FPGA based reconfigurable body area network using Nios II and uClinux
This research is focused on identifying an appropriate design for a reconfigurable Body Area Network (BAN). In order to investigate the benefits and drawbacks of the proposed design, a BAN system prototype was built. This system consists of two distinct node types: a slave node and a master node. These nodes communicate using ZigBee radio transceivers. The microcontroller-based slave node acquires sensor data and transmits digitized samples to the master node. The master node is FPGA-based and runs uClinux on a soft-core microcontroller. The purpose of the master node is to receive, process and store digitized sensor data. In order to verify the operation of the BAN system prototype and demonstrate reconfigurability, a specific application was required. Pattern recognition in electrocardiogram (ECG) data was the application used in this work and the MIT-BIH Arrhythmia Database was used as the known data source for verification. A custom test platform was designed and built for the purpose of injecting data from the MIT-BIH Arrhythmia Database into the BAN system. The BAN system designed and built in this work demonstrates the ability to record raw ECG data, detect R-peaks, calculate and record R-R intervals, detect premature ventricular and atrial contractions. As this thesis will identify, many aspects of this BAN system were designed to be highly reconfigurable allowing it to be used for a wide range of BAN applications, in addition to pattern recognition of ECG data.
Body Area Network, Reconfigurable, FPGA, Nios, uClinux
Master of Science (M.Sc.)
Electrical and Computer Engineering