IEEE Wireless Communications • February 201080 1536-1284/10/$25.00 © 2010 IEEE

Medical

Entertainmd i

Personal health care

uter games nce lessons

re detection

ECGEMGEEG

endoscopeefibrillators

nt ng re

ng on se ng

ACCE P TED FROM OP E N CALL

INTRODUCTION

An aging population and sedentary lifestyle arefueling the prevalence of chronic diseases suchas cardiovascular diseases, hypertension, anddiabetes. According to the World Health Orga-nization, cardiovascular disease causes 30 per-cent of all deaths in the world (about 17.5 millionpeople in 2005). Diabetes currently affects 180million people worldwide and is expected toaffect around 360 million by 2030. More than 2.3billion people will be overweight by 2015. Arapid rise in debilitating neuro-degenerative dis-eases such as Parkinson’s and Alzheimer’s isthreatening millions more.

Burgeoning healthcare needs are exertingenormous strain on the fragile healthcare deliv-ery system. Moreover, a shortage of skilled staff,overload, and tightening of healthcare budgets

have aggravated the impending healthcare crisis.These economical, social, and demographictrends highlight the need to exploit technologicalbreakthroughs to bring affordable and efficienthealthcare solutions to people that will improvetheir quality of life.

The advent of miniaturized sensors and actu-ators for monitoring, diagnostic, and therapeuticfunctions, and advances in wireless technologyhave opened up new frontiers in the race to con-quer healthcare challenges. Ultra-low-powerwireless connectivity among devices placed in,on, and around the human body is seen as a keytechnology enabling unprecedented portabilityfor monitoring physiological signs in the hospi-tal, at home, and on the move.

Strategically placed wearable or implanted(in the body) wireless sensor nodes sample,process, and transmit vital signs (e.g., heartrate, blood pressure, temperature, pH, respira-tion, oxygen saturation) without constrainingthe activities of the wearer. The gathered datacan be forwarded in real time to a hospital,clinic, or central repository over a local areanetwork (LAN), wide area network (WAN),cellular network, and the like. Physicians andcaregivers can remotely access this data toassess the state of the health of the patient.Additionally, the patient can be alerted usingSMS, alarm, or reminder messages.

The measured vital signs can be processedlocally for diagnosis of medical conditions, whichin turn could trigger a treatment procedure in aclose-loop bio-feedback system. For example,upon detecting high blood sugar level, animplanted blood glucose level sensing device canwirelessly trigger an insulin pump to inject arequired dose of insulin, thereby acting as anartificial pancreas.

Body area networking (BAN) technology hasthe potential to revolutionize healthcare deliveryin ambulances, emergency rooms, operation the-aters, postoperative recovery rooms, clinics, andhomes. The benefits of untethered, unobtrusive,and continuous monitoring/treatment includelong-term trend analysis, detection of transientabnormalities, prompt alerting of a caregiver to

MAULIN PATEL AND JIANFENG WANG, PHILIPS RESEARCH NORTH AMERICA

ABSTRACTAdvances in wireless technology and support-

ing infrastructure provide unprecedented oppor-tunity for ubiquitous real-time healthcare andfitness monitoring without constraining the activ-ities of the user. Wirelessly connected miniatur-ized sensors and actuators placed in, on, andaround the body form a body area network forcontinuous, automated, and unobtrusive moni-toring of physiological signs to support medical,lifestyle and entertainment applications. BANtechnology is in the early stage of development,and several research challenges have to be over-come for it to be widely accepted. In this articlewe study the core set of application, functional,and technical requirements of the BAN. We alsodiscuss fundamental research challenges such asscalability (in terms of data rate, power con-sumption, and duty cycle), antenna design, inter-ference mitigation, coexistence, QoS, reliability,security, privacy, and energy efficiency. Severalcandidate technologies poised to address theemerging BAN market are evaluated, and theirmerits and demerits are highlighted. A briefoverview of standardization activities relevant toBANs is also presented.

APPLICATIONS, CHALLENGES, AND PROSPECTIVE INEMERGING BODY AREA

NETWORKING TECHNOLOGIES

Advances in wirelesstechnology and supporting infrastructure provideunprecedentedopportunity for ubiquitous real-timehealthcare and fitness monitoringwithout constrainingthe activities of the user.

WANG LAYOUT 2/9/10 1:53 PM Page 80

IEEE Wireless Communications • February 2010 81

intervene in case of an emergency, regulation oftreatment regimes, reduction of errors, reductionof hospital stays, extending independent livingfor seniors, and improved patient comfort. BANoffers a paradigm shift from managing illness toproactively managing wellness by focusing onprevention and early detection/treatment of dis-eases. A recent report by ON World titled “WSNfor Healthcare” estimates that wireless sensornetworks can reduce annual healthcare costs byUS$25 billion by 2012.

APPLICATIONS

BAN is envisaged to unleash a wave of innova-tive, personalized, and integrated medical,lifestyle, fitness, gaming, entertainment, and con-sumer electronics applications. BAN can enablewireless connectivity to implanted cardioverterdefibrillators (ICDs), implanted drug delivery,swallowed camera pills, wearable ECG/EMG/EEG/BP/SpO2/temp monitoring, high risk preg-nancy monitoring, sleep analysis, gait analysis,emotion detection, media players, and headsets.Figure 1 illustrates some of the applicationsBAN technology is intended to support.

TECHNICAL REQUIREMENTS

Table 1 shows the technical requirements of afew typical BAN applications. Notice the widevariation in data rate, bit error rate (BER),delay tolerance, duty cycle, and lifetime, whichrequires scalable solutions with quality of service(QoS) provisions.

Despite significant technological advances inwireless technology, BAN poses unique technicalchallenges primarily due to the diversity of appli-cations and their stringent requirements, as canbe seen from Table 2.

KEY RESEARCH CHALLENGESBAN brings forward several research issues thatneed to be taken into account in the design ofradio frequency (RF) wireless systems.

FREQUENCY BAND SELECTIONTable 3 shows the characteristics of the frequen-cy bands that could potentially be used for aBAN radio. Many BAN devices such as ICDsand hearing aids are expected to be carried glob-ally by their users. Hence, it is desirable thatBAN radio can operate legally worldwide.

The MedRadio core (402–405 MHz) band’ssuperior propagation characteristics for implants,quiet channel properties, and worldwide avail-ability are primary reasons for its popularity forimplant applications.

The industrial, scientific, and medical (ISM),wireless medical telemetry service (WMTS),ultra-wideband (UWB), or MedRadio wingbands (401–402 MHz and 405–406 MHz) can beexploited to support on-body communications.Low-power BAN devices would suffer fromsevere performance degradation in the presenceof high-power technologies in crowded ISMbands thus making them less appealing for high-fidelity medical applications. WMTS bands arealso heavily used, and their use is restricted tohealthcare facilities in the United States. Exploit-ing UWB for wearable applications brings for-ward the issue of coexistence with high-data-ratemultimedia applications. MedRadio wing bandrules are very stringent and restrictive.

These factors have prompted the FCC toconsider opening up 2360–2400 MHz spectrumfor medical BANs. (See the FCC notice of pro-posed rulemaking 09-57, June 2009). The FCC isalso considering allocating up to 24 MHz ofspectrum in the 413-457 MHz range for a “Med-

Figure 1. BAN applications.

Medical

Emergency services

Defence

Entertainment and gaming

Personal health care

Consumer electronics

Sports and fitness

Lifestyle

Firefighters’vital sign monitor

Soldiers’ vital sign monitor Blast dosimeter

Cell phone Music player Headphones Hearing aids

Sport safetyBPHeartrate monitorActivity monitorPedometerWeight scale

Emotion detectionAmbient intelligencePosture detection

Computer games Dance lessons

Gesture detection

ECGEMGEEG

Capsule endoscopeCardiac defibrillators

Patient monitoring Health care

for aging population

Glucose monitoring

Despite significanttechnological

advances in wirelesstechnology,

BAN poses uniquetechnical challengesprimarily due to the

diversity of applications and

their stringentrequirements.

WANG LAYOUT 2/9/10 1:53 PM Page 81

IEEE Wireless Communications • February 201082

ical Micropower Network” service. (See the FCCnotice of proposed rulemaking, 09-20, March2009.) This is intended to support widebandimplanted microstimulator devices that can serveas an artificial nervous system to restore sensa-tion, mobility, and function to paralyzed/impaired limbs and organs.

CHANNEL MODELINGThe channel model plays a crucial role in thedesign of PHY technologies. Experimental chan-nel modeling for implants and wearable devicesis difficult due to involvement of human subjectsand healthcare facilities, both governed by regu-lations. The dynamic environment due to twist-ing, running, multipath, and mobility furthercomplicates the empirical validation of channelmodels. Some channel models and relevant ref-erences can be found in [2].

ANTENNA DESIGNAntenna design for BAN applications is also achallenging problem due to restrictions on thesize, material, and shape of the antenna and hos-tile RF environment. The RF environmentchanges with the wearer’s age, weight gain orloss, and posture changes [3]. Only non-corrosiveand biocompatible material such as platinum ortitanium can be used for implants, which results

in poorer performance when compared to a cop-per antenna. The shape and size of an implantantenna depends on its location and organ,which further limits the freedom of the designer.

PHY PROTOCOL DESIGNPhysical layer protocol design requires minimiz-ing power consumption without compromisingthe fidelity. Ideally, power consumption shouldscale linearly as the data rate is increased from afew kilobits per second to 10 Mb/s, therebyattaining constant energy per bit of information.Operation of BAN devices in the presence ofother high-power devices over unlicensed bandsrequires robust interference-agile protocols.Seamless connectivity should be maintained indynamic environments with the least possibleperformance degradation in terms of latency,data loss, and throughput. Quick turnaroundtime from transmit to receive and fast wakeupfrom sleep mode would contribute significantpower savings.

ENERGY-EFFICIENT HARDWAREExisting wireless technologies draw relativelyhigh peak current and mainly rely on duty cyclingthe radio between sleep (standby) and activemodes to minimize the average current drawn.Innovations in integrated circuits, radio hard-

Table 1. Technical requirements of selected BAN applications [1].

Application Target data rate No. ofnodes Topology Setup

timeP2Platency BER Duty

cycleDesired bat-tery lifetime

Deep brainstimulation 1 Mb/s 2 P2P < 3 s < 250 ms < 10–3 < 50% >3 years

Hearing aid 200 kb/s 3 Star < 3 s < 250 ms < 10–10 < 10% >40 hours

Capsule endoscope 1 Mb/s 2 P2P < 3 s < 250 ms < 10–10 < 50% >24 hours

Drug dosage < 1 kb/s 2 P2P < 3 s < 250 ms < 10–10 < 1% >24 hours

ECG

72 kb/s(500 Hz sample,12-bit ADC, 12channels)

< 6 Star < 3 s < 250 ms < 10–10 < 10% >1 week

EEG

86.4 kb/s(300 Hz sample,12-bit ADC, 24channels)

< 6 Star < 3 s < 250 ms < 10–10 < 10% >1 week

EMG

1.536 Mb/s(8 kHz sample,16-bit ADC, 12channels)

< 6 Star < 3 s < 250 ms < 10–10 < 10% >1 week

O2/CO2/BP/temp/respiration/glucose monitoring,accelerometer

< 10 kb/s < 12 Star < 3 s < 250 ms < 10–10 < 1% >1 week

Audio 1 Mb/s 3 Star < 3 s < 100 ms < 10–5 < 50% >24 hours

Video/med imaging < 10 Mb/s 2 P2P < 3 s < 100 ms < 10–3 < 50% >12 hours

WANG LAYOUT 2/9/10 1:53 PM Page 82

IEEE Wireless Communications • February 2010 83

ware, sensing technologies, and miniaturizationare expected to dramatically lower the peak cur-rent drawn [4]. Hence, devices can operate onlow peak pulse-discharge current supplied bythin film (paper) batteries thereby enabling newmarkets of disposable sensor patch type applica-tions.

Researchers are exploring several promisingtechniques such as low-power listening [5] andwake-up radios [6], which are intended to mini-mize power consumed by idle listening. Sincecrystal is one of the most expensive, bulky, powerhungry, and fragile components, developing acrystal-less radio [7] could further reduce thecost, size, and power requirements.

MAC PROTOCOL DESIGN

BANs are intended to support lifesaving medicalapplications. Hence, safety, security, QoS, andreliability are important metrics besides energyefficiency. Harmonized coexistence of multiplecollocated BANs in crowded places such as hos-pital elevators and wards needs a robust MACprotocol. Efficient duty cycling methods have tobe developed to minimize power consumptionwithout compromising QoS. The MAC protocolshould be able to cope with topology and densitychanges induced by nodes moving in and out ofrange due to body movements.

Message prioritization for real-time vital sign

Table 2. Technical requirements for BAN technology [1].

Characteristic Requirement Desired Range

Operating space In, on or around the body Typically 0–3 m and extendable up to 5 m

Network siz Modest < 64 Devices per BAN

Data rate Scalable From sub kb/s up to 10 Mb/s

Target Lifetime Ultra-long for implantsLong for wearable

Up to 5 year for implantsUp to 1 week for wearable

Target frequency bands Global Unlicensed and Medical bands MedRadio, ISM, WMTS, UWB

Peak Power consumption Scalable e.g., Between 0.001–0.1mW in stand-by mode up to30mW in fully active mode

MAC Scalable, reliable, versatile, self-forming Low power listening, wake up, turn-around andsynchronization

Topology Star, Mesh or Tree Self-forming, distributed with multi-hop support

Device duty cycle Adaptive, Scalable From 0.001% up to 100%

Coexistence Coexistence with legacy devices and self-coexistence

Simultaneous co-located operation of up to 10independent BANs

QoS support and differentiationReal-time waveform data, periodic para-metric data, episodic data and emergencyalarms

• BER: from 10–10 to 10–3

• P2P latency: from 10ms – 250ms• Reservation and prioritization

Fault tolerance No single point of failure Ability to isolate and recover from failures. Self-heal-ing capability

Dynamic Environment Body shadowing (twisting, turning, run-ning), attenuation

Seamless operation of multiple nodes moving in andout of range of each other

Security Many levels, long term, short term, lightweight

Authentication, Authorization, Privacy, Confidentiali-ty, Encryption, Message integrity

Safety/Biocompatibility No harmful effects of long term continu-ous use

Meet regulatory requirements. e.g., FDA, SAR andHIPPA

Setup time and procedure Not to be perceived as a slow or tedious Up to 3 sec

Ergonomic consideration Size, shape, weight and form factorrestricted by location and organ

Non-invasive, unobtrusive, small size, weight andform-factor

Reprogramming, Calibration,Customization

Personalized, integrated, configurable andcontext aware services

Ability to reprogram, recalibrate, tune and configuredevices wirelessly

WANG LAYOUT 2/9/10 1:53 PM Page 83

IEEE Wireless Communications • February 201084

monitoring and guaranteed delivery of alarm mes-sages in emergency situations need to be support-ed [1]. Adaptive frequency agility and channelmigration protocols have to be developed in orderto migrate to a quiet channel when heavy interfer-ence is detected. A simple network setup proce-dure, self-organizing, and self-healing are essentialfor the convenience of unskilled users [8].

QOS AND RELIABILITYQoS and reliability of wireless BAN technologyshould be at par (if not better) with currentwireline technologies to be adopted in clinicalsettings. The QoS framework should be flexibleso that it can be dynamically configured to suitapplication requirements without unduly increas-ing complexity or decreasing system perfor-mance. Real-time life-critical applications ofBANs are not only delay-sensitive but also loss-sensitive. Lost or corrupt alarm/alert packets dueto unreliable wireless networks have serious con-sequences. Fair bandwidth sharing among collo-cated BANs and graceful degradation of serviceare highly desirable. BAN devices have limitedmemory, which means there is little room tostore and retry unacknowledged data. Therefore,strong error detection and correction schemes,and efficient acknowledgment and retransmis-sion mechanisms have to be defined [1].

REAL-TIME CONNECTIVITY OVERHETEROGENEOUS NETWORKS

The full potential of BAN technologies can onlybe realized if the promise of anytime, anywhere,automatic, and continuous connectivity to infra-

structure networks is fulfilled. Connectivity of aBAN to infrastructure networks can be realizedusing a gateway device (e.g., a cell phone orPDA) that transfers data between the BAN andinfrastructure networks such as WLAN, WPAN,or cellular networks. Low-cost limited-rangehigh-capacity WLAN and WPAN infrastructurescan be leveraged for indoor connectivity (e.g.,inside a hospital or at home), whereas lower-capacity longer-range cellular infrastructure canbe leveraged for outdoor connectivity. Thisbrings forward the issues of integration of het-erogeneous wireless networking technologies tosupport seamless roaming and end-to-end QoS.

In non-real-time applications the gatewaymay store data locally and upload it when thegateway is connected to the Internet. Real-timewireless connectivity to infrastructure enableslocation freedom and universal mobility whilebeing monitored. A wireless-enabled ICD canautomatically call an ambulance through a cellphone when it detects cardiac arrest. Similarly, afall detector can automatically send an alarm orcall an emergency center/caregiver upon detect-ing a fall. However, precise location determina-tion technologies are needed to be able toprovide quick assistance to the person in case ofan emergency.

SECURITY AND PRIVACYBAN applications have significant legal, finan-cial, privacy, and safety implications. Hence, pri-vacy, confidentiality, authentication,authorization, and integrity are fundamentalrequirements. Conventional security and privacymechanisms are not suitable for BANs due to

Table 3. Properties of potential frequency bands for BAN.

Frequency(MHz) Acronym

Suitability to BAN applications

Merits Demerits

401~406 MedRa-dio

Worldwide availability, good propagationcharacteristics, quiet channel, medical only

Secondary usage, body-worn applications notallowed in 402–405 MHz core band, large anten-na size, limited bandwidth, stringent rules

433.05~434.79 GeneralTelemetry Good propagation characteristics EU/AU/NZ/SA only, crowded spectrum, large

antenna, limited bandwidth

608~6141395~14001427~1432

WMTS Good propagation characteristics, medical onlyLicensed secondary use limited to healthcare pro-viders inside healthcare facilities in US, limitedspectrum, heavily used

868~870 GeneralTelemetry Good propagation characteristics EU only, limited spectrum, heavily used

902~928 ISM Good propagation characteristics US/Canada only, crowded spectrum

2400~2483.5(2400~2500) ISM Worldwide availability, small antenna, large

bandwidthCrowded spectrum, many standards and tech-nologies

5725~5850 ISM Worldwide availability, small antenna, largebandwidth

Existing standards and technologies, severeattenuation

4200~48007250~8500 UWB Worldwide availability, short range, low power,

huge bandwidthCoexistence with high data rate multimediaapplications, severe attenuation

WANG LAYOUT 2/9/10 1:53 PM Page 84

IEEE Wireless Communications • February 2010 85

limited processing power, memory, and energy,lack of user interface, unskilled users, longevityof devices, and global roaming. Hence, novellightweight and resource-efficient methods haveto be developed for BANs [8].

There are inherent trade-offs among security,availability, and utility goals [9]. A highly securesystem (say an ICD) may preventmedics/paramedics from accessing critical physi-ological information in case of an emergency(cardiac arrest), thereby endangering the life ofthe person. On the other hand, any vulnerabilityin an ICD could be exploited by an adversary toinduce heart failure. Striking the right balanceamong these conflicting goals is very challenging.Global roaming over heterogeneous infra-structure networks further complicates the end-to-end security provisions.

REGULATORY COMPLIANCEMedical devices are subject to stringent regula-tions to promote the safety and well being ofusers. Compliance to applicable regulations setforth by the FCC, U.S. Food and Drug Adminis-tration (FDA), European TelecommunicationsStandards Institute (ETSI), and other regulatoryagencies is essential [10]. For example, in theUnited States the FDA regulates real-time con-nectivity requirements for patient monitoring.Similarly, HIPPA governs the privacy, integrity,and access control requirements for patient datain the United States. The FCC and other author-ities impose limit on specific absorption rate (RFenergy absorbed into human tissue) [11]. Regu-latory compliance is complicated by the fact thatthe device wearer is free to roam globally acrossdifferent jurisdictions.

CANDIDATE WIRELESS TECHNOLOGIES

In this section we review several candidate wire-less technologies that are leading contenders inthe emerging market of BANs.

Table 4 summarizes mainly the PHY charac-teristics of these technologies, and Table 5 high-

lights their merits and demerits. Note that end-to-end performance is determined by the com-plete protocol stack (i.e., including PHY andupper protocol layers).

BLUETOOTH CLASSICBluetooth (http://www.bluetooth.com) is a short-range wireless communication standard thatdefines the link and application layers to supportdata and voice applications. Bluetooth isdepoloyed in more than 2 billion devices, includ-ing cell phones and laptops. Bluetooth’s linklayer employs adaptive frequency hopping spreadspectrum full-duplex signal at a nominal rate of1600 hops/s to reduce interference between wire-less technologies sharing the 2.4 GHz spectrum.Up to eight Bluetooth devices form a short-range network called a piconet.

Bluetooth SIG has developed the BluetoothHealth Device Profile (HDP) that defines therequirements for qualified Bluetooth healthcareand fitness device implementations. This profileis used for connecting application data sourcedevices such as blood pressure monitors, weightscales, glucose meters, thermometers, and pulseoximeters to application data sink devices suchas mobile phones, laptops, desktop computersand health appliances without the need forcables.

BLUETOOTH LOW ENERGYBluetooth Low Energy (BTLE) is an upcomingstandard that provides ultra-low-power idlemode operation, simple device discovery, andreliable point-to-multipoint data transfer withpower save and encryption functionalities. BTLEwill have two implementation alternatives: stand-alone and dual mode. Small devices like watchesand sports sensors will be based on a standaloneBTLE implementation. Dual-mode implementa-tions will use parts of the Bluetooth classic hard-ware, sharing one physical radio and antenna.While connected to a standalone BTLE device(without utilization of classic Bluetooth), thedevice will enjoy the low power consumption

Table 4. Characteristics of candidate technologies for BAN.

Technology Spectrum Modulation Channels Data rate Operatingspace

Peakpower nJ/b Topology Join

time

Bluetoothclassic 2.4 GHz GFSK 79 1–3 Mb/s 1–10 m

on-body only~45mA@3.3V 50 Scatternet ~3 s

BluetoothLow Energy 2.4 GHz GFSK 3 1 Mb/s 1–10 m

on-body only~28mA@3.3V 92 Piconet

Star<100ms

ZigBee 2.4 GHz O-QPSK 16 250 kb/s 10–100 mon-body only

~16.5mA@1.8V 119 Star,

Mesh 30 ms

ANT 2.4 GHz GFSK 125 1 Mb/s 10–30 mon-body only

~22mA@3.3V 73 Star, tree,

or Mesh

Sensium 868 MHz915 MHz BFSK 16 50 kb/s 1–5 m

on-body only~3mA@1.2V 72 Star < 3 s

ZarlinkZL70101

402–405 MHz433–434 MHz 2FSK/4FSK 10 MedRa-

dio, 2 ISM200–800kb/s

2 m in-bodyonly

~5mA@3.3V 21 P2P < 2 s

WANG LAYOUT 2/9/10 1:53 PM Page 85

IEEE Wireless Communications • February 201086

advantages. The key advantages of BTLE arethe strength of the Bluetooth brand and thepromise of interoperability with Bluetooth radiosin mobile phones.

ZIGBEEZigBee (http://www.zigbee.org) defines a net-work, security, and application layer protocolsuite on top of the PHY and MAC layers definedby the IEEE 802.15.4 WPAN standard. ThePHY exploits the direct sequence spread spec-trum technique for interference tolerance andMAC exploits carrier sense multiple access withcollision avoidance (CSMA/CA) for channelaccess. Zigbee supports flexible network forma-tion. For example, full-function ZigBee devicesform a mesh network to which low-duty-cyclereduced-function ZigBee devices connect as leafnodes to form a hybrid topology.

Also, ZigBee is highly optimized for low-dutycycle operation of sensing devices (i.e., a sensorcan shut off the radio most of the time). This isin contrast to Bluetooth, where a slave needs tokeep synchronization to the master, resulting inmuch longer radio on time, and hence muchhigher average power consumption.

The Zigbee Alliance has developed the Per-sonal Health and Hospital Care (PHHC) profilefor interoperable wireless devices enabling secureand reliable monitoring of non-critical, low-acuityhealthcare services targeted at chronic diseasemanagement, obesity, and aging. It provides fullsupport for IEEE 11073 devices including glu-cometers, pulse oximeters, electrocardiographs,weight scales, thermometers, blood pressuremonitors, and respirometers.

ANTANT (http://www.thisisant.com) is a proprietarytechnology designed for general-purpose wirelesssensor network applications. ANT features sim-

ple design, low latency, the ability to trade offdata rate against power consumption, and a netdata rate of 20 kb/s (over-the-air data rate is 1Mb/s). ANT addresses interference issues in the2.4 GHz ISM band by employing a time-divisionmultiple access (TDMA)-like adaptiveisochronous scheme. ANT incorporates low-levelsecurity features.

ANT+ is an open alliance of over 100 mem-ber companies that defines health and fitnessdevice profiles and manages network keys. Cur-rent ANT+ device profiles include Heart RateMonitor, Stride-Based Speed and Distance Mon-itor (footpod), Bicycle Speed and Cadence, andBicycle Power. Some pending ANT+ device pro-files are Weight Scale, Music Player Controls,Multi Sport Speed and Distance, and Environ-ment Sensor.

SENSIUMSensium (http://www.toumaz.com) is a propri-etary ultra-low-power transceiver platform cus-tom designed for healthcare and lifestylemanagement applications. The network adopts amaster-slave architecture, where a body-wornslave node periodically sends sensor readings toa central master node. Joining a network is cen-trally managed, and all communications are sin-gle-hop. To reduce energy consumption, allsensor nodes are in standby or sleep mode untilthe centrally assigned time slot. Sensium fea-tures the leading ultra-low power solution (3 mA@ 1.2 V) for low-data-rate on-body applications.

ZARLINKZarlink (http://www.zarlink.com) has developedan ultra-low-power RF transceiver, ZL70101, formedical implantable applications. It uses a Reed-Solomon coding scheme together with cyclicredundancy check (CRC) error detection toachieve an extremely reliable link. For data

Table 5. Merits and demerits of candidate technologies for BAN.

TechnologySuitability for BAN

Merits Demerits

Bluetoothclassic

Established standard, widespread adoption in cellphones and laptops, health device profile defined,sufficient data rate, low cost

Higher power, limited scalability, limited QoS, coexistence withISM band technologies, limited security, on-body only

BluetoothLow Energy

Interoperable with Bluetooth, lower power than Blue-tooth, leverage Bluetooth brand

Compatibility requirements limit design freedom, limited scala-bility, limited QoS, coexistence with ISM band technologies,on-body only

ZigBeeEmerging standard, healthcare profile defined, lowerpower than Bluetooth, scalable, smaller memoryfootprint

Low data rate, limited QoS, coexistence with ISM band tech-nologies, on-body only

ANT Simple protocol, low power, healthcare device pro-files defined, smaller footprint

Proprietary, limited throughput, limited QoS, coexistence withISM band technologies, general-purpose design, on-body only

Sensium Ultra-low-power, custom designed for BANs Proprietary, low data rate, limited QoS, coexistence with ISMband technologies

ZarlinkZL70101

Ultra-low power, MedRadio compliant, customdesigned for implants Proprietary, implants only

WANG LAYOUT 2/9/10 1:53 PM Page 86

IEEE Wireless Communications • February 2010 87

blocks, a maximum BER of less than 1.5 × 10–10

is provided assuming a raw radio channel qualityof 10–3 BER.

When a ZL70101 transceiver is configured asan implantable medical device (IMD), thetransceiver is usually asleep and in a very lowcurrent state. The IMD transceiver can be wokenup by a specially coded 2.45 GHz wakeup mes-sage or an IMD processor to send an emergencymessage.

The key features of Zarlink ZL70101 areextremely low power consumption (5 mA, con-tinuous TX/RX, 1 mA low power mode), ultra-low-power wakeup circuit (250 nA), andMedRadio compliance.

OTHER TECHNOLOGIESProprietary RF technologies such as BodyLAN(www.fitlinxx.com) and Z-Wave (www.z-wave.com) are also emerging on the horizon.Inductive coupling (IC) and body coupled com-munications (BCC) technologies are also promis-ing. In IC an external coil, held close to thebody, powers an implanted coil by a coupledmagnetic field. Data transfer takes place byaltering the impedance of the implanted loop,which is detected by the external device. Thedata rate of IC is limited, and it cannot initiatecommunication from inside the body.

BCC transceivers are capacitively coupled tothe skin and use the human body as a channel toexchange data. BCC is energy efficient, and miti-gates interference and coexistence issues. BCCcan also be used for user identification and auto-matic formation of BANs, as explained in [8].

STANDARDIZATION

The success of many wireless technologies suchas Wi-Fi and Bluetooth is driven by standardiza-tion. Standardization enables interoperabilityand seamless user experience, and drives downthe cost by exploiting economies of scale. Stan-dardization frees the consumer from vendordependence and empowers them to buy whatbest suits their needs rather than what works.

Interoperability, low cost, and user conve-nience are key enablers for the mass market,which is why there has been growing interest instandardizing healthcare technologies. TheIEEE 802.15.6 Task Group [1] is developing thefirst industrial standard encompassing PHY andMAC layers for BAN. This standard is expectedto fill the critical gap in the peak power vs. datarate graph shown in Fig. 2. Advances in low-power RF technology are likely to lower peakpower consumption significantly, thereby makinglow-cost small disposable sensor patches a reali-ty. It remains to be seen whether the upcomingIEEE 802.15.6 standard outperforms other stan-dardized technologies such as ZigBee and Blue-tooth, and succeeds in penetrating the market.

To enable true plug-and-play interoperability,all layers of the protocol stack, application pro-files, and data exchange formats have to be stan-dardized, which is currently underway in thefollowing groups.

The ISO/IEEE 11073 Personal Health DataWorking Group defines standards and protocolsthat facilitate exchange of health information

between peripheral area network devices andapplication hosting devices such as cell phones,personal computers, and gateways. The groupdefines transport-independent applications andinformation profiles including data formats,exchanges, and terminology.

The Continua Health Alliance has been devel-oping interoperability guideline, testing, and cer-tification program for the emerging personaltelehealth ecosystem targeting disease manage-ment, aging independently, health, and fitness.Continua’s interoperability guideline defines pro-files over existing or upcoming standards. Contin-ua defines interoperability goals around fournetwork interfaces [12]. The Peripheral AreaNetwork Interface connects body area sensorsand actuators to application hosting devices. Forinstance, Bluetooth is a candidate technology forlower-layer connectivity, and IEEE 11073 proto-cols are candidates for higher layers. Similarly, aLAN or WAN interface connects applicationhosting devices to LAN or WAN devices, respec-tively. The Alliance has endorsed ZigBee health-care as Continua’s low-power LAN standard.

SUMMARY

BAN is a promising technology that can revolu-tionize next-generation healthcare and entertain-ment applications. BAN brings out a new set ofchallenges in terms of scalability, energy efficien-cy, antenna design, QoS, coexistence, interfer-ence mitigation, security, and privacy to name afew, which are highlighted in this article. Wealso discuss the state-of-the-art technologies andstandards relevant to BANs, and their meritsand demerits. Developing a unifying BAN stan-dard that addresses the core set of technicalrequirements is the quintessential step to unleashthe full potential of BAN, and is currently under

Figure 2. Target position of emerging BAN technology.

Data rate (kbps)

100

<10 mW

<1 μJ/b

<100 n

J/b

<10 nJ

/b

<1 nJ/b

10

Peak

pow

er c

onsu

mpt

ion

(mW

)0

100

1000

10,000

10 1000 10,000 100,000

BluetoothBluetooth LEZigBeeANTSensiumZarlink

802.15.6

WANG LAYOUT 2/9/10 1:53 PM Page 87

IEEE Wireless Communications • February 201088

discussion in the IEEE 802.15.6 Task Group.In the end several non-technical factors will

also play crucial roles in the success of BANtechnology in the mass market such as afford-ability; legal, regulatory, and ethical issues; anduser friendliness, comfort, convenience, andacceptance. BAN technology needs the blessingof key stakeholders in the healthcare domainincluding the medical electronics industry,patients, physicians, caregivers, policy makers,patient advocacy groups, and payers (insurancecompanies) for it to become the pervasive tech-nology. We, engineers, researchers, and practi-tioners from multiple disciplines, must cometogether and strive hard to overcome technicalroadblocks in order to bring the vision of a ubiq-uitous healthcare network to reality.

REFERENCES[1] B. Zhen et al., “TG6 Technical Requirements Docu-

ment,” IEEE P802.15-08-0644-09-0006; https://mentor.ieee.org/802.15/

[2] K. Yazdandoost et al., “Channel Model for Body AreaNetwork (BAN),” IEEE P802.15-08-0780-06-0006, Mar.2009; https://mentor.ieee.org/802.15/

[3] H. Higgins, “Wireless Communication,” in Body SensorNetworks, G.-Z. Yang, Ed., Springer, 2006, pp. 117–44.

[4] M. A. Hanson et al., “Body Area Sensor Networks: Chal-lenges and Opportunities,” Computer, vol. 42, no. 1,Jan. 2009, pp. 58–65.

[5] J. Polastre, J. Hill, and D. Culler, “Versatile Low PowerMedia Access for Wireless Sensor Networks,” Proc.ACM SenSys, Nov. 2004.

[6] L. Gu and J. Stankovic, “Radio-Triggered Wake-Up forWireless Sensor Networks,” Real-Time Sys., vol. 29,2005, pp. 157–82.

[7] F. Sebastiano et al., “Impulse Based Scheme for Crystal-less ULP Radios,” Proc. IEEE ISCAS, May 2008, pp.1508–11.

[8] J. Espina et al., “Towards Easy-to-use, Safe, and SecureWireless Medical Body Sensor Networks,” in MobileHealth Solutions for Biomedical Applications, P. Ollaand J. Tan, Eds., IGI Global, 2009

[9] D. Halperin et al., “Security and Privacy for ImplantableMedical Devices,” IEEE Pervasive Comp., Jan.–Mar.2008, pp. 30–39.

[10] H.-B. Li et al., “IEEE 802.15.6 Regulation Subcommit-tee Report,” IEEE P802.15-08-0034-10-0006; https://mentor.ieee.org/802.15/

[11] FCC, “Cellular Telephone Specific Absorption Rate(SAR)”, 2009; http://www.fcc.gov/oet/rfsafety/sar.html

[12] R. Carroll et al., “Continua: An Interoperable PersonalHealthcare Ecosystem,” IEEE Pervasive Comp., vol. 6,no. 4, Oct.–Dec. 2007, pp. 90–94.

BIOGRAPHIESMAULIN PATEL [M] (maulin.patel@philips.com) is a seniormember of research staff at Philips Research North Ameri-ca, Briarcliff Manor, New York, since 2006. He received hisM.S. degree in computer science and his Ph.D. degree intelecommunications engineering from the University ofTexas at Dallas in 2002 and 2006, respectively. His researchhas been focused on developing efficient protocols andalgorithms for sensor networks, body area networks, andcognitive radios. His research has led to publications ofnumerous journal and conference articles, a book chapter,and several patent applications. He has been an active con-tributor to the IEEE 802.15.6 Body Area Networking TaskGroup. He has also been the industrial co-principal investi-gator of a project jointly funded by the National ScienceFoundation and Philips. He has served on Technical Pro-gram Committees of GLOBECOM, AINA, ICCCN, BodyNets,CHINACOM, and PIMRC.

JIANFENG WANG [M] (jianfeng.wang@philips.com) receivedhis Ph.D. in electrical and computer engineering from theUniversity of Florida in 2006. He is now a senior memberof research staff at Philips Research North America. Hisresearch focuses on developing system architecture, net-work protocols, and algorithms for wireless healthcare andcognitive radios. He has published over 30 technical papersin the areas of wireless sensor networks (WSN), mobile adhoc networks (MANET), wireless PAN (WPAN), wireless LAN(WLAN), and WiMAX. He currently leads the developmentof the Ecma international standard for cognitive radiooperating in TV whitespace in his capacity as the chieftechnical editor. He is an associate editor of Advances inMultimedia Journal and has served as a TPC member formultiple international conferences including CrownCom,SECON, ICC, GLOBECOM, and ICCCN.

We, engineers,researchers, andpractitioners frommultiple disciplines,must come togetherand strive hard toovercome technicalroadblocks in orderto bring the vision ofubiquitous healthcarenetwork to reality.

WANG LAYOUT 2/9/10 1:53 PM Page 88