Positions

  • Present 2013

    Research Manager

    UCLA Wireless Health Institute

  • Present 2013

    Director of Engineering

    WearSens

  • Present 2012

    Lecturer & Teaching Assistant

    UCLA

  • 2010 2003

    Researcher & Project Manager

    Toyon Research Corporation

  • 2011 2010

    Adjunct Faculty

    Oxnard Charter College

Education & Training

  • Ph.D. 2015 (Expected)

    Computer Science

    System optimization in remote health monitoring (RHM)

  • M.Sc. 2010

    Computer Science

    Artificial life, robotics and global optimization

  • B.A.2003

    Computer Science

    Minor: Sociology

    Technical Minor: Electrical Eng.

    Summa cum laude from UCLA

Research Projects

  • image

    Wanda: An End-to-End Remote Health Monitoring System

    Wanda is an end-to-end remote health monitoring and analytics system comprising a smartphone-based data collection gateway, Internet-scale data storage and search system, and a backend analytics engine for diagnostic and prognostic purposes. The system supports the collection of data from a wide range of sensory devices that measure patients’ vital signs and physical activity, as well as collect data from self-reported questionnaires. The analytics engine is designed to learn and predict future events by examining patient readings. Wanda has been used in several studies to help reduce cardiovascular disease risk factors, predict heart failure outcomes and hospital readmission, and enable early detection of key clinical symptoms.

    image

    Table of Prior and Current Wanda Studies

    Name of Study Participant Information Purpose of Study Collaborating Partners
    Wanda-Heart Failure for the Elderly 1500 participants (age > 50) with Heart Failure. Predicting Heart Failure, Hospital Readmission. UCLA Dept. of Medicine, UC-Davis, UCSF, UCI, UCSD, and Cedar Sinai Hospital.
    Wanda-CVD for African American Women 90 participants (African American women ages 25-45) with 2 Cardiovascular disease (CVD) risk factors. Reducing CVD risk factors through self-management, social support, and automated wireless coaching. UCLA School of Nursing, and Local Churches.
    Wanda-2009 26 participants (68% male; 40% White, 13% Black, 32% Latino, and 15% Asian/Pacific Islander, age > 65). Enable early detection of key clinical symptoms indicative of CHF-related decomposition. UCLA School of Nursing, UCLA Ronald Reagan Med. Center, Harbor-UCLA Medical Center, Providence Holy Cross Medical Center.
    Wanda-Heart Failure for Latinos 18 participants (Latinos) with heart failure. Reduce CVD risk factors through remote health monitoring. UCLA School of Nursing.
  • image

    Data Analytics for Reduction of Hospital Readmission

    As of October 2012, the US government began implementing the Hospital Readmission Reduction Program, which levies financial punishment on hospitals with high readmission rates. Statistics show that nearly 20% of insured patients are readmitted to hospitals within 30 days after discharge, incurring approximately $17 billion in charges in 2009. Recent advances in wireless sensors, mobile technologies, and cloud computing have made continuous remote monitoring of patients possible. These remote health monitoring systems provide a continuous stream of patient physiological data that allows nurses and doctors to make timely decisions and help patients manage their chronic conditions while minimizing hospital readmission rates. Conventional Remote Health Monitoring (RHM) systems rely on threshold based alerts. That is, thresholds based on medical expertise are put in place that alert nursing staff when physiological data deviate from those thresholds. Analytics-based RHM systems on the other hand employ machine learning algorithms to predict the risk of medical adverse events. We have proposed a novel way to design an analytics engine based exclusively on electronic health records (EHR) from the Ronald Reagan UCLA Medical Center between 2005 and 2009. These records correspond to 913 unique patients out of which 169 had more than one CHF related admissions. We focus our efforts on Congestive Heart Failure (CHF) patients although our approach could be extended to other chronic conditions. Our goal is to construct statistical models that predict a CHF patient’s length of stay and by extension the severity of his/her condition. We show that it is possible to predict length of hospital stay based on physiological data collected from the first day of hospitalization. Using statistical models constructed from EHR and 10-fold cross validation we can achieve very accurate predictions with RMS error of 3.3 days for hospital stays that are less than 15 days in duration. We also propose a clustering of patients that organizes them in risk groups according to their estimated severity of condition as shown in the Figure below. These models can be used to analyze daily information collected from an RMS and allow a nurse / doctor to prioritize his/her intervention. The risk factor information could also be used to provide personalized advice to the patient on how to improve their condition in the short and long term as to avoid readmission.

    image
  • image

    Activity Recognition for Exergaming

    To encourage activity throughout the day, not only when playing the video game, we developed a video gaming system that maps specific activities performed throughout the day to the avatar in a video game. To test this we modified an existing open source role-playing game (RPG) video game named FreedroidRPG, to receive input from a wearable accelerometer that is embedded in a belt designed to be placed around the waist. The game uses our classification framework to detect activity type and time spent in each activity type to convert energy exerted in the real world to potential energy of the avatar of the game. The goal of the video gaming system is to promote exercise throughout the day, thus encouraging a more active lifestyle. Our activity recognition framework will aid such video games in classifying a given activity independent of its intensity level.

    image
  • image

    Energy Expenditure Calculation

    Activity monitoring systems need to be able to improve their accuracy when calculating Energy Expenditure (EE), as this accuracy is critical to the ability to determine relationships between energy expenditure and related variables. Crouter et al. showed improvement in EE estimates by using two regressions one for sedentary activities and another for non-sedentary activities. Albinali et al. found it useful to perform separate regressions on each specific activity. We reaffirm Albinali’s claim, and show that detecting the activity type first, and then using a separate regression model yields more accurate EE calculations.

    image
  • image

    Stochastic Approximation Modeling for Activity Recognition

    Many classifiers that train on an activity at a subset of intensity levels fail to recognize the same activity at other intensity levels. This demonstrates weakness in the underlying classification method. Training a classifier for an activity at every intensity level is also not practical. We tackle a novel intensity-independent activity recognition problem where the class labels exhibit large variability, the data is of high dimensionality, and clustering algorithms are necessary. We propose a new robust Stochastic Approximation framework for enhanced classification of such data. Experiments are reported using two clustering techniques, K-Means and Gaussian Mixture Models. The Stochastic Approximation algorithm consistently outperforms other well-known classification schemes which validates the use of our proposed clustered data representation.

    image
  • image

    Anticheating and Context Aware Algorithms for Activity Monitors

    In remote health monitoring of patient's physical activity, ensuring correctness and authenticity of the received data is essential. Although many activity monitoring systems, devices and techniques have been developed, preventing patient cheating of an activity monitor has been primarily an unaddressed challenge. Patients can manually shake an activity monitor device (sensor) with their hand and watch their physical activity points or rewards increase; this is what we define to be "self-inflicted" cheating. A second type of cheating, we name "impersonator" cheating, is when subjects hand the activity sensor over to a friend or second party to wear and perform physical activity on their behalf. In this paper, we propose two novel methods based on classification algorithms to address the cheating problems. The first classification framework improves the correctness of our data by detecting self-inflicted cheatings. The second technique is an advanced classification scheme that extracts and learns unique patient-specific activity patterns from prior data collected on a patient to distinguish the true subject from an impersonator. We tested our proposed techniques on Wanda, a remote health monitoring system used in a Women's Heart Health study of 90 African American women at risk of cardiovascular disease. We were able to distinguish cheating from other physical activities such as walking and running, as well as other common activities of daily living such as driving and playing video games. The self-inflicted cheating classifier achieved an accuracy of above 90% and an AUC of 99%. The impersonator cheater framework results in an average accuracy of above 90% and an average AUC of 94%. Our results provide insight into the randomness of cheating activities, successfully detects cheaters, and attempts to build more context-aware remote activity monitors that more accurately capture patient activity.

    image
  • image

    Battery Optimization

    The ubiquitous nature of wearable sensors, such as the readily available embedded accelerometers in smart phones, provides physicians with an opportunity to remotely monitor their patient’s daily activity. There have been several developments in the area of activity recognition using wearable sensors. However, due to power constraints, resource efficient algorithms are necessary in order to perform accurate realtime activity recognition while consuming minimal energy. In this paper, we present a two-tier architecture for optimizing power consumption in such systems. While the first tier relies on a hierarchical classification approach, the second one manages the activation and deactivation of the classification system. We demonstrate this using a series of binary Support Vector Machine classifiers. The proposed approach, however, is classifier independent. Experimenting with subjects performing different daily activities such as walking, going up and down stairs, standing and sitting, our approach achieves a power savings of 87% while maintaining 92% classification accuracy.

    image
  • image

    Pressure Sensitive Bedsheet

    The pressure sensitive bedsheet has a number of important medical applications. This system enables the evaluation of pressure ulcer risk, detection of sleep apnea, respiration rate measurement, sleep stage and sleep posture analysis. The bedsheet comprises e-textile material which employs resistive characteristics to produce pressure maps of the human body. With its 128 x 64 dense pressure sensor array, this system allows high resolution analysis and fine grain pressure point mapping. Since the sheet is made out of fabric, the system is comfortable and unobtrusive to the user. Other applications include fall risk and prevention analysis, as well as on-bed physical exercise rehabilitation.

    image
  • image

    Nutrition Monitoring Necklace

    Studies have shown that caloric beverage consumption is implicated in the obesity epidemic because of the weaker energy compensation response it elicits compared with solid food forms. Other studies show that the number of swallows during a day correlated more highly with weight gain on the following day than did estimates of caloric intake. Eating patterns and body hydration have been known to be a direct factor that impacts obesity in the United States. We have developed a novel wearable system in the form of a necklace, which aggregates data from an embedded piezoelectric sensor capable of detecting skin motion in the lower trachea during ingestion. We propose an algorithm based on spectrogram analysis of piezoelectric sensor signals to accurately distinguish between food types such as liquid and solid, hot and cold drinks and hard and soft foods. The necklace transmits data to a smartphone, which performs the processing of the signals, classifies the food type, and provides visual feedback to the user to assist the user in monitoring their eating habits over time. Experimental results demonstrate high classification accuracy of the proposed method, and validate the use of a spectrogram in extracting key features representative of the unique swallow patterns of various foods.

    image
  • image

    Autonomous Learning of Locomotion

    Evolution has produced organisms whose locomotive agility and adaptivity mock the difficulty faced by robotic scientists. The problem of locomotion, which nature has solved so well, is surprisingly complex and difficult. We explore the ability of an arti¯cial eight-legged arachnid, or animat, to autonomously learn a locomotive gait in a three-dimensional environment. We take a physics-based approach at modeling the world and the virtual body of the animat. The arachnid-like animat learns muscular control functions using simulated annealing techniques, which attempts to maximize forward velocity and minimize energy expenditure. We experiment with varying the weight of these parameters and the resulting locomotive gaits. We perform two experiments in which the first is a naive physics model of the body and world which uses point-masses and idealized joints and muscles. The second experiment is a more realistic simulation using rigid body elements with distributed mass, friction, motors, and mechanical joints. By emphasizing physical aspects we wish to minimize, a number of interesting gaits emerge.

    image

Filter by type:

Sort by year:

“Designing a Robust Activity Recognition Framework for Health and Exergaming using Wearable Sensors,” IEEE Journal of Biomedical and Health Informatics (JBHI) 2014.

Nabil Alshurafa

“Improving Compliance in a Remote Health Monitoring System through Smartphone Battery Optimization,” IEEE Journal of Biomedical and Health Informatics (JBHI) Special Issue 2015.

Nabil Alshurafa

“BreathSens: A Continuous On-Bed Respiratory Monitoring System with Torso Localization using an Unobtrusive Pressure Sensing Array,” IEEE Journal of Biomedical and Health Informatics (JBHI) Accepted 2014.

Nabil Alshurafa

"Sleep Posture Analysis using a Dense Pressure Sensitive Bedsheet," Pervasive and Mobile Computing (PMC) Journal Special Issue 2014.

Nabil Alshurafa

“Using Pressure Map Sequences for Recognition of On Bed Rehabilitation Exercises,” IEEE Journal of Biomedical and Health Informatics (JBHI) 2014.

Nabil Alshurafa

“Using Electronic Health Records to Predict Severity of Condition for Congestive Heart Failure Patients,” ACM International Joint Conference on Pervasive and Ubiquitous Computing (Ubicomp) 2014.

Nabil Alshurafa

"A Dense Pressure Sensitive Bedsheet Design for Unobtrusive Sleep Posture Monitoring," IEEE Int. Conf. on Pervasive Comp. and Comm (PerCom) 2013.

Nabil Alshurafa

“On-bed Monitoring for Range of Motion Exercises with a Pressure Sensitive Bedsheet,” IEEE Body Sensor Networks (BSN) 2013. --Best Paper!

Nabil Alshurafa

“A Framework for Predicting Adherence in Remote Health Monitoring Systems” ACM Conference on Wireless Health (WH) 2014.

Nabil Alshurafa

“Non-Invasive Monitoring of Eating Behavior using Spectrogram Analysis in a Wearable Necklace” IEEE EMBS Conference on Healthcare Innovation & Point-of-Care Healthcare Technologies (PHT) 2014.

Nabil Alshurafa

“Remote Health Monitoring: Predicting Outcome Success based on Contextual Features for Cardiovascular Disease,” IEEE Engineering in Medicine and Biology Society (EMBC) 2014.

Nabil Alshurafa

"Anti-Cheating: Detecting Self-Inflicted and Impersonator Cheaters for Remote Health Monitoring Systems with Wearable Sensors,” IEEE Body Sensor Networks (BSN) 2014.

Nabil Alshurafa

“Battery Optimization in Smartphones for Remote Health Monitoring Systems to Enhance User Adherence,” PErvasive Technologies Related to Assistive Environments (PETRA) 2014.

Nabil Alshurafa

“Robust Human Intensity-Varying Activity Recognition using Stochastic Approximation in Wearable Sensors,” IEEE Body Sensor Networks (BSN) 2013.

Nabil Alshurafa

“Beyond Dr. Google: Early Results of a Personalized Weight-Tracking Smartphone Application and Alert System for Patients with Ascites,” 65th Annual Meeting of the American Association for the Study of Liver Disease (AASLD) 2014.

Nabil Alshurafa

“Staying Connected: A CVD Risk Intervention for Young Black Women," American Heart Association (AHA) 2014.

Nabil Alshurafa

“Support Vector Regression for METs of Exergaming Actions” IEEE EMBS Conference on Healthcare Innovation & Point-of-Care Healthcare Technologies (PHT) 2014.

Nabil Alshurafa

“Multiple Model Analytics for Adverse Event Prediction in Remote Health Monitoring Systems” IEEE EMBS Conference on Healthcare Innovation & Point-of-Care Healthcare Technologies (PHT) 2014.

Nabil Alshurafa

“Spectrogram-Based Audio Classification of Nutrition Intake” IEEE EMBS Conference on Healthcare Innovation & Point-of-Care Healthcare Technologies (PHT) 2014.

Nabil Alshurafa

"A Wearable Nutrition Monitoring System,” IEEE Body Sensor Networks (BSN) 2014.

Nabil Alshurafa

"Determining the Single Best Axis for Exercise Repetition Recognition and Counting on SmartWatches,” IEEE Body Sensor Networks (BSN) 2014.

Nabil Alshurafa

“Remote monitoring Systems: What Impact Can Data Analytics Have On Cost?,” IEEE Body Sensor Networks (BSN) 2013.

Nabil Alshurafa

“MET Calculations from On-Body Accelerometers for Exergaming Movements,” IEEE Body Sensor Networks (BSN) 2013.

Nabil Alshurafa

"Improving accuracy in E-Textiles as a platform for pervasive sensing," IEEE Body Sensor Networks (BSN) 2013.

Nabil Alshurafa

“Inconspicuous On-Bed Respiratory Rate Monitoring.” PErvasive Technologies Related to Assistive Environments (PETRA) 2013.

Nabil Alshurafa

“Dynamic Task Optimization in Remote Diabetes Monitoring Systems,” Int. Conf. on Healthcare Informatics, Imaging and Systems Biology 2012.

Nabil Alshurafa

“Opportunistic Hierarchical Classification for Power Optimization in Wearable Movement Monitoring Systems,” IEEE Int. Symposium on Industrial Embedded Systems (SIES) 2012.

Nabil Alshurafa

“WANDA: An End-to-End Remote Health Monitoring and Analytic System for Heart Failure Patients,” Wireless Health 2012.

Nabil Alshurafa

“Artificial Spider: eight-legged insect and autonomous learning of locomotion,” SPIE-Unmanned Systems Technology VIII 2006.

Nabil Alshurafa

Currrent Teaching

  • Present 2013

    CS180 Algorithms & Complexity

    Lecturer, Summer 2013

    image

    This course involves teaching techniques for the design and analysis of efficient algorithms, emphasizing methods useful in practice. Topics include sorting; search trees, heaps, and hashing; divide-and-conquer; dynamic programming; greedy algorithms; graph algorithms; shortest paths; network flow; NP-Complete and NP-Hard problems.

  • 2014 2012

    CS152A/EE116 Introduction to Digital Design

    Teaching Assistant, Winter 2012, and Winter 2014

    image

    This course involves the implementation of digital logic designs. Students will use Spartan-III Xilinx boards and work with the ISE design software to implement designs using schematic and HDL editors. Students will learn how to use discrete components, breadboards, oscilloscopes, voltmeters, datasheets, and instruction manuals. They will be able to: design a digital system, implement their designs with a schematic editor and VHDL, program and use an FPGA, and work with a partner to produce working and well-designed results.

  • 2013 2012

    CS180 Algorithms & Complexity

    Teaching Assistant, Fall 2012, Fall 2013

    This course involves teaching techniques for the design and analysis of efficient algorithms, emphasizing methods useful in practice. Topics include sorting; search trees, heaps, and hashing; divide-and-conquer; dynamic programming; greedy algorithms; graph algorithms; shortest paths; network flow; NP-Complete and NP-Hard problems.

  • 2011 2010

    Oxnard Charter College

    Adjunct Faculty (2010-2011)

    • Courses: College Mathematics, Advanced Algebra, Calculus, Statistics

    • Courses: Network Essentials, Wireless Essentials, Computer Essentials, Network Security, C++

    • Student Mentor: Mentored several students in Mathematics and Computer Science