Sensor networks have been gaining more and more importance in scientific research and development in the past years. Along with this the extent of installed networks has also increased. In this diploma work wireless sensor networks are analysed. The most critical factor in such networks is the power consumption. Usually the sensor network is deployed for a longer time, sometimes it can thus be necessary to update the software on the participating sensor nodes, which is another critical operation. The method that we use to solve this problem has an effect on power consumption and network lifetime, but also on the used human resources.
Significant power can be saved, if the chance of resending the program due to a transmit failure is minimized, along with the number of nodes needed for resending. This can be achieved by designing a robust transfer and correcting code to increase efficiency. The other purpose of the task is to decrease the usage of human resources by implementing distributed and parallel program loading.
The structure of this dissertation is the following:
At first I give a summary about the sensor networks in general, emphasizing the problems in wireless sensor networks, and presenting solutions to some of these. This is followed by the introduction of the specific problem this work tends to solve. The second chapter details the used instruments, their build-ups and peculiarities. The next section deals with the Java user application program, which is used for controlling the process of the program update. The fourth part introduces how to configure and program the used instruments for the defined purpose. In the fifth chapter the probability of the failure is calculated by measuring three different parameters. Finally, according to the results, the designed error correcting process is presented.