CrowdWatch: enabling in-network crowd-sourcing

  • Authors:
  • Robin Kravets;Hilfi Alkaff;Andrew Campbell;Karrie Karahalios;Klara Nahrstedt

  • Affiliations:
  • University of Illinois, Urbana, IL, USA;University of Illinois, Urbana, IL, USA;Dartmouth College, Hanover, NH, USA;University of Illinois, Urbana, IL, USA;University of Illinois, Urbana, IL, USA

  • Venue:
  • Proceedings of the second ACM SIGCOMM workshop on Mobile cloud computing
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Proliferation of mobile smartphones has opened up possibilities of using crowd-sourcing to gather data from and so monitor large crowds. However, depending on the size of the crowd, current solutions either put unpredictable stress on the infrastructure and energy-constrained smartphones or do not capture the crowd behavior accurately. In response, we present CrowdWatch, a scalable, distributed and energy-efficient crowd-sourcing framework. CrowdWatch achieves its goal through off-loading some of the processing to the devices and establishing a hierarchy of participants by exploiting devices with multiple radios (i.e. WiFi (high-power) and BlueTooth (low-power)). CrowdWatch can outperform traditional crowd-sourcing frameworks by reducing the stress on the infrastructures to 10% of that of a traditional crowd-sourcing solution, while only requiring each phone to use their Wi-Fi radios 15% of the time in a dense environment.