Research On Privacy Protection Strategy For Crowdsourcing Task Data In Internet Of Things

As a new data acquisition and processing technology,crowdsourcing technology has the advantages of low cost,high efficiency and wide application scenarios.It plays an important role in the Internet of Things and provides efficient and convenient data acquisition services for various Internet of Things applications.However,while crowdsourcing technology provides convenient data collection services,it also faces many challenges.Among them,the privacy protection of the crowdsourcing task data is one of the most noteworthy privacy protection issues.Aiming at the problem of data privacy leakage of crowdsourcing tasks in the Internet of Things,this paper comprehensively uses blockchain,deep reinforcement learning technologies to conduct research on privacy protection strategies for crowdsourcing task data.The main research work includes:(1)Privacy protection strategy for crowdsourcing task data based on improved blockchain.An improved blockchain-based privacy protection strategy for crowdsouring task data is proposed to address the problems of tampering and deletion of task data privacy and malicious workers conspiring to obtain task data privacy in centralised crowdsourcing platforms.Firstly,the blockchain is used as a distributed crowdsourcing platform to effectively prevent malicious people from tampering with and deleting task data by taking advantage of the untampered and decentralized characteristics of blockchain.Secondly,a new block structure based on privacy level is designed to realize hierarchical management of tasks with different privacy levels.Finally,a sensitive task decomposition and crowdsourcing worker grouping method is proposed to prevent malicious workers from conspiracy to obtain task data privacy.Simulation experiments show that the strategy can effectively enhance the protection of task data privacy and prevent internal collusion attacks during the execution of crowdsourcing tasks.At the same time,it has better performance in terms of transaction latency and throughput.(2)Privacy protection strategy for crowdsourcing task data based on multiple sub-chains.Aiming at the problem that malicious crowdsourcing workers steal task privacy data by applying for tasks,a multi-subchain based crowdsourcing task data privacy protection strategy is designed.Firstly,a multi-subchain architecture is constructed for the hierarchical storage of crowdsourcing tasks.Different levels of crowdsourcing tasks are stored in different levels of task subchains to ensure that only crowdsourcing workers with credit requirements can obtain task data and realize the protection of task data privacy.Secondly,a task allocation strategy based on deep reinforcement learning is proposed,which dynamically allocates crowdsourcing tasks,selects trusted crowdsourcing workers to perform tasks,and realizes the protection of task data privacy by reducing the probability of malicious workers to obtain task data.Finally,a performance improvement strategy based on deep reinforcement learning is designed and implemented to achieve a balance between performance and data privacy protection.The simulation results show that the strategy can effectively prevent malicious workers from acquiring task privacy data and fully ensure data privacy.At the same time,compared with the existing related strategies,the strategy has good throughput and high task allocation accuracy.(3)Privacy protection strategy for crowdsourcing task data based on cross chain authentication.Aiming at the problem that crowdsourcing platform cannot effectively authenticate crowdsourcing workers based on their historical task completion,a data privacy protection strategy for crowdsourcing tasks based on cross-chain authentication is designed.The strategy reduces the number of malicious workers joining crowdsourcing platforms through cross-chain authentication,ensuring that most crowdsourcing workers in crowdsourcing platforms are credible to achieve the goal of protecting data privacy of crowdsourcing tasks.Firstly,a cross-chain architecture supporting cross-chain query is designed.Different crowdsourcing platforms can query the data of crowdsourcing workers' historical task completion through cross-chain architecture,and provide credible data sources for crowdsourcing workers ' certification.Secondly,when crowdsourcing workers join the crowdsourcing platform,smart contracts are used to read the relevant data queried through the cross-chain architecture to determine whether the crowdsourcing workers are eligible to join the crowdsourcing platform,ensuring that the crowdsourcing workers joining the crowdsourcing platform are trustworthy.Simulation experiments show that the strategy can be useful for effectively controlling the number of malicious workers in a crowdsourcing platform and safeguarding task data privacy.At the same time,the strategy has good performance and low resource consumption compared to existing related strategies.