Research On Trust Evaluation Mechanism For Trusted Data Collection In Internet Of Things

The Internet of Things(Io T)utilizes a large number of randomly distributed smart devices to collect data in real-time.With characteristics of universal perception,low cost and intelligent processing,Io T has been widely applied in many fields such as environmental monitoring and intelligent transportation.Trusted data collection is the foundation and prerequisite for constructing high-quality Io T services,and malicious data not only deteriorate service quality,but also lead to wrong decisions,resulting in human and property losses.The trust evaluation mechanism can predict data quality and cooperation probability of devices before purchasing data or establishing cooperation,so as to select trusted devices for data perception and interaction,ensure high-quality services from the source.However,there are still some problems in current trust mechanism,such as the difficulty in obtaining trust evidence,inaccurate evaluation results,and limited application scenarios,which make these methods cannot be well applied to trusted data collection in Io T.To address the above issues,this thesis focuses on trust evaluation mechanism for the trusted data collection in Io T,and conducts a series of research on trust evidence acquisition,trust value calculation,and trust relationship application.First,for the difficulty in obtaining trust evidence and lack of evidence in sparse interaction scenarios,a UAV-assisted active trust evaluation framework is proposed.Then,for the problems of few reference factors and inaccurate evaluation results in trust value calculation,a trust evaluation system based on sequence extraction and evidence reasoning and a blockchain-based multi-tier trust calculation framework are proposed.Finally,for the trust construction and practical application,a trusted task assignment and trust verification strategy based on discrete particle swarm optimization is proposed.Compared with existing methods,the proposed methods have the characteristics of more comprehensive trust evidence,stronger scenario availability,and higher evaluation accuracy.Specifically,the main work and contributions are as follows:(1)To address issues such as insufficient evidence in sparse scenarios and improve the comprehensiveness of trust evidence,a UAV-assisted trust evaluation framework is proposed.First,a global trust evaluation model is proposed for the data interaction between data collectors and data center.By dispatching drones to initiate on-demand and verifiable active evidence collection,it solves problems such as sparse interaction and node collusion.Then,a local trust evaluation model is proposed for the communication interaction between data collectors.It establishes local trust relationships by accumulating interaction feedbacks,thus helping data collectors choose reliable data exchange objects.(2)For the problems with few reference factors and inaccurate results in trust calculation,a trust evaluation system based on sequence extraction and evidence reasoning is proposed.First,it collects more comprehensive trust evidence based on active detection of drones,interactive feedbacks from interacted objects,and recommendations from trusted third parties.Then,based on the sequence type,length,time decay,etc.,the positive,negative and tendentious trust values are assigned to credible,untrusted and uncertain evidence sequences,and the normalized trust is calculated by integrating the three values,so the trust evaluation results are more accurate and objective.(3)For the differentiation trust identification of devices in at different Io T layers,a blockchain-based multi-tier trust computing framework is proposed,which constructs a trust relationship in “data collection-network routing-service construction”.First,a two-tier trust evaluation model is proposed.At data collection layer,it conducts evaluation on data reporters based on data submission and communication interactions.At network layer,it evaluates trust of routers by path backtracking verification,multiservice analysis and coincident path analysis.Finally,a differentiated trust detection is initiated for common and abnormal nodes.(4)For the trust construction and application issue,a trusted task allocation and trust verification strategy based on discrete particle swarm optimization is proposed,which utilizes the remaining time of workers to conduct trust evaluation,and constructs trust relationships with low cost.First,the discrete particle swarm optimization algorithm is improved through particle encoding,initialization,fitness evaluation,velocity and position update,and validity correction to complete the task allocation in the first stage.Then,the remaining time and resources of workers are fully utilized to initiate the second-phase redundant task assignment and trust verification,and combine the two-stage results to establish workers’ trust relationships,thus guiding subsequent trusted task assignments.Extensive theoretical analysis and simulation experiments have fully proved the effectiveness of our methods.These methods can effectively improve the comprehensiveness of evidence acquisition,the accuracy of trust calculation and the availability of trust relationships,and provide important theoretical and technical support for security application of Io T in different fields.