Download Biomedical Measurements in Embedded Applications

Electronic Measurements in Biomedical Engineering
Biomedical Measurements in
Embedded Applications
Enderle, Blanchard, Bronzino, Introduction to Biomedical Engineering, Elsevier (2005).
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
1
Electronic Measurement Systems
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
2
Dr. Martin Land
4
Physiological Measurements Analog domain systems
Sensor (transducer)
Mechanical
General physical measurement
Examples
Physiological property of tissue → electrical signals
Amplifier
Length, weight, pressure, temperature, motion
Sensor output → input requirements of measuring system
Analog filter
Electrical / chemical
Contact with tissue → electrical signal
Examples
Amplifier output → information of interest
Analog to digital converter (A/D)
Signal → binary coded sample sequence
EKG, EEG, blood glucose
Digital domain systems
Imaging
Stimulate tissue — pressure, sound, electromagnetic waves
Detect induced emissions
Examples
Microcontroller
Process automation: measurement → display process
Digital signal processing (DSP)
Digital domain filtering
Storage
Ultrasound, X-ray, infrared, MRI
Interface → information systems
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
3
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Electromagnetism
Electric Circuit Concepts
Force on charged particle
Functions of time
F = q[E + u × B ] , charge q, electric field E, magnetic field B , velocity u
Measured from ground
v1 = v1G , v2 = v2G
Voltage (potential)
Measure of work performed against electric force F
A
Pushing current → perform work → loss of voltage
B
Voltage source
Reference point for measuring voltage
Power supply = source of energy for work = – drop
vB = voltage from ground to point B
Kirchhoff laws (graph theory)
Current
Sum of currents entering any node = 0
Measure of motion of charged particles
i1 – i2 – i3 = 0
i = total charge crossing unit area per second
Sum of voltage drops around any loop = 0
Analogous to flow in a water pipe
Embedded Systems — Hadassah College — Spring 2012
– v1 + vdrop + v2 = 0
Biomedical Applications
Dr. Martin Land
5
Electrical Circuit Elements
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
6
Frequency Domain Analysis
Conductor
Carries current with voltage drop → 0
Resistor
General circuit
Ordinary differential equations with respect to time
i1
Example
Inductor LA and resistor RB
Current i(t) causes voltage drop v(t) = R × i(t)
Resistance R
i1 = i2 = i
i1
Capacitor
i2
v1 = v A + vB = L A
vA
Current i = C × dv / dt
Capacitance C
v1
v1
∞
v ( t ) = ∫ V ( ω ) e jωtdt ⇔ V ( ω ) =
Voltage drop v = L × di / dt
∞
i ( t ) = ∫ I( ω ) e jωtdt ⇔ I( ω ) =
−∞
Inductance L
v1 ( t ) = L A
Example — resistor elements RA and RB
i1 = i2 = i
v1
RB
v
RB
v1 = v A + vB = iRA + iRB ⇒ i =
⇒ vB = iRB =
v1 ⇒ B =
RA + RB
RA + RB
v1 RA + RB
Dr. Martin Land
vB
di
+ iRB
dt
−∞
Biomedical Applications
i2
vA
Fourier transform
vB
Inductor
Embedded Systems — Hadassah College — Spring 2012
i3
Voltage drop
Force F
Ground
v1
i2
v2
Current i(t)
vAB = work required to push unit charge (q = 1) from A to B
Analogous to pressure in a water pipe
vdrop
i1
Voltage v(t)
VB ( ω ) =
7
1 ∞
v ( t ) e‐jωtdω
2π ∫−∞
1 ∞
i ( t ) e‐jωtdω
∫
−∞
2π
di
+ iRB → V1 ( ω ) = jωI( ω ) L A + I( ω ) RB
dt
RB
V1 ( ω )
RB + jω
Embedded Systems — Hadassah College — Spring 2012
⇒
VB ( ω )
V1 ( ω )
=
Biomedical Applications
ω
2π
frequency f =
⇒ I( ω ) =
V1 ( ω )
RB + jωL A
RB
R + ω2L2A
2
B
Dr. Martin Land
8
Frequency and Analog Filters
Amplifiers
Low pass filter
H( ω ) =
Transfer function
VB ( ω )
V1 ( ω )
RB
=
R +ω L
2
B
2 2
A
1
=
1+
Higher frequency ω ⇒ lower transfer
R
ω0 = B
LA
,
2
ω
ω20
v out = A ( v + − v − )
iin =
Simplified model
High pass filter
A, Rin → ∞ ⇒ v + − v − =
Resistor RA and inductor LB
Transfer function
Operational amplifier (op amp)
Analog integrated circuit
Differential amplifier
H( ω ) =
VB ( ω )
V1 ( ω )
=
ωLB
R +ω L
2
A
2 2
B
=
ω
ω +ω
2
0
2
,
ω0 =
H( ω ) =
Resistor R, inductor L, and capacitor C
ω = ω0 ⇒ highest transfer
⎛
ω2 ⎞
ω + ω ⎜1 − 2 ⎟
⎝ ω0 ⎠
2
Biomedical Applications
2
2
1
R2
R1
Dr. Martin Land
9
Analog to Digital Conversion — Sampling
v-
v
R2
R
v in ⇒ out = − 2
R1
vin
R1
Embedded Systems — Hadassah College — Spring 2012
v+
i1
vin
v out = i2R2 = −i1R2 = −
vout
-
v out
v − v−
⎯⎯⎯
→ 0, iin = +
⎯⎯⎯
→0
A →∞
Rin →∞
A
Rin
v
i1 = in
R1
ω
+
v+ − v−
Rin
iin → 0 ⇒ i1 + i2 = 0 ⇒ i2 = −i1
Band pass filter
iin
v-
Feedback amplifier
v + − v − → 0 ⇒ v + = 0
RA
LB
Higher frequency ω ⇒ higher transfer
Embedded Systems — Hadassah College — Spring 2012
v+
i2
+
vout
Set R1 and R2 to any convenient values for arbitrary amplification
Biomedical Applications
Dr. Martin Land
10
Convert Samples to Digital Form
Nyquist Theorem
Rounding-off
n-bit integer codes 2n levels
Round-off samples to n-bit integer
Filter data signal to bandwidth fmax
Sample data signal at sample rate fsample ≥ 2 × fmax
Distorts data
Equivalent to added noise
Reproduce data signal from samples without distortion
161
Larger n ⇒ more levels ⇒ higher
resolution ⇒ less noise
data signal d(t)
160
159
158
157
Sequence of sample values
rounding
158
159
160 160
159 159
t
sampling signal S(t)
Example
t
sampled signal S(t)d(t)
Sampled values
158.276
158.879
159.724
159.821
159.312
158.791
Rounded values
158
159
160
159
159
159
t
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
11
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
12
Sampling for Standard Telephony
Sampling for Standard CD Audio
Telephone line
Filter audio frequencies 300 Hz to 3300 Hz
CD audio
Filter audio frequencies 20 Hz to 22,000 Hz
Sample voice channel
Sample voice channel
fsample = 8000 samples / second > 2 × 3300 Hz
fsample = 44,100 samples / second > 2 × 22,000 Hz
161
160
Round-off samples
Scale = 28 = 256 levels (0 to 255)
Each sample encoded as 8-bit byte
159
Round-off samples
Scale = 216 = 65,536 levels (0 to 65,535)
Each sample encoded as 16-bit word
158
157
158
160 160 159
159
159
161
160
159
158
157
158
DS-0 voice channel
8000 samples/second × 8 bits/sample = 64 kbps
10011110
158
10011111
159
Embedded Systems — Hadassah College — Spring 2012
10100000
160
10100000
160
10011111
159
160 160 159
159
CD audio channel
44,100 samples/second × 16 bits/sample = 705,600 bps
MP3 encoding → 5 times compression rate
705,600 bps / 5 ~ 140 kbps ~ 17,508 bytes / sec ~ 1 MB / minute
10011111
159
Biomedical Applications
Dr. Martin Land
13
Biomedical Sensors
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
14
Dr. Martin Land
16
Electrochemical concepts Classification by interaction type
Interaction → analog electrical signal (voltage / current)
Ion
Atom / molecule with electrons ≠ protons
Cation
Physical / electrical / optical / chemical
Biosensor
Electrons < protons ⇒ net + charge
Biological element — enzyme / antibody / receptor
Biochemical reaction → optical / electrical / physical signal
Anion
Electrons > protons ⇒ net – charge
Packaging
Safe — biocompatibility
Reliable — long operational lifetime
Isolate sensor from body
Metal
Material containing free electrons ⇒ electrical conductor
Electrolyte
Material containing free ions ⇒ electrical conductor
Typically ions in solution
Polymer (plastic) covering / barrier layers
Host body affects sensor function
Sensor affects implantation site
Ionization current
Environmental factors
Interactions affected by temperature / pressure / noise ...
Biomedical Applications
+
- electrons
-
Primary reference: Yitzhak Mendelson, "Biomedical Sensors," chapter 9 in Enderle, Blanchard, Bronzino, Introduction to Biomedical Engineering, Elsevier (2005).
Embedded Systems — Hadassah College — Spring 2012
159
Dr. Martin Land
15
Embedded Systems — Hadassah College — Spring 2012
-
electrolyte
Biomedical Applications
+
+
cations +
+
Half‐Cell Potential
Common Electrode Types
Metal / electrolyte interface
Electrons flow to / from metal
Charge distribution near surface
Voltage between metal and electrolyte — half-cell potential
Battery (cell)
2 half-cells with different metals
Electrons flow from metal 1 ⇒ net +
Electrons flow into metal 2 ⇒ net –
Biopotential electrodes
Electrolyte = host tissue
2 half-cells with same metal
Electrocardiogram (ECG) electrodes
Common electrode — flexible polymer + carbon / metal powder
Pre-pasted electrolyte gel for application to skin
Electromyographic (EMG) electrodes
Sense signals from muscles
Surface EMG recording
1 cm circular discs of silver / platinum
Direct recording
Percutaneous (skin puncture) needle electrodes
Electroencephalographic (EEG) electrodes
Sense signals from brain
Cup electrodes
Equal half-cell potentials (ideal model) ⇒ half-cell potentials cancel
5 – 10mm discs of platinum / tin with conducting gel attached to scalp
Example
Subdermal
Two similar electrodes taped to chest near heart
Measure electrical potentials generated by heart
Differential amplifier → electrocardiogram (ECG = EKG) signals
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Platinum / stainless-steel needle electrodes
10mm long by 0.5mm wide
Dr. Martin Land
17
Magnetism
Biomedical Applications
Dr. Martin Land
18
Electromagnetic Effects
Magnetic field
Induced by accelerating charge
Example — electron in stable orbit around nucleus
N
Electromagnet
Electric current in wire coil (winding) → magnetic field
Field configuration
+
S
Identical to permanent bar magnet
Many atoms with electrons in stable parallel orbits
Compass aligns with earth's magnetic field
Earth's North Pole is a magnetic S
Mechanical forces
Opposite poles (N ↔ S) attract
Like poles (N ↔ N / S ↔ S) repel
S
N
S
N
S
N
N
S
S
N
N
S
N
S
S
N
S
F = q (u × B )
y
z
http://hyperphysics.phy‐astr.gsu.edu/hbase/magnetic/elemag.html
Biomedical Applications
N
Current level + direction → field strength + polarity
Induction
Varying magnetic field → varying current in conductor
Mutual induction
Varying current in coil → magnetic field → current in second coil
Hall effect
Current in magnetic field
Charges spread to edges of conductor → induced voltage VHall
Permanent ferromagnet
Metal (typically iron) with permanent magnetic field
Embedded Systems — Hadassah College — Spring 2012
Embedded Systems — Hadassah College — Spring 2012
Dr. Martin Land
19
Embedded Systems — Hadassah College — Spring 2012
u along x
B along z
x F along y
+
+
+
+
charge
separation
u
-
-
-
Biomedical Applications
induced
voltage
Dr. Martin Land
20
Physical Measurements — 1
Physical Measurements — 2
Linear Variable Differential Transformer (LVDT)
Mutual induction effect
L = displacement of core from exact center
Secondary voltage VS = kgeometry × VP × L
VS
VP
Precise measure of position
Elastic resistive transducer
Thin elastic tube containing conductor
Electrical resistance depends on length
Breathing → expanding chest → higher resistance → voltage change
Strain gauge
Similar to elastic resistive transducer
Stress (force) → strain (change in size) → higher resistance
Smaller range of expansion
Electromagnetic flow transducer
Magnetic field surrounds blood vessel
Hall effect → electrolytes separate to walls
Transverse voltage measures blood flow
Bonded
Single thin wire in flexible frame
Unbonded
Multiple thin wires in two-part frame
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
21
Variable Capacitance
Biomedical Applications
Dr. Martin Land
22
Pressure Transducers
Capacitor
Parallel metal plates
A = area of plate
D = distance between plates
Apply voltage → separate charges
Capacitance = charge stored per volt
q = C×V
cations +
+
+
+
+
Piezoresistance effect
Pressure on crystal → contraction → higher resistance
Used in portable blood pressure monitors
- electrons
insulator
-
Piezoelectric effect
Voltage on crystal → contraction
Pressure on crystal → contraction → voltage
Sensitive to short mechanical pulses
Used in sensitive cardiac monitors
voltage
C = constant × A × D
Capacitor current
dq
dV
dC
i=
=C
+V
dt
dt
dt
Measure small changes in displacement
Ultra-sound
Apply high frequency (1 to 10 MHz) voltage to crystal
Crystal vibrates at voltage frequency → pressure wave → ultra-sound
Varying D with constant V
dD
i = constant × V × A ×
dt
Embedded Systems — Hadassah College — Spring 2012
Embedded Systems — Hadassah College — Spring 2012
Apply ultra-sound to crystal
Pressure wave → crystal vibrates → voltage at ultra-sound frequency
Biomedical Applications
Dr. Martin Land
23
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
24
Temperature Measurement
Blood Pressure Set
High Level Design Contact thermometer
Measure body temperature in direct contact with skin
Thermistor
cuff
Metallic mixtures change resistance with temperature T
controller
display
switches
air pump
air hose
⎡ ⎛ 1 1 ⎞⎤
R = R0exp ⎢β ⎜ − ⎟ ⎥
⎣ ⎝ T T0 ⎠ ⎦
air hose
connector
Noncontact thermometer
Measure temperature of ear canal near tympanic membrane
Infrared (IR) radiation guided to sensor
Thermopile sensor
cuff
air hose
motor control
pump
Two metal plates in contact
Heated plates → voltage depending on temperature
Used in standard thermostat
Pyroelectric sensor
pressure
sensor
Heated crystal → voltage depending on temperature
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
LCD display
LCD control
switches
digital input
amplifier
filter
A/D
microcontroller
Dr. Martin Land
25
Blood Pressure Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
26
Sphygmomanometer
Pressure
Force per unit area
Measured as mmHg
vacuum
weight of Hg
External pressure balances weight of Hg in column
air pressure
Pulse
Heart muscle contracts
Internal surface area decreases ⇒ pressure increases
Cuff
Pump air into cuff
Pressure measured in gauge
Pressure on artery → restrict blood flow
Blood flow
Unrestricted artery
Laminar (smooth) flow → silent
Restricted (occluded) artery
Systolic pressure
Turbulent flow → BP oscillations → Korotkoff sounds
Maximum BP at peak contraction
Normal range: 90 – 120 mmHg
Diastolic pressure
Minimum BP at minimum contraction
Normal range: 60 – 79 mmHg
L. A. Geddes, Handbook of Blood Pressure Measurement
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
27
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
28
BP Measurement with Sphygmomanometer
BP Measurement in Digital Test Set
Inflate cuff on upper arm
Pump cuff to 160 – 200 mmHg
Brachial artery occluded → no flow → silent
Oscillometric method
Filter + amplify oscillations
Cuff inflation
Cuff deflation
Pressure in cuff
Gradually deflate cuff
Listen to pulse with stethoscope
Systolic BP
Onset of Korotkoff sounds
Maximum contraction BP > cuff pressure
High pass
filter (at 1 Hz)
SBP / DBP
criteria
determined
by comparison
with clinical
data
Korotkoff sounds
Minimum contraction BP < cuff pressure < maximum contraction BP
Diastolic BP
Silent flow ⇒ minimum contraction BP > cuff pressure
Amplify
Chua and Hin, Digital Blood Pressure Meter, Freescale Semiconductor
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
29
Embedded Systems — Hadassah College — Spring 2012
Biomedical Applications
Dr. Martin Land
30