A Trusted Distributed Crowdsourcing Framework Based on User Preferences

As a new computing paradigm to solve large-scale group collaboration problems, crowdsourcing has attracted more and more attention. However, malicious users’ participation in crowdsourcing tasks will affect the completion of crowdsourcing tasks or generate malicious evaluations that are inconsistent with the facts, which will reduce the user satisfaction of ordinary users and even lose their trust in the system. In addition, most of the existing crowdsourcing systems rely on the central server and are vulnerable to a single point of failure, affecting users’ trust in the system. To solve the above problems, this paper proposes a trusted distributed crowdsourcing framework based on user preferences. Firstly, we propose a trust model of identifying malicious users (IMU) based on reputation value, which can quickly identify all kinds of malicious users. Secondly, the framework is based on an open, transparent, and tamperproof consortium blockchain to ensure the security and reliability of transaction information, and has developed a complete service process for it. Finally, this paper also takes into account the different preferences of users, and gives priority to the tasks that best meet users’ preferences to improve user satisfaction. The proposed framework is deployed on the IBM Hyperledger Fabric. The average transaction confirmation time is 1.4424s and the average system throughput is 186tps. The experimental results show that the framework can quickly identify malicious users.