Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Developing for Simband

Developing for Simband

Simband is a new open platform for innovation in digital health. We will describe the principles that drove our design decisions around building a sensor module to read bio-signals from your wrist. These biometric sensors have never been combined and synchronized before. We will explain how engineers can create new sensor modules based on our reference design, how to integrate with our hardware from a mechanical and electrical point of view. We will show you how we created algorithms that read the bio-signals and turns them into meaningful information for the user and how developers can do the same and push the boundaries of personal health and wellness.

Andrea Trasatti

November 13, 2014
Tweet

More Decks by Andrea Trasatti

Other Decks in Technology

Transcript

  1. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com Developing

    for Simband Insight into Building Breakthrough Sensors Jim Schuessler Andrea Trasatti
  2. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 2

    • Simband hardware overview  Mechanical  Electrical • Simband software overview  Sensor data  Visualize data  Live demo  SAMI integration • Working with Simband, working with us • Q&A
  3. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 4

    Inside Simband • Base unit provides the “infrastructure” to support a wide variety of sensor modules  Communications to SAMI  User interfaces  Internal power and data interfaces • The “Bucket” is the mechanical and electrical interface to sensor modules Simband is a modular PLATFORM Base Clasp The Bucket Active Band
  4. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 5

    Simsense Reference sensor module at the anterior side of the wrist Samsung’s Reference Sensor Module: Simsense Photo Plethsmography (PPG) Bio-impedance (BioZ) Galvanic Skin Response (GSR) Accelerometer Skin Temperature Electro-cardiogram (ECG) All time synchronized
  5. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 6

    Simsense • Sensor position designed with anatomy in mind  Optical sensors over ulnar artery • Multiple channels of photo- plethysmography (PPG) • Colors tend to reflect at different depths  Bio-Impedance (Bio-Z) targets the radial artery  Active clasp closes ECG loop • Finger touch from opposite hand Reference sensor module at the anterior side of the wrist Radial artery Ulnar artery
  6. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 7

    Mechanical Construction • Athena = active sensor layer  Flexible PCB with electrodes, optical elements and other sensors • Apollo = front-end processing electronics  Over 600mm2 surface area available Materials and Assembly Stack-up
  7. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 8

    Powerful processing resources 1. Dual processor architecture  Bio-processor front end (M0)  Floating point, secure backend (M4F) 2. Serial Peripheral Interconnect (SPI) Data Interface + Interrupt line from Sensor Module to Base Unit + Reset line from Base Unit to Sensor Module 3. Battery power from Base Unit Versatile architecture offers power optimization Full System Diagram
  8. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 9

    Simple, Standard Electrical Interfaces • SPI is available on every MCU in the world  Easy to develop for • Bandwidth  We use 2Mb/s • ECG lead 2 from clasp SPI and direct battery power The Bucket Simband S_MOSI S_MISO S_CLK S_RST S_INT S_CS ECG2 Res.(3) V+ GND Sensor Module Connector Host System Band Flex Cable
  9. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 10

    Power Interface • Essentially? How direct?  Coulomb counter monitors SM power usage  Switch allows Base Unit to control SM power • Limits are a balance between SM freedom and safety to Base Unit  100mA peak  50mWhr (averaged over 24hrs.) Essentially a direct connection to the battery Base Unit Sensor Module
  10. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 13

    Designed to support different data from different sensors
  11. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 15

    Sensor data • We define the data coming from each sensor as a stream • Each stream has a set of characteristics that depend on the sensor and the data it generates • Everything is time-coded and synchronized
  12. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 16

    Blood pressure Algorithm gettimestamps send Writing an algorithm
  13. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 17

    Data streams • Heart rate • Heart rate variability • Blood pressure • Confidence Indicator • Steps (increments) • Steps (daily) • Activity (walking, running, etc.) Simband streams from algorithms
  14. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 18

    Confidence Indicator • Rate the curve of the signal from 1 to 4 • Analyze the stream and compare with our model All streams from Simsense are graded Grade Definition 4 All features visible, very low noise 3 Key features visible, low noise 2 Marginal, noisy 1 Anomalies and a lot of noise
  15. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 19

    Confidence Indicator Confidence Indicator 1 Confidence Indicator 4
  16. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 22

    Working with Simband, Working with us • Yesterday we officially opening our ecosystem to the world • Our agenda is very simple: work together to build a better digital health platform • Simband source code will be available to everyone
  17. © 2014 Samsung Developer Conference. All rights reserved. www.samsungdevcon.com 23

    Let’s keep talking Read about Simband at http://www.voiceofthebody.io Andrea Trasatti – @AndreaTrasatti Jim Schuessler – @Jim_Schuessler Follow us on Twitter @voiceofthebody