Alireza Vahdatpour, Navid Amini, Majid Sarrafzadeh
We are interested in the design of automatic and unsupervised methods to extract valuable information from the data collected from various sensors which are worn or implanted in the human body. Our methods realize the relation between different sources of information such as blood pressure, heart rate, and body motion and enable field experts (medicine, healthcare) to understand the irregularities in subjects' habits and frequent actions and their impact on their vital signs.
Mahsan Rofouei, Myung-kyung Suh, Ani Nahapetian, Majid Sarrafzadeh
High intensity interval (HIT) training involves a series of high intensity periods of exercise followed by a recovery period. The time required for an HIT session is less than a continuous one yet yields greater benefits. A correct scheduling of exercise periods and level of workload can have huge effects in the result. This scheduling can have applications in precision fitness for individuals and also in activity scheduling for first responders . It is critical to know and schedule the level of exertion not to exceed a threshold. This problem maps nicely to the conventional DVS approaches to energy minimizations with the addition of several constraints. We use similar methods to model individual's heart rate response taking into consideration several factors such as fatigue level and heart rate thresholds. Through this guidance and exercise protocol, you can exercise safely and efficiently in minimum time.
Alireza Vahdatpour, Mars Lan, Majid Sarrafzadeh
Falls are currently a leading cause of death from injury in the elderly. The usage of conventional assistive cane devices is critical in reducing the risk of falls and is relied upon by over 4 million patients in the U.S. While canes provide physical support as well as supplementary sensing feedback to patients, at the same time, these conventional aids also exhibit serious adverse effects that contribute to falls. The SmartCane system combines advances in signal processing, embedded computing, and wireless networking technology to provide capabilities for remote monitoring, local signal processing, and real-time feedback on the cane usage. This system aims to reduce risks of injuries and falls by enabling training and guidance of patients in proper usage of assistive devices.
Navid Amini, Alireza Vahdatpour, Foad Dabiri, Majid Sarrafzadeh
The skin care product market is growing due to the threat of ultraviolet (UV) radiation caused by the destruction of the ozone layer, increasing demand for tanning, and the tendency to wear less clothing. Accordingly, there is a potential demand for a personalized UV monitoring device, which can play a fundamental role in skin cancer prevention by providing measurements of UV radiation intensities and corresponding recommendations. This project aims at developing and validating of a wireless and portable embedded device for personalized UV monitoring which is based on a novel software architecture, a high-end UV sensor, and conventional PDA (or a cell phone). In terms of short-term applications, by calculating the UV index, it informs the users about their maximum recommended sun exposure time by taking their skin type and sun protection factor (SPF) of the applied sunscreen into consideration. As for long-term applications, given that the damage caused by UV light is accumulated over days, it displays the amount of UV received over a certain course of time, from a single day to a month.
Tammara Massey, Tia Gao, Leo Selavo, David Crawford, William Bishop, Jamie Macbeth, Foad Dabiri, Bor-rong Chen, Konrad Lorincz, Victor Shnayder, Logan Hauenstein, David White, Matt Welsh , and Majid Sarrafzadeh.
While previously medical examinations could only extract localized symptoms through snap shots, continuous monitoring can now discretely analyze how a patients lifestyle affects his/her physiological conditions . Through data driven and experimental techniques, a qualitative approach to system design consisted of participatory design techniques and a quantitative analysis of an actual deployment. Through participatory design, several iterations of the medical system was developed through feedback from paramedics, first responders, medical doctors, and nurses. In addition, quantitative methods, specifically n on-parametric statistical modeling techniques, were used on experimental data to enhance resource management on the embedded systems.
H. Noshadi, S. Ahmadian, H. Hagopian, F. Dabiri, A. Vahdatpour, M. Sarrafzadeh
Lightweight smart shoe called Hermes aimed at extending fall risk analysis and human balance monitoring outside of a lab environment. The goal is to combine embedded sensing, signal processing, and balance modeling techniques to create this scientific tool capable of accurately determining fall risk assessment and walking behavior patterns. The model which determines these incorporates variability and correlation of features extracted while walking with the shoe. These variances and correlations of features which have been identified by geriatric motion study as precursors to falls and/or balance abnormalities, can then be used as a way of warning the user or the user's doctor for further investigation. Other important goals of Hermes are to provide an affordable, durable, mobile, reliable, and customizable solution to monitor balance outside of the lab.
K. Dorman, A. Nahapetian, N. Amini, M. Sarrafzadeh
The Nutrition Monitor is an easy way for people to manage their diets. Users scan barcodes of products and receive immediate nutritional feedback on their cell phones. Users can also track their dieting progress throughout the course of a day. By setting up a personalized diet, users can see immediately whether or not they are following their diet.
F. Dabiri, A. Vahdatpour, H. Noshadi, M. Sarrafzadeh
A new model of computation for distributed systems, called constant computation constructs the best possible solution for the problem while using a bounded amount of computation, communication, and memory. Lightweight sensor networks usually form a topology after deployment and do not possess a pre-defined structure. Therefore, global information about the network is either costly or impossible. Current distributed algorithms share a fair amount of global information. However, in these algorithms, the computation is asymptotically in the order of f(n), where f(n) is non-constant function of n. In our constant model of computation, each lightweight embedded system is such that, memory usage, communication iterations and data processing are in O(1).
T. Massey, F. Dabiri, H. Noshadi, G. Marfia, A. Chang, J. Chen, S Jung, R. Tan, C. Lin, D. Jea, J. Schmidt, M. Gerla, M. Sarrafzadeh
Advancements in nanotechnology make it possible to fabricate a Body Sensor Network (BSN) with miniature embedded sensors that monitor physiological activities occurring inside of the body and simultaneously probe the outside environment for harmful chemicals, dangerous radiation levels, and a more general score of other hostile events. Wireless Embedded medical SysTemS (WESTS) will allow for the efficient monitoring of physiological occurrences in the body. HealthNet will allow the communications of WESTS with the infrastructure for remote monitoring.
F. Dabiri, T. Massey, A. Nahapatian, H. Noshandi, S. Chaudhry, R. Jafari, M. Sarrafzadeh
CustoMed is a new architecture that will reduce the customization and reconfiguration time for systems that use reconfigurable embedded systems through "Med Nodes", stand-alone components consisting of a processing unit, a battery and various sensors for physiological reading from the human body. Research challenges in the area of reconfiguration, security, power awareness, communication, software partitioning, fault tolerance, and adaptability are explored.
R. Jafari, F. Dabiri, F. Chen, V. Raghunathan, T. Schoellhammer, D. Sievers, B. Wu, D. Estrin, G. Reinman, M. Sarrafzadeh, M. Srivastava, Y. Yang
RFab consists of a medical vest for sensor-driven personalized transdermal drug-delivery for medical treatment. RFab research explores environmental dynamics, physical coupling, resource constraints, infrastructure support, and robustness requirements in this wearable application.
M. Moazeni, M. Rofouei, M. Sarrafzadeh, A. Bui
With the introduction of inexpensive, single-chip, massively parallel platforms such as the new generation of GPU architectures, more developers will be creating real-life applications for these platforms. However, traditional programming models are not entirely suitable for development on such many-core architectures and unreasonable loss of performance can arise from direct porting of applications to them in spite of significant computational power of these platforms. We are currently studying and developing structured transformation methods that could remove existing bottlenecks in applications to make them more suitable for high performance many-core architectures. We are focusing on medical imaging and video processing applications as case studies. In addition, through collaboration with the UCLA Department of Radiology, novel high-performance medical imaging algorithms are being designed and studied for these platforms.
Z. Karimi, T. Taghavi, B. Choi, X. Yang, M. Wang, M. Sarrafzadeh
Dragon is a fast, effective standard cell placement tool for both variable-die and fixed-die ASIC design. Dragon enables wirelength and routable optimization by combining power hypergraph partitioning package (hMetis) with simulated annealing techniques.