18 Mobile Scenario Analysis
18.1 Learning Objectives
By the end of this chapter, you will be able to:
- Design Agricultural Sensor Networks: Evaluate sub-GHz vs 2.4 GHz trade-offs and construct power budgets for multi-year battery life deployments across large areas
- Diagnose 2.4 GHz Interference: Interpret RSSI measurements to identify Zigbee/Wi-Fi channel overlap and recommend coexistence strategies using frequency separation analysis
- Compute Indoor Link Budgets: Derive received power and link margin for same-floor and multi-floor sensor placements, incorporating wall and floor penetration losses
- Justify Technology Selections: Defend wireless protocol choices by synthesising range, power, capacity, and cost constraints into a structured decision framework
18.2 Prerequisites
Required Chapters:
- Mobile Wireless Technologies Basics - Core concepts
- Cellular Network Architecture - Cellular IoT selection
- Networking Fundamentals - Basic networking
Technical Background:
- Path loss and link budget concepts
- Frequency band characteristics
- ISM band regulations (duty cycle, power limits)
Estimated Time: 45 minutes
- Scenario-Based Analysis: Applying wireless technology knowledge to real-world IoT deployment decisions rather than abstract specifications
- Technology Selection Scenarios: Matching application requirements (range, power, throughput, mobility, cost) to appropriate wireless technology
- Urban vs Rural Coverage: Urban areas have dense cellular and Wi-Fi infrastructure; rural areas may require LoRaWAN, satellite, or cellular with reduced coverage
- Environmental Constraints: Temperature extremes, humidity, vibration, and EMI affecting hardware selection and antenna design
- Regulatory Compliance Scenarios: Designing within duty cycle limits, power restrictions, and frequency allocations for target deployment regions
- Business Case Analysis: Total cost of ownership including hardware, data plans, deployment, and maintenance over device lifetime
- Failure Mode Analysis: Identifying single points of failure and designing redundancy or graceful degradation
- Scalability Planning: Designing systems that can grow from pilot (100 devices) to full deployment (100,000 devices) without architectural changes
18.3 For Beginners: How to Use These Scenarios
What are these scenarios? Real-world wireless deployment problems that require trade-off analysis. Each scenario presents constraints and asks you to reason through solutions.
How to approach them:
- Read the scenario and constraints carefully
- Think about the questions before revealing the answer
- Study the “Key Insight” sections for important principles
- Use “Verify Your Understanding” to test your reasoning
Why scenarios matter: Multiple-choice questions test recall. Scenarios test understanding - the ability to apply principles to new situations you haven’t seen before.
18.4 Scenario 1: Large-Area Agriculture Wireless Design
Scenario: You’re deploying soil moisture sensors across a 200-hectare farm (1.4 km x 1.4 km). Sensors must transmit 100-byte readings every 15 minutes and run on batteries for 5+ years without replacement. The farm has crops, equipment, and varying terrain that will obstruct line-of-sight.
Think about:
- How does radio frequency affect range when penetrating vegetation and soil?
- What battery capacity is needed for 5 years if transmitting 96 times per day?
- Why might infrastructure cost matter less than battery replacement labor over 5 years?
Key Insight: Range vs Frequency
Rules of thumb:
- At the same distance, 868/915 MHz has approximately 9 dB less free-space path loss than 2.4 GHz
- Lower frequencies are often more forgiving with foliage and non-line-of-sight paths, but range is still site-dependent (antenna height, terrain, noise floor, regulations)
Path Loss Calculation:
Using the free-space path loss formula: FSPL(dB) = 20log(d_km) + 20log(f_MHz) + 32.45
| Frequency | FSPL at 1 km | Difference |
|---|---|---|
| 868 MHz | 91.2 dB | baseline |
| 915 MHz | 91.7 dB | +0.5 dB |
| 2.4 GHz | 100.0 dB | +8.8 dB |
The approximately 9 dB advantage means sub-GHz can reach similar distances with roughly 8x less transmit power (or achieve greater range with equal power).
Key Insight: Battery Life Reality Check
- 100 bytes every 15 minutes averages approximately 0.9 bps, so bandwidth is not the limiting factor
- Battery life is dominated by time-on-air, receive windows, and retries - not just payload size
- A useful estimate:
I_avg = (I_tx x t_tx + I_rx x t_rx + I_sleep x t_sleep) / 24h, then compare against battery capacity (and include temperature/aging margins)
Example Power Budget (LoRaWAN Class A):
| State | Current | Duration | Energy/Day |
|---|---|---|---|
| Transmit (14 dBm) | 120 mA | 80 ms x 96 = 7.68 s | 0.26 mAh |
| RX Windows | 12 mA | 500 ms x 96 = 48 s | 0.16 mAh |
| Sleep (PSM) | 2 uA | 86,344 s | 0.05 mAh |
| Total | 0.47 mAh/day |
With a 19,000 mAh lithium battery (common D-cell): 19,000 / 0.47 = 40,000+ days theoretical
Reality factors (temperature, self-discharge, aging): 5-10 year battery life is achievable.
Key Insight: Likely Architectures
Choose based on your specific constraints:
Sub-GHz LPWAN (LoRaWAN):
- Good fit when you can deploy/operate gateways and tolerate shared-spectrum constraints
- Typical: 1-3 gateways cover 200 hectares with proper antenna placement
- No per-device subscription fees
Licensed Cellular LPWAN (NB-IoT/LTE-M):
- Good fit if coverage exists and subscriptions/lock-in are acceptable
- Reduces your gateway operations burden
- Carrier manages network infrastructure
2.4 GHz Mesh (Zigbee/Thread):
- Can work if you can place powered/solar routers
- Avoid making battery sensors route traffic
- Range limitations in outdoor/vegetated environments
Wi-Fi:
- Typically needs power and denser infrastructure
- Best when throughput is the priority
- Not ideal for multi-year battery operation
Verify Your Understanding:
If 2.4 GHz adds approximately 9 dB of FSPL vs 868/915 at the same distance, what does that imply for range in free space (n=2) vs a cluttered environment (n > 3)?
Which deployment choices increase link margin without raising transmit power (gateway height, antenna choice, payload interval, data rate)?
Where do you expect the operational cost to land: field visits for batteries vs installing/maintaining gateways?
Answer (sketch): Approximately 9 dB corresponds to roughly 8x power. In free space that’s roughly 2.8x range (since range scales with sqrt(power) for n=2), and less in cluttered environments. Mesh routing shifts cost to always-on routers and maintenance. The payload rate is tiny, so the real battery drivers are airtime, retries, and idle current.
18.5 Scenario 2: 2.4 GHz Interference Mitigation
Scenario: Your Zigbee smart building deployment on Channel 20 (2450 MHz) is failing. RSSI measurements show: - Desired Zigbee signal: -65 dBm - Wi-Fi Channel 6 (2437 MHz): -55 dBm (10 dB stronger!) - Wi-Fi Channel 11 (2462 MHz): -60 dBm - Microwave oven: periodic -40 dBm spikes (25 dB stronger than Zigbee!)
Think about:
- How wide is a Wi-Fi channel vs a Zigbee channel in MHz?
- Which Zigbee channels avoid Wi-Fi Channel 6 and 11 overlap?
- Can you eliminate microwave interference by changing channels?
Key Insight: Channel Bandwidth Mismatch
- Wi-Fi channels: 22 MHz wide, centered at 5 MHz intervals
- Zigbee channels (15-26): 2 MHz wide, centered at 5 MHz intervals
- Wi-Fi Channel 6 (2437 MHz) spans 2426-2448 MHz
- Zigbee Channel 20 sits between Wi-Fi channels 6 and 11, but strong Wi-Fi on both sides can still cause adjacent-channel interference. It’s also near the microwave oven center frequency (approximately 2.45 GHz).
Key Insight: Frequency Separation Analysis
| Zigbee Ch | Center (MHz) | Delta to Wi-Fi Ch6 (2437) | Delta to Wi-Fi Ch11 (2462) | Practical note |
|---|---|---|---|---|
| 15 | 2425 | 12 MHz | 37 MHz | Often good when Wi-Fi uses 6/11 heavily |
| 20 | 2450 | 13 MHz | 12 MHz | “Squeezed” between 6 and 11 when both are active |
| 25 | 2475 | 38 MHz | 13 MHz | Above Wi-Fi 11; often a solid first choice |
| 26 | 2480 | 43 MHz | 18 MHz | Also above Wi-Fi 11, but near band edge (power/regulatory constraints vary) |
Key Insight: Best Practice - Wi-Fi Coexistence
- Use a spectrum scan to pick a Zigbee channel with the lowest observed interference (common candidates: 15, 20, 25)
- In this scenario (strong Wi-Fi on 6 and 11 + microwave spikes near 2.45 GHz), Zigbee channel 25 is often a good first try
- If you control the Wi-Fi network, moving high-throughput traffic to 5 GHz reduces 2.4 GHz congestion for Zigbee
Key Insight: Microwave Reality
- Microwave ovens leak noise centered around approximately 2.45 GHz that can impact multiple nearby channels (often worst around Zigbee channel 20)
- Channel selection can reduce impact, but you should still expect periodic retries during microwave use
- Mitigation: Zigbee’s CSMA/CA automatically retries when clear
- Microwaves typically run less than 5 minutes; Zigbee tolerates brief outages
Verify Your Understanding:
- Why does Wi-Fi Channel 6 span 2426-2448 MHz if it’s “centered” at 2437 MHz?
- If you can’t change Zigbee channel, what Wi-Fi channels would reduce interference?
- Why is -55 dBm Wi-Fi worse for Zigbee than -65 dBm desired signal?
Answer: Wi-Fi uses approximately 20-22 MHz channels, so +/-10-11 MHz from center. If Zigbee must stay on channel 20, move Wi-Fi away from channels 6/11 (or move Wi-Fi to 5 GHz). The -55 dBm Wi-Fi signal is approximately 10 dB stronger than Zigbee, so the receiver’s signal-to-interference ratio is poor even if the Zigbee link budget is “fine.”
18.6 Scenario 3: Indoor Link Budget Calculation
Scenario: You’re deploying Wi-Fi-based sensors in a 3-story office building (50m x 30m per floor). One access point per floor, centered. Specs: - TX Power: 20 dBm (100 mW) - Antenna Gain (TX/RX): 2 dBi each - Frequency: 2.4 GHz - RX Sensitivity: -85 dBm - Required Fade Margin: 10 dB (for reliability) - Floor penetration: 15 dB loss - Wall penetration: 5 dB each (max 2 walls to any corner)
Think about:
- What’s the distance from center to corner of 50m x 30m floor?
- How much does one concrete floor reduce signal strength?
- Is 10 dB fade margin enough for production deployment?
Key Insight: Same-Floor Corner Sensor (Scenario 1)
Distance calculation: Distance = sqrt((50/2)^2 + (30/2)^2) = sqrt(625 + 225) = 29.2 meters
Path loss calculation:
- Free Space Path Loss at 2.4 GHz, 29.2m: 69.3 dB
- Wall loss (2 walls): 10 dB
- Total path loss: 79.3 dB
Same-floor link budget summary: TX (+20 dBm) + Antennas (+4 dBi) - Path Loss (79.3 dB) = RX Power (-55.3 dBm). With -85 dBm sensitivity, margin is +29.7 dB (excellent).
Key Insight: Adjacent Floor Corner Sensor (Scenario 2)
3D distance calculation: Distance = sqrt(29.2^2 + 4^2) = 29.4 meters
Path loss calculation:
- FSPL at 29.4m: 69.4 dB
- Wall loss: 10 dB
- Floor loss: 15 dB
- Total path loss: 94.4 dB
Adjacent-floor link budget summary: TX (+20 dBm) + Antennas (+4 dBi) - Path Loss (94.4 dB) = RX Power (-70.4 dBm). With -85 dBm sensitivity, margin is +14.6 dB (marginal - only 4.6 dB excess).
Key Insight: Reality Check
- Same floor: 29.7 dB margin - Excellent, handles multipath, interference, movement
- Different floor: 14.6 dB margin - Only 4.6 dB excess after 10 dB requirement
- 4.6 dB excess can disappear from: furniture, metal filing cabinets, interference, actual floor construction
Key Insight: Production Recommendation
- A single AP per floor is often insufficient for reliable multi-floor coverage
- Plan for additional APs and/or dedicated coverage per floor (and validate with a site survey)
- Floor penetration losses can be a dominant limitation in practice
- Design with an appropriate fade margin (often on the order of 10-20 dB, depending on requirements and environment)
Verify Your Understanding:
- Why does a same-floor sensor at 29.2m perform far better than an adjacent-floor sensor at 29.4m despite nearly identical distance?
- What happens if actual concrete floor has 20 dB loss instead of assumed 15 dB?
- How much fade margin would you target for a high-reliability enterprise deployment, and why?
Answer: Floor penetration adds 15 dB extra loss (huge!). 20 dB floor would give only +9.6 dB total margin (margin falls below the 10 dB requirement) - deployment would be fragile. Many enterprise designs target substantial fade margin (often approximately 15-20 dB), but the right number depends on the environment, traffic, and reliability goals.
For a 3-story building with 50m x 30m floors and center-placed APs, calculate link budget to a corner sensor on adjacent floor. Distance: \(\sqrt{(50/2)^2 + (30/2)^2 + 4^2} = \sqrt{625 + 225 + 16} = 29.4\) m = 0.0294 km. At 2.4 GHz:
\[ \text{FSPL} = 20\log_{10}(0.0294) + 20\log_{10}(2400) + 32.45 = -30.63 + 67.60 + 32.45 = 69.4 \text{ dB} \]
Worked example: TX: +20 dBm, antenna gains: +4 dBi, total path loss: 69.4 dB (FSPL) + 10 dB (walls) + 15 dB (floor) = 94.4 dB. RX power: \(20 + 4 - 94.4 = -70.4\) dBm. With -85 dBm sensitivity, margin is \(-70.4 - (-85) = 14.6\) dB. After subtracting 10 dB fade margin, only 4.6 dB excess remains – marginal, consistent with the main analysis above showing multi-floor coverage is risky.
18.7 Scenario Analysis Framework
When analyzing wireless deployment scenarios, use this systematic framework:
18.7.1 Step 1: Identify Constraints
| Category | Questions |
|---|---|
| Range | Maximum distance? Obstacles? Indoor/outdoor? |
| Power | Battery life requirement? Power source available? |
| Data | Payload size? Transmission frequency? Latency tolerance? |
| Environment | Interference sources? Regulatory region? |
| Cost | Per-device budget? Infrastructure investment? |
18.7.2 Step 2: Calculate Link Budget
Received Power = TX Power + TX Antenna Gain - Path Loss + RX Antenna Gain
Link Margin = Received Power - RX Sensitivity
Required Margin = Fade Margin + Interference Margin18.7.3 Step 3: Select Technology
Match constraints to technology capabilities:
| If You Need… | Consider… |
|---|---|
| Long range, low power, infrequent data | Sub-GHz LPWAN (LoRaWAN, Sigfox) |
| Long range, moderate data, mobility | Cellular (LTE-M, NB-IoT) |
| Short range, low power, mesh | 802.15.4 (Zigbee, Thread) |
| Short range, high data | Wi-Fi (2.4/5/6 GHz) |
| Indoor, deep penetration | Sub-GHz or NB-IoT |
18.7.4 Step 4: Validate and Iterate
- Conduct site surveys before deployment
- Measure actual path loss vs calculated
- Test under realistic interference conditions
- Plan for worst-case scenarios
Sammy Sensor: “Wireless problems are like detective cases! You gather clues (measurements), analyze evidence (link budgets), and solve the mystery!”
Lila the Light Sensor: “When Wi-Fi and Zigbee fight over the same frequencies, it’s like two people trying to talk at the same time - someone needs to move to a different conversation!”
Max the Motion Detector: “Buildings are like mazes for radio waves. Walls slow them down, floors really slow them down, and metal stops them almost completely!”
Bella the Button: “The best wireless detective always checks the crime scene (site survey) before guessing what happened!”
18.8 Knowledge Check
Q1: A warehouse deploys Zigbee sensors on channel 20 (2450 MHz). A spectrum scan reveals strong Wi-Fi on channels 6 and 11. Which Zigbee channel would provide the best interference avoidance?
- Channel 11 (2405 MHz) – below all Wi-Fi channels
- Channel 18 (2440 MHz) – between Wi-Fi 6 and 11
- Channel 25 (2475 MHz) – above Wi-Fi channel 11
- Channel 20 (2450 MHz) – keep the current channel
C) – Channel 25 at 2475 MHz sits above Wi-Fi channel 11 (which spans up to ~2473 MHz), providing the best frequency separation. Channel 11 risks overlap with Wi-Fi channel 1, channel 18 sits between two active Wi-Fi channels, and channel 20 is the current problematic channel.
18.9 Knowledge Check
Q2: A sensor at the corner of a 50m x 30m floor is 29.2 meters from a central Wi-Fi AP. If FSPL at 2.4 GHz is 69.3 dB and two walls add 10 dB, what is the received power with 20 dBm TX and 4 dBi total antenna gain?
- -45.3 dBm
- -55.3 dBm
- -65.3 dBm
- -75.3 dBm
B) – Received power = TX power + antenna gain - path loss = 20 + 4 - 79.3 = -55.3 dBm. This gives a 29.7 dB margin above -85 dBm sensitivity, meaning same-floor corner coverage is robust.
When facing a wireless IoT deployment, avoid the trap of choosing technology based on familiarity or single metrics. Use this systematic framework to make data-driven decisions.
18.9.1 Framework Overview: 7-Step Process
1. Define Requirements
↓
2. Calculate Link Budget
↓
3. Map Requirements to Technology Capabilities
↓
4. Evaluate Trade-offs
↓
5. Calculate Total Cost of Ownership
↓
6. Prototype and Validate
↓
7. Monitor and Iterate18.9.2 Step 1: Define Requirements (Quantitative, Not Qualitative)
Bad: “We need long range and low power” Good: “Coverage across 5 km² with 10-year battery life on CR2032, reporting every 15 minutes”
Requirement Template:
| Category | Metric | Your Value | Units |
|---|---|---|---|
| Coverage | Maximum distance | _____ | meters |
| Coverage | Indoor/outdoor | _____ | indoor / outdoor / mixed |
| Coverage | Obstacles | _____ | walls / floors / metal / foliage |
| Power | Battery type | _____ | CR2032 / AA / rechargeable |
| Power | Target lifetime | _____ | years |
| Data | Payload size | _____ | bytes |
| Data | Update interval | _____ | seconds / minutes / hours |
| Data | Latency tolerance | _____ | ms / seconds |
| Mobility | Device speed | _____ | stationary / walking / vehicle |
| Cost | Module budget | _____ | $ per device |
| Cost | Infrastructure budget | _____ | $ total |
| Regulatory | Deployment region | _____ | EU / US / Global |
18.9.3 Step 2: Calculate Link Budget
Formula:
Received Power (dBm) = TX Power + TX Gain - Path Loss + RX Gain
Link Margin = Received Power - RX Sensitivity
Viable if: Link Margin ≥ Fade Margin (typically 10-20 dB)Example Calculation for 868 MHz at 1 km:
TX Power: +14 dBm (25 mW, typical for 868 MHz)
TX Antenna Gain: +2 dBi (dipole)
Path Loss (FSPL 868 MHz, 1 km): 20log(1) + 20log(868) + 32.45 = 91.2 dB
Environmental Loss: +10 dB (vegetation, non-LOS)
RX Antenna Gain: +2 dBi
RX Sensitivity: -137 dBm (LoRa SF12)
Received Power = 14 + 2 - 91.2 - 10 + 2 = -83.2 dBm
Link Margin = -83.2 - (-137) = 53.8 dB
Excess Margin = 53.8 - 10 = 43.8 dB ✓ EXCELLENTRepeat for all candidate technologies - whichever fails link budget is immediately eliminated.
18.9.4 Step 3: Map Requirements to Technology Capabilities
| Requirement | Zigbee (2.4 GHz) | LoRaWAN (868 MHz) | NB-IoT | LTE-M | Wi-Fi |
|---|---|---|---|---|---|
| Range (outdoor) | 100 m | 2-5 km | 10 km | 10 km | 50 m |
| Data Rate | 250 kbps | 0.3-50 kbps | 250 kbps | 1 Mbps | 100+ Mbps |
| Latency | 50 ms | 1-5 sec | 5-10 sec | 50 ms | 10 ms |
| Battery Life | 5-10 years | 10+ years | 10+ years | 5-10 years | Days |
| Mobility | Stationary/slow | Stationary | Poor | Excellent | Stationary |
| Duty Cycle (EU) | None | 1% (868 MHz) | None | None | None |
| Infrastructure | Mesh/Gateway | Gateway | Carrier | Carrier | AP |
| Module Cost | $3 | $10 | $8 | $12 | $5 |
Elimination matrix:
- Need 2 km range? → Eliminate Zigbee, Wi-Fi
- Need < 1 sec latency? → Eliminate LoRaWAN, NB-IoT
- Need mobility > 50 km/h? → Eliminate NB-IoT, Zigbee
- No infrastructure budget? → Eliminate LoRaWAN, NB-IoT (if no carrier coverage)
18.9.5 Step 4: Evaluate Trade-offs
For each remaining technology, calculate:
A. Power Budget
Example: LoRaWAN sensor transmitting every 15 min
TX: 120 mA for 1 second (SF7) = 0.033 mAh
RX windows: 12 mA for 1 second = 0.003 mAh
Sleep: 2 uA for 899 seconds = 0.0005 mAh
Per cycle: 0.0365 mAh
Per day: 96 cycles × 0.0365 = 3.5 mAh/day
Battery: 2200 mAh (CR2032) / 3.5 = 629 days = 1.7 years
Apply derating:
- 70% voltage efficiency: 1.7 × 0.7 = 1.2 years
- 3% self-discharge: 1.2 / 1.03 = 1.16 years
Is 1.2 years acceptable for your application?B. Network Capacity
Example: 500 sensors in one building
LoRaWAN gateway capacity:
- Uplink: 8 channels × 1% duty cycle (EU) = 28.8 seconds/hour total airtime
- At 1 sec per packet: 28 packets/hour
- For 500 sensors: 500 / 28 = 17.8 hours to collect all readings
- Problem: Cannot support 15-minute intervals!
Zigbee mesh:
- 250 kbps shared
- 100-byte packets: 3.2 ms each
- 500 sensors × 4 packets/hour = 2000 packets/hour
- Airtime: 2000 × 3.2 ms = 6.4 seconds/hour = 0.18% duty cycle
- No problem! Even with overhead, < 1% utilizationC. Coverage Reliability
Test: Can 95% of locations achieve minimum RSSI?
LoRaWAN: Need RSSI > -130 dBm
- Deploy gateway at center
- Simulate coverage (Python tool from Chapter 6)
- Result: 88% coverage (fail)
- Solution: Add 2nd gateway ($500)
Zigbee: Need RSSI > -100 dBm
- Mesh routing extends coverage
- Place 5 mains-powered routers
- Result: 97% coverage (pass)18.9.6 Step 5: Calculate Total Cost of Ownership (7-Year TCO)
Example: 500 Parking Sensors
Option A: LoRaWAN
Modules: 500 × $10 = $5,000
SIM cards: 500 × $5 = $2,500 (if using cellular backhaul)
Gateways: 3 × $500 = $1,500
Installation: $50/gateway × 3 = $150
Gateway backhaul: $30/month × 3 × 12 × 7 = $7,560
RF survey: $5,000 (one-time)
Battery replacement (10-year life): $0
Maintenance: $500/year × 7 = $3,500
Total: $25,210 over 7 years
Cost per sensor: $50.42Option B: Zigbee Mesh
Modules: 500 × $3 = $1,500
Routers: 5 × $30 = $150 (mains-powered, no SIM needed)
Installation: $50/router × 5 = $250
Backhaul: $0 (uses building Ethernet)
RF survey: $1,000 (simpler than LoRaWAN)
Battery replacement (5-year life): 500 × $5 (labor) = $2,500 (at year 5)
Maintenance: $200/year × 7 = $1,400
Total: $6,800 over 7 years
Cost per sensor: $13.60 (73% savings vs LoRaWAN!)Hidden costs to include:
- Gateway maintenance and replacement
- RF survey for optimal placement
- Battery replacement labor (not just battery cost)
- Subscription fees for cellular options
- Firmware update costs (easier with mesh than gateways)
18.9.7 Step 6: Prototype and Validate
Don’t trust calculations alone! Deploy 10-unit pilot:
Week 1-2: Baseline
- Deploy pilot in representative environment
- Measure RSSI at all planned locations
- Record packet loss, latency, battery drain
Week 3-4: Stress Test
- Add interference sources
- Test at peak occupancy times
- Simulate failed nodes/gateways
Week 5-6: Long-term Stability
- Monitor battery voltage weekly
- Check for configuration drift
- Validate link budget matches reality
Go/No-Go Criteria:
| Metric | Target | Measured | Pass/Fail |
|---|---|---|---|
| Coverage (% locations > RX sensitivity) | > 95% | 97% | ✓ |
| Packet loss | < 5% | 3% | ✓ |
| Latency (95th percentile) | < 1 sec | 850 ms | ✓ |
| Battery life (extrapolated) | > 5 years | 6.2 years | ✓ |
| Interference tolerance | No outages > 5 min | 2 min max | ✓ |
If pilot fails: Iterate on design before full deployment
18.9.8 Step 7: Monitor and Iterate (Post-Deployment)
Continuous monitoring dashboard:
Daily metrics:
- Packet loss per sensor (alert if > 10%)
- Battery voltage (alert if < 2.7V on CR2032)
- Gateway/router uptime (alert if < 99%)
- RSSI heatmap (detect coverage degradation)
Weekly analysis:
- Identify sensors with high retry rates
- Check for new interference sources
- Validate link margin still > 10 dB
Monthly optimization:
- Adjust channels if congestion increases
- Fine-tune transmission intervals
- Update firmware if bug fixes available18.9.9 Common Pitfalls and How to Avoid Them
Pitfall 1: “Technology X is always better than Y”
- Reality: Every technology has trade-offs
- Solution: Use this framework to evaluate for YOUR specific requirements
Pitfall 2: “Our requirements are unique”
- Reality: Most IoT deployments fit standard patterns
- Solution: Start with similar case studies, adapt to your specifics
Pitfall 3: “We’ll optimize later”
- Reality: Post-deployment fixes are 10x more expensive
- Solution: Invest in proper design upfront (pilot, link budget, TCO)
Pitfall 4: “Marketing says this is best”
- Reality: Vendor marketing focuses on strengths, hides weaknesses
- Solution: Measure yourself (pilot) and talk to peer companies
Pitfall 5: “Budget is the only factor”
- Reality: Cheap upfront often means expensive long-term
- Solution: Calculate 7-year TCO, not just module cost
18.9.10 Quick Reference: When to Use Each Technology
| Use This | When Your Requirements Include |
|---|---|
| LoRaWAN | Range > 1 km, updates < 1/min, stationary, 10+ year battery, no existing infrastructure |
| NB-IoT | Range > 5 km, updates < 1/min, deep indoor, carrier coverage available, 10+ year battery |
| LTE-M | Range > 5 km, updates > 1/min OR mobility > 50 km/h, carrier coverage, latency < 1 sec |
| Zigbee | Range < 100 m, updates > 1/sec, mesh self-healing, no gateways, 5-10 year battery |
| Thread | Same as Zigbee + need IPv6 end-to-end, Matter compatibility, Apple/Google integration |
| Wi-Fi | Range < 50 m, throughput > 1 Mbps, latency < 100 ms, power available, existing AP infrastructure |
| BLE | Range < 10 m, smartphone interaction, coin cell battery, pairing UX important |
18.9.11 Real-World Decision Example
Requirement: Smart agriculture sensors across 200 hectares (1.4 km × 1.4 km), 200 sensors, 10-year battery, report every 30 minutes
Step 1-3: Requirements → Link Budget → Technology Map
- Range: 1.4 km → Eliminates Zigbee, Wi-Fi, BLE
- Candidates: LoRaWAN, NB-IoT, LTE-M
Step 4: Trade-offs
- LoRaWAN: 1-3 gateways ($1,500), self-operated
- NB-IoT: Carrier coverage uncertain in rural area
- LTE-M: Higher subscription cost, overkill for stationary sensors
Step 5: TCO
- LoRaWAN: $15,000 over 10 years
- NB-IoT: $35,000 (if coverage exists)
- LTE-M: $55,000
Step 6: Pilot
- Deployed 10 LoRaWAN sensors
- Coverage: 98% with 1 gateway centrally placed
- Battery: Projected 12 years
Step 7: Decision → LoRaWAN selected (range + TCO + pilot success)
Key Insight: Systematic evaluation prevented expensive mistakes (NB-IoT would have required carrier build-out; LTE-M was unnecessary overkill).
18.10 Summary
This chapter practiced scenario-based wireless analysis:
Agriculture Deployment:
- Sub-GHz frequencies provide 8-9 dB link budget advantage over 2.4 GHz
- Battery life depends on time-on-air, not payload size
- LoRaWAN or cellular LPWAN are typical solutions for multi-year battery life
Interference Mitigation:
- Wi-Fi channels are 22 MHz wide; Zigbee channels are 2 MHz wide
- Channel selection requires understanding frequency overlap, not just channel numbers
- Move 2.4 GHz traffic to 5 GHz when possible
Indoor Link Budget:
- Floor penetration adds significant loss (15+ dB)
- Multi-floor coverage from single AP is often marginal
- Design with appropriate fade margin (10-20 dB)
General Principles:
- Always calculate before deploying
- Validate with site surveys
- Plan for worst-case scenarios
18.11 Knowledge Check
18.12 What’s Next
| If you want to… | Read this |
|---|---|
| Take the comprehensive mobile quiz | Mobile Wireless: Comprehensive Quiz |
| Review cellular architecture in depth | Cellular Architecture for IoT |
| Study frequency band selection | IoT Wireless Frequency Bands |
| Understand technology comparison | Mobile Wireless Technologies Basics |
| Explore design considerations | Design Considerations & Labs |