الفهرس 
Data ReliabilityOnly 14 pages are availabe for public view
 |  | from | 231 |  |  |
| | | from | 231 | | |
Abstract
Participatory sensing is an emerging paradigm in which citizens voluntarily use their mobile phones to capture and share sensed data from their surrounding environment in order to monitor and analyze some phenomena. In these applications, no restrictions are usually imposed about participants’ experience, concern, trustworthiness, and interest. Moreover, participants are not usually paid for their participation in sensing campaign. Thus, they usually do not have strong incentives to comply with the tasks’ requirements. That is, they are not concerned about some parameters which may improve the quality of their contributions (e.g. time, location and the position of the device). As a consequence, these applications are vulnerable to erroneous and malicious participants. We define erroneous and malicious participants as those who mislead and disrupt the system measurements by reporting false, corrupted or fabricated contributions either intentionally or non-intentionally. Malicious participants can launch different types of attacks including corruption, collusion, and on-off attack, etc. All these attacks disrupt the collected data and negatively affect the reliability of the data analysis. Different reputation systems have been proposed to monitor participants’ behavior and to estimate their honesty. Reputation is an accumulative score that describe participants’ behaviors. The core objective of these systems is to identify those who provide the system with bad contributions. Subsequently, those bad contributions are excluded in order to improve the reliability of the aggregated data.
Another strategy to improve the reliability of the aggregated data is to increase the number of the correct contributions received. However, participants are usually prohibited to join sensing campaign for different reasons. First, they do not have some strong incentives to join sensing campaign. Oppositely, joining such campaign may generate battery drain as it causes high power consumption. Second, participants are asked to provide their sensed data including time, location, accelerometer, barometric, and biometric data. Such data help attackers to accurately re-identify participants’ original identities. It has been shown in different works that up to ۹٥% of participants identities are re-identified through sharing multi-sensor data. Sharing just four contributions including time and location data is sufficient to reveal such a high ratio of
participants’ identities. Therefore, participants can be physically tracked, hacked, or robbed based on these sensitive data. That is why participants may not be motivated to join sensing campaign unless they have some strong privacy guarantees. As a result, privacy preservation of participants is a major challenge that needs to be fulfilled. Privacy preservation does not only motivate participants to join sensing campaigns but it also leads to higher levels of reliability. Subsequently, the integration between privacy preserving systems and reputation systems is a crucial need for building secure and reliable participatory sensing applications. This integration requires the assurance of seemingly conflicting objectives. Indeed, reputation systems monitor participants’ behaviors along subsequent interactions. One of the major objectives, of privacy preserving systems, is to detach the link between subsequent interactions. The purpose of this thesis is to study the ability of constructing a comprehensive privacy preserving and reputation.