This paper studies the bundled task assignment problem within the context of mobile crowdsensing. It builds upon the traditional task assignment problem by introducing the concept of task bundles. In this scenario, each requester submits a bundle of two tasks to the platform, which can choose to accept both, accept one, or reject both. The study adopts a disaggregate approach to ensure that the overall quality of workers assigned to each selected task meets a predefined quality threshold. T Additionally, the problem considers both premium and discount pricing strategies for task bundles. The study develops two integer programming formulations for the new version of the problem. The second formulation, in particular, is found to be amenable to solution via column generation, providing near-optimal solutions for problem instances that are larger than those used in the literature. The research contributes to the field of mobile crowdsensing by providing a more realistic variation of the traditional task problem and developing an effective solution method.