Wearable Sensor Systems Flashcards
Wearable technology
Any technology you can wear
Wearable sensors
- Often synonymous with the term wearable technology
- Measuring some physical quantity
Examples of measurable physical quantities
Movement, force, light, temperature, chemicals
Examples of Sensors
Inertial sensors, force sensors, light sensors, thermocouples, chemical sensors
Measuring Movement
- Inertial Sensors - Utilize the principle of inertia
- Variety of applications - drones and vehicles, smart phones, electronics, industry measurements, human movement
Application of inertial sensors for human movement
- Activity trackers
- Gait analysis
- Balance assessment
- Fall detection
- sleep analysis
- sport performance
What are movement constructs
Physical activity
Mobility
Gait
Metrics
Things that measure movement constructs
PA: step counts, activity tracking, sedentary time, activity intensity
Mobility: Timing functional tasks, turning velocity, balance
Gait: gait speed, stride time, impact accelerations, knee flexion angle
Primary types of inertial sensors
Accelerometer (linear acceleration) and gyroscope (rotational movement)
- Integrated together as an inertial measurement unit (imu)
- may also integrate magnetometer or microcontroller unit
Newton’s laws of motion
- First Law: Object remains at rest unless acted on by a force
- Second Law: F=ma
- Third Law: Equal and opposite reaction
How do accelerometers work
- Case containing proof mass (known)
- Force exerted by proof mass movement during acceleration is measured
- m and F used to calculate a
- at rest measured as 9.81
- during free fall measured as 0
- Aligned orthogonally to measure acceleration in 3D
How do you calculate the resultant acceleration
SQRT (ax^2+ay^2+az^2)
How do you convert voltage to gs
- Know resting data is 1g and upside down -1g
- change in g over change in mV = conversion factor
Range
Min/max values that can be accurately measured
Sensitivity
How much the output changes in relation to the input
- How well you can measure changes in movement
How are range and sensitivity related
Inversely related
Linearity
The output does not follow a straight line with input
- Change in voltage is not linearly proportional to change in acceleration
- Usually more common near end-range
- Try to stay with away from end-ranges
Tips when selecting an accelerometer
Select a sensor that offers the greatest sensitivity within a safe range for your movement/placement
GENERAL RECOMMENDATIONS
- +/-2g lower back during walking or ADL
- +/-8g lower limbs during walking or ADL
- +/-16-32g lower limbs during running
- +/- 100-200g high impact activities
Magnetometers
An electrical compass
-Work based on the hall effect (voltage difference produced across an electrical conductor due to a perpendicular magnetic field
- Uses earth’s magnetic field to measure heading
- Outputs can be significantly distorted by other magnetic fields, electrical signals, or devices
What can linear acceleration and angular velocity be used for
- Assessing overall amount of movement or type of activity
- determining timing of gait events
- segment motion, pelvic stability or between limb asymmetry
- Impacts magnitudes
- Can be integrated to determine location or position and segment or joint angles
Using accelerometers as tilt sensors
- Uses gravity and simple trigonometry
- Words very well when there is little to no movement
Sampling Frequency
- The number of samples taken in 1 second
- Balance amount of data collected with enough to not miss events and get full pattern view
What are sensor reactions to dynamic changes in input tied to?
- The type of input
- Step, ramp, impulse, sinusoidal - The type of mechanical system
- Zero, first, second order
-n+1 coefficients characterized the system
Input types
Changes in the input, where we can measure the time response of the system
1. Step input: Sudden change in input that is held constant
2. Ramp input: linear increase in input
3. Impulse input: sudden spike in input
4. Sinusoidal input: sine wave input
Zero order systems
- The output of a zero order system is proportional to the input
- Depends only on displacement and 1 constant k
- F(t)=kx(t)
k= spring constant, resistance
First order system
- The output depends on differential equation
- Depends only on displacement (x) and velocity (x’) and 2 constants (c and k)
- F(t) = c x’(t)+ kx(t)
- c = damping, viscosity, friction coefficient
- Experience a time delay between input and output dependent on velocity
Second order system
- Output depends on differential equation
- Depends on displacement (x), velocity (x’) and acceleration (x’’) and 3 constants
- F(t) = mx’‘(t) + cx’(t) + kx(t)
- involves oscillations related to natural frequency and damping ratio
- Mass can overshoot and carries own velocity
Natural frequency
Frequency the system would oscillate if there was no damping
w (omega)
How system responds relates to the relationship between the 3 constants
w= sqrt(k/m)
Damping ratio
Describes how oscillations decay after an input (zeta)
zeta = c/2squrt(mk)
Undamped system
Nothing to damp out oscillations and oscillations occur at w
Underdamped system
Some damping but not enough to eliminate all oscillations
Optimal damping
Does not eliminate oscillations but settles very quickly
Critically damped system
No overshoot or undershoot (damping just enough to eliminate oscillations)
Overdamped system
No overshoot but reaches final value slower than critically damped
Second order system car example
- Underdamped suspension will have large oscillations that take multiple cycles to settle
- Overdamped suspension will not give quickly
Accelerometers as second order systems
While minor oscillations could occur due to intrinsic properties, these are often more related to properties extrinsic to the sensor
- Sensor mounting in unit
- sensor attachment to body system
- soft tissue artifact
Filtering Data
Filtering the outputs can remove unwanted noise
LOW-PASS FILTER
- removes unwanted high-frequency component
HIGH-PAS FILTER
- Removes unwanted low-frequency components
