Test 5 Flashcards
Additional sesnors and connectivitiy
Rings
-Tracks sleep, HR
Hearing aids/earphones
-Measure motion and adjust autitory algorythms
-“spatial audio” with apple air pod pros using inertial sensors
Can we integrate all these data– next step
Optiocal sensors in smart watches
technology is not new- but now can package it up smaller
Optical emitters
-2+ LEDS send light into the skin (to measure HR)
Photodectecor
-Captures refracted light dependent on the volume and oxygenation of blood
Accuracy of HR from wearables
-Most recordings within ~2%
best at rest- not as good when moving
Accuracy of blood oxygen readings from wearables
-Less accurate than HR but still within 3-4%
-but consider the small windows for “healthy” ranges
-Small difference is clincally meaningful (100-95=good) vs (90-95=ok) vs (<90=seek medical)
inconclusive in personal testing
Optical sensors in smart watches: what’s next- continous BP monitoring
-often just predicitve equations based on demographic, spo2 and HR
Optical sensors in smart watches: what’s next- blood, sweat and tears
-Biofluid sampling of interstitial fluid (ISF), sweat, tears, saliva, urine, ect
-Glucose, lactate, hormones, metabolites, blood alcohol levels
Optical sensors in smart watches: what’s next- data driven decision making
can collect lots of imformation but need to take extreme caution wehn metrics can personally inform or alter health or medical decisions
Other advancements- U of A
-Self regulating mock-aorta tube- can adapt to cahnges in PSI
-Smart jaw prothetics
-Two-layered dry electrode patch for wearable sEMG and EEG Devices- flexible sensor
-Skin with pressure/temperat re sensitivity (hydrogels to make e-skin smarter)
-Gold coated material for biolumenscence
Force sensing insoles
-inertial sensores only measure impacts and motion
-Ground reaction forces can be inferred when assessed near centre of mass, but not lower limbs
-ie RPYXSI- created for diabetic neuropathy
Motion capture solutions- markers
conventional marker-based technology- marks on specific landmarks
very time consuming
Motion capture solutions- inertial sensor solutions- Xsens
suit with many inertial sensors on rigid bodies
-still knee to put on many sensors and calibrate but less bas vs traditional markings
Motion capture solutions- markerless motion capture
-quicker vs marker sensor based measurement
-relies entirely on “computer vision”- ability for computers to see
– build upon large database with objects labelled/segmented
–many different uses, algorythms, cameras/image types - train to be able to identify different things
markerless motion capture process
image classification: person, sheep, dog
Object detection: bounding box
Segmentation: segment out things
motion capture solutions: Theia Markerless
-Puts virtual markers on specific land markers to track individual
-still limited by space in which the camera can see- don’t ger real world- sepace for inertial sensors
Virtual reality- need many things at once
Map surrounding while constantly updating the position of the user
-simultaneous localization and mapping (SLAM)
-Same technology required for self driving cars
Track position/ orientation of headset
-computer vision + inertial sensor in headset
Track positioning of controllers or hands
-computer vision and/or inertial sensors in controllers
What is 3D printing
A process that creates physical objects from digital designs by adding material, one layer at a time
A quick history on 3D printing
-Invented in 1980s
-Grew in the 1990s
-forst commerical printer in 2006
-Became more affordable in 2010s
-Now multibillion dollar business
3D printing Applications
Wide variety of aplications
-From scuptures, trinkets, and parts to prothetics and even food
Basics of 3D printing- 3 general steps:
1-Design an object you want to print (or one someone else has made)
2-IMport to a program to “slice” it for printing
3-Upload to your design to print
Tips to think about when 3D printing
1-Test and optimize your printer settings for each filament
-Temperature, speed, layer depth, infill density, ect
-Specific to each machine/ filament
II-Make it as easy as possible to rpint
-Flat layer on bottom that can be easily sliced (if possible)
-minimal overhang or add supports
