A High Quality Task Assignment Mechanism in Vehicle-Based Crowdsourcing Using Predictable Mobility Based on Markov

In recent years, crowdsourcing has become a research hotspot. How to formulate a reasonable task allocation mechanism to recruit the most suitable participant for the current perceptual task, and maximize the benefits of the platform has become a problem that most researchers focus on. Great efforts have been invested on task assignment mechanisms from the perspective of the platform or requesters, i.e., quality-sensitive, budget-sensitive, time-sensitive, and location-sensitive. Especially for the location-sensitive task assignment mechanism, many studies motivate users to participate by some coverage estimation methods, i.e., minimizing the traveling distance. Most existing methods statically estimate the distance between the current location of the participant and the task destination, without giving any consideration about the movement track of the participant, which may result in the failure of task for the misallocation. In this paper, we propose a location-sensitive task assignment mechanism using predictable mobility based on Markov model for the vehicle-based crowdsourcing platform. Specially, we present a location transfer prediction model based on Markov model named Markov-TPM by analyzing the positional regularity of task participants during a period of time, to predict where the participant will appear in the next time period first. In addition, we propose a task assignment mechanism based on Markov-TPM that is helpful for the platform to select the most suitable participant to complete the task. Finally, experiments are carried out by using the data set about the taxi trajectory which is collected in Shanghai, and it is shown that the proposed algorithm can improve the accuracy of the task-delivered, which is evidently superior to two algorithms compared, i.e., random prediction algorithm and prediction algorithm based on neighbor relation.