Mobile CrowdSensing (MCS) has become a powerful sensing paradigm for information collection recently. As sensing becomes more complicated, it is beneficial to deploy edge servers between users and the cloud center with a so-called mobile edge computing. Instead of directly offloading the sensing data to the cloud center, mobile users offload the sensing data to the edge servers. Then, the edge server processes and transmits the data to the cloud center in a distributed and parallel manner. It’s however critically important to balance cost, such as energy consumption, and the stability of the queues on both mobile users and edge servers. Therefore, to minimize the data offloading cost while maintaining system stability, we should carefully design the sensing data offloading strategy for edge-based crowdsensing.

To solve the problems, a research team led by Dongming Luan published their new research on 15 November 2025 in Frontiers of Computer Science co-published by Higher Education Press and Springer Nature.

The team formulated a double-queue Lyapunov optimization problem and proposed a sensing data offloading strategy. This problem is challenging due to the following reasons. First, since a user can accept several tasks simultaneously and the time for the user to perform tasks is unknown, it is hard to predict the size of the data a user collects. Hence, the traditional deterministic control approaches do not work here. Second, as mentioned above, there is a trade-off between the queue stability and the cost minimization. It is hard to minimize the total offloading costs while maintaining the double-queue stability by taking only one control action. Third, in the multiple data type scenario, the cost minimization problem is subject to more constraints. Hence, it becomes more difficult to make the offloading decision dynamically in each time slot compared with the uniform-data case.

In this paper, the team proposed a double-queue Lyapunov optimization based data offloading approach in the edge-based MCS scenario. The approach minimized the cost without priori knowledge of the data producing rate. The trade-off turned into minimizing the upper bound of drift-plus-penalty term. Furthermore, in the multiple data type scenario, the problem was formulated as a minimum weight bipartite graph complete matching problem and a Kuhn-Munkres algorithm based data offloading approach was proposed.