Distributed Communication System
in Biomedical Applications

Péter Várady

Dept. of Control Engineering and Information Technology, Technical University of Budapest, Hungary
1111 Budapest, Mûegyetem rkp. 9, Hungary E-mail: varady@iit.bme.hu

Abstract

Our project is a distributed demonstrative intensive patient monitoring system for educational purposes. Today medical communication standards exist only for the higher level of medical care, like various databases and hospital information systems, and not for low level communication between various diagnostic devices. Our system brings a new approach in the standardless and vendor-specific world of the bed side medical equipments. With the usage of a vendor-independent, open, broadly supported industrial fieldbus it will be possible to interconnect bed side medical devices of various vendors at an economical cost. The system is so planned that also the already existing devices are connectable with external link modules. To reduce the amount of data to be transmitted and to get a reasonable transmission speed, various signal preprocessing algorithms and prediagnostic will be used in the fieldbus link modules. The distributed demonstrator system consists of the medical equipments linked to the fieldbus with such modules, two simulated actuators and a central monitoring station. The patient is realized artificially with a PC. The main goals are to ensure a real-time, deterministic and secure data exchange between the linked devices, and to present a user-friendly visualization of the patient's state.

Keywords: medical monitoring, medical instruments, fieldbus, demonstrator, Profibus

Preface

Today one of the most evolving areas of information technology are the various biomedical applications like patient and diagnostic databases, hospital information systems, clinical monitoring and the use of computers in intensive acute patient care. In these tasks we can distinguish between two levels of hierarchies.

For the purposes of various databases and information systems are off-the-shelf solutions available. There are already many standards for platform independent representation and interchange of static or quasi-static medical records (patient data, insurance data, medical images etc.) [1]. These communication standards are based on the upper layers of the OSI model, and the lower layers, like the physical transmission, the media access and the routing methods are not defined. This approach allows the adaptation of such high level protocols on many existing commercial or semi-professional transmission technologies, like 802.x LANs. According to its economical price lots of hospitals and medical centers used and use the commercial 802.3 Ethernet to interconnect their departments.

Clinical monitoring and intensive patient care mean a much different problem. In these applications the various bed side medical equipments are to be networked at a lower level of the communication hierarchy. This requires a dynamic, real-time, deterministic, fault-tolerant and secure data exchange. The 802.x standards are lack of some or more of these features [2]. According to a survey in 1989 [3] the main requirements of bed side medical monitoring systems are:

As it can be established, to fulfill all of these requirements we must define the lower layers of the OSI model. Although many vendors are offering their monitoring systems, they are in lack of an open communication technology so in lack of the interchangeability of devices various vendors. Since 1992 IEEE has a draft proposal of the Medical Instrumentation Bus (MIB) [4]. The final standard is not yet approved. Another problem is that MIB is very poor hardware supported. Although a standard MIB device interface is in the range of $100, only a couple of vendors offer a MIB connectable equipment. Until MIB is not an approved standard vendors are not interested in the producing of MIB devices. On the customers' side it can be observed that no one like to invest money in devices described in an evolving, subject to change draft standard. So the old and unique communication solutions stay and mean a stable and more economical solution for both manufacturer and customer.

Concept

It can be noticed, that the hierarchy of the industrial communication systems are very similar to the medical ones. The level of the hospital information systems and databases is analogue to the process or manufacturing control level, and the task of the clinical monitoring and patient care can be identified with the field level of the industry standard communication. Our basic idea was to use a standardized industrial fieldbus transmission technology which satisfies all of the expected requirements mentioned in the preface. These standards are supported since many years with the appropriate hardware and software components. So various medical devices can be interfaced to a fieldbus using an existing know-how.
Interfacing can be done in two different ways: a device itself can contain the interface or with help of an external interface module. The last approach is more economical, since already existing equipments can be easily interfaced. But how? Well, all of the bed side devices have a quasi-standard analogue signal output. These analogue outputs can be measured and aquisted by the external link modules, then the corresponding information can be sent through the fieldbus data link (FDL) to the central monitoring station.

To introduce the ideas described above, we decided to realize a distributed demonstrator system based on an open fieldbus communication technology. The demonstrator system will be installed in the Biomedical Engineering Laboratory at the TU Budapest, Department of Control Engineering and Information Technology. Due to its open architecture the system will serve optimally our educational purposes. This work runs under the INCO Copernicus #960161 project of the European Union.

Overview of the demonstrator

We decided to integrate the components as it follows:

As central monitoring station:

As sensors: As actuators: As patient:

The sensory and actuary components representate the bed side euipments of a single patient. Beside them the system can be expanded with additional elements due to future enhancement.

The main goal of the demonstrator is the ensure the real-time, deterministic and secure data exchange demand of the monitoring purpose. To achieve a "real-time" monitoring and to keep the required data transmission rate under a reasonable value we agreed to reach 10 polling cycle per second pro device. The transmission rate can be reduced using various preprocessing and signal-compressing algorithms in the external interface modules.

Vital signals

To allow a modular and economical concept, two bigger category of various vital signals can be distinguished:

"Slow" vital signals
(temperature, pulse, CO2)
"Fast" vital signals
(EEG, ECG, blood pressure)
Bandwidth0 - 10 Hz0 - 1000 Hz
Required resolution12 bit12 bit
Data / polling cycleapprox. 10 bytesapprox. 100 bytes
Signal preprocessingnoneintensive

Components

The "slow" signals require no signal preprocessing so they can be acquisted with simple external DAQ units (marked as Type I slaves).

"Fast" signals have a spectra of up to 1 kHz, they supply a much larger amount of data. To get a reasonable amount to be transmitted, intensive signal preprocessing is required. This task can be fulfilled only with the use of more powerful hardware, like industrial PCs (marked as Type II slaves).
To be economical signals of more than one bed side device can be measured with both a Type I and a Type II slave. In addition the slaves should have an appropriate fieldbus link.

Due to the educational purposes of the demonstrator an artificial patient will be used. The artificial patient is a desktop PC. This PC contains D/A cards to produce the required vital signals, and A/D cards to close the control loop from the side of the actuators. The various physiological or pathological vital signals to be generated can be selected on a user-friendly graphical interface.

To easily measure their effect, the actuators are motorized potentiometers with appropriate FDL.

A powerful standard desktop PC plays the role of the central master monitoring station. Its task is to manage the communication, to gather the information form the Type I and Type II devices, to make decisions, to control the actuators, to visualize the status and vital signals of the cared patient on a user-friendly graphical interface and to store the data in a database. These tasks require a group of parallel executed real time activities. So the monitoring station will run a graphical user interface based operating system capable of multitask process handling with preemptive scheduling and a secure file system.

The fieldbus in medical applications

As already described, an open communication protocol with a deterministic time-characteristic is required for our purposes. The main parameters are as follows:

Number of devices in the system
Due to the specification of the demonstrator the system consists of a central monitoring station (a master unit) and 6 distributed devices. These are the external modules, that interfacing the vital signals onto the fieldbus, and the actuators (infusion pump and gas inhalator represented by the motor-stepped potentiometers). For the proper capability to expand the system in the future, but keep the required transmission speed under an affordable level the maximum number of the slaves is fixed to a reasonable number of 10.

Data transfer rate
For a satisfactory reaction time at least 10 telegrams per second from each slave should be transmitted. Calculations have shown, that due to the intensive signal preprocessing a serial data transfer rate of 100 kbit/s is enough for the full system.

Data security
The transferred data should be protected at least with a CRC code. The main controller should realize all kind of eventual defects in the communication media or in the slaves and manage the defects in a plug & play manner.

Distance and media
The distributed system should support distances as high as maximal 100 meters and the transmission media should be due to the ease of installation a single twisted pair.

One of the most powerful open industrial communication technology is the family of PROFIBUS standards. Since this is the market-leader European fieldbus and supported by hundreds of vendors, we decided to use it in our demonstrator.

The family of PROFIBUS standards

PROFIBUS is a vendor independent, open field bus standard for various applications like industrial automation, production control, building automation and in our case, as a new area: biomedical applications. The original PROFIBUS standard was the German standard DIN 19245 Part I and Part II [5]. The bus is specified based upon the OSI layer model. The Part I describes the functionality of the lower two layers, and Part II deals with the complex functionality of the layer 7 (Fieldbus Message Specification, FMS).
By the year 1990 Part III of DIN 19245 published [6], and a simple, but yet powerful application extension over the layer 2 - without the FMS layer - and a standardized user interface was described. This new PROFIBUS standrard is the so called PROFIBUS DP (Decentralized Periphery). In the middle of the 90's a new Part IV of the DIN 19245 was released. The branch of PA (Process Automation) devices based upon the DP, and this standard offers a transmission media for the explosion secured (Ex) areas, but at a lower transmission speed. Since 1996 PROFIBUS is a European standard EN 50170.
Since PROFIBUS hasn't got the OSI layers 2-6, it is impossible to realize a complex system with routing and multiple sessions, but it is not required at all, since medical monitoring as other "on-the-field" tasks, interconnect a limited number of devices with the main aspects of deterministic and reliable data transmission.
As our main aim is to use a vendor-independent, fully standardized, and easy-to-use communication protocol, we decided to use the PROFIBUS DP.

PROFIBUS DP cares a patient

The PROFIBUS DP was developed for its simple standardized user interface and its main goal is to realize a fast cyclical data transfer, upto 246 bytes/data per message and upto 12 Mbit/s between the various devices connected to the bus. The transmission media can be either a single twisted pair or an optical fiber at higher speeds. A mono-master DP structure will be used. All system components except the artificial patient are linked to the fieldbus. The central monitoring station is linked as the DP master, the other devices are the slave ones.

Summary

Nowadays low level communication between bed side medical equipment is based only on vendor-specific and closed standards. The efforts of IEEE since 1992 to introduce the MIB is not yet succeeded. Even if MIB would be end-released the wide support of MIB by various vendors would take many years of time. Our aim was to introduce the existing, broadly supported know-how of industrial fieldbuses into the world of medical communication. We described an educational demonstrator system that interconnects various bed side medical devices with the support of existing ones. The link components and technology are based on off-the-shelf products.

References

  1. S. Sengupta: Computer Network in Health Care, The Biomedical Enginnering Handbook, IEEE Press, 1995, ISBN 0-8493-8346-3, pp. 2642-2649.
  2. http://stdsbbs.ieee.org/groups/mib/tech/1073vsEthernet.html
  3. http://stdsbbs.ieee.org/groups/mib/overviews/SCAMC94.html
  4. http://stdsbbs.ieee.org/groups/mib/main.html
  5. PROFIBUS Standard Process Field Bus DIN 19245-1 Translation of the German Standards, Profibus Nutzerorganisation e.V., 1991
  6. PROFIBUS Dezentrale Peripherie (DP) DIN 19245-3 Standard Deutsche Elektrotechnische Komission im DIN und VDE, Ausgabe 1993-10
  7. E. A. Woodruff: Clinical Care of Patients with Closed-Loop Drug Delivery Systems, The Biomedical Enginnering Handbook, CRC Press, IEEE Press, 1995, ISBN 0-8493-8346-3, pp. 2447-2458.