Actis: WBAN Demo

Actis: WBAN Demo

Emil [email protected]

Electrical and Computer Engineering Department University of Alabama in Huntsville

Outline

Introduction WBAN System Architecture TinyOS environment Actis application WBAN Wireless Communications Time Synchronization Data format and processing … play time

WBAN for Health Monitoring

Wireless Body Area Network for Ambulatory Health Monitoring Mobility / Ubiquitous System Real-time on-sensor processing Warnings Increased Quality of Life Multisensor Monitoring (Synergy)

Hierarchical Multi-tier Telemedicine System

WBAN Implementation

WBAN for ambulatory health monitoring Activity Monitor / Motion Sensor ECG Sensor Network Coordinator Personal Server

Challenges Sensor Fusion On-Sensor Processing Ubiquitous Communications Power Efficiency/Battery Life

Outline


User2

nc

Internet

Tier 1:WBAN

A A

A E

Tier 2:PS

Tier 3:MS

WBANWWAN(GPRS)

Emergency

Informal caregiver

Healthcareprovider

UserN

Medical Server

WWANWLAN

WLAN(Wi-Fi)

…

User2User1 UserN…

System Architecture

ZigBee orBluetooth

User1

A. Milenkovic, C. Otto, E. Jovanov, "Wireless Sensor Networks for Personal Health Monitoring: Issues and an Implementation," Computer Communications, Vol. 29, No. 13‑14, August 2006, pp. 2521-2533.

Outline


TinyOS An open-source OS designed for

embedded WSN (limited resources) Component-based architecture application =

scheduler + graph of components event-driven architecture NO kernel, process/memory

management, virtual memory

Component A

Component B

ComponentD

Component C

Application

configuration

configuration

Component E

ComponentF

Components

A component has: Frame (internal state) Tasks (computation) Interface (events, commands)

Frame: one per component statically allocated fixed size

Tasks

ComponentFrame

EventsCommands

Commands and Events are function calls Application: linking/gluing interfaces (events, commands)

Commands/Events

commands: deposit request parameters into the frame are non-blocking need to return status

=> postpone time consuming work by posting a task can call lower level commands

events: can call commands, signal events, post tasks, can not be signaled by

commands preempt tasks, not vice-versa interrupt trigger the lowest level events deposit the information into the frame

Scheduler

two level scheduling: events and tasks scheduler is simple FIFO a task can not preempt another task events preempt tasks (higher priority)

Hardware

Interrupts

even

ts

commands

FIFOTasks

POSTPreempt

Time

commands

Outline


Network Coordinator

Sensor Nodes

Typical Message Flow

NC

ACTIS_CFG_STARTMSG

ACK

ACTIS_EVENT_MASKMSGevent = HRV, R Peak

ACTIS_CALMSG

ACK

Calibration Time

ACK

Session Started

eActiS (ECG)Sensor

Personal Server

ACTIS_EVENT_MSGevent = R Peak

Heartbeat (R peak)Detected

TIME

NC

ACTIS_CFG_STARTMSG

ACK

ACTIS_EVENT_MASKMSGevent = HRV, R Peak

ACTIS_CALMSG

ACK

Calibration Time

ACK

Session Started

eActiS (ECG)Sensor

Personal Server

ACTIS_EVENT_MSGevent = R Peak

Heartbeat (R peak)Detected

TIME

NC

eActiS (ECG)Sensor

Personal Server

ACTIS_EVENT_MSGevent = R Peak Heartbeat (R peak)

DetectedACTIS_EVENT_MSGevent = R Peak Heartbeat (R peak)

Detected

ACTIS_EVENT_MSGevent = Arrhythmia Arrhythmic event

Detected

.

.

.

.

.

.

.

Stream Raw DataEvent Log

augmented with contextual raw data

ACTIS_CFG_STOPMSG

Session stopped

ACK

TIME

NC

eActiS (ECG)Sensor

Personal Server

ACTIS_EVENT_MSGevent = R Peak Heartbeat (R peak)

DetectedACTIS_EVENT_MSGevent = R Peak Heartbeat (R peak)

Detected

ACTIS_EVENT_MSGevent = Arrhythmia Arrhythmic event

Detected

.

.

.

.

.

.

.

Stream Raw DataEvent Log

augmented with contextual raw data

ACTIS_CFG_STOPMSG

Session stopped

ACK

TIME

Typical Message Flow (2)

Feature Extraction

46 46.5 47 47.5 48 48.5 49 49.5 50 50.5 51

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Accelerometer based step recognition

Time[s]

Vy[m/s]Alpha[rad]accY[g]StepFoot switch

Step detected

Raw data

Events

MS

Internet Gateway

PS

Sensors

NC

WBAN

WLAN/WAN

S S S

Internet

WBAN Messaging

Session Files

Database

Data Flow and Analysis

Healthcare Access

EMR

Outline


Typical WSN Application

Periodic Data Collection Network Maintenance Majority of operation

Triggered Events Detection/Notification Infrequently occurs

But… must be reported quickly and reliably

Long Lifetime Months to Years without

changing batteries Power management is the

key to WSN success

sleep

wak

eup

processingdata acquisitioncommunication

Pow

er

Time

From Polastre et al: “The Mote Revolution: Low Power Wireless Sensor Network Devices”Hot Chips 2004 : Aug 22-24, 2004

Overhead of switching from Sleep to Active Mode

Wakeup

Microcontroller Radio (FSK)

10ns – 4ms typical1– 10 ms typical

2.5 ms

292 ns


Power Efficient TDMA

Super Cycle

Beacon ……Slot#3Slot#2Slot#1Beacon Beacon ……Slot#3Slot#2Slot#1Beacon

Listen Tr. Listen

Time50 ms 100 ms 150 ms 1000 ms

85% Sensor Power from Radio Significant Power Savings From Disabling Radio Timeslots for Communication

Distributed Events Concentrated Bursts Allows Radio to be disabled

Extended Battery Life / Lower Weight

TDMA Means Low Power

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

5

10

15

20

25

Time [sec]

I [m

A]

0.65 0.7 0.75 0.8 0.85 0.90

5

10

15

20

Time [sec]

I [m

A]

Listen Transmit

Idle

superframe

Deterministic RF timeslots

Radio can be disabled Extend battery life

ZigBee

An industry consortium that promotes the IEEE 802.15.4 standard (www.zigbee.org)

Low-cost, low-power features for multi-year operation on standard batteries

Low data throughput (250 kbps) Star and peer-to-peer network topologies

http://www.zigbee.org/

Design Principles

Key to Low Duty Cycle Operation:Sleep – majority of the timeWakeup – quickly start processingActive – minimize work & return to sleep

Sleep

Majority of time, node is asleep Typically >99%

Minimize sleep current through Isolating and shutting down individual circuits Using low power hardware

Need RAM retention Run auxiliary hardware components from low speed

oscillators (typically 32kHz) Shut down all unused clocks Perform ADC conversions, DMA transfers, and bus operations

while microcontroller core is stopped

Active

Microcontroller Fast processing, low active

power Avoid external oscillators

Radio High data rate, low power

tradeoffs Narrowband radios

Low power, lower data rate, simple channel encoding, faster startup

Wideband radios More robust to noise, higher

power, high data rates

External Flash (stable storage) Data logging, network code

reprogramming, aggregation High power consumption Long writes

Radio vs. Flash 250kbps radio sending 1 byte

Energy : 1.5J Duration : 32s

Atmel flash writing 1 byte Energy : 3J Duration : 78s


Outline


Time Synchronization

Crucial service in WSNs Group operations Source localization Data aggregation Distributed sampling Communication channels sharing

Metrics for synchronization protocols Precision Longevity of synchronization Time and power budget available for synchronization Geographical span Size and network topology

?

??

?

?

Time Synchronization - Motivation

Wall Clock versus WBAN (Jiffy) Time Two Issues

Offset Drift

TDMA Efficient Sharing of Communication Channels Timeslot Assignments Beacon Prediction (Maximize Radio off)

Correlating Intra-WBAN events Relative timing is important Synchronizing start time

Clock Offset, Skew

1 2 3 4

Beacons

Jiff

y Ti

me

Local TimeGlobal Time

Offset2

SkewOffset2

Sending Time

Receiving Time

Time Synchronization - Issues

Time Synchronization #2

Propagation

Data transmitted over RF

Preamble SFD Length MAC Protocol Data

SFD Capture Timer

Timestamp

Preamble SFD Length MAC Protocol Data TimestampData received

over RF

SFD Capture Timer

Synchronize local time(TinyOS)

NetworkCoordinator

Inserting the Timestamp

Network coordinator Starts the transmission

(time sync header) Captures timer and converts

to a global timestamp Inserts it into the message

(sends over SPI) Is this enough time not to

underrun the TxFIFO in CC2420? Time capture and calculate

timestamp: 150 s Send timestamp: 300 s Sync header transmission: 700 s

Outline


Demonstration

System configurationPosition sensor (3 ACC axes – raw data)

Values Acc = 8000 + acc_comp*2000Heart rate sensor

RR intervals in ticks (32 KHz – jiffy periods) Heart rate – 60 * 32768 / RRint

EMR

…Update

Session 1Update

Session 2

Contiguous Data Blocks

System organization - Sessions

Accelerometer sensor

X

Y

Z

Acc = 8000 + acc_comp*2000

Session file format

Byte Field Description 4 Timestamp 32-bit global jiffy time 2 numValues Number of values used in

data field 2 Not used Reserved for future param. 2 LastSampNum 4 SessionSignalID Unique identifier

formed from concatenation of session number with signal type

20 data Signal dependent for raw data this is 10 16-bit samples AEE, the first 4 bytes are little endian AEE

Total: 34 bytes/record

Test data

Web page: http://www.ece.uah.edu/~jovanov/whrms/data_proc.html

Experimental data Example processing (.M) What information can you extract?

Change of position Classification of the physical activity HR as a function of physical activity

http://www.ece.uah.edu/~jovanov/whrms/data_proc.html

Outline


Documents

Actis: WBAN Demo