Introduction

We are exploring applications of wireless sensor network technology to a range of medical applications, including pre-hospital and in-hospital emergency care, disaster response, and stroke patient rehabilitation. Recent advances in embedded computing systems have led to the emergence of wireless sensor networks, consisting of small, battery-powered "motes" with limited computation and radio communication capabilities. Sensor networks permit data gathering and computation to be deeply embedded in the physical environment. This technology has the potential to impact the delivery and study of resuscitative care by allowing vital signs to be automatically collected and fully integrated into the patient care record and used for real-time triage, correlation with hospital records, and long-term observation. This project is supported by grants from the National Science Foundation, National Institutes of Health, U.S. Army, as well as generous gifts from Sun Microsystems, Microsoft Corporation, Intel Corporation, Siemens AG, and ArsLogica. AID-N and CodeBlue featured on CNN, March 14, 2007 (the story does not mention CodeBlue and Harvard, although the CodeBlue software runs the devices).

Wireless Vital Sign Sensors

We have developed a small, wearable wireless pulse oximeter and 2-lead EKG based on the Mica2, MicaZ, and Telos sensor node platforms. These devices collect heart rate (HR), oxygen saturation (SpO2), and EKG data and relay it over a short-range (100m) wireless network to any number of receiving devices, including PDAs, laptops, or ambulance-based terminals. The data can be displayed in real time and integrated into the developing pre-hospital patient care record. The sensor devices themselves can be programmed to process the vital sign data, for example, to raise an alert condition when vital signs fall outside of normal parameters. Any adverse change in patient status can then be signaled to a nearby EMT or paramedic. Specifications: Our wireless vital sign sensors consist of a low-power microcontroller (Atmel Atmega128L or TI MSP430) and low-power digital spread-spectrum radio (Chipcon CC2420, compliant with IEEE 802.15.4, 2.4 GHz, approximate range 100 meters, data rate about 80 Kbps). The devices have a small amount of memory (4-10 KB) and can be programmed (using the TinyOS operating system) to sample, transmit, filter, or process vital sign data. These devices are powered by 2 AA batteries with a lifetime of up to several months if programmed appropriately. The basic hardware is based on the MicaZ and Telos sensor nodes, described above, and a custom sensor board integrating the pulse oximeter or EKG circuitry is attached to the mote devices. Wireless pulse oximeter sensor. Wireless two-lead EKG. Accelerometer, gyroscope, and electromyogram (EMG) sensor for stroke patient monitoring. The Pluto mote, designed here at Harvard, is a scaled-down version of the Telos designed to be small, lightweight, and wearable. The Pluto incorporates a tiny, rechargeable Li-ion battery, small USB connector, and 3-axis accelerometer. It will be used initially for monitoring physical activity and motor functions. The Harvard "Pluto" mote, designed to be small and wearable. Pluto mote with case and wriststrap. Pluto mote in case. The Digital Health Group at Intel has developed a second-generation wearable mote, called SHIMMER, that incorporates a TI MSP430 processor, CC2420 IEEE 802.15.4 radio, triaxial accelerometer, and rechargeable Li-polymer battery. SHIMMER includes a MicroSD slot supporting up to 2 GBytes of Flash memory. This allows SHIMMER to store significant amounts of data (2GB can store more than 80 days of continuous triaxial accelerometer data sampled at 50Hz). SHIMMER can also be configured with an optional Bluetooth radio. The Intel SHIMMER mote, including a triaxial accelerometer. The SHIMMER mote connected to its programming board. CodeBlue is also being used to by the AID-N project at Johns Hopkins Applied Physics Laboratory, which is investigating a range of technologies for disaster response. The AID-N wireless sensors (which run the CodeBlue software) include an electronic "triage tag" with pulse oximeter, LCD display, and LEDs indicating patient status; a packaged version of our two-led EKG mote, and a wireless blood pressure cuff. The ETag sensor hardware was developed by Leo Selavo at University of Virginia. UVa/AID-N "eTag" wireless triage tags, with pulse oximeter, LEDs to indicate patient triage status, and control buttons. UVa/AID-N wireless two-lead EKG (same as above, but with case). UVa/AID-N wireless blood pressure cuff. In collaboration with the Motion Analysis Laboratory at the Spaulding Rehabilitation Hospital, we are developing a seperate sensor board for monitoring the limb movements and muscle activity of stroke patients during rehabilitation exercize. These boards, consisting of 3-axis accelerometer, gyroscope, and electromyogram (EMG) sensors, will permit researchers to capture a rich data set of motion data for studying the effect of various rehabilitation exercizes on this patient population. Our sensor hardware designs are available under an "open source" license to research groups that are interested in experimenting with these devices. We are actively pursuing research collaborations with other medical groups, disaster response teams, and companies interested in this technology. Please contact us at the email address below for more information.

CodeBlue Software Platform

CodeBlue architecture for emergency response. Monitoring limb movement in stroke patient rehabilitation. PDA displaying real-time vital signs of multiple patients. In addition to the hardware platform, we are developing a scalable software infrastructure for wireless medical devices, called CodeBlue. CodeBlue is designed to provide routing, naming, discovery, and security for wireless medical sensors, PDAs, PCs, and other devices that may be used to monitor and treat patients in a range of medical settings. CodeBlue is designed to scale across a wide range of network densities, ranging from sparse clinic and hospital deployments to very dense, ad hoc deployments at a mass casualty site. CodeBlue must also operate on a range of wireless devices, from resource-constrained motes to more powerful PDA and PC-class systems. For more information, please see the IEEE Pervasive Computing article about CodeBlue or this technical report with more details. Part of the CodeBlue system includes MoteTrack, a system for tracking the location of individual patient devices indoors and outdoors, using radio signal information. In MoteTrack, a hospital, clinic, or other area is outfitted with a set of fixed radio beacon nodes that are used to calculate the 3D position of the wireless sensors, which may be attached to patients, carried by physicians or nurses, or attached as "location tags" to medical equipment. MoteTrack has been demonstrated in a building-wide deployment at Harvard and yields an 80th percentile error of about 2 meters, which is more than adequate for many location-tracking applications. The CodeBlue system is currently under development and we anticipate a source code release soon. The MoteTrack system is currently available for download at the link above. Our research focuses on the following areas: Integration of medical sensors with low-power wireless networks Wireless ad-hoc routing protocols for critical care; security, robustness, prioritization Hardware architectures for ultra-low-power sensing, computation, and communication Interoperation with hospital information systems; privacy and reliability issues 3D location tracking using radio signal information Adaptive resource management, congestion control, and bandwidth allocation in wireless networks We are also investigating wide-area event delivery infrastructures for medical care, as part of the Harvard Hourglass project. Such a system will allow seamless access to patient care data by EMTs, emergency department personnel, and other physicians through a variety of interfaces, including handheld PDA and Web-based clients. In collaboration with 10Blade, we are integrating Vital Dust sensors into iRevive, a PDA-based patient care record database. The combined system will allow real-time vital sign capture and triage, automatically inserting time-stamped vital sign data in the patient care record (PCR) prepared by EMTs. This will lead to more accurate reporting and a significant reduction of paperwork for EMSs. Our PDA-based triage application displays vital signs for multiple patients and immediately alerts the EMT to a change in patient status.

People

Faculty Prof. Matt Welsh, Harvard University Prof. Gu-Yeon Wei, Harvard University Dr. Steve Moulton, Denver Children's Hospital and 10Blade Dr. Paolo Bonato, Spaulding Rehabilitation Hospital Dr. Philip Anderson, Beth Isreal Deaconess Medical Center Dr. Jason Tracy, Beth Isreal Deaconess Medical Center Collaborators Dr. Lucila Ohno-Machado, Brigham and Women's Hospital Prof. Mark Gaynor, Boston University Tia Gao, AID Networks Leo Selavo, University of Virginia Nhedti Colquitt, CIMIT Benjamin Kuris, Intel Digital Health Advanced Technology Group Fernando Pianegiani, University of Trento Prof. Jack Stankovic, University of Virginia Students Bor-rong Chen, PhD student, Harvard University Konrad Lorincz, PhD student, Harvard University Jason Waterman, PhD student, Harvard University William Christopher Newman, Undergraduate research assistant, Harvard University Tammara Massey, PhD student, UCLA Alex Alm, Johns Hopkins Applied Physics Laboratory Research Staff Dan Myung, 10Blade Project Alumni Thaddeus Fulford-Jones (now at MIT) David Malan Keith Berkoben (now at Intermed Advisors) Mervin John (now at Intermed Advisors) Nada Hashmi Victor Shnayder Stephen Dawson-Haggerty (now at UC Berkeley)

Software Release

You can download a prototype release of the CodeBlue software and hardware description files here: http://www.eecs.harvard.edu/~mdw/proj/codeblue/release We have also created a public mailing list for users of the CodeBlue software. Use this list for questions, comments, and discussion about the system as well as to receive notification of updates. To subscribe, visit this page: https://www.eecs.harvard.edu/mailman/listinfo/codeblue-users

Publications, Posters, and Abstracts

Wireless Medical Sensor Networks in Emergency Response: Implementation and Pilot Results, Tia Gao, Christopher Pesto, Leo Selavo, Yin Chen, JeongGil Ko, JongHyun Lim, Andreas Terzis, Andrew Watt, James Jeng, Bor-rong Chen, Konrad Lorincz, and Matt Welsh. In Proceedings of the 2008 IEEE International Conference on Technologies for Homeland Security, Waltham, MA, May 2008. (PDF) LiveNet: Using Passive Monitoring to Reconstruct Sensor Network Dynamics, Bor-rong Chen, Geoffrey Peterson, Geoff Mainland, and Matt Welsh. To appear in Proceedings of the 4th IEEE/ACM International Conference on Distributed Computing in Sensor Systems (DCOSS 2008), Santorini Island, Greece, June 2008. (PDF) The Advanced Health and Disaster Aid Network: A Light-weight Wireless Medical System for Triage, Tia Gao, Tammara Massey, Leo Selavo, David Crawford, Bor-rong Chen, Konrad Lorincz, Victor Shnayder, Logan Hauenstein, Foad Dabiri, James Jeng, Arjun Chanmugam, David White, Majid Sarrafzadeh, and Matt Welsh. IEEE Transactions on Biomedical Circuits and Systems, in press, 2007. Analysis of Feature Space for Monitoring Persons with Parkinson's Disease With Application to a Wireless Wearable Sensor System, Shyamal Patel, Konrad Lorincz, Richard Hughes, Nancy Huggins, John H. Growdon, Matt Welsh, and Paolo Bonato. In Proceedings of the 29th IEEE EMBS Annual International Conference, Lyon, France, August 2007. Participatory User Centered Design Techniques for a Large Scale Ad-Hoc Health Information System, Tia Gao, Tammara Massey, Leo Selavo, Matt Welsh, and Majid Sarrafzadeh. In First International Workshop on Systems and Networking Support for Healthcare and Assisted Living Environments (HealthNet'07), San Juan, Puerto Rico, June 2007. (PDF) Design of a Decentralized Electronic Triage System, Tammara Massey, Tia Gao, Matt Welsh, and Jonathan Sharp. To appear in Proceedings of the American Medical Informatics Association Annual Conference (AMIA 2006), Washington, DC, November 2006. Ad-Hoc Multicast Routing on Resource-Limited Sensor Nodes, Bor-rong Chen, Kiran-Kumar Muniswamy-Reddy, and Matt Welsh. In Proceedings of the Second ACM/Sigmobile workshop on Multi-hop Ad Hoc Networks: from theory to reality (REALMAN'06), Florence, Italy, May 2006. (PDF) MoteTrack: A Robust, Decentralized Approach to RF-Based Location Tracking, Konrad Lorincz and Matt Welsh. Personal and Ubiquitous Computing, Special Issue on Location and Context-Awareness, Springer-Verlag, 2006. In press. Communicating Data from Wireless Sensor Networks using the HL7v3 Standard, S. Baird, S. Dawson-Haggerty, D. Myung, M. Gaynor, M. Welsh, and S. Moulton. In Proceedings of the International Workshop on Wearable and Implantable Body Sensor Networks (BSN 2006), April 2006. Vital Signs Monitoring and Patient Tracking Over a Wireless Network, Tia Gao, Dan Greenspan, Matt Welsh, Radford R. Juang, and Alex Alm. In Proceedings of the 27th IEEE EMBS Annual International Conference, September 2005. (PDF) Improving Patient Monitoring and Tracking in Emergency Response, Tia Gao, Dan Greenspan, and Matt Welsh. In Proceedings of the International Conference on Information Communication Technologies in Health, July 2005. (PDF) A Wireless, Low-Power Motion Analysis Sensor for Stroke Patient Rehabilitation. Mervin John, Thaddeus R.F. Fulford-Jones, Paolo Bonato, and Matt Welsh. Abstract 143281, Biomedical Engineering Society (BMES) 2005 Annual Fall Meeting, Baltimore, MD, September 28-October 1, 2005. Sensor Networks for Medical Care, Victor Shnayder, Bor-rong Chen, Konrad Lorincz, Thaddeus R. F. Fulford-Jones, and Matt Welsh. Harvard University Technical Report TR-08-05, April 2005. (PDF) Sensor Networks for Emergency Response: Challenges and Opportunities, Konrad Lorincz, David Malan, Thaddeus R. F. Fulford-Jones, Alan Nawoj, Antony Clavel, Victor Shnayder, Geoff Mainland, Steve Moulton, and Matt Welsh. In IEEE Pervasive Computing, Special Issue on Pervasive Computing for First Response, Oct-Dec 2004. (PDF) iRevive, a Pre-hospital Mobile Database for Emergency Medical Services, Will Tollefsen, Marissa Pepe, Dan Myung, Mark Gaynor, Matt Welsh, and Steve Moulton. To appear in International Journal of Healthcare Technology and Management (IJHTM), Summer 2004. A Portable, Low-Power, Wireless Two-Lead EKG System, Thaddeus R. F. Fulford-Jones, Gu-Yeon Wei, and Matt Welsh. In Proceedings of the 26th IEEE EMBS Annual International Conference, San Francisco, September 2004. (PDF) CodeBlue: An Ad Hoc Sensor Network Infrastructure for Emergency Medical Care, David Malan, Thaddeus Fulford-Jones, Matt Welsh, and Steve Moulton. International Workshop on Wearable and Implantable Body Sensor Networks, April 2004. (PDF) Wireless Sensor Networks for Emergency Medical Care. Presented at GE Global Research, March 8, 2004. Vital Dust: Wireless sensors and a sensor network for real-time patient monitoring. Dan Myung, Breanne Duncan, David Malan, Matt Welsh, Mark Gaynor, and Steve Moulton. 8th Annual New England Regional Trauma Conference, November 20-21, 2003, Burlington, MA (Poster). Resuscitation Monitoring With a Wireless Sensor Network, Matt Welsh, Dan Myung, Mark Gaynor, and Steve Moulton. American Heart Association, Resuscitation Science Symposium. To appear in Supplement to Circulation: Journal of the American Heart Association, October 28, 2003. (Poster), (Abstract)

Related Projects and Links

University of Virginia - LCD and button module for Mica2/MicaZ motes

Contact:
Prof. Matt Welsh
Harvard University
Phone: (617) 495-3311