Mobile data acquisition and tele-transmission by PDA.

Nedelcu, Adrian Valentin ; Sandu, Florin ; Borza, Paul Nicolae 等

1. INTRODUCTION

The problem addressed by the authors is the integration of a system for data acquisition, logging and/or tele-transmission that can be personal and portable--e.g. by a patient (Perednia, 1995). Most of the pre-existing comparable solutions are highly specialized and proprietary, being not always very affordable. The accomplished configuration is using common PDA-s that have a plugged-in micro DAQ card--general purpose and with common interface (CF--"Compact Flash"), a high-capacity common memory card (SD--"Secure Digital") and run an universal instrumentation software--NI LabVIEW (LV). LV can perform not only data acquisition and logging but also processing (e.g. digital filtering, identification-classification, compression etc) and communication.

If the PDA has also "SmartPhone" capabilities, it can also transfer the data via mobile communications (e.g. towards a central server). If not, an inexpensive alternative (that was integrated by the authors) is the connection to the owner's mobile phone that can be controlled, for instance, via AT commands (for dial-up, GPRS and/or SMS transmissions). Local data-logging can be used for symptoms' collection and detailed diagnosis, while mobile tele-transmission can be used mainly for updates and alarms (e.g. in case of patient's status deterioration).

2. STATE OF THE ART

As demonstrated (e.g. by Philips "Motiva" Interactive Healthcare Platform--2007), home tele-monitoring is the basis of remote patient management technology and can be enhanced with two-way communications that allow better personalization and support. Personal means of communication, like mobile phones and SmartPhones or PDA-based "communicators" are more and more involved in different ways of monitoring, logging (Boxwala et al, 2004), tracing but also messaging (including alarms when threshold values are reached) and notifications of clinical decisions (De Clercq, 2006)--all these in a way that should be tailored to the patient's capability and needs.

For remote health-care, following the famous monitor invented by Norman Holter, many portable devices were developed for acquisition, logging, processing and transmission of bio-potentials (for ECG, EEG) as well as pulse, oximetry etc. Digital signal processing (DSP) enable morphologic analysis (e.g.--for the ECG--analysis of the P wave, of the QRS complex, of the T and U waves) and allow diagnosys of heart rate variability, bundle branch blocks etc.

3. RESEARCH COURSE

The authors developed different tele-monitoring systems in the last 8 years. A "twin-microcontroller" solution had one microcontroller allocated to the control of data acquisition and processing and another one to the management of communications (interfacing and control of a GSM modem). Such dual-processor solutions are common even to SmartPhones (only recently DSP and general purpose processing were brought to a common core).

3.1 System's architecture

The most recent system implemented by the authors, with the architecture depicted in Fig. 1, consists of a 4 channels, 200kS/s DAQ card, namely NI's CF 6004, connected in the Compact Flash slot of a HP iPAQ 2210 PDA. The authors have used in their test configuration a portable ECG probe consisting of electrodes and signal conditioning circuitry. Because most Compact Flash PDAs currently available on the market do not embed mobile communications technologies such as GSM/UMTS or WiFi, which would allow the system to transfer the acquired data directly to a remote server, the authors decided to use a common mobile phone (e.g. the one of the patient) as a GSM/GPRS modem. The PDA is connected to the mobile phone through Bluetooth and is using standard AT commands to transfer a message (that incorporates the acquired data and/or specifically parameters computed out of them) to the mobile phone that should transmit these data to a remote server using GPRS (and/or at least SMS). In the author's vision this server could be a Hospital Server which stores data about patients with cardiac problems in a database (MySQL and Java were used for the pilot programming).

3.2 Technical Solution and Results

The software for both data acquisition and transmission was implemented using LV's PDA Module, a subset of NI LV functions specifically designed for PDAs. The data acquisition is performed using NI DAQmx Base, a set of drivers and functions for the DAQ boards made by NI. A DAQ task has to be defined, specifying attributes such as sampling rate, used channels, number of samples etc. This is accomplished using the DAQ Configuration Utility included in LV.

[FIGURE 1 OMITTED]

A measurement cycle (Fig. 2) consists of starting a specific DAQ task (using DAQmxBase Start Task.VI), reading data from the DAQ card (using DAQmxBase Read.VI) in a loop until the user stops the system or until an error occurs and then stopping the DAQ task (using DAQmxBase Stop Task.VI). The acquired data can be viewed both online, on the PDA screen using a Waveform Chart (see the VI Panel on the screen, in Fig.3), or offline by logging the data into spreadsheet files (XLS format). The maximum frequency in the spectrum of an ECG signal is 100Hz (Rijnbeek, 2001); a sampling rate of 250 S/s has been chosen to meet Nyquist's criterion. A relevant measurement cycle implies acquiring 1000 samples which can be stored in 10 KB spreadsheet files. The log file obtained after 24 hours of continuous data acquisition (required for monitoring cardiac activity over a longer period) would be 216 MB large. Generated files which are as big as 10 KB can be directly transferred over GPRS to a remote server (which in the authors' vision could be a Hospital Server). If the acquired signal needs to be monitored over a longer period of time, thus resulting in larger files which are not fit for transfer over GPRS, these files can be stored on the SD card of the PDA and then "downloaded" to a PC in order to be analysed by a cardiologist. The transfer of the acquired data over Bluetooth to the patient's mobile phone (which acts as a GPRS modem) is handled by a subVI (Fig. 4) which opens a Bluetooth connection to that phone (using the LV function Bluetooth Open Connection.VI). The Bluetooth Open Connection function receives two main parameters, namely the MAC address of the mobile phone to which one wants to connect and the channel number (which in the case of the Bluetooth dial-up service is 0). The subVI then sends a sequence of standard AT commands for initiating and performing a GPRS transfer.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

These commands are sent over Bluetooth by calling LV's Bluetooth Write.VI function. They are as follows:

AT+CGATT=1--attaches the mobile unit to the GPRS network

AT+CGDCONT=1,"IP","internet"--defines a mobile operator specific PDP context

AT+CGACT=1,1--activates that context

ATD *99***1#--connects to the GPRS network

4. CONCLUSIONS AND FURTHER RESEARCH

The present implementation proved the possibility of using powerful instrumentation hardware and software (NI CF6004 and NI LV) embedded in a very miniaturized format (for PDA) for mobile tele-monitoring. The research was financed by the Romanian Minstry of Education and Research, through its National Center for Programs' Management, in the frame of the 4th Program--"Partenerships in Prioritary Domains". The project is named "BIOMED-TEL" ("Acquisition of Biomedical Signals and their Tele-transmission via Mobile Computing Equipment")--Contract PNII--P4, nr. 11-057/2007--Code PO-04-Ed1-R0-F5. It is co-ordinated by "Transilvania" University of Brasov-Romania. Further development is centered on AVR 32 bit microcontrollers produced by Atmel (3 times more computational powerful compared to the standard ARMv5--due to its pipeline architecture, with 50% more efficient in memory usage, with 3 times less power consumption). Linux OS would ease programming (and NI LV for Linux could also be used, to overcome specific DSP complexity and to offer some protocol stacks) and would allow the development of an open source solution, which could be further extended by other researchers.

5. REFERENCES

Boxwala, A.A., et al. (2004), GLIF3. A Representation Format for Sharable Computer-Interpretable Clinical Practice Guidelines, Journal of Biomedical Informatics, Vol. 37, Issue 3, June 2004, pp 147-161--ISSN: 1532-0464

De Clercq, P.A., et al (2006). Approaches for Creating Computer Interpretable Guidelines that Facilitate Decision Support, Journal of Artificial Intelligence in Medicine, Vol. 31, Issue 1, pp 1-27--ISSN 0933-3657

Perednia, D.A.; Allen, A. (1995). Telemedicine Technology and Clinical Applications, Journal of the American Medical Association, Vol. 273, Issue 6, February 8, 1995--pp 483-488, ISSN 0098-7484

Philips Medical Systems (2007) "Motiva" Interactive Healthcare Platform, Available on: http://www.medical.philips.com/main/products/telemonitori te/products/motiva/; Accessed on: 2008-04-12

Rijnbeek, P.R., et al (2001). Minimum Bandwidth Requirements for Recording of Pediatric Electrocardiograms, Circulation, Journal of the American Heart Association, Vol. 104, No 25, pp 3087-3090--ISSN 0009-7322