Armando Bellini, - phdindustrialengineering
Transcript of Armando Bellini, - phdindustrialengineering
Real time processing of biomedical and biomechanical signals
• Armando Bellini, Full Professor
• Vincenzo Bonaiuto, Associate Professor
• Fausto Sargeni, Associate Professor
• Stefano Bifaretti, Researcher
• Giuseppe Annino, Researcher
• Nicola De Simone, PostDoc Grant
• Luca Federici, PhD Student
• Luca Tarisciotti, PhD Student
People
Real time processing of biomedical and biomechanical signals
The NA62 Experiment – High Energy Phisycs - CERN • Design of the Veto Trigger System for LKr Calorimeter
Artificial Neural Networks – Analogue VLSI Circuit implementation • Cellular Neural Networks • I&F Neuromorphic Neural Networks
Research Activities
Sport Engineering Technologies • Extraction of functional parameters for performance evaluation of
high-level athletes
Real time processing of biomedical and biomechanical signals
Main Academic and Industry Partners
C.H.O.S.E
TELETECNICA
Apparecchiature elettriche,
elettroniche e per
teletrasmissioni
Real time processing of biomedical and biomechanical signals
Data Acquisition
Data Transmission
Data Processing
ASIC (Analogue or Digital)
FPGA
Software on PC
Hardware Design for Signal Processing
EMG Sensor
IMU Sensor
Stereo Vision Camera
Time Constrain Problem
Real time processing of biomedical and biomechanical signals
Hardware vs Software Solution
Real time?
Wired or Wireless?
Battery powered? Low Power
System Dimensions?
Algorithm Complexity
Costs
Engineering problem to solve
Time Constrain
Real time processing of biomedical and biomechanical signals
Computational Power
Pro
cess
ing
Tim
e R
equ
irem
en
ts DSP Processors
ASIC (Analogue or Digital)
FPGA
ARM Microcontroller
16bit Microcontroller
8bit Microcontroller
Hardware vs Software Solution
Software on PC
Real time processing of biomedical and biomechanical signals
Hardware design
Design of mixed-signal electronic boards based on microcontroller (C2000, MSP430, TMS320, STM, Cortex-M3, Arduino, PIC, ecc.)
Design of mixed-signal electronic boards based on programmable electronic devices (FPGA) and/or Digital Signal Processor (DSP)
Design of analogue or mixed-signal ASIC
Software design
Software application design using object-oriented languages on Android,
Linux as well as Windows platforms (Delphi, Visual C++, Visual C#)
Main skills
Real time processing of biomedical and biomechanical signals
Cellular Neural Networks Applications: Stereovision System for Autonomous Robotics
Disparity Map Extraction
16x64 Stereo
Vision CNN Chip
Dual ADSP-TS201S
TigerSHARC®
Processors 6x24 Stereo Vision DPCNN Board
DSP Processors ASIC (Analogue)
Real time Image Processing Problem
Real time processing of biomedical and biomechanical signals
Multi-chip I&F Neuromorphic Architecture
Specific Algorithm
ASIC (Analogue)
Real time processing of biomedical and biomechanical signals
Cellular Neural Networks Applications: Support Systems to the Clinical Diagnosis of the Eye Diseases
Morphometry Analysis of the Corneal Endothelium
Salerno, M. ; Sargeni, F. ; Bonaiuto, V. ; Amerini, P. ; Cerulli, L. ; Ricci, F.: A new CNN based tool for an automated morphometry analysis of the
corneal endothelium - Proceedings of Fifth IEEE International Workshop on Cellular Neural Networks and Their Applications, 1998
Image Processing Problem
Software on PC
•Density of cells
•Morphometry analysis
NO Real Time requirement
Real time processing of biomedical and biomechanical signals
Retinal Images Registration and Clinical Parameter Extraction (PRIN 2004)
Cellular Neural Networks Applications: Support Systems to the Clinical Diagnosis of the Eye Diseases
Image Processing Problem
FAST Processing Time requirement
Software on PC
Real time processing of biomedical and biomechanical signals
Design of electronic systems by using new materials and new technologies to acquire and study the physiological and biomechanical parameters of human movement
Sport Engineering
Easy to use in Athletic as well as soccer field (outdoor)
or basketball and volleyball court (indoor)
Smallest system, Minimally invasive, Wireless (short or long distance),
MUST NOT MODIFY THE ATHLETE’S PERFORMANCE
Analysis of Athlete’s Performances
The data of the performance have to be easy to
understand
Support to the trainer in fully recovery of the athlete in
the post rehabilitation phase
Real time processing of biomedical and biomechanical signals
Training and performance evaluation of swimmers
Ben Hur: Electrical machine for training swimmers and measuring their force and speed
Goldeneye: Underwater motion capture and evaluation
Kz: A system to measure the thrust of a swimmer Speed RT: a system to measure the
displacement and speed of a swimmer
Real time processing of biomedical and biomechanical signals
Bottoni, A., Lanotte, N., Bifaretti, S., Gatta, G., Bonifazi, M., Boatto, P. : “Direct measurement of stroke propulsion in real swimming by means of a non
invasive gauge” - Biomechanics in Swimming – Oslo June 16th -19th 2010
KZ: a system for measure the thrust of a swimmer
The athlete can swim freely and
the data are stored in the system.
They can be acquired by a PC via
a Bluetooth radio link
It is based on two differential pressure sensors
receiving the input from two special mini paddles on
the hands of the swimmer.
These sensors measure the pressure difference
between palm and back of the hand, which determines
the thrust.
16bit Microcontroller
Real time processing of biomedical and biomechanical signals
• Quantitative analysis (maximum thrust, avg. thrust, frequency,…)
• Qualitative analysis (shape of the curve, faults, symmetry…)
• Study of efficiency
• Effects of training
• Effects of fatigue
• Integration with other sensors (accelerometers, gyroscopes, GPS,..)
• Real time data transmission
“Technical skill differences in stroke propulsion between high level athletes in triathlon and top level swimmers“: A. Bottoni, N. Lanotte, P.
Boatto, S. Bifaretti, M. Bonifazi - JOURNAL OF HUMAN SPORT & EXERCISE - VOLUME 6 | ISSUE 2 - 2011
KZ: a system for measure the thrust of a swimmer
Software on PC
Real time processing of biomedical and biomechanical signals
TriodeTM
Electrodes sEMG
MCU and Wireless TX Unit
MP430 Texas Instr
IMU MPU6050
Wireless IMU-sEMG System
EMG Accelerometer
16bit Microcontroller
Real time processing of biomedical and biomechanical signals
Wireless IMU-sEMG System
Software on PC
Real time processing of biomedical and biomechanical signals
Counter Movement Jump (CMJ)
countermovement ( squat) 390 ms
Rise Time 350ms
FLy Time 490 ms Contact
Time
Acceleration VT
Peak Concentric Phase
Peak Eccentric Phase
Flight
Contact
EMG
Accelerometer
Real time processing of biomedical and biomechanical signals
Corsa
Run Parameters evaluated
from acquired data
•4
Steps
•19,54
Rhythm (steps/min)
•2,5
Average Step Lenght (m)
•2,71
Speed (m/s)
•0,34
Contact Time (s)
•0,66
Fly Time (s)
Walking 1st Step
EMG
Wireless IMU-sEMG System