Use wireless sensor networks for patients in home monitoring--data flow.

Bujdei, Catalin ; Moraru, Sorin Aurel ; Dan, Stefan 等

1. INTRODUCTION

The wireless sensors network (WSN) is a new tehnology developed since few years ago which have impact over different areas of interest. Being a new technology different functionalities and characteristics need to be developed or improved. One of the major interests of using the WSN is into the medical domain, and especially for monitoring the patients which suffer of different ills. The patient could stay at home without being disturbed by the aspect of a hospital. He could walk from one room to another, to develop different activities, during the monitoring process. The only requirement is to keep the monitoring equipments attached to them. Nowadays, the monitoring equipments become smaller and smaller in size, but this also depends on their functionalities.

We have under developement a system which permits the monitoring of the level of oxygen from blood and the pulse for a patient using a small medical equipment, capable to monitor this parameters. The system records this data and stores all the data into a database. The data from the database could be accesed real-time for being vizualized and analyzed. Different supplementary functional modules could be attached to the system for let generating different alarms depending on the measured data variation, but this will be for future implementing. Into the system under development another requirement for the patient is to not exit the area of wireless communications, and which, in our system case, covers the entire living space.

Since the system record a large amout of data, which need to be stored into the database, it must to process and transmit the data as quikly as possible. It is now allowed to exist locks of the system or lose of the data. Also, the energy consumed by the wsn must be minimum possible. Since the monitoring have to run continuously the optimization methods of the system are different than the system which measure values from time to time. The most systems developed for monitoring the patients consider more the mobility of the patient or the analyze of the recorded data (Proulx et al., 2006; Chen et al., 2004) and less the possibilities of minimize the quantity the data transmited and in this way, the consumed energy. Since it is an important system for the medical domain we want get the best solution possible for optimization. In this way we want to ensure an stable and long life monitoring system.

2. SYSTEM ARCHITECTURE

The main functionalities of the system are: to record the necessary information (parameters), to transport the information securely and with minimum delay to a central database and to permit the visualization of the data, in different formats on different users computers. The system architecture is presented into the figure 1.

[FIGURE 1 OMITTED]

The first part of the system is represented by the monitoring network, represented by the top part of the figure 1. The monitoring network is formed by the end node, attached to the patient and which is connected to the medical equipment, router node, necessary for increase the area of wireless signal and the coordinator node which transfer the data from the wireless network to a Driver computer.

The Driver is the second important part of the system. Its role is to get the recorded information by the network and transfer it to the Database Server computer. The Server computer has the role of processing the data of the database and from the system. Different users could connect to the system, using a personal computer or a mobile phone, and vizualize the recorded data. They could see the instant values or history graphs with data.

Similar systems, for monitoring of the patients, have been developed by different researchers or centers of research. Some of them preferred to use instead the WSN, a PDA equipment which to transmit the data using Bluetooth or mobile phones which to transmit the data using the existing GPRS service (Proulx et al., 2006; Chen et al., 2004). The Bluetooth technology consumes more power energy than the ZigBee, which is used by the most wireless sensor networks. Also, the Bluetooth range is smaller than the ZigBee range. Using the GPRS service the mobile consumes its energy power very quikly and the system could not run for a long term. The system must run continuously without errors or interruptions.

3. THE FLOW OF DATA IN WSN

In our system we want to record only the heart-rate and the blood oxygen level. The analogic signals are transferred to the network end node for being proceesed. First the analog signal is transformed in digital and when this signal is transmited through the network to the coordinator node. This requires lot energy to be consumed for processing and transmiting of the data. The most part of energy is consumed at data transmission.

For this reason the rate of transmission and the quantity of data transmited should be minimized.

For sending the data into the wireless network the data must be packed in telegrams which to respect a establish format.

The format of the telegram is presented into the table 1.

For minimize the quantity of data transmited, into the telegram are not inserted all real measured values but only the first value measured into a period of time, and then the differences of the measured values and the first value for that period of time (these represent the offset values). A measured value is a real value represented in memory by a number of octets. This number is determined by the precision which we want for storing data and the range of values for the monitored parameters. The offset values are smaller than the real values recorded for that moment of time, and this means that they can be represented on a smaller number of octets. In this way we can reduce the number of total octets for a telegram; the telegram become shorter and the quantity of data transmited into the network become smaller, but without losing precious information. For a further development we want to pack the information from the telegram using different archiving alghoritms (Fullford-Jones et al., 2004).

4. THE FLOW OF DATA IN INTERNET

The Driver computer sent the data, using the Internet network, to the Server computer. The data must be sent without delay and safety throught the Internet network. If the Internet network is not accessible for a period of time the data will be buffered on the Driver computer and send to the Server computer when the Internet network becomes available. For the transmission of data in this direction. Net software application have been implemented. The data is packed using a zip tool existent into the. Net platform. For security reason 2 types of encrypting the data could be used: using symmetric keys or using asymmetric keys. Both, implies a collaboration between the users of the application for exchange the public keys (Northrup et al., 2006).

5. THE DATABASE

The database is an important part of the system which could affect the performances of the entire system. It is necessary to have an optimization of the database for increase the performances of the system. No paper was found till yet which to discuss the optimization of a database inside such a system.

The system will have to be functional not for a year or less but for many years. The database has the role to store a lot of information (millions of record) and the processing of the data from the database must be done in the shortest time possible. If an alarm signal must be generated it should not be delayed.

We had to consider the optimization of the database especially for the insert and select operations. The optimization of the database is essential for such big number of records for allow the system to function in normal conditions. It is not desired that when a user want to view a graphic with the recorded information for a specific period of time he have to wait a while, about minutes, since the information is retrieved from the database, processed and then sent to the user's application interface for view. Also, since the record of the information is done continuously is necessary that the insert operation of the information into the database to be done in the shortest time possible. In this way there will not be system locks, by creating unlimited waiting queue of information which to be inserted.

The most important methods, which have been tested and which could lead at a database's performances improvement:

* Normalization;

* Using indexes; recreate indexes; indexes defragmentation;

* Statistics update;

* Manage concurrency;

* Table and index partitioning (Kimberly, 2005);

* SQL tuning;

* Disk defragmentation (Henderson, 2003).

6. CONCLUSION

The optimization of the data flow has a major impact over the system. In this way the system could execute rapidly different function without having disturbing delays. The success of the system depends on this optimization. The further tests will permit us to discover and other optimization solutions.

7. REFERENCES

Chen, S.; Gaur, A.; Mutthukrishnan, S. & Rosenbluth, D. (2004), Wireless in loco Sensor Data Collection and Applications, Available from: http://www.cs.rutgers.edu/~muthu/inloco.pdf, Accesed: 2008-07-02

Fullford-Jones, T.R.F.; Wei, G.Y. & Welsh, M. (2004), A Portable, Low-Poer, Wireless Two-Lead EKG System, Proceedings of the 26th Annual International Conference of the IEEE EMBS, pp 2141-2144, 0-7803-8439-3, San Francisco, CA, USA, 1-5 september 2004

Henderson, K. (2003), Proceduri stocate in SQL Server. XML. HTML.(Stored Procedures in SQL Server. XML. HTML.), Teora, 973-20-0613-7, Bucharest

Kimberly, L. (2005), Partitioned tables and indexes in SQL Server 2005, Available from: http://msdn2.microsoft.com/ en-us/library/ms345146.aspx, Accessed 2008-04-11 Northrup, T.; Wildermuth, S. & Ryan, B. (2006), Microsoft .Net Framework 2.0 Application Development Foundation, Microsoft Press, 0-7356-2277-9, Redmond, Washington

Proulx, J.; Clifford, R.; Sorensen, S.; Lee, D.J. & Archibald, J. (2006), Development and Evaluation of a Bluetooth EKG Monitoring Sensor, Proceeding of the 19th IEEE Symposium on Computer-Based Medical Systems, pp 507-511, Salt Lake City, Utah, 22-23 june 2006

Tab. 1. Telegram format

Begin ID Date, First Offset -- Offset End
mark Patient hour data value 1 value n mark
 value