University of Southern California WiDeS - Wireless Devices and Systems Group The USC Andrew and Erna Viterbi School of Engineering USC

WIRELESS HEALTHCARE

 

BACKGROUND

Wireless Body Area Networks (WBAN)  are communication networks that allow wireless connectivity among devices operating within very close proximity to the human body. They are formed by sensors and/or actuator devices with limited power and buffering capacity that communicate to the outside world through a body wearable ``local hub'' node with significantly larger buffering and processing capacity. The local hub acts as the WBAN controller, data fusion center and gateway for external connectivity. We live in a world where wireless data connectivity is ubiquitous, through WLANs and 3G/4G cellular networks. For the data rates typically involved in healthcare applications, the rapid evolution of cellular devices from conventional phones  to multi-mode terminals with huge storage capacity and processing power (e.g., iPhone or Android-based devices) clearly points out that the functions of local hub can and should be implemented by such smart-phones.

Another important aspect is health monitoring by means of external sensors, i.e., sensors not worn on the body. For example, an "intelligent house" could be able to monitor vital signs of a handicapped person.
 

CURRENT RESEARCH 


Body Area Networks Channel Measurement

In recent years, there has been increasing interest in Wireless Body Area Networks (WBAN) , mostly because of its potential application in Healthcare. Vital signs of pa- tients such as heart rate, body temperature and blood pressure can be obtained through wireless biomedical sensors attached to the human body. The Ultrawideband(UWB) ra- dio, due to its low-power, high data rate and robustness to fading as been suggested as the technology of choice. In addition to this is the introduction of a multiple input multiple output (MIMO) antenna setup, which would significantly improve channel capacity and the system’s robustness to fading.

Developing any reliable wireless system with optimal performance in a particular chan- nel requires the measurement and modeling of the channel. Unlike typical wireless prop- agation channels which have been exhaustively treated in several literature, there are fewer models available for the WBAN, especially with UWB MIMO setup. In WBAN, electromagnetic (EM) wave transverses the human body either through surface waves or through a diffraction mechanism, its is expected that the channel characterization in this case would be particular different perhaps even unique.

Our research is focused on using a UWB MIMO channel sounding setup (which we specifically designed and built for the WBAN) to investigate the channel responses when tranmsitters and receiver antennas are placed at different parts of the human body. Char- acterization of the channel responses on these parts of the body would help in develop- ing wireless systems designed for these parts of the human body. We hope to create a measurement-based WBAN Model for these on-body propagation channels.


Bluetooth Networks for Healthcare

Smart-phones have a built-in WBAN capability for short-range communication with other devices. Today, this is implemented by Bluetooth, though in the future, this may use the IEEE 802.15.6 standard for body-area networks, which is currently in the process of being standardized.  The sensing signals of interest in healthcare applications require data speeds that range from very low rates (some tens of bit/s) to relatively higher rates (hundreds of kbit/s). Given the random, possibly impulsive, nature of the sensor-generated data, the sensor node must be equipped with a transmission queue to buffer the data. The rate and power of channel from the sensor node to the local hub (uplink) must be such that the queue is stable. Otherwise, eventually the queue buffer overflows and some data will be lost. On the other hand, power efficiency of the sensor nodes is critically important in order to maximize battery life. 

In our work, we have been analyzing this tradeoff, and are providing an algorithm and closed-form results for picking the best data rate and transmit power in a bluetooth (or similar) protocol. We show that smart power control and scheduling strategy can greatly reduce power consumption and thus increase battery lifetime of sensors in WBANs, Such improvements will allow a more widespread use of vital-sign monitoring and collection/transmission of other medical data in wireless health applications.


Radar for Vital Sign Detection


Experiments involving the remote detection of Human Vital signs were conducted using Ultrawideband MIMO Radar. The implementation of such a minimally intrusive respiration/breathing monitoring technique comes with several advantages primarily in healthcare especially for hospital patients in circumstances where body-mounted sensors on patient must be avoided. The use of Ultrawideband signal comes with several advantages such as its robustness to fading and the fine delay resolution. These advantages combined with the MIMO antenna configuration aid the detection, tracking and localization of the human beings and their breathing patterns.

The Ultrawideband MIMO Radar experiment was initially performed with an Artificial Breathing Object (ABO) placed on a precision positioner for a + 15mm displacement in order to replicate a sinusoidal breathing pattern of a Human subject. The critical assumption in this procedure is that the channel remains static throughout the measurement. This experiment was performed with a frequency domain channel sounding SIMO setup using a Vector Network Analyzer (VNA, Agilent 8720ET) in an Anechoic Chamber. The measurement setup is shown in the Fig 1.1 below. 

     

Fig1.1  Breathing detection experimental setup in an Anechoic Chamber


The UWB Channel, which includes the test subject as well, is modeled as a superposition of propagation multipath components (MPC). Estimation of path parameters such as its delay and direction aids in the detection and tracking of the breathing pattern of the Human subject. The measurements results were processed by decomposing the observed channel response into a superposition of plane wave (far-field) propagation paths. For this purpose, a Maximum Likelihood (ML) based path detection framework is formulated and the detected propagation paths parameters are jointly estimated using a high-resolution parameter extraction algorithm known as RIMAX. Results obtained are shown in the figures below

         Fig1.2 PADP inhale for LOS                   Fig. 1.3 PADP inhale for LOS



      Fig1.3 PADP inhale for NLOS                   Fig. 1.4 PADP inhale for NLOS

                  

Another measurement was implemented in a similar fashion to the previous setup except that the Artificial-breathing object (ABO) was replaced with a Human subject and the measurements were limited to two snapshots (“inhale” and “exhale”) for a Line-of-sight (LOS) and Non Line-of-sight (NLOS) scenarios. The Human subject sat in a comfortable chair at about 5m distances from the receiver array while holding breath (“inhale”) and emptying lungs (“exhale”) for each VNA frequency sweep per antenna element position in the virtual linear array at the receiver antenna.

The lengthy sweep time of the Vector Network Analyzer used for the above experiment precludes the observation of dynamic changes in an environment. These measurements were done mainly as a proof-of-concept since an actual monitoring of human respiratory activity would require a real-time observation. In a real-time implementation of the above measurement procedures, the capture of all multipath components (MPC) can be achieved as the channel undergoes even the most minute variations. A real-time channel sounding measurement setup with MIMO arrays has been implemented and introduced. Results for the real-time experiment is currently been processed.


SPONSORS


Under Constructions

PUBLICATIONS:


Journal Papers


E. Rebeiz, G. Caire, and A. F. Molisch, “Energy-Delay Tradeoff and Dynamic Sleep Switching for Bluetooth-Like Body-Area Sensor Networks”, IEEE Trans. Comm., in press.

J. Salmi and A. F. Molisch, “Propagation Parameter Estimation, Modeling and Measurements for Ultrawideband MIMO Radar”, IEEE Trans. Antennas and Propagation, 59, 4257 – 4267 (2011).

Conference Papers


E. Rebeiz, G. Caire, and A. F. Molisch, “Dynamic Power and Rate Control for Bluetooth-Like Body-Area Sensor Networks”, Mobile Health Summit, Washington, DC, Nov. 2010.

J. Salmi, J. Poutanen, K. Haneda, A. Richter, V.-M. Kolmonen, P. Vainikainen, and A. F. Molisch, “Incorporating Diffuse Scattering in Geometry-based Stochastic MIMO Channel Models”, EuCAP 2010, 2010.

J. Salmi, S. Sangodoyin,  and A. F. Molisch, “High Resolution Parameter Estimation for Ultra-Wideband MIMO Radar”, Asilomar 2010.