Authors: Mingzhong Wang
Addresses: University of the Sunshine Coast, 90 Sippy Downs Drive, Sippy Downs, Queensland, Australia
Abstract: Mobile crowd sensing becomes a promising solution for massive data collection with the public participation. Besides the challenges of user incentives, and diversified data sources and quality, the requirement of sharing spatial-temporal data drives the privacy concerns of contributors as one of the top priorities in the design and implementation of a sound crowdsourcing platform. In this paper, FollowMe is introduced as a use case of mobile crowd sensing platform to explain possible design guidelines and solutions to address these challenges. The incentive mechanisms are discussed according to both the quantity and quality of users' contributions. Then, a k-anonymity based solution is applied to protect contributors' privacy in both scenarios of trustworthy and untrustworthy crowdsourcers. Thereafter, a reputation-based filtering solution is proposed to detect fake or malicious reports, and finally a density-based clustering algorithm is introduced to find hotspots which can help the prediction of future events. Although FollowMe is designed for a virtual world of the popular mobile game Pokémon Go, the solutions and discussions are supposed to be applicable to more complex applications sharing spatial-temporal data about users.
Keywords: mobile crowd sensing; spatial-temporal data; crowdsourcing; privacy; k-anonymity; hotspot; reputation; incentive mechanism.
International Journal of High Performance Computing and Networking, 2019 Vol.14 No.4, pp.416 - 424
Received: 30 May 2017
Accepted: 02 Mar 2018
Published online: 19 Sep 2019 *