A Quality-Aware Attribute-Based Filtering Scheme for Participatory Sensing

Abstract

Participatory sensing systems can gather and process personal data to produce some knowledge about the world. In this setting, sensing data quality has been the primary focus of widespread adoption of many systems. To encourage more participants to contribute high-quality data, many incentive mechanisms have been proposed to motivate them. Unfortunately, this brings about some other new issues. Firstly, the application server must have to analyze and process a large-scale of sensing data for the incentives, but most of them might not satisfy the essential data requirements. Moreover, the application server cannot determine whether those important data originate from the legal participants, thus has to check at the expense of efficiency loss. Even worse, since anyone who knows the sensing task can report sensing data, the application server would face the denial-of-service attacks from malicious participants. In this paper, we propose a Quality-aware Attribute-based Filtering (QAFilter) scheme to address these problems. In QAFilter, the application server can distribute a Receiver Filtering Policy along with the sensing task, so that only the participants whose data quality satisfies the policy can report sensing data to the server. Furthermore, in order to maintain the data confidentiality, the participant can encrypt the sensing data under his requirement such that only the receivers who hold the related attributes can decrypt it. In particular, the costs of computation and communication are constant and independent of the number of required data attributes, thus making our scheme more adapt to personal mobile devices and the wireless network environment. Our security analysis and experiment evaluation demonstrate the presented scheme’s security strength and performance efficiency, respectively.