Design Aspects of BSNs

From Capsil Wiki

Jump to: navigation, search

Contents

Overview

The ultimate promise of body sensor networks is that they will become pervasive and operate unobtrusively behind the scenes without significant human set-up or intervention. While a lot has been written about the technical aspects of body sensor networks, this section will be concerned with the design aspects of body sensor networks. The focus here is less on hardware and technical specifications but more on how 'practical' and useable these devices are and how easily they can integrate in to peoples lives unobtrusively. Examples will be presented both from academia and industry to demonstrate these different design aspects.

Key Design Considerations

There has been a realisation that if these sensors are to get truly smaller according to Moore's Law then there exists the possibility of these sensors to become ubiquitous and operate reliably without any human set up. The early work regarding wireless sensor network has been very focussed on technical details such as routing algorithms, quality of service and power optimisation and this work continues. A parallel strand of research now is underway as to how in practical terms these sensors can interface with humans. For example sensor systems need to be mobile and reliable. As an example, there is little point in system that fails if the person moves too much or if body temperature or sweat corrupts measurements. Also, the wearing comfort of the sensor system is important, again there is little value in a system that needs to be so tight (to maintain good electrical contact) that it is uncomfortable to wear or cuts off blood circulation. The ideal scenario of a body sensor network is one that is embedded in to an everyday item that we use such as furniture, clothing, bed or jewellery. However these present design challenges which will be the topic of discussion in the following sections.

Usability

A requirement of a truly ubiquitous and pervasive monitoring system is that it is 'usable' i.e. it fits in to peoples lives unobtrusively and is 'practical'. This section is focussed on initiatives that concentrate on these aspects of body sensor networks. Quite a lot of recent research has been focussed on wearable sensors and sensors that can be integrated in to garments (clothes).

Research Efforts Concerning Useability

An early project in this area was the EU project WEALTHY Project - Wearable Healthcare Systems for vital signs monitoring [1] which aimed to weave fabric material with piezoresistive materials to form a garment capable of measure ECG of patients undergoing cardiac rehabilitation. Further work was undertaken by the EU funded MyHeart Project [2] which was a consortium of 33 partners in 11 countries led by Philips. The focus here again was to achieve a 'practical' wearable sensor solution for measuring ECG reliably. A further EU project called Project STELLA - Stretchable Electronics for Large Area Applications [3] was launched in 2006 and will run through 2010. It is entirely focussed on the concept of bendable and stretchable electronics i.e. substrates and PCBs, and how these could be integrated in to wearable devices. The CONTEXT Project - Contactless Sensors for Body Monitoring Incorporated in Textiles [4] project was a EU funded STREP project which completed in mid 2008 and developed more garment-type prototypes for measuring physiological vital signs. A particular application investigated here is the area of body sensor networks in sports applications. The OFSETH - Optical Fibre Sensors Embedded in to technical Textiles for Healthcare monitoring [5] project used a slightly different principle to the previous projects which mainly looked at the merging of electrical devices and textiles. The OFSETH project was concerned with using pure optical methods (instead of electrical) of interfacing with textiles.

Commercial Products Concerning Usability

For a round up of some of what has been happening commercially follow this link Commercial Products Concerning Usability

Reliability and Stability

A major challenge with wireless sensor networks in general to date is how to ensure that communications is reliable and robust. Reliability can be defined as the ability of the network to ensure reliable data transmission in a state of continuous change of network structure.

The Reliability Dilemma - In typical wireless ad hoc networks, reliability and scalability are always inversely coupled. In other words, it becomes more difficult to build a reliable ad hoc network as the number of nodes increases. This is due to the network overhead that comes with the increased size of the network. In ad hoc networks, the network is formed without any predetermined topology or shape. Therefore, any node wishing to communicate with other nodes needs to generate more packets than just the data packets i.e. control packets. These extra packets are generally called "control packets" or "network overhead". Increased overhead is unavoidable in a larger scale wireless sensor network to keep the communication paths intact. In typical ad hoc networks, the overhead increases exponentially as the network size grows - therefore reliability and scalability are closely related - they act against each other. Other issues such as responsiveness and mobility are also on an inverse relation with network efficiency. For a network to be more responsive (e.g. a mobile network) more control packets needed, therefore more overhead (also more battery power). Some techniques to address these issues are available including Dynamic Routing.

For smaller networks such a personal health monitoring and networks where there is a relative amount of static nodes, these effects are minimised, however for larger networks such as in a hospital environment, where many nodes may be present, and there may be a lot of mobility, the issues of security and reliability are of tremendous importance in ensuring reliable data transfer.

Security of Sensor Networks

Of major concern is the security aspects of sensor networks. As networks grow, the vulnerability of network nodes to physical and software attack increases. Attackers can also obtain their own commodity sensor nodes and induce the network to accept them as legitimate nodes, or they can claim multiple identities for an altered node. Once in control of a few nodes inside the network, the adversary can then mount a variety of attacks—for example, falsification of sensor data, extraction of private sensed information from sensor network readings, and denial of service attack. Therefore routing protocols must be resilient against compromised nodes that behave maliciously. Ensuring that sensed information stays within the sensor network and is accessible only to trusted parties is an essential step toward achieving security. Data encryption and access control is one approach. Another is to restrict the network’s ability to gather data at a detail level that could compromise privacy. For example, in a healthcare environment storing data in an anonymous format and removing any personal referencing information. Another approach is to process queries in the sensor network in a distributed manner so that no single node can observe the query results in their entirety. This approach guards against potential system abuse by compromised malicious nodes.

Some of the more common security considerations of any sensor network include the following; Eavesdropping and Denial of Service Attacks. See the Privacy & Security Capsil for more information.

Examples of Systems Demonstrating Security and Reliability

Follow this link for details of some detailed Examples of Systems Demonstrating Security and Reliability

Biocompatibility

The interface of the sensor node to the human is all important in body sensing networks. The material from which the sensor is made must be capable of performing the sensing robustly while at the same time causing no side effect problems to the human interface (i.e. skin or other surface). Examples of such problems could be itching or rash or possibly even cell tissue damage. There is a lot of material that discusses the materials and treats the various chemical/physical processes, however there is not a lot of material that deals explicitly with biocompatibility and presents demonstrable solutions with scalable robust working prototypes.

In the book "Body Sensor Networks" [6] there is an excellent treatment of the issues of biocompatibility of sensors. The key property of biocompatibility is for the surface area to be as large as possible. Fundamental issues such as sensor fouling, sensor adsorption, and the use of micro sensor arrays to overcome these effects are discussed. The development of novel electrode made from Boron doped diamond films is also discussed as one promising advance in the area of sensor biocompatibility. Diamond is also attractive from a number of fronts in these applications due to its stability, chemical inertness and low surface oxygen content, but again it is too early to say if this is truly a breakthrough development.

Examples of Systems Demonstrating Biocompatibility

This linked section will detail Examples of Systems Demonstrating Biocompatibility. This biocompatibility is either explicit or inferred (no available literature).

RF Emissions and Interference Aspects

Further issues to be considered here are the body effects on RF signals i.e. how the body effects RF signals. Higgins [7] has shown that antenna design is crucial in body sensor networks and the increase in dielectric constant from having proximity to a human body actually works in the designers favour in that smaller antenna can be used.(physically small antennas though do have their drawbacks also!). Other authors such as Alomainy et al [8] have shown that antenna design must be carefully investigated with the proposed sensor and issues such as sensor size, orientation impedance matching, gain and efficiency need to be carefully considered. Less reported also are the effects of RF on tissue, the so called heating effect of RF signals on human cells. It is generally assumed that if power levels are kept low that and that if 'exposure time' i.e. time the device in 'on' is kept low (low duty cycle), then this heating effect is negligible. However low power, low duty cycle can lead to unreliable networks so a compromise is needed. Also to be considered are the potential sources of interference that can take place between similar wireless devices operating in proximity to each other and other non-wireless devices operating in the area. For example an implanted pacemaker cannot be allowed to interfere with a wireless ECG monitor worn on the body, or a 'noisy' fluorescent light can not be allowed to interfere with a wireless pulse oxidation monitor.

Follow this link for more information on RF Emissions and Interference Aspects.

Privacy & Security

As sensor systems become more and more pervasive and truly start to operate in the background unobtrusively, issues of human privacy become a major concern. Radio Frequency Identification (RFID) has been leading the charge in the deployment of 'intelligent' network nodes being widely disseminated and many of the privacy issues brought about by RFID are common to body sensor network nodes. Many examples of body sensor networks treat privacy through security mechanisms i.e. encryption and data protection throughout the hardware and software layers, however there is a whole area of privacy that needs to be proactively addressed if sensor networks are to reach their true potential in a timely manner. Some of the key Privacy concepts include:

*Fundamentals of Freedom A popular dictionary defines privacy as: "The quality or condition of being secluded from the presence or view of others. The state of being free from unsanctioned intrusion: a person's right to privacy". The commonly accepted definitions of Privacy that are built in to much legislation use the concepts of a person’s right to be free from unreasonable search and seizure and intrusion. It also states that the protection of personal information is a fundamental right. The United Nations Universal Declaration of Human Rights [9] which is exactly 50 years old, enunciates the fundamental right to privacy and can be viewed at www.un.org/overview/rights.html . The EU Directive 95/46/EC "on the protection of personal data" [10] is the guiding policy document within then EU. It affirms the right to privacy, transparency, legitimacy of purpose and proportionality. It also deals with transfer of data between boundaries including third party countries (non-EU).

*OECD Guidelines on Privacy The Organisation for Economic Cooperation and Development (OECD) released Guidelines for Data Protection and Privacy [11] in 1980 which was based on the US initiated Fair Information Practises (FIPS) [12] policy. These Guidelines were reaffirmed in 1998 as still relevant and form the basis of much legislation worldwide. These key principles form the basis of much privacy legislation world-wide and all wireless sensor network systems must at a minimum comply with these guidelines.

The are major issues around privacy raised and ethics raised by the evolution towards a so called ubiquitous computing society and as the evolution progresses, these issues become more important. In fact privacy/ethics will be a fundamental barrier to adoption unless handled proactively, as presently technology has been outpacing policy.

For a more detailed treatment on these topics see the Privacy & Security Capsil.

Not Everyone Likes this Technology

With the above concerns in mind, many groups have sprang up that lobby and campaign against the ubiquitous deployment of wireless technology. As RFID systems have been ahead of commercial wireless sensor networks they have been in a way the lightning rod for a lot of potential privacy issues that wireless sensor networks will encounter and the issues are almost identical (actually they will be tougher for sensor networks as we move to truly ubiquitous networks!). An example of a lobby group who are opposed to this kind of technology is called CASPIAN (Consumers Against Supermarket Privacy Invasion And Numbering) [13]. They refer to RFID devices as spychips and are concerned with personal information being used in an unauthorised manner. Quoting from their website (spychips.com)... "We do believe, however, that these technologies pose serious risks to consumers, and we have called on the world's shoppers to reject them. CASPIAN hopes to see both technologies (RFID and supermarket loyalty cards) ultimately fail in the marketplace as a result of consumer opinion. In the long run, outright market failure would offer more effective consumer protections than temporary legislative band-aids. (What the legislature grants, the legislature can easily take away, limiting the field of consumer espionage to itself."


This gives a flavour of some of the difficulties in the privacy debate and shows the need for it to be handled proactively and not 'bolted on' once the technology is ready i.e. as an afterthought. The area of healthcare is one where the 'hearts and minds' debate can be easier to argue i.e. people would be willing to trade off some privacy if their wellbeing or quality of life is improved. However for normally healthy people, the argument can be lost if for example as a result of remote monitoring, a persons home is broken in to and medication stolen or if information on a sensitive medical condition is disseminated, or if a person finds that their medical insurance premiums are fluctuating based on data they collected at home (i.e. one bad day sees the premium skyrocket....). The linkage of personal health data to 'other systems' such as insurance databases, national security etc (databases that are ostensibly in the consumer interest) needs to be handled very carefully to avoid a backlash from consumers and ultimately market failure.

Examples of Body Sensor Networks Concerned with Design Aspects

Academic

Examples of Body Sensor Networks from Academia aimed at addressing some the Design considerations

Commercial

Here are some examples of body sensor networks that that have been commercialised, thereby overcoming at least some of the design constraints that we have been discussing. Obviously going from a lab prototype is a long journey in terms of certification and reliability and some of these systems to date have proven very successful.

  • Medtronic Reveal Insertable Loop Recorder [14]
  • Cardionetics C.Net5000 - 24-Hour Ambulatory ECG Monitor with Instant Analysis [15]
  • Kiwok AB - BodyKom SeriesTM ECG - Kiwok AB [16]
  • CardioMEMS - EndoSure [17]
  • Tunstall Fall Sensor [18]

Related EU Projects

Summary and Recommendations For Further Work

There is quite an amount of information on aspects of sensor networks in healthcare applications and many of them are detailed here. There does appear to be quite a lot of institutions doing similar work in this area for example there are many published details on ECG, EMG, Pulse Ox etc monitoring systems and many different approaches to the problem of reliable and robust monitoring. Interestingly there are very few published end-to-end examples of systems that have been designed all the way to the patient (and back again). More work is certainly needed on aspects of end to end reliability and security and demonstrating robustness in large sensor networks where parameters change frequently (data rates, number of nodes, latency, mobility etc). There needs to be agreement on a 'standard' approach to such platforms where the underlying architecture, policies and aspects of hardware are defined according to requirements such as HIPAA and the EU privacy directive.

Given the amount of information also on aspects of sensor network design, there is very little information on biocompatibility of sensor materials. Much of the efforts here are at the basic research level (materials science, garment fabrication etc) and this is appropriate as the promise of wearable devices is quite considerable. However there needs to be some initiatives aimed at investigating the relationship between the sensor interface and the human body/skin, long term. Some of the commercial organisations with implantable devices have obviously performed a lot of work in this area, however information is usually protected and so not much can be gleaned form lessons learned here.

More work also needs to be done of the RF effects of sensor networks on the human body. For example RF produces a heating effect which could possibly damage human cells. Even with low emitted and radiated power levels, it remains to be proven what the effect on human tissue over time (and with many sensors on the body) would actually be.

If wireless sensor networks are to become truly pervasive and ubiquitous, the privacy debate will be a very tricky one indeed and one that needs to be handled proactively. If it is not handled proactively then the technology can be developed and available but will sit on the shelf with no market demand! Research needs to change course slightly from the technical areas and move into the societal, ethnographic and demographic areas. Concerns such as profiling, 'big brother', 'one big database' etc need to be addressed up front and policies developed and agreed ahead of the technology becoming mature.

References

  1. ↑ http://www.wealthy-ist.com/
  2. ↑ http://www.hitech-projects.com/euprojects/myheart/home.html
  3. ↑ http://www.stella-project.eu
  4. ↑ http://www.context-project.org
  5. ↑ http://www.ofseth.org
  6. ↑ "Body Sensor Networks" By Guang-Zhong Yang, Magdi Yacoub, Contributor Guang-Zhong Yang, Magdi Yacoub Published by Birkhäuser, 2006 ISBN 1846282721, 9781846282720
  7. ↑ http://212.67.202.176/~armms/images/conference/1128506900.pdf
  8. ↑ Akram Alomainy, Yang Hao and Frank Pasveer "Modelling and Characterisation of a Compact Sensor Antenna for Healthcare Applications" http://www.elec.qmul.ac.uk/people/akram/papers/Philips_SensorAntenna_BSN2007.pdf
  9. ↑ http://www.unhchr.ch/udhr/
  10. ↑ http://ec.europa.eu/justice_home/fsj/privacy/law/index_en.htm
  11. ↑ "OECD Guidelines on the Protection of Privacy and Transborder Flows of Personal Data". http://www.oecd.org/document/18/0,3343,en_2649_34255_1815186_1_1_1_1,00.html
  12. ↑ "Fair Information Practice Principles". US Federal Trade Commission. http://www.ftc.gov/reports/privacy3/fairinfo.shtm
  13. ↑ http://www.spychips.com/
  14. ↑ http://www.medtronic.com/physician/reveal/
  15. ↑ http://www.cardionetics.com/cnet5000.php
  16. ↑ http://www.kiwok.se/index.php?lang=2
  17. ↑ http://www.cardiomems.com/
  18. ↑ http://www.tunstall.co.uk/assets/Literature/6_1_41falls_management_solutions_sheet.pdf

Navigation

Retrieved from "http://www.capsil.org/capsilwiki/index.php/Design_Aspects_of_BSNs"
Personal tools