Introduction to Genetic Algorithms1 Assaf Zaritsky Ben-Gurion University, Israel assafza.
1 SmartSpaghetti: Use of Smart Devices to Solve Health Care Problems Mostafa Uddin,A. Gupta, T....
-
Upload
veronica-terry -
Category
Documents
-
view
213 -
download
0
Transcript of 1 SmartSpaghetti: Use of Smart Devices to Solve Health Care Problems Mostafa Uddin,A. Gupta, T....
1
SmartSpaghetti: Use of Smart Devices to Solve Health Care Problems
Mostafa Uddin,A. Gupta, T. Nadeem, K. MalySandip Godambe, Arno Zaritsky
BIBM/BIH Shanghai Dec. 18 - 21, 2013
2
Contents
• Mobile technology opportunities in medical field
• LEAN : spaghetti problem• Smart Spaghetti System• Prototype & Experiments
• Future Work
BIBM/BIH Shanghai Dec. 18 - 21, 2013
3
Opportunities*
• Role of Smart Devices• The amount of research in the use of the smartphone in medicine
is rapidly growing.• Smartphones have a very bright future in the world of medicine,
while doctors, engineers, and others alike continue to contribute more ingenuity to this dynamic field.
• Given the numerous ways in which the smartphone can be used in healthcare, smartphones will be recognized as a diagnostic and therapeutic tool that is as irreplaceable as the stethoscope has been in the practice of medicine.
* Ozdalga E, Ozdalga A, Ahuja N, “The Smartphone in Medicine: A Review of Current and Potential Use Among Physicians and Students”, J Med Internet Res 2012;14(5):e128
BIBM/BIH Shanghai Dec. 18 - 21, 2013
4
Vision
• Create collaborative systems for specific niches such as childhood obesity that:• For individuals
• Collect streams of sensor data continuously• Integrate with information in electronic health records• Enable interactions from stakeholders• Provide action lists for patients, caretakers and doctors• Track costs associated with data collection and actions
• For researchers• Create anonymous database of disease patterns• Create models and explore hypotheses• Develop cost/benefit models
BIBM/BIH Shanghai Dec. 18 - 21, 2013
5
Approach• Given that:
User-Friendly Remote Patient Monitoring
Availability of new inexpensive gadgets that monitor your health and fitness ranging from heart monitors to biosensors that read body temperature and motion
More than 100 million wearable health-related devices sales annually by 2016 (ABI Research).
Projected to reach 80 million wearable sports and fitness-related monitoring devices sales by 2016.
Will build: Flexible optimized system integrating smartphones,
gadgets, patients, doctors and EHR*.
*EHR: Electronic Health Records
BIBM/BIH Shanghai Dec. 18 - 21, 2013
6
Lean
• Lean Healthcare is the application of concepts, tools and management prescriptions aimed at furthering the organizational mission by strengthening operating processes.
• Characteristics of a Lean Healthcare organization
• More Efficient (operationally & capital-wise)• Faster & more reliable• Delivers higher quality• More Responsive• Performs way above the rest with more satisfaction
Plenty of room for improvement!
Sandip Godambe, MD, PhD, MBA; Quality Improvement and Safety Team (QuIST)
BIBM/BIH Shanghai Dec. 18 - 21, 2013
7
Spaghetti Problem
BIBM/BIH Shanghai Dec. 18 - 21, 2013
Findings: • Layout not visual control friendly• Many isolated islands• Workstation layout not standardized
8
Smartphone Approach
• Obtain room layout, targets, feasible paths
• Use smartphone with accelerometer, gyroscope sensors
• For a starting point have individual walk to a target• Obtain raw data from sensors• Extract information such as strides, directions, and pauses• Compute final path
BIBM/BIH Shanghai Dec. 18 - 21, 2013
9
Basic Scheme
Human movement path can be segmented into units of strides and turns.
Stride length
Turn
BIBM/BIH Shanghai Dec. 18 - 21, 2013
10
Basic Scheme We use the sensors (Accelerometer+Gyroscope) reading to
count the stride.
Detected Stride
GyroscopeReading
BIBM/BIH Shanghai Dec. 18 - 21, 2013
11
Basic Scheme
We use the orientation and magnetic field sensor to detect the turns.
Stride
Orientation sensor reading
Turn/ change of angle
BIBM/BIH Shanghai Dec. 18 - 21, 2013
13
Tracking User Movement Path – Basic Approach Step1: Collecting sensor reading using user's smartphone
Start application; select start location; walk & collect data; end walk phase; send data to backend
Step2: Offline process on the sensor reading to estimate user's movement
• Smooth data• Model parameters
• Stride – average length of an individual’s steps• Turn angles – a step function of angles approximating the angle of a turn
between two adjacent segments
BIBM/BIH Shanghai Dec. 18 - 21, 2013
22
Conclusions & Challenges
• Smartphones can be used in an automated, non-intrusive manner to generate spaghetti diagrams but needs:
• User/Device - independent• Orientation/position -independent• Error correction schemes
• Ideal paths: Select the best path among all possible paths.
• Location confirmation: • Fusing with other technologies
• WiFi, sound, Bluetooth• WiFi MSE CISCO infrastructure
• Use of anchor point• Enhanced machine learning scheme for estimating location
Obstacle
Target 1
Target 2
Start point
BIBM/BIH Shanghai Dec. 18 - 21, 2013