We are delighted to announce the CITRIS Seed Funding Program awards for 2017. This year, 45 highly competitive teams from the four CITRIS campuses at UC Berkeley, Davis, Merced, and Santa Cruz submitted proposals for collaborative research projects within the University of California system.
Eleven teams will receive a one-time award, averaging about $56,000 each, for interdisciplinary work that can lead to larger research programs and extramural grant proposals. The winning proposals incorporate sensors, drones, data analytics and data visualization to advance climate change management, civic expression through gaming, geologic hazard assessment, and other timely applications in healthcare, energy, and agriculture. Of the successful proposals, 91% of lead investigators are receiving their first CITRIS Seed Grant and 73% of the projects include pre-tenured faculty members.
CITRIS has awarded annual seed grants to spur multi-campus, multidisciplinary collaborations among four University of California campuses since 2007. Numerous companies and centers have resulted from the program, including Cellscope, an imaging device that transforms ordinary cell phones into otoscopes for ear exams in the field; and the Center for Autonomous and Interactive Systems at UC Merced, which focuses on robotics, virtual characters, cognitive science, and human-computer interfaces.
Congratulations to faculty leaders of the 2017 CITRIS Seed Funding awards below.
National parks, forests, and public lands provide opportunities for people to interact with nature that can improve human well-being through mental health, identity, and connectedness. However, forest systems are vulnerable to disturbances such as wildfire, drought, and disease, and as these landscapes change so do the values, activities, and the management approaches that sustain them. In the southern Sierra Nevada ecoregion, warming and variability in the amount and timing of precipitation have led to decreased snowpack and mega-fires with historically unprecedented severity resulting in unexpected succession outcomes. With these climate change impacts, the legal landscape is evolving to reflect the importance of our reliance on the resilience of the ecosystem services provided by the forests and watersheds of the State. For example, watersheds are now legally defined as part of the water infrastructure in California. Different agencies use different management approaches with regard to climate change, visitor use, fire suppression, prescribed fire, surface water storage projects, and pro-active resources management practices. These approaches are valued differently by those who manage public lands or those who recreate on these lands, while at the same time, we all rely on the natural infrastructure and ecosystem services of these areas. We propose to develop the Comparative Adaptive Management and Ecosystem Response Assessment (CAMERA) to photographically survey the preferences of the forest community associated with resilient or vulnerable landscapes, the management projects used to sustain them, and to communicate the impacts of climate change on regional public lands and watersheds.
Digital Refuge: aggregating and visualizing asylum seeker and host community concerns
Principal investigators: Katherina Linos (UC Berkeley) and Anupam Chander (UC Davis)
Unexpected crises, frequent policy reversals, and miscommunication have characterized the European response to the largest refugee crisis since World War II. In this crisis-driven policy space, information sharing has become increasingly difficult: governments and aid organizations do not know what refugees need most until the last minute, while asylum seekers fail to understand and benefit from key protections. Across Europe, a variety of rumor-trackers, hoax maps, and social media sites have been established in response to this crisis. However, this information is dispersed; much of it is posted, for example, in the form of individual refugee queries or within weekly NGO newsletters. Our objective is to scrape this information from a variety of social media and websites, organize it chronologically and geographically, and synthesize it and translate it into Greek and Arabic. Our end product will serve to communicate with local stakeholders of diverse types including refugees, NGOs aid agencies, and the Greek authorities. Our data visualizations will facilitate these stakeholders understanding of which concerns are becoming more salient at particular times and places and plan accordingly.
Democratizing civic expression video games
Games are an increasingly important medium for our society, especially for youth. 97% of teens play video games. The time youth spend playing video games nearly tripled in the years 1999-2009 (from 26 minutes a day to 1:13) while overall media use increased only fractionally (from 6:19 to 7:38). Unfortunately, while many other forms of media have seen their creation and distribution radically democratized, video games have not. Because even the easiest to use video game-making tools require significant engagement with the details of behavior programming and the creation of images and animations, only a small handful of the population (and very few youth) can use games to express themselves. So as games become an increasingly prominent medium, the overall democratizing trend in media is being reversed, and video games about political topics are largely only produced by those with significant resources. This project will provide compelling evidence that it is possible to reverse this trend, and make civic expression through video games broadly accessible, especially for the youth who are making games an increasingly central part of their media lives. We will do this by creating the first prototype of a game creation system that drives all three key elements of game expression (mechanics, structure, and surface expression) from an underlying rhetorical structure specified by authors. In doing so, we will build upon key successes of past PI research. We believe this work has high potential for follow-on funding from research funders, foundations, and media platforms.
HealthHub: An in-home platform to integrate communications, health monitoring, and intervention for an aging population
According to the US Department of Health and Human Services, the number of people in the USA age 65+ was 46 million in 2014 (15% of population), but will be 98 million by 2060 (22% of population). One goal of future aging-related programs will be to assist seniors to stay in their homes as long as possible, while providing dependable health monitoring and needed intervention. Technological tools are poised to play a major role in lower-cost, efficient in-home monitoring. The goal of the current project is to provide technology-based in-home monitoring and intervention for the (relatively) healthy senior population who is technologically-shy. The hypothesis is that in-home monitoring and intervention can be used to identify and treat potential health challenges early-on so as to delay major health problems and encourage independent living. We propose a new home-based health monitoring platform called the UC HealthHub, which supports the following functionality: 1) communications with family members, friends, doctors, and caregivers, as well as a portal for health information/videos 2) passive non-contact measurement of health-related vitals and markers, and clear feedback to the user, 3) machine learning to understand a user’s baseline over time and alert to user-specific unusual activity, 4) health information integration, and 5) customized modes to support caregivers and health providers.
To cath or not to cath: timely alerts for patients with spinal cord injury
Millions of patients suffer from the consequences of spinal cord injury (SCI) and congenital spinal anomalies. Although many of these patients have obvious limitations in mobility, unbeknownst to the general public is that nearly all have neurogenic bladder dysfunction and lack control of their bladder. Since SCI patients are unable to sense bladder fullness, they are recommended to catheterize every 2 to 4 hours throughout the day. This high frequency of emptying adds “insult to injury”. A common problem is making the trip to the bathroom and only finding a small amount of urine in the bladder. Or worse, not getting to the bathroom in time and leaking because the bladder was too full. To address this problem, we aim to build a non-invasive, patch-like device that would be worn by SCI patients to receive timely alerts for starting to look for a bathroom to perform catheterization. The device would utilize an array of light-emitting diodes (LEDs) and photodetectors to infer spatial expansion of the bladder. The underlying physical principle exploited by our device is measurement of backscattered light at wavelengths for which water has high absorption coefficient (e.g., ~950nm) via an array of light source and detectors with fixed distances. We will develop machine learning algorithms to identify patterns in light absorption maps generated by the sensor array, and to personalize the alert to better match individual patientâ€™s body characteristics and preferences. Extensive empirical studies with bladder replicas, swine bladder and healthy human volunteers will be carried out.
Development of community-based gait monitors for children with Duchenne muscular dystrophy
Duchenne muscular dystrophy (DMD) is the most common neuromuscular disease of childhood affecting 1 in 3,500-5,500 males with an estimated prevalence in the U.S. of 15,000. DMD causes increased muscle fragility, and males with DMD develop progressive loss of muscle fibers severe disabling weakness, loss of ambulation and self-care skills, respiratory and cardiac failure, and premature death. Treatment in DMD will have the greatest impact on mobility and longevity if initiated as early as possible. Unfortunately, clinical trials have not typically included patients from toddler to preschool age because of lack of reliable clinical endpoints for longitudinal follow-up from this age range to middle childhood and beyond. In this proposal, the overall goal is to develop new and innovative methodologies based on state-of-the-art accelerometers and GPS technologies with the objective of providing important new tools to allow clinicians and researchers to better monitor community mobility and effectively evaluate clinical status and therapeutic benefit in the clinical and home environment. To address the remote monitoring of gait and activity in a home environment, we will develop a monitoring system based on the Berkeley Telemonitoring Project. The project consists of a general-purpose framework for health monitoring that allows for the design and implementation of research-grade health telemonitoring apps for Android smartphones. The framework includes tools for fault-tolerant data storage, secure client-server communication and access to health data via three specific modules. We will compare data from the new monitoring devices to commonly-used clinical endpoints in boys with DMD and age-matched peers.
Predictive modeling of surgical site infections using electronic health record data with machine learning tools
The objective of this proposal is to develop machine learning models using Electronic Health Record (EHR) Data to predict patients who may develop Surgical Site Infection (SSI) post-operatively. SSI are one of the most important post-operative complications that increase mortality, morbidity, increased length of hospital stay, poor surgical outcomes and healthcare expenditure. Utilizing all available resources to predict patients who are more prone to SSI is critical for taking preventative measures, both before and after surgery. EHR data such as laboratory test results, diagnosis/ comorbidities, procedures, medications, vitals and patientsâ€™ age are critical resources that guide healthcare providers make decisions about all aspects of surgical patient management, including SSI prevention, both in elective and emergency scenarios. Providing Clinical Decision Support (CDS) with predictive models developed with Machine Learning (ML) tools and EHR data about who is at risk of SSI at the point-of-care may help reduce SSI. We plan to use EHR data at UC Davis Health System, an early adaptor of EHR and work with experts in ML from UC Berkeley to develop the models. After pre-processing the data we will use ML algorithms such as Random Forest (RF), Support Vector Machine (SVM) and Logistic Regression (LR) to develop training models and probe the model with a testing set of data. We will further validate the model prospectively for its ability to accurately predict those who may or may not get SSI post-operatively. Future directions include generalizing the model to include other hospital acquired infections.
Consequences-aware co-piloting system for human-in-the-loop drone operations
Small unmanned aerial systems (UAS) are becoming more and more prevalent, driven by consumer interest and their potential for revolutionizing aspects of commercial applications, such as delivery of urgent goods. The expected ubiquity of such systems raises concerns about their safety, and the ability of such autonomous systems to operate safely in densely populated areas (where their value will be greatest). We propose to develop a system which adds an additional layer of safety to aerial systems operated by a human pilot, by monitoring the UAVs environment for visual cues, and monitoring the human pilot for signs of distraction. The system will endow a UAS with the ability to reason about its safety, and the consequences of safety failures during its operation. The UAS will furthermore continuously reason about possible safety maneuvers in response to likely failures — in the event of an emergency, the vehicle can then execute its last safe maneuver, thus reducing the system’s danger. We will exploit the capabilities offered by combining expertise from UC Berkeley and UC Merced: prior experience with rotorcraft, and safety thereof from Berkeley will be combined with experience on human factors, general UAS safety, and the drone safety center at UC Merced.
Bioinspired burrowing robot for natural and human-made geologic hazard assessment
Geologic hazards can have catastrophic consequences. One notable example is the failure of Teton Dam produced by the unknown presence of fissures underneath the dam, which created a downstream flood that caused damages estimated to 2 billion dollars. The natural variability of geologic deposits poses a significant challenge to engineers because current characterization technologies only provide information of discrete points within the subsurface, requiring engineers to interpolate the data available to estimate the subsurface profile. The need for interpolation introduces significant uncertainty in the assessment of geologic hazards, often resulting in either inefficient or unsafe infrastructure. The objective of this project is to improve current subsurface exploration technologies by developing a bioinspired robot with the unique capability of burrowing in any direction to reach locations within the subsurface that drive the risk of failure. This work will obtain bioinspiration from three burrowing organisms (ghost crab, European nightcrawler, and razor clam), assess the mechanical efficiency of their burrowing strategies, and implement the most efficient strategies in a robotic prototype design plan. This robot will improve the quality and quantity of data provided by subsurface characterization tools, allowing engineers to significantly reduce the uncertainty in geologic hazard assessments. This work is well-aligned with two of CITRIS initiatives, and it will make use of the CITRIS invention lab. The successful completion of this project will benefit society by providing engineers enhanced capabilities to assess the risk of geologic hazards, which in turn will allow for development of safer and more resilient infrastructure systems.
Million Hands: prosthetic hands for children through an open source platform, 3D printers and sensors
Our team proposes to build a low-cost modular prosthetic platform for children with upper limb deficiency, ectrodactyly, and other conditions that affect the structure and utility of the hand at a young age. We will be focusing on children with partially developed hands who have a limited number of digits – a condition that can lead to physical barriers in daily life. These children are typically outside of the scope of the prosthetics market due to the constant growth of children’s hands. Million Hands was motivated by our prototype, Sophie’s Super Hand (Berkeley Engineer magazine) , which took the form of a custom-fitted, 3D printed prosthetic hand. Because the previous model was designed for people without any fingers, our team encountered difficulty fitting the prosthetic hand to the more complex handshapes of users with partial digits. The device tended to satisfy users aesthetically, but not functionally, as such participants found the strength, comfort, and consistency of the prosthesis lacking. We propose to develop a modular platform that is 1) customizable to the many handshapes that are possible as a result of the above conditions, 2) capable of natural movement, and 3) strong enough to perform most daily tasks. We will also explore the use of adding sensors for daily monitoring and use in myoelectric augmentation in hands. We are partnering with the Oakland and San Francisco’s Children’s Hospital, including “Hand Camp,” a program offered by the hospital to support kids with upper limb deficiency.
A multispectral and thermal imaging sUAS system for monitoring crop water use and detecting water stress
California’s growers face great water use challenges due to the expansion of perennial crops and a warmer and drier climate. The accurate and timely estimate of crop consumptive water use and water stress at a field scale is the missing link in the current on-farm irrigation management. Field-based methods are limited by the high cost, while remote sensing approaches are constrained by accuracy, spatial, and temporal resolution. We here propose to develop and test a robust and cost-effective measurement and analytical approach to fill this gap, using the emerging drone and imaging technology. Calibrated multi-spectral and thermal imaging cameras will be integrated on a leading drone platform. We will use this fully integrated sUAS system to collect the aerial imageries over several walnut and almond sites in California, where we have both on-farm and plot irrigation experiments and ground measurements of evapotranspiration (ET) and water stress ongoing. Analytical approaches will be developed and validated to estimate ET and quantify water stress based on multispectral and thermal imagery. This project put together the best combination of talents to figure out the best cost-effective way of advancing the plant water monitoring technique, with scientific rigor, not only in data acquisition stage but also in the post-flight data processing and analytical phase. The resulting capability will provide growers observation-based guidance for site-specific and time-sensitive irrigation management, and thus ensure agriculture sustainability. The ET mapping tool will also improve the consistent estimate of the water budget and thus support state and local water agencies for both water planning and regulatory/compliance purposes.