Incentive Mechanisms In Crowdsensing

With the rapid development of Internet technology,the Internet of Things has gradually emerged,and smart mobile devices have sprung up.These smart mobile devices not only have computing capabilities but also have powerful storage capabilities.A large number of users can have sensing devices and crowd sensing has emerged.Crowd Sensing combines the advantages of crowdsourcing and mobile sensing networks to recruit a large number of users with sensing devices to participate in sensing activities to collect and analyze sensing data.However,users are unwilling to participate in the process of collecting sensing data due to resource constraints and potential privacy leakage risks.Therefore,we need to design an incentive mechanism to appropriately reward users to promote enough users to participate in sensing activities and further improve the quality of sensing data.We mainly studies the multi-task problem and incentive mechanism in the crowd sensing systems.Using the auction theory in game theory,a multi-task incentive method based on sensing quality and a double auction model based on task similarity are designed.On the basis of ensuring the completion of multi-task assignment,proposed method both have the characteristics of individual rationality,budget balance,computational efficiency and truthfulness.And our algorithms also can improve the quality of data collected by users.The specific research content is as follows:(i)In actual multi task application scenarios,the sensing task can be divided into several sub tasks.The similarity between the sub tasks is not considered in the existing methods,and its overall utility needs to be improved.To solve this problem,we propose a double auction model based on task similarity.First,LDA model is used to analyze the task.Then we cluster the sub tasks by cosine similarity,and complete the auction process through the matching after filtering.Then through theoretical analysis and experimental proof,it is proved that the proposed auction mechanism can realize individual rationality,budget balance,computational efficiency and truthfulness.Finally,the algorithm proposed is compared with baseline method.The experimental results show that the incentive mechanism improves the overall effectiveness of platform and providers on the basis of satisfying the preference of providers.(ii)In crowdsensing system,in order to reward providers who provide high-quality data,and further improve the quality of data collection,we propose a multi task incentive mechanism based on sensing quality.We propose a multi-task incentive mechanism based on sensing quality.First,we use the double auction model to establish a model of crowd sensing system.Then,we propose a new evaluation algorithm.The evaluation algorithm not only considers the influence of many factors on the providers to complete the task,but also considers the influence of the relationship between the data quality provided by providers and the task quality requirements on the completion of the task.Through theoretical analysis,it is proved that the proposed auction mechanism can achieve individual rationality,budget balance,computational efficiency and truthfulness.Finally,the proposed algorithm is compared with the baseline method.The experimental results show that when there are more tasks,the platform utility,provider utility,requester utility and social welfare of the proposed algorithm are higher than the baseline method,and the number of matching between high and medium regions is higher than the baseline method.