Crowdsensing¶
Collective measurement of a shared phenomenon from situated sensor observations contributed by multiple mobile carriers.
Core Idea¶
Crowdsensing combines sensor observations from multiple mobile carriers into a shared account of a phenomenon of common interest. Each carrier encounters only part of the target, contributes situated readings or detections, and a common system integrates those contributions into a map or event set. Ganti, Ye and Lei's original definition covers people with sensing and computing devices collectively sharing data and extracting information; smartphones are common, but in-vehicle devices also fit the device scope.[^ref-c5d01f6a7de9]
The crowd need not be huge, and user-triggered versus background sampling is a design choice. The constitutive structure is shared target, multiple mobile carriers, situated sensor observations, contribution, and collective integration—not a promise of universal coverage or accuracy.
Scope of Application¶
NoiseTube's 2009 phone/server prototype used microphone, location and optional tags to create a collective noise map from shared measurements. Its authors reported device calibration and indoor-positioning limits and said a larger public experiment was still planned.[^ref-f9d98522c6c6]
Pothole Patrol's seven specially instrumented Boston taxis detected road anomalies with accelerometers and GPS and sent events to a server for clustering. This small vehicular predecessor performs the mobile collective-sensing pattern without proving mass citizen-phone participation or citywide road coverage.[^ref-1c7df74cfca9]
Clarity¶
Many sensors alone are not enough: a fixed network lacks mobile carriers, and many private phone logs lack contribution and joint inference. Nor are crowd opinions the same as instrument readings. A system must say which phenomenon is being measured, what the devices actually observe, and how observations become a shared claim.[^ref-c5d01f6a7de9]
An observed bump is not automatically a pothole, and a phone loudness reading is not automatically a calibrated noise-map value. Validation and coverage determine what the output can support.[ref-f9d98522c6c6][ref-1c7df74cfca9]
Manages Complexity¶
The method partitions observation across devices carried through different routes, then brings their partial readings into a common frame. This can create a usable joint view without placing a dedicated stationary sensor everywhere. Integration compresses many local records into map locations or event candidates, but it can conceal unvisited places, uneven times, device differences and false detections.[ref-f9d98522c6c6][ref-1c7df74cfca9]
Abstract Reasoning¶
Define the shared target and desired area/time. Check which participants or vehicles carry sensors, what signal and context they record, how records are contributed, and how the system combines them. Then compare reported coverage to actual routes and validate the sensor-to-target inference. More contributors cannot by itself correct miscalibration, a biased route pattern or a detector that confuses braking with road damage.[ref-f9d98522c6c6][ref-1c7df74cfca9]
Knowledge Transfer¶
The same five-role structure applies to NoiseTube's sound map and Pothole Patrol's road-anomaly map, although microphones and accelerometers need different validation. The live Data collection supplies the broader observation-acquisition prerequisite; crowdsensing adds mobile distribution and collective synthesis. The term transfers literally only where a shared sensing target and multi-carrier contribution are present, not to any activity involving a crowd.[^ref-c5d01f6a7de9]
[^ref-c5d01f6a7de9]: Raghu K. Ganti, Fan Ye and Hui Lei, “Mobile crowdsensing: Current state and future challenges”, IEEE Communications Magazine 49(11) (2011), 32–39, original IBM Research publication abstract. [^ref-f9d98522c6c6]: Nicolas Maisonneuve et al., “NoiseTube: Measuring and mapping noise pollution with mobile phones”, Information Technologies in Environmental Engineering (2009), 215–228, original paper §§3–7. [^ref-1c7df74cfca9]: Jakob Eriksson et al., “The Pothole Patrol: Using a Mobile Sensor Network for Road Surface Monitoring”, MobiSys (2008), original paper §§1–2.
Relationships to Other Abstractions¶
Current abstraction Crowdsensing Domain-specific
Parents (1) — more general patterns this builds on
-
Crowdsensing presupposes Data collection Prime
Crowdsensing requires recording and contributing observations for a defined shared phenomenon, then adds mobile carriers and collective integration.
Hierarchy path (1) — routes to 1 parentless root
- Crowdsensing → Data collection → Measurement
Neighborhood in Abstraction Space¶
Crowdsensing sits in a sparse region of the domain-specific corpus (85th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (2551 abstractions)
Nearest neighbors
- Event detection for WSN — 0.83
- Kushner–Stratonovich Equation — 0.82
- Wireless triangulation — 0.81
- Map Matching — 0.81
- Heartbeat message — 0.81
Computed from structural-signature embeddings · 2026-10-08