A Real-Time Route Prediction-Based Multiobjective Task Allocation for Opportunistic Mobile Crowdsensing

With the widespread use of mobile networks and smart devices, opportunistic mobile crowdsensing (MCS) has emerged as one of the most promising sensing paradigms for intelligent data. In Opportunistic MCS, the real-time mobility of participants and requesters is a crucial feature, as it significantly impacts the quality of MCS services. However, most existing task allocation approaches focus on optimizing the overall system performance while disregarding the mobile attribute of participants and requesters. To remedy this issue, this article proposes a real-time route prediction-based multiobjective task allocation for Opportunistic MCS, called RRP-MOTA, which presents the participants' route-considered task allocation scheme to maximize social welfare comprehensively. Specifically, instead of merely optimizing system performance, a two-stage mechanism is designed to comprehensively enhance task allocation efficiency by estimating and leveraging participants' routes. Moreover, by utilizing participants' spatio-temporal location information, an improved graph convolutional network-based participant route prediction method is developed to provide more accurate participant location information for task allocation. Furthermore, a reference vector-based multiobjective task allocation method is suggested to cater to diverse usage preferences by balancing quality of service and task cost. To validate the performance of our proposed method, extensive simulations are performed on synthetic and real datasets in two scenarios. Experimental results demonstrate that the proposed RRP-MOTA significantly outperforms the chosen existing designs.