Portable Heart Attack Detector (PHAD) Final Presentation

1
Portable HeartAttack Detector(PHAD)Final
Presentation
Technion - Israel Institute of Technology Departme
nt of Electrical Engineering High-Speed Digital
Systems Lab

• Nov 27, 2005

2
AGENDA

• Project objective
• System block diagram
• ECG signal analysis algorithm
• Firmware block diagram
• LabVIEW development platform
• VI blocks and GUI
• Summary and conclusions
• System demonstration

3
Project objective

• Develop a system that detects R characteristic
  point and measures ST-elevation and QRS duration.
• Implement the system on National-Instrument()
  Real-Time FPGA development environment using
  LabVIEW graphical programming language.

() website http//ni.com
4

• Basic ECG complex

5

• ST Elevation

6
System block diagram
7
ECG signal analysis algorithm

• ECG characteristic points are detected with DWT
  (Discrete Wavelet Transform).
• DWT is implemented with Algorithme à trous
  (implementation without decimation).

Source A wavelet-based ECG delineator
evaluation on standard databases, IEEE
Transaction on biomedical engineering, April
2004.
8
Algorithm (cont.)

• Why can ECG characteristic points be detected
  with Wavelet transform?

• The Wavelet transform (WT) is proportional to the
  derivative of the filtered version of the signal.
• Zero-crossing of the WT corresponds to the local
  maxima or minima of the filtered signal.
• Maximum absolute values of the WT are associated
  with the maximum slopes in the filtered signal.

9
Algorithm (cont.)

• ECG waves are composed of slopes and local maxima
  or minima.
• Therefore, QRS complex produces an unique pattern
  (max-min-max).

10

• ECG characteristic pointsdetection flow chart

11

• Q, S detection

12

• R detection

13
Firmware block diagram

• Algorithm has been implemented in VHDL.
• Firmware includes 3 main blocks
• Wavelet decomposition.
• d4 signal processing for QRS complex detection
  and calculation of QRS duration.
• ST elevation calculation.

14
Firmware (cont.)
Top level
15
LabVIEW developmentplatform
What is LabVIEW?

• Graphical programming language with built-in
  functions for I/O, control, analysis and data
  presentation.
• LabVIEW advantages
• Intuitive graphical development similar to
  flowcharting.
• Bulit-in tools for design, control, data
  acquisition and data presentation.

16
LabVIEW developmentplatform (cont.)

• Platform includes two independent modules
• LabVIEW for Windows (Host)
• ? Floating-point calculations.
• ? Data presentation.
• ? Off-line data acquisition.
• LabVIEW for FPGA
• ? Fix-point signal processing.
• ? Real-time data acquisition.
• ? VHDL integration.

17
VI blocks and GUI
FPGA

• Data trans-ceiving between FPGA (signal
  processing) and host (data presentation) is based
  on synchronization interrupts.
• FPGA synchronization interrupts demands
  sequential framing operations
• FPGA VI includes three frames
• I/O and signal processing modules (VHDL core).
• Sampling time delay.
• IRQ to host.
• HDL clock is synthesized from while loop
  indexs LSB.

18
VI blocks and GUI (cont.)
FPGA VI
19
VI blocks and GUI (cont.)
Host

• Host includes two independent sub VIs
• Test mode for system verification (off-line ECG
  analyzing).
• Real-time controlling mode for analyzing on-line
  ECG signals.
• Hosts GUI graphically presents both sub VIs
  outputs and controls FPGA module.

20
VI blocks and GUI (cont.)
Test mode VI
21
VI blocks and GUI (cont.)
Real-time controlling mode VI
22
VI blocks and GUI (cont.)
Systems GUI
23
Summary and conclusions

• Project involves a system development for a
  medical application.
• The system is based on a firmware implementation
  for a sophisticate signal processing algorithm
  (DWT).
• ECG real-time DWT analysis is feasible for HW
  implementation.
• This project has familiarized us with new
  development tools and techniques, such as
• LabVIEW, HDL designer, ModelSim, Matlab/Simulink.
• Real-time system development.
• HW-SW integration.

24
Summary and conclusions

• System performance
• Both QRS complex and R characteristic point FP
  (False Positive false alarm) rates are very
  low.
• QRS complex TP (True Positive) rate is very high
  (gt95).
• R characteristic point TP rate is lower than in
  Matlab/Simulink model, because implemented
  algorithm doesnt use d2 and d3 (in addition to
  d4).
• System can be used as STEMI detector, because
  QRS complex TP rate is high enough to detect
  irregular ST level variations on time.

25
Summary and conclusions

• LabVIEW platform advantages
• Rapid prototype system.
• Dedicated hardware and software.
• I/O easy access.

26
Summary and conclusions

• LabVIEW platform disadvantages
• Development environment is non-conventional -
  design extraction to other non NI environments is
  NOT possible.
• VHDL code is hidden from user.
• Lack of debugging tools.
• Unfriendly VHDL interface.
• Emulator supports FPGA simulations but doesnt
  support IRQ simulation (FPGA-Host data
  trans-ceiving).

27
System demonstration
28
Questions?