User Willingness-Based participant selection strategy of Crowdsensing

Mobile crowd sensing network has become a new type of perceptual datas collection method, and the quality of datas will have a great impact on the subsequent use of perceptual datas. The quality of datas upload is closely related to the willingness of participants, but the existing participant selection method ignores the user's willingness to a certain extent, which may cause users to quit the task midway and maliciously upload false datas. In order to select the appropriate task participants, this paper combines the actual scenario, considers multiple attributes such as user, task, and surrounding environment, and establishes a regression model based on user willingness using a fully connected deep neural network to quantitatively evaluate the user's willingness to perform the perceptual task. After that, a participant selection strategy that takes into account user willingness and user utility is designed for perceptual scenarios with different requirements. Through simulation experiments, this method is verified to be superior to the baseline method. Its rationality and effectiveness are illustrated by simulating two application scenarios with different degrees of urgency,.