Title page for etd-0827113-191922


[Back to Results | New Search]

URN etd-0827113-191922
Author Chen-ming Nien
Author's Email Address No Public.
Statistics This thesis had been viewed 5351 times. Download 534 times.
Department Electrical Engineering
Year 2013
Semester 1
Degree Master
Type of Document
Language English
Title Implementation of Front-end Compression for a Wireless BAN Sensor Node
Date of Defense 2013-08-26
Page Count 76
Keyword
  • Body area network (BAN)
  • Curvature compression
  • Huffman coding
  • Microcontroller coding
  • FPGA digital circuit design
  • Abstract With the aging of the world population, the importance of long-term medical monitoring has increased along with the need for home care services for the elderly. To serve these needs, a body area network (BAN) is proposed as a possible solution. A system for long-term monitoring with wireless data transmission consumes a considerable amount of power, a significant portion of which is dissipated by the wireless transmitter. Therefore, front-end compression circuits are designed and evaluated in this study to reduce the required data rate of wireless transmission. Compression circuits based on curvature detection (providing lossy compression) and Huffman coding (lossless compression) are selected due to their low complexity and resulting low-power implementation. Huffman coding for ECG compression is designed using a pre-defined coding table. It is shown that truncating the table at a very shallow depth yields a compression factor (CF) of 2.16, very close to the optimum result of 2.53 using the complete table. This truncated table yields a compact hardware, implemented and tested in this study using a FPGA. Furthermore, the method of curvature detection for ECG, EMG and gait pattern compression is examined and an algorithm including target CF tracking is implemented in software on a microcontroller. The target CF is user programmable. Measured results demonstrate a CF of 2.23 and PRD error of 1.7%, and CF of 10 with PRD of 5.5% for ECG compression. Finally, the curvature algorithm is implemented on a wireless transceiver to demonstrate its suitability for BAN application. It is concluded that a CF of 10 yields system power reduction between 3% and 75% depending on the relative power consumption of the data processor and the transmitter. This power advantage is achieved by the low complexity algorithms described in this thesis, yielding low computation energy per compressed sample as compared with many other algorithms for ECG/EMG compression.
    Advisory Committee
  • Jia-Jin Chen - chair
  • Ching-Hsing Luo - co-chair
  • Tsang-Ling Sheu - co-chair
  • Chung-Yao Kao - co-chair
  • Robert Rieger - advisor
  • Files
  • etd-0827113-191922.pdf
  • indicate access worldwide
    Date of Submission 2013-09-28

    [Back to Results | New Search]


    Browse | Search All Available ETDs

    If you have more questions or technical problems, please contact eThesys