Date of Award

Fall 2014

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computing Sciences - (Ph.D.)

Department

Computer Science

First Advisor

Cristian Borcea

Second Advisor

Reza Curtmola

Third Advisor

David Nassimi

Fourth Advisor

Guiling Wang

Fifth Advisor

Nishkam Ravi

Abstract

People-centric sensing with smart phones can be used for large scale sensing of the physical world by leveraging the sensors on the phones. This new type of sensing can be a scalable and cost-effective alternative to deploying static wireless sensor networks for dense sensing coverage across large areas. However, mobile people-centric sensing has two main issues: 1) Data reliability in sensed data and 2) Incentives for participants. To study these issues, this dissertation designs and develops McSense, a mobile crowd sensing system which provides monetary and social incentives to users.

This dissertation proposes and evaluates two protocols for location reliability as a step toward achieving data reliability in sensed data, namely, ILR (Improving Location Reliability) and LINK (Location authentication through Immediate Neighbors Knowledge). ILR is a scheme which improves the location reliability of mobile crowd sensed data with minimal human efforts based on location validation using photo tasks and expanding the trust to nearby data points using periodic Bluetooth scanning. LINK is a location authentication protocol working independent of wireless carriers, in which nearby users help authenticate each other’s location claims using Bluetooth communication. The results of experiments done on Android phones show that the proposed protocols are capable of detecting a significant percentage of the malicious users claiming false location. Furthermore, simulations with the LINK protocol demonstrate that LINK can effectively thwart a number of colluding user attacks.

This dissertation also proposes a mobile sensing game which helps collect crowd sensing data by incentivizing smart phone users to play sensing games on their phones. We design and implement a first person shooter sensing game, “Alien vs. Mobile User”, which employs techniques to attract users to unpopular regions. The user study results show that mobile gaming can be a successful alternative to micro-payments for fast and efficient area coverage in crowd sensing. It is observed that the proposed game design succeeds in achieving good player engagement.

Share

COinS