65  WSN Human-Centric & DTN

In 60 Seconds

WSN Human-Centric & DTN covers the core principles and practical techniques essential for IoT practitioners. Understanding these concepts enables informed design decisions that balance performance, energy efficiency, and scalability in real-world deployments.

MVU — Minimum Viable Understanding

Human-centric wireless sensor networks turn everyday people carrying smartphones into mobile sensing infrastructure, enabling crowdsourced data collection at scales impossible with fixed sensor deployments alone. When continuous connectivity is unavailable – in wildlife reserves, disaster zones, or rural areas – Delay-Tolerant Networks (DTNs) use a store-carry-forward approach where mobile nodes physically transport data across disconnected network segments, trading latency for delivery in environments where traditional networking fails entirely.

65.1 Learning Objectives

By the end of this section, you will be able to:

  • Design Human-Centric Systems: Plan sensing systems where humans serve as targets, operators, or data sources
  • Implement Participatory Sensing: Build crowdsourced data collection applications using smartphones
  • Apply DTN Concepts: Select appropriate Delay-Tolerant Networking protocols for intermittent connectivity scenarios
  • Evaluate Privacy Risks: Assess re-identification threats and implement privacy-preserving techniques for human-centric sensing
  • Assess Crowdsourcing Quality: Measure data reliability and apply multi-report validation mechanisms
  • Design Incentive Mechanisms: Create gamification and reward systems to sustain participation in sensing campaigns

Sammy the Sensor says: “Did you know that YOUR phone is a sensor too? It can measure sounds, light, movement, and even where you are!”

Imagine this: the Sensor Squad needs to find out which streets in town are the bumpiest. But they only have four sensors – that is not enough to cover the whole town!

Lila the LED has a brilliant idea: “What if we ask EVERYONE in town to help? People drive and walk on all the streets every day. Their phones can feel the bumps!”

So the squad creates an app. When people drive over a pothole, their phone’s motion sensor detects the big bump and marks it on a map. Soon, hundreds of people are reporting bumps, and the town has a complete pothole map!

Max the Motor points out another cool trick: “But what about the forest where there’s no phone signal?”

Bella the Button explains: “We can use the ‘bucket brigade’ method! A hiker walks through the forest, and their phone collects data from sensors along the trail. When the hiker gets back to town and finds Wi-Fi, the phone uploads everything. The data hitched a ride!”

That is exactly how a Delay-Tolerant Network works – data rides along with people and animals until it finds a way home!

Fun Fact: The Waze traffic app works exactly like participatory sensing – millions of drivers share their speed and location, so everyone can avoid traffic jams!

What is human-centric sensing? Traditional sensor networks use fixed devices bolted to walls, bridges, or poles. Human-centric sensing flips this – instead of bringing sensors to locations, we use sensors that are already everywhere: the smartphones in people’s pockets. Your phone has a GPS receiver, accelerometer, microphone, barometer, light sensor, and camera, making it a powerful mobile sensing platform.

Think of it like weather reporting. Decades ago, we relied on a few official weather stations. Today, millions of personal weather stations and phone-based apps contribute to a crowdsourced weather picture that is far more detailed.

There are three ways humans participate:

  • As targets: Wearable health sensors monitor the person wearing them (heart rate, steps, sleep)
  • As operators: People actively use their phones to collect data (snapping photos of potholes, measuring noise levels)
  • As passive sources: Apps collect data in the background without requiring action (reporting traffic speed via GPS)

What is a Delay-Tolerant Network (DTN)? Imagine you need to send a letter to a friend in a remote village with no postal service. Instead, you give the letter to a traveling merchant who passes through your town. The merchant carries it for weeks until they reach your friend’s village and deliver it. The letter arrived, but with a long delay.

DTNs work the same way. In places with no continuous internet – deep forests, oceans, disaster zones, or even outer space – data is stored on a mobile node (a person, animal, vehicle, or drone), physically carried to a connected location, and then forwarded. This “store-carry-forward” approach trades speed for reliability. A message might take hours or days, but it gets there.

Real-world example: ZebraNet attached GPS collars to zebras in Kenya. When zebras wandered near a base station, their stored location data was automatically uploaded. No cell towers needed in the savanna!

65.2 Prerequisites

Before diving into this section, you should be familiar with:

  • WSN Overview: Fundamentals: Understanding of wireless sensor network basics, communication patterns, and energy constraints provides foundation for human-centric extensions
  • Wireless Sensor Networks: Knowledge of network topologies, data aggregation, and routing protocols helps understand how human mobility affects network dynamics
  • Networking Basics: Familiarity with networking fundamentals is essential for grasping delay-tolerant networking and intermittent connectivity challenges
  • Sensor Fundamentals and Types: Understanding of sensor capabilities in mobile devices (GPS, accelerometers, cameras) is necessary for participatory sensing applications

65.3 Section Overview

This section covers human-centric wireless sensor networks, where humans play active roles in data collection, and delay-tolerant networking for scenarios where continuous connectivity is unavailable. The content is organized into three focused chapters.

65.3.1 Human-Centric and Participatory Sensing Architecture

The following diagram illustrates how human-centric sensing systems are organized, from individual participants through data collection platforms to actionable analytics:

Architecture diagram showing the four layers of a participatory sensing system: participants with smartphones at the bottom, a data collection layer with quality filtering, a server processing layer for aggregation and analysis, and an analytics output layer producing heatmaps, alerts, and reports at the top.

65.3.2 DTN Store-Carry-Forward Mechanism

Delay-Tolerant Networks use a fundamentally different communication paradigm from traditional networks. Instead of establishing end-to-end paths, DTNs relay data through opportunistic contacts:

Sequence diagram showing the DTN store-carry-forward mechanism: a source sensor stores data locally, transfers it to a mobile carrier (data mule) when they come into contact range, the carrier physically moves to a new location, then forwards the data to a relay node, which eventually delivers it to a base station with internet connectivity.

Chapter Navigation

65.3.3 Human-Centric Sensing: Roles and Paradigms

Topics covered:

  • Human roles as sensing targets, sensor operators, and data sources
  • Participatory vs. opportunistic vs. people-centric sensing paradigms
  • Privacy challenges and the 95% re-identification risk from location data
  • Privacy-preserving mechanisms: differential privacy, k-anonymity, spatial cloaking
  • Worked examples: Urban noise mapping and crowdsourced pothole detection

Key concepts: Human-centric sensing leverages smartphones and wearables to create large-scale sensing systems where humans participate actively or passively in data collection.

65.3.4 Participatory Sensing: Platforms and Applications

Topics covered:

  • Platform architecture: mobile app, data collection, server processing, analytics layers
  • Core principles: democratic data collection, data sharing, authenticity verification
  • Applications: environmental monitoring, urban sensing, health, social sensing
  • Data quality challenges and solutions: sensor heterogeneity, user error, malicious submissions
  • Case studies: NoiseTube (noise mapping) and Waze (traffic sensing)

Key concepts: Participatory sensing enables citizen-driven data collection at scale through crowdsourced mobile applications.

65.3.5 Delay-Tolerant Networks for IoT

Topics covered:

  • DTN characteristics: intermittent connectivity, long delays, resource constraints
  • Store-carry-forward mechanism for disconnected environments
  • Routing protocols: Epidemic (90-95% delivery, ~100x overhead), Spray-and-Wait (80-85% delivery, 6x overhead), PRoPHET (70-80% delivery, 3-5x overhead)
  • Protocol selection guidelines based on criticality and mobility patterns
  • Case studies: ZebraNet (wildlife tracking) and DakNet (rural connectivity)

Key concepts: DTNs enable communication in environments where traditional end-to-end connectivity assumptions fail.

65.4 Comparing Sensing Paradigms and DTN Routing

65.4.1 Sensing Paradigms Comparison

The following diagram compares the three main human-centric sensing paradigms across key dimensions:

Comparison diagram of three sensing paradigms -- participatory, opportunistic, and people-centric -- showing trade-offs in user effort, data quality, privacy risk, coverage, and scalability for each approach.

65.4.2 DTN Routing Protocol Comparison

Decision flowchart for selecting a DTN routing protocol based on application criticality and network resources: Epidemic routing for critical messages with high resources, Spray-and-Wait for moderate resources, and PRoPHET for resource-constrained environments with predictable mobility.

65.5 Knowledge Check

Test your understanding of human-centric sensing and DTN concepts:

Common Pitfalls

  1. Assuming participatory sensing data is reliable by default. Crowdsourced data suffers from sensor heterogeneity (different phone models have different sensor quality), user error (incorrect manual reports), spatial bias (popular areas are over-sampled, remote areas under-sampled), and even malicious submissions. Always implement multi-source cross-validation, outlier detection, and reputation scoring before trusting participatory data for decision-making.

  2. Ignoring the privacy-coverage trade-off. Aggressive privacy protection (coarse spatial cloaking, extreme differential privacy noise) can destroy the spatial resolution that makes participatory sensing valuable. Conversely, collecting fine-grained location data enables 95% re-identification from just 4 data points. Design privacy mechanisms that match the application’s actual spatial resolution needs rather than applying one-size-fits-all settings.

  3. Using Epidemic routing when resources are constrained. Epidemic routing achieves the highest delivery ratio (90-95%) but at enormous cost: every node stores and forwards every message to every contact, creating O(N) message copies and draining both storage and battery. In resource-constrained DTN deployments (wildlife collars, remote sensors), use Spray-and-Wait or PRoPHET instead.

  4. Treating DTN as a replacement for connected networking. DTNs are a last resort, not a preferred architecture. Message latency ranges from hours to days, and delivery is probabilistic, not guaranteed. If any form of continuous or periodic connectivity is available (even low-bandwidth cellular or satellite), it will almost always outperform DTN in latency, reliability, and simplicity. Reserve DTN for genuinely disconnected environments.

  5. Forgetting incentive design in participatory sensing. Without proper incentives – gamification, micropayments, community recognition, or altruistic framing – participation rates drop to near zero within weeks. Building the sensing platform without a sustainable incentive mechanism is the most common reason participatory sensing projects fail after their initial pilot phase.

65.6 Summary

Human-centric sensing and delay-tolerant networking extend traditional WSN capabilities to leverage human mobility and cope with intermittent connectivity:

Concept Key Principle Trade-off
Participatory Sensing Active human data collection via smartphones High quality but requires sustained motivation
Opportunistic Sensing Background collection with no user effort Broad coverage but higher privacy risk
People-Centric Sensing Mining social media and public posts Massive scale but unstructured, noisy data
Epidemic Routing Flood messages to all contacts 90-95% delivery but ~100x overhead
Spray-and-Wait Distribute limited copies (L) 80-85% delivery with 6x overhead
PRoPHET Forward based on encounter history 70-80% delivery with 3-5x overhead
Privacy Protection Spatial cloaking, k-anonymity, differential privacy Stronger privacy reduces data resolution

Key takeaways:

  • Human Roles: Humans serve as sensing targets (health monitoring), sensor operators (crowdsourced data collection), or passive data sources (social media) in modern sensing systems
  • Participatory Sensing: Democratic data collection paradigm where individuals use mobile devices to actively contribute georeferenced data, enabling community-driven environmental and urban monitoring
  • Opportunistic Sensing: Automatic background data collection from smartphones with minimal user intervention, trading higher coverage for potential privacy concerns
  • Privacy Mechanisms: Spatial cloaking, differential privacy, k-anonymity, and edge processing protect personal information while enabling valuable aggregate insights – but 4 spatio-temporal points can re-identify 95% of people
  • Delay Tolerant Networks: Networking paradigm designed for intermittent connectivity where traditional end-to-end assumptions fail, enabling communication through store-carry-forward and opportunistic contacts
  • DTN Routing Protocols: Epidemic routing maximizes delivery through flooding, Spray-and-Wait limits copies for efficiency, and PRoPHET uses encounter history to make probabilistic forwarding decisions
  • Incentive Design: Sustainable participation requires gamification, micropayments, or community recognition – without incentives, most participatory sensing projects fail after the pilot phase

65.7 Worked Example: Incentive Budget for a City-Wide Air Quality Campaign

A city health department wants to measure PM2.5 air quality at street level across 200 km of roads using citizen smartphones with low-cost particulate sensors ($12 clip-on units). They need 500 active participants daily for 6 months.

Three Incentive Models Compared

Model Per-User Cost 6-Month Budget Retention Rate Data Quality
Fixed payment ($0.50/day) $0.50/day 500 x $0.50 x 180 = $45,000 60% (many quit after 2 weeks) Low (users game system with fake readings)
Quality-weighted ($0.10 base + $0.50 bonus for validated readings) $0.10-$0.60/day $32,000 45% (only motivated users persist) High (GPS + timestamp validation filters spoofs)
Gamification + lottery (leaderboard + $10 monthly gift card for top 50) $0.06/day 50 x $10 x 6 = $5,400 70% (social competition sustains participation) Medium-High (leaderboard rewards consistency)

Decision: Gamification costs 88% less than fixed payments ($5,400 vs $45,000) while achieving higher retention. Intrinsic motivation (seeing your neighborhood’s air quality data, competing with friends) outperforms cash payments for sustained participation.

Total Campaign Cost: $5,400 (incentives) + $7,200 (sensors) + $4,000 (data validation backend) + $2,400 (cloud) = $19,000 for 6 months of city-wide data. Equivalent professional monitoring stations cost $15,000-50,000 each; the city would need 40+ stations for similar spatial coverage – a $600,000+ investment. Participatory sensing achieves 97% cost reduction through human mobility as infrastructure.

Participatory Sensing vs Fixed Infrastructure Cost Comparison: For 200 km of city streets requiring air quality coverage with ±50m resolution:

Professional fixed stations: \(N = \frac{200{,}000 \text{ m}}{50 \text{ m}} = 4{,}000 \text{ monitoring points}\)

With overlapping coverage (each station monitors 100m radius), need ~40 stations:

\[\text{Cost}_{fixed} = 40 \times \$30{,}000 = \$1{,}200{,}000\] (capital) \[\text{Cost}_{maintenance} = 40 \times \$2{,}000/\text{year} = \$80{,}000/\text{year}\]

Participatory sensing with 500 volunteers carrying $12 clip-on sensors:

\[\text{Cost}_{participatory} = 500 \times \$12 + \$5{,}400 + \$4{,}000 + \$2{,}400 = \$19{,}000\] (6 months)

ROI calculation: Fixed stations provide 40 points of coverage for $1.2M. Participatory provides 500 mobile sensors covering all 200km for $19K = 63× cost reduction. The human mobility network provides 12.5× denser spatial coverage at 1/63rd the cost.

65.8 What’s Next

Topic Chapter Description
Human-Centric Sensing Human-Centric Sensing Three human roles and sensing paradigms (participatory, opportunistic, people-centric)
Participatory Sensing Participatory Sensing Platform architecture, FixMyStreet, and data quality validation
Delay-Tolerant Networks DTN for IoT Epidemic, Spray-and-Wait, PRoPHET routing for disconnected environments

Deep Dives:

Protocols:

Security & Privacy:

Learning Hubs: