Multi-Task Assignment Strategy for Vehicular Crowdsensing with Clustering Characteristic

Recently, with a large number of on-board sensors, vehicles have been widely used for Vehicular Crowdsensing (VCS). Appropriate multi-task assignment strategy is crucial for VCS. However, sensing efficiency and benefit of the system of the existing works need to be improved due to the following challenges. On one hand, many sensing tasks have clustering characteristics in terms of their geographic distribution and sensing requirements, but current works are often ignored. On the other hand, existing multi-task assignment strategies often only design optimization problem for the benefit of one of the platform or participants, and fail to maximize the overall benefit of the system. To remedy that, this paper proposes a multi-task assignment for tasks with clustering characteristics. First, we propose a task combination algorithm, which can greatly improve the task assignment efficiency and reduce the sensing cost. Next, we design a two-stage task assignment scheme, in which the benefits of the platform and participants are optimized respectively in two stages to maximize the benefit of the system. Finally, we carry out extensive simulations, and the simulation results verify the effectiveness of our proposal from clustering validity and sensing cost.