A Time Synchronization Protocol for Wireless Body Sensor Networks

Overview: Patients suffering from chronic lung disease can follow a prescribed physical exercise at home to monitor their lungs during physical activity or ambulant rehabilitation. Equipped with a Wireless Body Area Sensor Network (WBASN), sensor nodes attached to the patient’s body sense, process and communicate vital signs to a base station. These sensors consist of an ECG sensor, activity sensors (accelerometers), a pulse oximeter, a blood pressure and a breathing rate sensor. The individual vital signs can be collected for a long period. Through the internet, a physician from his office can review the collected data and check the adherence of the patient to the program and his physiological response. He can also provide a feedback indicating whether actual physical strain is too low, right or overextending and suggest new prescribed physical exercises.

Health monitoring system architecture

Time Synchronization is a crucial issue for such a WBASN. The detected events between sensors during a sampling interval need to be correlated. Applying a higher level multi-sensor data fusion provides a means to the physician to analyze the collected data. Hence, it is often necessary to fuse sensor values that occur at the same time. Time synchronization is also important to ensure accurate coordination of duty cycles and minimize energy consumption. The heterogeneous wireless sensors distributed on the body measure time using local clocks and have different sample frequencies. Local time readings of sensors can easily mismatch because of random phase shifts and drift rates of clock oscillators. The error is cumulative as the sensors operate for long periods and they can loose synchronization if we do not address this issue by a time synchronization protocol.

Project: This project is in collaboration with Peter L. Reichertz Institute for Medical Informatics. The goal of this thesis is to implement a time synchronization protocol for this WBASN scenario. The basic task of the student is to investigate current synchronization algorithms used in common Wireless Sensor Networks and to determine a suitable algorithm that can be used in the context of this medical application. Afterwards this algorithm has to be implemented and evaluated. The performance of the algorithm has to be analyzed in terms of the energy consumption of the sensors and the synchronization delay error.

Prerequisites: To accomplish this task, the student should have:

• Basic Knowledge of Wireless Networks
• Good C/C++ programming skills

Supervisors: Sana Saadaoui(IBR) , Klaus-Hendrik Wolf (MI)