Online Task Allocation Based on Lyapunov Optimization and Fuzzy Control System in Vehicular Mobile Crowdsensing

As a powerful and promising paradigm, vehicular mobile crowdsensing (VMCS) recruits sensing devices on vehicles to collect sensory data, offering advantages such as low cost, wide coverage, and high data quality. Task allocation plays a crucial role in determining the efficiency and performance of vehicular mobile crowdsensing systems. However, existing task allocation methods are mostly designed for offline, single-task crowdsensing scenarios and rarely consider privacy protection, which cannot meet the dynamic and complex requirements of real-world applications. Therefore, we propose an online task allocation and privacy preservation scheme called Lya-FCGA (Lyapunov optimization with Fuzzy Controller-based Genetic Algorithm). To maintain long-term system stability, we utilize Lyapunov optimization to transform long-term objectives into per-time-slot queue control problems. Additionally, we introduce a fuzzy control system to adaptively adjust the genetic algorithm parameters, enhancing the accuracy and efficiency of task allocation. Moreover, differential privacy technology is employed to protect workers' privacy and increase their participation willingness. Extensive experiments on real datasets demonstrate that our scheme reduces sensing costs while ensuring long-term system stability and effectively protecting privacy, showcasing great advantages and potential.