Connected Health and Wearables · Study deck

Wearable IoT: Validation and Design Trade-offs

A wearable fits the chosen body location and users are willing to wear it.

Ada is your guide for this deck.

applicationdomainswearables
iotclass.org

After studying this chapter

Learning objectives

You will be able to:

  • Explain: Everything else, from battery to compute to audio, must fit into the arms and the front housing and still be light enough to wear for an hour.
  • Explain: Below the strip the dark visor curves across both eyes, and the display optics have to live inside that thin shell without blocking sight.
  • Explain: The light reaching the detector depends on how firmly the cuff presses, and a swinging arm changes that pressure many times a second.
  • Explain: The mistake: Expecting consistent biometric readings without accounting for how skin conditions, device positioning, and environmental factors cause sensor drift.
iotclass.org

Major section

Start With the Story

That still does not prove its readings, battery promise, or sleep and health claims.

  • The team now needs tests that expose motion, drift, sampling, and validation limits before release.
  • The mistake: Selecting wearables based on advertised battery life without understanding that GPS, always-on displays, and continuous heart rate dramatically reduce runtime.
  • This explains why the same watch can last "7 days" (basic mode with infrequent HR sampling) or "1.5 days" (all features enabled).
iotclass.org

Major section

PPG Motion Artifact Pitfall

The light reaching the detector depends on how firmly the cuff presses, and a swinging arm changes that pressure many times a second.

  • At rest the reading looks clean; in motion, the pressure change can be larger than the pulse itself.
  • Why it happens: Arm movement causes sensor shift against skin.
  • Running cadence (150-180 steps/min) matches typical exercise heart rates.
This wrist-worn pulse oximeter places an optical PPG sensor against moving skin. The wearable form makes the central pitfall concrete: changes in contact pressure and motion can modulate the received light more strongly than the blood-volume pulse the algorithm is trying to isolate. Photo: Peter H Charlton, CC BY 4.0
This wrist-worn pulse oximeter places an optical PPG sensor against moving skin. The wearable form makes the central pitfall concrete: changes in contact pressure and motion can modulate the received light more strongly than the blood-volume pulse the algorithm is trying to isolate. Photo: Peter H Charlton, CC BY 4.0
iotclass.org

Major section

Skin Contact Sensor Drift

The mistake: Expecting consistent biometric readings without accounting for how skin conditions, device positioning, and environmental factors cause sensor drift.

  • Why it happens: Skin hydration changes, sweat accumulation, device loosening, sunscreen/lotions create optical barriers.
  • Scenario: A health technology company is launching a sleep tracking feature for their fitness band.
  • They must determine appropriate accuracy claims based on validation against polysomnography (PSG), the clinical gold standard.
iotclass.org

Major section

Wearable Sensor Technologies

Step counts, oxygen saturation, ECG, and bioimpedance each use different measures and reference methods.

  • The clinical column is not a software upgrade label.
  • Below the strip the dark visor curves across both eyes, and the display optics have to live inside that thin shell without blocking sight.
  • Everything else, from battery to compute to audio, must fit into the arms and the front housing and still be light enough to wear for an hour.
Consumer and clinical accuracy differ across optical heart rate, movement, oxygen saturation, skin temperature, ECG, and bioimpedance sensors.
Consumer and clinical accuracy differ across optical heart rate, movement, oxygen saturation, skin temperature, ECG, and bioimpedance sensors.
iotclass.org

Deck summary

Key takeaways

That still does not prove its readings, battery promise, or sleep and health claims.

  • The light reaching the detector depends on how firmly the cuff presses, and a swinging arm changes that pressure many times a second.
  • The mistake: Expecting consistent biometric readings without accounting for how skin conditions, device positioning, and environmental factors cause sensor drift.
  • Step counts, oxygen saturation, ECG, and bioimpedance each use different measures and reference methods.
iotclass.org

Retrieval practice

Recall check

Ada says: answer from memory, then check your reasoning.

Q1In a wear test, people stop using a tracker because it irritates skin, needs inconvenient charging, and disrupts daily routines. Which design principle has the clearest failure?

AComfort and daily-routine fit, including skin tolerance and charging burden.
BUse sensor accuracy as the primary wear-test acceptance measure.
CAdd sensor features to improve the tracker’s acceptance in daily use.
DPrioritize a sharper display and clearer screen prompts as the main response to poor wear-test retention.
Show answer

Answer: A The wear test identifies discomfort, charging burden, and routine disruption.

Q2During a morning jog, a user's wearable shows a heart rate spike to 185 BPM during what feels like moderate effort. The user's actual heart rate (chest strap verified) is 142 BPM. What is the MOST likely explanation?

AArm-swing motion artifacts corrupting the optical PPG heart-rate signal.
BTreat optical-sensor hardware drift as the cause and schedule sensor replacement.
CEscalate the wearable spike as a confirmed arrhythmia event.
DTreat the chest-strap value as erroneous and retain the wearable reading.
Show answer

Answer: A The chest-strap reference and exercise-only wrist spike make motion/contact artifact a plausible explanation.

Q3A consumer sleep tracker reports 74% epoch-by-epoch accuracy against polysomnography (PSG), with no repeated-night validation. Which communication strategy is MOST appropriate?

AMarket clinical validation from 74% single-night accuracy without repeat tests.
BPresent wellness estimates with limits; validate weekly trends separately.
CReject all wellness use because 74% is insufficient for any sleep estimate.
DShow raw nightly numbers alone, implying they are exact stage measurements.
Show answer

Answer: B The study supports a bounded wellness statement about epoch agreement.

Q4A wearable advertises seven days under basic use (100/7% battery drain per day). Continuous heart-rate monitoring adds 12 percentage points of daily drain and one GPS hour adds 10 points; an always-on display then raises that subtotal by 30% relative to the subtotal. Approximately how long will the battery last?

AAbout six days, treating added daily GPS, HR, and display drain as negligible.
BAbout two to three days, allowing for the combined added daily drain.
CAbout four to five days, undercounting the listed extra daily drain.
DLess than one day, because all enabled features draw the full battery capacity each day.
Show answer

Answer: B Under the stated illustrative model, daily drain is (100/7 + 12 + 10) × 1.30 ≈ 47.2%, so nominal runtime is 100/47.2 ≈ 2.1 days. Measure the real usage profile before making a battery-life claim.

Q5What turns qualified wearable readings into a bounded trend or coaching action?

AOn-Body Sensor and Local Signal
BTrend Insight and User Action
CPhone Relay and Consent Boundary
Show answer

Answer: B The application turns qualified readings into a bounded trend, prompt, or coaching action and states when the evidence is not diagnostic.

Q6Complete the wearable heart-rate confidence gate:

Aif sample['motion_artifact'] > 0.3: return False
Bif sample['motion_artifact'] < 0.3: return False
Cif sample['heart_rate_bpm']: return False
Dif sample['battery_pct'] > 5: return False
Show answer

Answer: A Wearable metrics should be withheld when battery, skin contact, motion artifact, or model confidence make the sample unreliable.

iotclass.org

Print reference

Answers 1 of 2

Answer key.

  1. A · The wear test identifies discomfort, charging burden, and routine disruption.
  2. A · The chest-strap reference and exercise-only wrist spike make motion/contact artifact a plausible explanation.
  3. B · The study supports a bounded wellness statement about epoch agreement.
  4. B · Under the stated illustrative model, daily drain is (100/7 + 12 + 10) × 1.30 ≈ 47.2%, so nominal runtime is 100/47.2 ≈ 2.1 days. Measure the real usage profile before making a battery-life claim.
iotclass.org

Print reference

Answers 2 of 2

Answer key.

  1. B · The application turns qualified readings into a bounded trend, prompt, or coaching action and states when the evidence is not diagnostic.
  2. A · Wearable metrics should be withheld when battery, skin contact, motion artifact, or model confidence make the sample unreliable.
iotclass.org