
correctness guarantees.

The educational agenda of the proposal focuses on the development of a cross-
departmental, undergraduate, signals and systems course that broadens the
definition of systems in order to capture software and hardware systems in
addition to traditional control or communication systems. This course has been
carefully coordinated with recent curriculum changes and aims at the
educational uniformity of signals and systems concepts in both the discrete and
the continuous world.

The HSCC 2004 {Seventh International Workshop on Hybrid Systems: Computation and Control, and a satellite workshop devoted to foundations and applications of Robustness, Abstractions and Computations, were held March 25-28, 2004 in Philadelphia on the campus of the University of Pennsylvania. The annual workshop on hybrid systems attracts researchers from academia and industry interested in modeling, analysis, and implementation of dynamic and reactive systems involving both discrete and continuous behaviors. The emphasis is on the interdisciplinary nature of the topic, bringing together researchers in computer science and control engineering. HSCC 2004 includes tracks on the use of hybrid systems in computational biology and a special session devoted to foundations of Robustness, Abstractions and Computation.

NSF support enables participation by U.S. graduate and undergraduate students in hybrid systems.

The workshops supported by NSF as part of the Robotics: Science and Systems conference will focus on human-oriented robotics including rehabilitation, exoskeleton and prosthetics, human-robot interaction, robots manipulating in human environments, and socially assistive robotics. The breadth of human-oriented robotics is large, and is becoming more important for the viable transfer of robot technology to real human applications as well as understanding the science of the interaction between the two. The workshops will bring together a large number of researchers in these related fields. The NSF support will enable students to attend these workshops, learning issues in these growing fields and training them to be researchers. Robotics: Science and Systems is a new conference that brings together researchers working on algorithmic or mathematical foundations of robotics, robotics applications, and analysis of robotic systems.

The University of Pennsylvania plans to become the fourth research site of the existing Industry / University Cooperative Research Center (I/UCRC) for Safety, Security, and Rescue Research Center (SSR-RC), composed of the University of South Florida as the lead university with the University of Minnesota as a partner. The University of Pennsylvania will follow the same policies as the existing center.

The main goal of the center and the research site is to create the infrastructure to facilitate engaging academic and industrial expertise for the direct and immediate benefit of our society at times and in situations during which it is most vulnerable. A planning meeting has been scheduled to determine the organization and viability of forming a research site for the existing I/UCRC.

As modern embedded systems gain more functionality and complexity, there is a need for a novel discipline for their design, development and deployment. In recent years, the idea of the model-based design paradigm is to develop design models and subject them to early analysis, testing, and validation prior to their implementation. Simulation-based testing ensures that a finite number of user-defined system trajectories meet the desired specification. Even though computationally inexpensive simulation is ubiquitous in system design, it suffers from incompleteness, as it is impossible or impractical to test all system trajectories. On the other hand, verification methods enjoy completeness by showing that all system trajectories satisfy the desired property. For embedded hybrid systems with an infinite number of possible behaviors, exhaustive verification seems to be very hard, and simulation-based testing seems to provide no confidence in our system design. In addition to the gap between testing and verification for embedded systems, there is even a more fundamental, and largely unaddressed, challenge. Uncertainty in the environment, errors in physical devices make overall system robustness one of most critical yet least understood challenges in embedded systems. There is a clear intellectual opportunity for laying the scientific foundations and developing methods and algorithms for analyzing and testing the robustness and safety of embedded hybrid systems. This project brings together leading experts in embedded control, hybrid systems, and software monitoring and testing to develop the foundations of a modern framework for testing the robustness of embedded hybrid systems. The central idea that this proposal is centered around is the notion of a robust test, where the robustness of nominal test can be computed and used to infer that a tube of trajectories around the nominal test will yield the same qualitative behavior. By computing the robustness margins of tests, this project explores how to infer how robust each test is, guide subsequent tests, estimate the robustness for the system, as well provide well-defined coverage metrics using finite number of tests. In addition, this project emphasizes cross-cutting, multi-departmental education of graduate students and emphasizes testing and robustness for embedded hybrid systems in relevant electrical engineering and computer science courses. The educational agenda is to expose computer science students to notions of robustness, and control students to software testing algorithms.

The University of Pennsylvania has joined the multi-university Industry/University Cooperative Research Center for Safety, Security and Rescue Research located at the University of South Florida and the University of Minnesota. The I/UCRC will bring together industry, academe, and public sector users together to provide integrative robotics and artificial intelligence solutions in robotics for activities conducted by the police, FBI, FEMA, firefighting, transportation safety, and emergency response to mass casualty-related activities.

The need for safety, security, and rescue technologies has accelerated in the aftermath of 9/11 and a new research community is forming, as witnessed by the first IEEE Workshop on Safety, Security and Rescue Robotics. The Center is built upon the knowledge and expertise of multi-disciplinary researchers in computer science, engineering, industrial organization, psychology, public health, and marine sciences at member institutions.

This project seeks to address fundamental gaps in the understanding of collective behavior, and to develop a methodology for software design for cyber physical systems that use sensing, communication, and actuation to accomplish tasks that are well beyond the capabilities of individual units. Specifically, the focus is on methodologies that will allow cyber physical systems to adapt to changing environmental conditions and be resilient to disturbances and attacks, and tools to translate design specifications for the group to software design specifications for individual units by essentially solving the inverse problem for networked cyber physical systems.

This research represents the cross pollination of research in molecular, cell and population biology, systems modeling, control theory and robotics. It brings novel modeling approaches and recent results in systems biology to bear on the problem of designing and architecting cyber physical systems. Specifically, it will establish a framework for designing and realizing algorithms for real-time, aggregated networked systems across multiple time-scales, and help develop the foundation for high-confidence software for reconfigurable, adaptive and resilient systems.

The project is expected to lay the foundation for a new community of researchers that include biologists, control theorists and roboticists through research workshops.

The objective of this research is to develop the scientific foundation for the quantitative analysis and design of control networks. Control networks are wireless substrates for industrial automation control, such as the WirelessHART and Honeywell's OneWireless, and have fundamental differences over their sensor network counterparts as they also include actuation and the physical dynamics. The approach of the project focuses on understanding cross-cutting interfaces between computing systems, control systems, sensor networks, and wireless communications using time-triggered architectures.

The intellectual merit of this research is based on using time-triggered communication and computation as a unifying abstraction for understanding control networks. Time-triggered architectures enable the natural integration of communication, computation, and physical aspects of control networks as switched control systems. The time-triggered abstraction will serve for addressing the following interrelated themes: Optimal Schedules via Quantitative Automata, Quantitative Analysis and Design of Control Networks: Wireless Protocols for Optimal Control: Quantitative Trust Management for Control Networks.

Various components of this research will be integrated into the PIs' RAVEN control network which is compatible with both WirelessHART and OneWireless specifications. This provides a direct path for this proposal to have immediate industrial impact. In order to increase the broader impact of this project, this project will launch the creation of a Masters' program in Embedded Systems, one of the first in the nation. The principle that guides the curriculum development of this novel program is a unified systems view of computing, communication, and control systems.

The objective of this research is to establish a new development paradigm that enables the effective design, implementation, and certification of medical device cyber-physical systems. The approach is to pursue the following research directions: 1) to support medical device interconnectivity and interoperability with network-enabled control; 2) to apply coordination between medical devices to support emerging clinical scenarios; 3) to ?close the loop? and enable feedback about the condition of the patient to the devices delivering therapy; and 4) to assure safety and effectiveness of interoperating medical devices.
The intellectual merits of the project are 1) foundations for rigorous development, which include formalization of clinical scenarios, operational procedures, and architectures of medical device systems, as well as patient and caregiver modeling; 2) high-confidence software development for medical device systems that includes the safe and effective composition of clinical scenarios and devices into a dynamically assembled system; 3) validation and certification of medical device cyber-physical systems; and 4) education of the next-generation of medical device system developers who must be literate in both computational and physical aspects of devices.
The broader impacts of the project will be achieved in three ways. Novel design methods and certification techniques will significantly improve patient safety. The introduction of closed-loop scenarios into clinical practice will reduce the burden that caregivers are currently facing and will have the potential of reducing the overall costs of health care. Finally, the educational efforts and outreach activities will increase awareness of careers in the area of medical device systems and help attract women and under-represented minorities to the field.

Computers have revolutionized our daily lives, and yet the way we program computers has changed little in the last several decades. Software development still remains a tedious and error-prone activity. ExCAPE aims to change programming from a purely manual task to one in which a programmer and an automated program synthesis tool collaborate to generate software that meets its specification.

A distinguishing feature of the ExCAPE approach is that the program description can involve a wide range of artifacts that are best-suited to the particular development task: incomplete programs; declarative specifications of high-level requirements; positive and negative examples of desired behaviors; and optimization criteria for selecting among alternative implementations. This diversity is aimed at allowing a programmer flexibility to express insights through a variety of formats, leading to a more intuitive and less error-prone way of programming.

The synthesis tool uses a range of computational approaches and developer interaction to compose these different views about the structure and functionality of the system into a unified, concrete implementation. The computational techniques include decision procedures for constraint-satisfaction problems; iterative schemes for abstraction and refinement; and data-driven learning. The methodology for programmer interaction moves verification from the back-end of the design cycle to the front-end, with the promise of a more reliable software product.

To develop the theory and practice of the proposed paradigm, the ExCAPE team brings together expertise in theoretical foundations (computer-aided verification, control theory, program analysis), design methodology (human-computer interaction, model-based design, programming environments), and applications (concurrent programming, network protocols, robotics, system architecture). Research will focus on developing new computational engines for transformation and integration of synthesis artifacts, and effective methods for programmer interaction and feedback.

While the benefits of the ExCAPE approach will apply broadly to software development, the ExCAPE team will focus its efforts by initially targeting four challenge problems: developing efficient concurrent data structures; developing protocols for on-chip interconnection networks; developing distributed routing network protocols; and end-user programming for autonomous robots. The ExCAPE approach will be a radical departure from the way these problems are solved today. For example, for the challenge problem on concurrent programming, the planned design tool will provide smart assistance for expert programmers to produce efficient and correct code, while the proposed tool for the robotics challenge problem will let end users program robots by demonstrating example behaviors. As ExCAPE aims to affect industrial practice, design tools for all four challenge problems will be developed and evaluated in close collaboration with industrial partners.

The technology developed by ExCAPE also has the potential to revolutionize the way computing concepts are taught. Building on the core technology used in program synthesis, the ExCAPE team plans to develop smart tutoring software that can analyze students? answers for conceptual errors and generate additional problems tailored to that student.. This tutoring software will be developed for representative high-school and undergraduate courses and will be made widely available. This outreach effort is aimed at attracting more students to computing disciplines by promoting a new and more appealing vision of what it means to program. ExCAPE will also nurture an inter-disciplinary community of researchers in computer-augmented programming, via an annual workshop, a biannual summer school, and a competition for synthesis tools, with associated challenge problems and benchmarks.

For more information visit http://excape.cis.upenn.edu

Cyber physical systems (CPSs) are merging into major mobile systems of our society, such as public transportation, supply chains, and taxi networks. Past researchers have accumulated significant knowledge for designing cyber physical systems, such as for military surveillance, infrastructure protection, scientific exploration, and smart environments, but primarily in relatively stationary settings, i.e., where spatial and mobility diversity is limited. Differently, mobile CPSs interact with phenomena of interest at different locations and environments, and where the context information (e.g., network availability and connectivity) about these physical locations might not be available. This unique feature calls for new solutions to seamlessly integrate mobile computing with the physical world, including dynamic access to multiple wireless technologies. The required solutions are addressed by (i) creating a network control architecture based on novel predictive hierarchical control and that accounts for characteristics of wireless communication, (ii) developing formal network control models based on in-situ network system identification and cross-layer optimization, and (iii) designing and implementing a reference implementation on a small scale wireless and vehicular test-bed based on law enforcement vehicles.

The results can improve all mobile transportation systems such as future taxi control and dispatch systems. In this application advantages are: (i) reducing time for drivers to find customers; (ii) reducing time for passengers to wait; (iii) avoiding and preventing traffic congestion; (iv) reducing gas consumption and operating cost; (v) improving driver and vehicle safety, and (vi) enforcing municipal regulation. Class modules developed on mobile computing and CPS will be used at the four participating Universities and then be made available via the Web.

Security and privacy concerns in the increasingly interconnected world are receiving much attention from the research community, policymakers, and general public. However, much of the recent and on-going efforts concentrate on security of general-purpose computation and on privacy in communication and social interactions. The advent of cyber-physical systems (e.g., safety-critical IoT), which aim at tight integration between distributed computational intelligence, communication networks, physical world, and human actors, opens new horizons for intelligent systems with advanced capabilities. These systems may reduce number of accidents and increase throughput of transportation networks, improve patient safety, mitigate caregiver errors, enable personalized treatments, and allow older adults to age in their places. At the same time, cyber-physical systems introduce new challenges and concerns about safety, security, and privacy. The proposed project will lead to safer, more secure and privacy preserving CPS. As our lives depend more and more on these systems, specifically in automotive, medical, and Internet-of-Things domains, results obtained in this project will have a direct impact on the society at large. The study of emerging legal and ethical aspects of large-scale CPS deployments will inform future policy decision-making. The educational and outreach aspects of this project will help us build a workforce that is better prepared to address the security and privacy needs of the ever-more connected and technologically oriented society.

Cyber-physical systems (CPS) involve tight integration of computational nodes, connected by one or more communication networks, the physical environment of these nodes, and human users of the system, who interact with both the computational part of the system and the physical environment. Attacks on a CPS system may affect all of its components: computational nodes and communication networks are subject to malicious intrusions, and physical environment may be maliciously altered. CPS-specific security challenges arise from two perspectives. On the one hand, conventional information security approaches can be used to prevent intrusions, but attackers can still affect the system via the physical environment. Resource constraints, inherent in many CPS domains, may prevent heavy-duty security approaches from being deployed. This proposal will develop a framework in which the mix of prevention, detection and recovery, and robust techniques work together to improve the security and privacy of CPS. Specific research products will include techniques providing: 1) accountability-based detection and bounded-time recovery from malicious attacks to CPS, complemented by novel preventive techniques based on lightweight cryptography; 2) security-aware control design based on attack resilient state estimator and sensor fusions; 3) privacy of data collected and used by CPS based on differential privacy; and, 4) evidence-based framework for CPS security and privacy assurance, taking into account the operating context of the system and human factors. Case studies will be performed in applications with autonomous features of vehicles, internal and external vehicle networks, medical device interoperability, and smart connected medical home.

The purpose of this grant is to plan and lay the groundwork for a new Electrical and Computer Engineering (ECE) Research Community organization. Over the past eighteen months consensus and critical mass have developed around the need for a stronger collective voice to advance the national agenda in research and innovation in areas of importance to ECE Departments. This national discussion was led by ECEDHA, the Electrical and Computer Engineering Department Heads Association, which represents more than 300 ECE departments and more than 7,000 faculty in the nation. The goal of this project is to develop an organization that will continuously bring the ECE research community together and develop visioning activities that align the research frontier with national priorities.

The intellectual merit of this proposal is the collective development of a strategic vision and a compelling scientific and technological agenda that will impact research and education in all electrical and computer engineering departments across the nation. This project will also support a workshop where the ECE community will deliberate on future grand challenges that align with national priorities, and will define the research frontier of ECE departments. The investigators will also set the roadmap for a fiscally sustainable ECE Research Community Organization that will ensure solutions, which are required to meet to grand challenges and for which the ECE community is a major and essential player, will continue to be found through wise investments from a scientific and technological perspective.

This effort is expected to have an enormous impact as the intellectual capital of ECE departments is collectively steered toward 21st century grand challenges for the nation. Through coordination with ongoing efforts on curriculum reform and branding of the discipline, this project will also address tremendous workforce development issues to support a larger and more diverse ECE workforce, which is now facing stronger competition from other fields.

The internet-of-things (IoT) revolution is bringing millions of physical devices online (e.g. cars, UAVs, homes, medical devices), enabling them to connect to each other in real-time, as well as to cloud services. Wireless communication will be critical in providing IoT connectivity. There are three main strategic directions that are envisioned in the future wireless networking. First, enhanced mobile broadband will result in even larger data rates for future 3D applications such as virtual reality. Second, Smart City/Community applications require large number of sensors that communicate sporadically over large urban or rural areas in a scalable, asynchronous, and energy efficient manner. The previous two directions, while important, are not the focus of our proposal. Instead, our proposal focuses on low-latency and ultra-reliable communications and networking that is critical for latency-sensitive, closed-loop control applications, like vehicle to vehicle communications, collaborative swam planning, and industrial control. In such latency sensitive applications, we need to rethink the networking stack, coding, networking architecture, and control design to enable communications and networking that can provide ultra-low latency (99.999%). This is far beyond what is currently possible. But even more importantly, we do not know what is possible and what are the fundamental limits for control system design over low-latency, high-reliability communications.

In this proposal, we will be rethinking the scientific foundations for ultra-reliable, low-latency wireless communications for latency sensitive control applications. We propose to achieve our scientific agenda by addressing three intellectual challenges: 1) Low-latency channel coding, where the goal is to focus on short packet codes for control loops 2) Control over low latency-aware communication channels, where the goal is to understand the what is the optimal tradeoff of latency to reliability for control loops and 3) Learning for Large Scale Wireless Control Networks, where machine learning will perform resource allocation for large numbers of control loops with competing latency/reliability requirements We intend to evaluate the proposed research agenda by leveraging our existing Intel Science and Technology Center (ISTC) on Wireless Autonomous Systems and demonstrate our ideas in future wireless protocols (IEEE 802.11ax) and experimentally demonstrate it in high-speed V2V and fast formation control with aerial swarms. On the educational front, the University of Pennsylvania is planning to offer a Micro-Masters program in Internet of Things (IoT) on the edX MOOC platform. Longer term, our goal is to create a new community of researchers that focus on control over low-latency wireless networks for IoT devices. Towards this goal, we plan on leveraging departmental efforts to increase and diversify the PhD students working on this project.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Studies show five of the top 10 most-gridlocked cities in the world are in the United States. Traffic congestion puts undue burden on transportation systems across the United States, raising transportation costs and the energy footprint. Vehicle automation creates an opportunity to reduce traffic and improve efficiency of the transportation infrastructure. In particular, this project aims to reduce the energy footprint of phantom traffic jams, where dense traffic comes to a halt for no apparent reason, and also stop-and-go-waves in congestion. The research team aims to reduce the overall energy footprint of stop-and-go congestion by up to 40% via a small portion of connected and autonomous vehicles (CAVs) inserted into normal traffic with drivers, also known as manned traffic. The work will build models of mixed autonomy (a combination of CAVs and manned traffic), and test the ability for this portion of CAVs to smooth the flow of traffic in a controlled manner, and thus reduce the energy footprint. The research combines mathematics, control theory, machine learning, and transportation engineering. The project includes four universities and engages industry and government partners. The project will also engage students and community stakeholders, including State and Federal transportation agencies and CAV manufacturers.

Specifically, the technical contributions enabling traffic smoothing and reduction in the environmental footprint include new mean-field optimal control formulations for sparse control settings where only a subset of vehicles are CAVs and can be controlled. Investigators will develop data-driven control algorithms based on deep reinforcement learning designed to enable control in settings where analytical approaches to derive explicit controllers are too complex (e.g., due to multi-lane, ramps, and high variation of human driving styles). They will also develop tools based on Satisfiability Modulo Convex optimization to enable safety and robustness of these controllers. The approach will first be validated using microsimulation tools to assess their efficiency and their validity. Once validated in simulation, the project will then field test the algorithm with manned vehicles following real-time control commands of the system, executed by 100 human drivers following control signals communicated via a phone app with target speeds and lanes. After which, the system will be tested with up to 20 CAVs inserted onto a freeway stretch in the Transition to Practice component of the project.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Recent advances in deep learning have led to many disruptive technologies: from automatic speech recognition systems, to automated supermarkets, to self-driving cars. However, the complex and large-scale nature of deep networks makes them hard to analyze and, therefore, they are mostly used as black-boxes without formal guarantees on their performance. For example, deep networks provide a self-reported confidence score, but they are frequently inaccurate and uncalibrated, or likely to make large mistakes on rare cases. Moreover, the design of deep networks remains an art and is largely driven by empirical performance on a dataset. As deep learning systems are increasingly employed in our daily lives, it becomes critical to understand if their predictions satisfy certain desired properties. The goal of this NSF-Simons Research Collaboration on the Mathematical and Scientific Foundations of Deep Learning is to develop a mathematical, statistical and computational framework that helps explain the success of current network architectures, understand its pitfalls, and guide the design of novel architectures with guaranteed confidence, robustness, interpretability, optimality, and transferability. This project will train a diverse STEM workforce with data science skills that are essential for the global competitiveness of the US economy by creating new undergraduate and graduate programs in the foundations of data science and organizing a series of collaborative research events, including semester research programs and summer schools on the foundations of deep learning. This project will also impact women and underrepresented minorities by involving undergraduates in the foundations of data science.

Deep networks have led to dramatic improvements in the performance of pattern recognition systems. However, the mathematical reasons for this success remain elusive. For instance, it is not clear why deep networks generalize or transfer to new tasks, or why simple optimization strategies can reach a local or global minimum of the associated non-convex optimization problem. Moreover, there is no principled way of designing the architecture of the network so that it satisfies certain desired properties, such as expressivity, transferability, optimality and robustness. This project brings together a multidisciplinary team of mathematicians, statisticians, theoretical computer scientists, and electrical engineers to develop the mathematical and scientific foundations of deep learning. The project is divided in four main thrusts. The analysis thrust will use principles from approximation theory, information theory, statistical inference, and robust control to analyze properties of deep networks such as expressivity, interpretability, confidence, fairness and robustness. The learning thrust will use principles from dynamical systems, non-convex and stochastic optimization, statistical learning theory, adaptive control, and high-dimensional statistics to design and analyze learning algorithms with guaranteed convergence, optimality and generalization properties. The design thrust will use principles from algebra, geometry, topology, graph theory and optimization to design and learn network architectures that capture algebraic, geometric and graph structures in both the data and the task. The transferability thrust will use principles from multiscale analysis and modeling, reinforcement learning, and Markov decision processes to design and study data representations that are suitable for learning from and transferring to multiple tasks.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Consider two future autonomous system use-cases: (i) a bomb defusing rover sent into an unfamiliar, GPS and communication denied environment (e.g., a cave or mine), tasked with the objective of locating and defusing an improvised explosive device, and (ii) an autonomous racing drone competing in a future autonomous incarnation of the Drone Racing League. Both systems will make decisions based on inputs from a combination of simple, single output sensing devices, such as inertial measurement units, and complex, high dimensional output sensing modalities, such as cameras and LiDAR. This shift from relying only on simple, single output sensing devices to systems that incorporate rich, complex perceptual sensing modalities requires rethinking the design of safety-critical autonomous systems, especially given the inextricable role that machine and deep learning play in the design of modern perceptual sensors. These two motivating examples raise an even more fundamental question however: given the vastly different dynamics, environments, objectives, and safety/risk constraints, should these two systems have perceptual sensors with different properties? Indeed, due to the extremely safety critical nature of the bomb defusing task, an emphasis on robustness, risk aversion, and safety seems necessary. Conversely, the designer of the drone racer may be willing to sacrifice robustness to maximize responsiveness and lower lap-time. This extreme diversity in requirements highlights the need for a principled approach to navigate tradeoffs in this complex design space, which is what this proposal seeks to develop. Existing approaches to designing perception/action pipelines are either modular, which often ignore uncertainty and limit interaction between components, or monolithic and end-to-end, which are difficult to interpret, troubleshoot, and have high sample-complexity.

This project proposes an alternative approach and rethinks the scientific foundations of using machine learning and computer vision to process rich high-dimensional perceptual data for use in safety-critical cyber-physical control applications. Thrusts will develop integration between perception, planning and control that allow for their co-design and co-optimization. Using novel robust learning methods for perceptual representations and predictive models that characterize tradeoffs between robustness (e.g., to lighting & weather changes, rotations) and performance (e.g., responsiveness, discriminativeness), jointly learned perception maps and uncertainty profiles will be abstracted as ``noisy virtual sensors? for use in uncertainty aware perception-based planning & control algorithms with stability, performance, and safety guarantees. These insights will be integrated into novel perception-based model predictive control algorithms, which allow for planning, stability, and safety guarantees through a unifying optimization-based framework acting on rich perceptual data. Experimental validation of the benefits of these methods will be conducted at Penn using photorealistic simulations and physical camera equipped quadcopters, and be used to demonstrate perception-based planning and control algorithms at the extremes of speed/safety tradeoffs. On the educational front, the research outcomes of this proposal will be used to develop a sequence of courses on safe autonomy, safe perception, and learning and control at the University of Pennsylvania. Longer term, the goal of this project is to create a new community of researchers that focus on robust learning for perception-based control. Towards this goal, departmental efforts will be leveraged to increase and diversify the PhD students working on this project.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

The proliferation of data-driven decision making, and its increased popularity, has fueled rapid emergence of data science as a new scientific discipline. Data science is seen as a key enabler of future businesses, technologies, and healthcare that can transform all aspects of socioeconomic lives. Its fast adoption, however, often comes with ad hoc implementation of techniques with suboptimal, and sometimes unfair and potentially harmful, results. The time is ripe to develop principled approaches to lay solid foundations of data science. This is particularly challenging as real-world data is highly complex with intricate structures, unprecedented scale, rapidly evolving characteristics, noise, and implicit biases. Addressing these challenges requires a concerted effort across multiple scientific disciplines such as statistics for robust decision making under uncertainty; mathematics and electrical engineering for enabling data-driven optimization beyond worst case; theoretical computer science and machine learning for new algorithmic paradigms to deal with dynamic and sensitive data in an ethical way; and basic sciences to bring the technical developments to the forefront of health sciences and society. The proposed institute for emerging CORE methods in data science (EnCORE) brings together a diverse team of researchers spanning the afore-mentioned disciplines from the University of California San Diego, University of Texas Austin, University of Pennsylvania, and the University of California Los Angeles. It presents an ambitious vision to transform the landscape of the four CORE pillars of data science: C for complexities of data, O for optimization, R for responsible learning, and E for education and engagement. Along with its transformative research vision, the institute fosters a bold plan for outreach and broadening participation by engaging students of diverse backgrounds at all levels from K-12 to postdocs and junior faculty. The project aims to impact a wide demography of students by offering collaborative courses across its partner universities and a flexible co-mentorship plan for truly multidisciplinary research. With regular organization of workshops, summer schools, and seminars, the project aims to engage the entire scientific community to become the new nexus of research and education on foundations of data science. To bring the fruit of theoretical development to practice, EnCORE will continuously work with industry partners, domain scientists, and will forge strong connections with other National Science Foundation Harnessing Data Revolution institutes across the nation.<br/><br/>EnCORE as an institute embodies intellectual merit that has the potential to lead ground-breaking research to shape the foundations of data science in the United States. Its research mission is organized around three themes. The first theme on data complexity addresses the complex characteristics of data such as massive size, huge feature space, rapid changes, variety of sources, implicit dependence structures, arbitrary outliers, and noise. A major overhaul of the core concepts of algorithm design is needed with a holistic view of different computational complexity measures. Faced with noise and outliers, uncertainty estimation is both necessary, and at the same time difficult, due to dynamic and changing data. Data heterogeneity poses major challenges even in basic classification tasks. The structural relationships hidden inside such data are crucial in the understanding and processing, and for downstream data analysis tasks such as in visualization and neuroscience. The second theme of EnCORE aims to transform the classical area of optimization where adaptive methods and human intervention can lead to major advances. It plans to revisit the foundations of distributed optimization to include heterogeneity, robustness, safety, and communication; and address statistical uncertainty due to distributional shift in dynamic data in control and reinforcement learning. The third and final theme of EnCORE proposes to build the foundations of responsible learning. Applications of machine learning in human-facing systems are severely hampered when the learned models are hard for users to understand and reproduce, may give biased outcomes, are easily changeable by an adversary, and reveal sensitive information. Thus, interpretability, reproducibility, fairness, privacy, and robustness must be incorporated in any data-driven decision making. The experience and dedication to mentoring and outreach, collaborative curriculum design, socially aware responsible research program, extensive institute activities, and industrial partnerships would pave the way for a substantial broader impact for EnCORE. Summer schools with year-long mentoring will take place in three states involving a large demography. Joint courses with hybrid, and fully online offerings will be developed. Utilizing prior experience of running Thinkabit lab that has impacted over 74,000 K-12 students so far, EnCORE will embark on an ambitious and thoughtful outreach program to improve the representation of under-represented groups and help create a future generation of workforce that is diverse, responsible, and has solid foundations in data science.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.