Yasamin Mostofi - US grants

The objective of this project is to build a design framework for robust functionality of mobile cooperative robotic router networks in realistic communication environments. Most of the existing work on mobile cooperative decision-making uses oversimplified models for communication links by assuming either a perfect link or no link between any two agents. Such models do not embrace realistic wireless communication effects such as fading, shadowing and path loss and therefore will not be suitable for robust operation of robotic router networks. We propose novel motion-planning strategies where the agents constantly reconfigure themselves to optimally route the information through the network. In addition, the theoretical results and algorithms will be implemented on actual autonomous vehicles engaged in a wide range of tasks. By considering realistic communication unreliability, our proposed framework will result in the robust flow of information in the network. Emergency response, security and surveillance are examples of applications that rely on robust and intelligent operation of autonomous networks in harsh conditions and can benefit tremendously from the proposed work. This project also has a significant educational impact on the female high school students of Albuquerque by organizing learning summer camps at the University of New Mexico.

The research objective of this Faculty Early Career Development (CAREER) project is to develop the foundations of sensing and navigation in mobile cooperative networks from a compressive sampling perspective. A mobile ooperative network faces an abundance of information in its environment. Since there is not enough time for direct measurement of the whole terrain, finding the fundamental minimum sensing required for high-integrity and cooperative reconstruction of the parameter of interest is considerably important and an open problem. Currently, there is no analysis and design theory for ooperative mapping based on a severely under-determined data set. onsequently, a avigation framework that can guide the vehicles to the locations better for sparse sensing is also lacking. Inspired by the recent breakthroughs in non-uniform sampling theory, this proposal shows how the network can exploit the sparse transformation of the parameter of interest for cooperative mapping based on a considerably small observation set. The proposed approach provides an answer to the question of "the next best positions for sensing" in mobile networks and guides the vehicles to the locations that are better for compressed sensing/mapping. To ensure uninterrupted cooperation, it furthermore complements this by proposing a foundation for communication-aware compressive mapping. Along this line, it shows how to build realistic communication objectives that are reflective of communication unreliability such as path loss, shadowing and fading, and integrate them with compressive sensing/mapping objectives. The proposed research is fundamental in nature as it seeks to unveil the minimum sensing and communication needed for the robust operation of cooperative mobile networks.

If successful, the proposed research will make a significant contribution to the understanding and optimization of mobile cooperative networks in realistic communication environments. Emergency response, exploratory missions, security and surveillance are a few examples of the applications that have to operate in an information-rich environment robustly and in a timely manner, and can therefore benefit tremendously from the proposed work. This proposal also has a significant educational impact on the Native Americans of New Mexico through partnership with Southwestern Indian Polytechnic Institute (a Native American community college) and Bernalillo High School (with 46% Native American and 46% Hispanic students). More specifically, it will l) contribute to stablishing a learning platform for the basics of control and communications at Southwestern Indian Polytechnic Institute, 2) create a echnical/educational remote site at Bernalillo High School and 3) develop a networked control course geared towards the aforementioned community college and high school students, in which these students will team up with UNM graduate students to do projects related to the learning platform.

Robotic networks can have a tremendous impact in many different areas such as disaster relief, emergency response, and national security. The recent disasters such as Hurricane Sandy of 2012 or Japan's earthquake of 2011 remind us of the crucial role that unmanned autonomous networks can play as part of our society. The goal of this project is to introduce a new multi-disciplinary design paradigm for the successful operation of mobile robotic networks through the co-optimization of sensing, communications and navigation. In the robotics/control community, most existing work does not deal with realistic communication issues (such as shadowing and multipath fading) and ideal links/disk models are assumed for predicting connectivity. On the other hand, the communication and networking communities are not typically concerned with path planning and navigation. In a robotic network, path planning not only affects sensing quality but also impacts connectivity maintenance. This multi-disciplinary nature makes designing robust decision-making strategies for a successful task accomplishment in robotic networks considerably challenging and an open problem. Furthermore, a separate optimization of the given sensing, communications and navigation resources may not suffice for a successful operation under resource constraints.

In this research effort, the focus is on the impact of limited energy (both motion and communications), time, and bandwidth resources and on laying the foundation of the corresponding optimum sensing, communication and navigation co-design policies, which includes trajectory, sensing, connectivity, motion speed/power, and communication transmission rate/power optimization. In this approach, realistic probabilistic connectivity metrics are properly co-optimized with sensing and navigation goals such that each robot chooses a trajectory that allows it to maximize its information gathering while maintaining the needed connectivity. This framework answers fundamental questions such as when to invest in motion and when to invest in communications. The project also addresses task feasibility and the fundamental limits of information generation, gathering and exchange, which can provide key insights for resource planning before deployment.

Overall, this new co-optimization foundation enables the successful operation of robotic networks under limited resources and can thus have a tremendous impact on our society. This project also has a significant educational impact on minority and under-represented students.

A team of unmanned vehicles can have tremendous impact in areas such as disaster relief, emergency response, and national security. Human operators are an integral part of these missions. Given the complexity of the tasks, harshness of the operation environments, and resource limitations, a fundamental understanding of how humans and robots can best interact is crucial to a successful operation. This proposal introduces a new multidisciplinary approach for efficient human-robot collaboration on tasks that require a visual search in the presence of uncertainty. The proposed methodology for characterizing human visual performance also contributes to the field of human perception and cognitive psychology. This proposal also has a significant educational component targeting under-represented students of the central California area and UC Santa Barbara.

A new methodology is proposed for predicting whether people can make a correct visual decision for a given sensory input. The approach is based on training convolutional neural networks using extensive input from Amazon Mechanical Turk workers. This foundational understanding of human visual performance will have significant implication for robotic field decision making. A new set of mathematical tools will be developed for robotic field operation, given a proper understanding of human visual capabilities. This enables a robot to properly decide when to seek human help, when to rely on itself, and when to sense more. Finally, the cost of communication is characterized, due to the space-varying nature of link quality, which can significantly affect seeking human help. The proposed theories and design paradigms will be validated using a robotic testbed.

The overall goal of this proposal is to introduce a new multi-disciplinary foundation for see-through imaging with everyday RF signals using unmanned autonomous vehicles. See-through imaging with everyday RF signals can considerably impact many different areas such as search and rescue operations, surveillance and security, detection/classification of occluded objects, infrastructure assessment, medical imaging, and archaeological exploration, just to name a few. Robotic networks, on the other hand, can have a tremendous impact in many different areas such as disaster relief, emergency response, environmental monitoring, surveillance, and security. This proposed work at the intersection of RF sensing and robotics can considerably advance the state-of-the-art in RF sensing by jointly and successively optimizing robotic path planning and RF imaging, and can thus have a transformative impact on our society. The proposal also has a significant educational component targeting under-represented students.

More specifically, in this research effort, a new multi-disciplinary paradigm is proposed to equip a number of unmanned vehicles with see-through imaging of completely unknown areas using everyday RF signals. Along this line, the first major task focuses on the interplay between motion patterns and RF imaging performance, in order to understand and mathematically characterize robotic motion patterns most informative for RF imaging. The second task then develops the foundation of jointly and successively co-designing the path planning and imaging of the robots while considering the dynamics of the vehicles, environmental navigation constraints, motion energy budget, and operation time. Finally, the proposed theories and design paradigm are extensively validated with ground and aerial vehicles. Overall, the proposed research can make a significant contribution to enhancing the state-of-the-art in both RF imaging and robotics.

Robust and accurate occupancy assessment is key to many applications. Optimization of heating/cooling in buildings, smart lighting, retail business planning, smart home applications, smart city planning, and evacuation/emergency planning are just a few examples. The overall goal of this proposal is to introduce a new paradigm for occupancy assessment using ubiquitous wireless signals. Not only are communication devices ubiquitous these days, but most smart devices whose main goal is not communication (e.g., smart speakers or thermostats) are now equipped with communication capabilities using cheap communication transceivers such as Wi-Fi or Bluetooth. This opens the possibility of using RF (Radio Frequency) signals for sensing and learning about the environment. This project proposes to use wireless links in a non-traditional manner: as a sensing mechanism for assessing the spatio-temporal occupancy dynamics in a large region. Spatio-temporal occupancy assessment refers to measuring several attributes of how people move and utilize their spaces, such as the total number of occupants in an area as a function of time, the average speed and dwell time of people, and the flow rate of people in between several areas, among other factors. More specifically, the project will develop the theoretical foundation and design tools for occupancy assessment with ubiquitous RF signals, with an emphasis on understanding the fundamental capabilities and limitations. The proposed framework does not rely on people to carry any device, has through-wall sensing capabilities, and preserves the privacy.

Accurate and robust occupancy assessment with everyday communication signals is considerably challenging. The proposed effort has three major tasks to address the underlying challenges. The first major goal brings a foundational understanding to how a single link can be turned into an occupancy assessment sensor whose capabilities and limitations are well characterized. More specifically, it proposes a mathematical probabilistic modeling that reveals the relationship between the key attributes of occupancy dynamics and the key attributes of the wireless measurements. The second major task then builds on the first one to develop the foundation of optimum joint processing of a number of links as an occupancy assessment sensor network. The third major task develops the mathematical tools for designing an efficient occupancy assessment system over a large region that consists of several areas, with only a small number of wireless links. Finally, the proposed theoretical foundation and design tools are extensively tested with several experiments in different areas, including local department stores.

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.

A robotic network can have a tremendous impact in many different areas such as emergency response, national security, and mobile service provisioning. In such systems, a team of unmanned vehicles are tasked with information gathering and cooperation to accomplish a given mission. Wireless communication is an integral part of such a system as an unmanned vehicle needs to connect to other nodes or to remote operators in order to transfer sensing data and/or receive control commands. Since a robot?s path directly affects its link quality, an unmanned vehicle needs to take the communication quality into account during path planning. This then requires an unmanned vehicle to have an accurate prediction of its link quality at unvisited locations. Accurately predicting the channel quality, however, is a considerably challenging problem and an open one, due to the high spatial variations of the channel power. This project enables the robust cooperation of a team of unmanned vehicles by developing new predictive wireless channel models and its corresponding co-optimization with sensing and path planning, which can significantly impact the area of robotics. The project also has an educational component, leveraging appeals of both robotics and communications, to reach out to the young minds.

This project develops a new multi-disciplinary paradigm for the robust networked operation of unmanned vehicles. The first research task shows it possible for the robot to reveal the makeup of the incoming wireless rays, based only on its onboard received channel power measurements. More specifically, it proposes a novel combination of path planning and information processing that can reveal the makeup of the rays, enabling the robot to learn considerably more about the channel than the measured received power. This new approach has a significant implication for robotic field operation as the project addresses. More specifically, the second major task builds on the first one in order to develop a new framework for robotic channel prediction. The third major task then develops a multi-disciplinary design that properly co-optimizes the new predictive channel models with sensing and path planning decisions, in order to enable the robust networked operation of unmanned vehicles. Finally, the overall new design paradigm is tested on a robotic testbed.

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.

Neurological/brain-related gait disorders affect many people and can be devastating to the mobility, independence, cognition, and self-esteem of an individual. Thus, early diagnosis and proper monitoring is key to giving an individual the proper care that can optimize their well-being. Yet, many individuals seek a medical opinion only after the disease has progressed for a while. Furthermore, once medical help is sought, many patients miss follow-up appointments to monitor the progress of their disease under therapy/medication. The situation can be much worse in impoverished/developing nations (or rural areas in US), due to the high cost of healthcare and/or lack of it for an average family. Wireless signals, on the other hand, are ubiquitous these days, which open up the possibility of using them for sensing and learning about the environment. This is the main motivation for the proposed work, to introduce a new mathematical foundation and design methodology that enables everyday RF signals, such as WiFi, to detect, classify, and monitor gait disorders. The proposed work can result in a considerable advancement towards developing an affordable home health system for gait disorder assessment. Such a system can further work in partnership with medical professionals, for the diagnosis and monitoring of gait disorders. The project also has an educational component, targeting K-12 as well as under-represented groups.<br/><br/>This research proposes a new foundation for gait disorder assessment with ubiquitous RF signals. This is a considerably challenging problem, which is divided into four major tasks. The first major goal proposes a new methodology that can translate the vast already-available online non-RF gait disorder datasets to RF data, enabling the creation of a large synthetic gait disorder RF datasets pertaining to different gait disorders. Such datasets are necessary for a methodical analysis/design, but currently lacking and prohibitively laborious to manually collect. The second major task then proposes a new processing foundation that can mathematically characterize, for the first time, the general full frequency content of the received signal, and its high-energy epochs, enabling proper analysis of more complex movements. The third objective then builds on the previous two to extract rich, efficient, and meaningful features from the received signal and design a robust machine leaning pipeline for the detection, classification, and monitoring of gait disorders, with an emphasis on understanding the feasibility and limitations. Finally, the last objective validates the proposed foundation with extensive experiments.<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.

Advances in low-cost low-power silicon radio frequency (RF) integrated circuits (ICs) in the last two decades have opened up the commercial applications for millimeter wave (mmWave) frequencies which are an order of magnitude beyond those used in WiFi and cellular today. Large-scale deployment of mmWave communication networks, such as NextG cellular infrastructure outdoors and NextG WiFi infrastructure indoors, implies that these resources can be leveraged for RF imaging at scales that are not otherwise possible. The project develops foundational algorithms, architectures and protocols for such Joint Communication and Imaging (JCAI) systems. Each sensor in such a system provides 4D measurements (range, Doppler, azimuth angle and elevation angle) whose resolution improves by going to higher frequencies. The project establishes US leadership in a critical technology by developing large-scale RF imaging using frequencies beyond 100 GHz. Outdoor applications include pedestrian and vehicular tracking for global situational awareness supporting vehicular autonomy, and addressing security challenges such as timely detection of illegal drones or unauthorized personnel. In indoor settings, the technology enables fine-grained inference/prediction of human actions for eldercare and smart home applications. RF imaging technologies are especially useful in low-light or high-smoke/fog conditions when visible light or infrared technologies are not effective.<br/><br/>The project develops and demonstrates a framework for JCAI at mmWave frequencies. A core aspect of the technical plan is to drastically improve resolution by synthesizing large apertures (Thrust 1). This employs a combination of novel approaches to single sensor design which utilize large antenna arrays developed for communication, and networked collaboration between multiple sensors. A complementary aspect (Thrust 2) is the strategic utilization of unmanned vehicles to image difficult-to-reach areas, utilizing the fixed infrastructure to reduce the robot payload. In Thrust 3, hardware at 140 GHz previously developed by the PIs for communication will be adapted to support demonstration of networked RF imaging at 100+ GHz. Thrust 4 develops a control plane for networked imaging, including a resource management framework based on imaging demand and imaging capacity, and protocols supporting collaborative imaging. The concepts and methods to be developed have potential impact in a vast array of applications, including vehicular autonomy and road safety, manufacturing automation, indoor and outdoor security, eldercare, and healthcare. The PIs will work closely with industry partners, building on their strong track record in transitioning mmWave research, and plan to incorporate this research into the undergraduate curriculum through courses, capstone projects, and REU projects.<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.