109  IoT Application Domains: Knowledge Checks

109.1 Knowledge Checks and Exercises

Estimated Time: 30 min | Complexity: Intermediate

Test your understanding of IoT application domains with these quizzes and scenario-based exercises.

109.2 Learning Objectives

By completing these exercises, you will:

  • Assess your understanding of domain requirements across latency, reliability, scale, and power
  • Apply domain selection frameworks to real-world scenarios
  • Identify appropriate technologies for specific application needs
  • Recognize common deployment pitfalls and their solutions

109.3 Quiz 1: Domain Requirements

Question 1: A hospital needs to monitor ICU patient vital signs with alerts reaching nurses within 5 seconds. Which requirement profile is most appropriate?

Healthcare patient monitoring requires sub-second latency for critical alerts, 99.99% uptime (52 minutes/year max downtime), clinical-grade sensor accuracy, and HIPAA compliance for patient data privacy. Option D describes autonomous vehicle requirements (even more stringent latency but different regulatory framework).

Question 2: A vineyard deploys soil moisture sensors across 100 hectares. Which connectivity technology is most appropriate?

Agricultural IoT requires long-range, low-power connectivity for battery-operated sensors. LoRaWAN provides 2-15 km range with 5-10 year battery life on a single gateway deployment. Wi-Fi requires too much infrastructure, cellular is expensive per-device, and Bluetooth has insufficient range.

Question 3: What is the primary reason healthcare IoT devices cost 3-5x more than equivalent consumer devices?

FDA Class II medical device approval (510(k)) requires clinical validation studies, extensive documentation, and typically adds 12-24 months and $50K-500K to development costs. This regulatory burden, not materials or manufacturing, drives the cost differential.

109.4 Quiz 2: Smart Cities

Question 4: A city deploys smart parking sensors but achieves only 50% coverage due to budget constraints. Why does this lead to adoption failure?

Smart parking systems require 80%+ coverage to be trusted by drivers. When the app shows “available” but spots are taken (because the empty spot was in the uncovered 50%), users quickly lose faith and abandon the app. NYC’s initial 40% pilot underperformed until expanded to 80%+ coverage.

Question 5: Barcelona’s smart city deployment saved 200M+ EUR annually. What was the primary success factor?

Barcelona’s success came from deploying Sentilo as a unified data platform, enabling cross-domain optimization (parking data improves traffic routing, traffic data triggers adaptive lighting). Siloed deployments miss the 30-50% additional value from data correlation.

109.5 Quiz 3: Transportation and V2X

Question 6: Why is DSRC (802.11p) preferred over standard Wi-Fi for V2V safety communication?

Two vehicles approaching at 100 km/h have only ~100ms to exchange safety warnings. Standard Wi-Fi’s WPA2 handshake takes 2-3 seconds. DSRC (802.11p) eliminates this handshake entirely, enabling immediate broadcast of safety messages. It actually uses higher frequency (5.9 GHz) with lower data rates (3-27 Mbps) - reliability over speed.

Question 7: What percentage of traffic accidents could potentially be prevented by full V2X deployment?

94% of crashes are caused by human error (distraction, fatigue, impairment). V2X addresses these through intersection collision warning, lane change assistance, rear-end collision warning, and other safety applications. Studies estimate 70-80% of non-impaired crash scenarios could be prevented, representing 2+ million crashes annually in the US.

109.6 Quiz 4: Agriculture and Healthcare

Question 8: A dairy farm implements rumen bolus temperature sensors with a factory-default threshold of >39.5C for illness alerts. Why does this approach generate excessive false alarms?

Individual cows have unique baseline temperatures (Cow #234 might be 38.8C, Cow #456 might be 39.2C). A flat 39.5C threshold treats the 39.2C cow as always “almost sick.” Per-animal baselines with deviation-based alerts (baseline + 0.5C) reduce false positives by 80% while improving detection lead time to 18-24 hours before visible symptoms.

Question 9: A wearable ECG device achieves 95% sensitivity and 90% specificity for atrial fibrillation detection. In a population with 15% AFib prevalence, what is the approximate Positive Predictive Value?

PPV = (Sensitivity x Prevalence) / [(Sensitivity x Prevalence) + (1-Specificity) x (1-Prevalence)] PPV = (0.95 x 0.15) / [(0.95 x 0.15) + (0.10 x 0.85)] = 0.1425 / 0.2275 = 62.6%

This means 37% of positive alerts are false positives. In healthcare IoT, 90% specificity sounds good but can cause significant harm (unnecessary anticoagulation) in real populations.

109.7 Scenario Exercises

109.7.1 Scenario 1: Smart City Waste Management

Context: A city currently spends $45M annually on garbage collection with trucks following fixed routes regardless of bin fill levels. Analysis shows 40% of stops find bins less than 50% full, wasting $18M in fuel and labor.

Question: The city considers deploying fill-level sensors to 25,000 bins across 500 square miles.

  1. Domain Fit: Would “Smart Environment” (air quality) or “Smart Waste Management” (part of Smart Cities) be the correct domain classification?

  2. Connectivity Choice: Which technology - Wi-Fi, cellular (NB-IoT), or LoRaWAN - makes most sense for 25,000 bins spread across 500 square miles?

  3. ROI Calculation: If sensors cost $3M to deploy and save $10M annually through route optimization, what is the payback period?

  4. Coverage Threshold: At what percentage of bin coverage would the system provide reliable route optimization?

Answers: 1. Smart Waste Management falls under Smart Cities infrastructure (not Smart Environment which focuses on air quality, fires, earthquakes) 2. LoRaWAN or NB-IoT - both provide city-scale coverage with 10-year battery life. Wi-Fi would require 5,000+ access points. 3. Payback = $3M / $10M = 0.3 years (3.6 months) - extremely fast ROI 4. 80%+ coverage required for reliable route optimization; below this, drivers can’t trust the system

109.7.2 Scenario 2: Healthcare Remote Patient Monitoring

Context: A hospital system wants to implement remote monitoring for 2,000 heart failure patients post-discharge to reduce 30-day readmissions (currently 25% readmission rate, costing $50M annually).

Question: Design the monitoring approach.

  1. Sensor Selection: What vital signs should be monitored for heart failure patients?

  2. Alert Threshold Design: How would you balance sensitivity (catching deterioration) vs. specificity (avoiding alert fatigue)?

  3. Connectivity: Should devices use Wi-Fi (patient home), cellular, or Bluetooth to smartphone?

  4. Regulatory Considerations: What FDA and HIPAA requirements apply?

Considerations: 1. Weight (daily), blood pressure, heart rate, SpO2, and symptoms questionnaire. Weight gain of >2 lbs/day indicates fluid retention. 2. Use trending (weight increase over 3 days) rather than single readings. Multi-parameter algorithms (weight + BP + symptoms) improve specificity. 3. Cellular or Bluetooth-to-smartphone with cellular backup. Wi-Fi only works if patient has reliable home internet. 4. FDA Class II for devices making clinical claims; HIPAA for all data transmission and storage; need BAA with cloud providers.

109.7.3 Scenario 3: Wearable Fitness Tracker Accuracy

Context: A user notices their fitness tracker shows heart rate of 175 BPM during a casual 3 mph walk (actual: ~100 BPM).

Question: Explain what’s happening and how to fix it.

  1. Root Cause: What causes PPG optical heart rate sensors to produce wildly inaccurate readings during movement?

  2. Cadence Confusion: If the user is walking at 170 steps per minute, how does this affect the heart rate reading?

  3. Solutions: What can the user do to get more accurate readings during exercise?

  4. Design Implications: How should fitness apps communicate heart rate data quality to users?

Answers: 1. Motion artifacts - arm movement causes the sensor to shift against skin, creating light intensity variations interpreted as pulse beats. 2. Walking cadence (170 steps/min) matches typical exercise heart rates (170 BPM), making it impossible for the algorithm to distinguish motion from pulse. 3. Wear device tighter; use chest strap for exercise; use arm bands instead of wrist; trust average HR over instantaneous readings. 4. Show confidence indicators; flag data as “motion-affected”; recommend chest strap for serious training; use accelerometer to detect high-motion periods and reduce HR display confidence.

109.8 Understanding Check: Building Automation

Scenario: A 200,000 sq ft office building has 180 VAV (Variable Air Volume) boxes. The building engineer suspects many are malfunctioning, causing simultaneous heating and cooling in the same zones.

Questions to Consider:

  1. What sensors would you deploy to detect VAV box faults?
  2. How would you identify “simultaneous heating and cooling” from sensor data?
  3. What is the energy impact of this fault?
  4. How would you prioritize repairs across 180 boxes?

Key Insights: - Supply air temperature + zone temperature + damper position + reheat valve position sensors per zone - Fault signature: reheat valve open (>10%) while damper fully open AND zone temperature below setpoint - Energy impact: 15-30% zone energy waste from the affected boxes - Prioritize by: (1) energy waste magnitude, (2) comfort complaints, (3) repair cost

109.9 Self-Assessment Checklist

Before moving to the next chapter, verify you can:

109.10 Summary

These exercises tested your understanding across all IoT application domains:

  • Domain requirements vary dramatically (10ms to hours latency, 95% to 99.99% reliability)
  • Technology selection must match domain constraints (power, range, data volume)
  • Healthcare IoT requires clinical validation and regulatory compliance
  • Smart cities need unified platforms and 80%+ sensor coverage
  • V2X communication prioritizes reliability and instant messaging over data rates
  • Wearables must balance accuracy claims with realistic sensor limitations

109.11 What’s Next

Return to the domain that interests you most:

Return to Overview