Privacy-preserving task assignment in mobile crowdsensing: a bilateral location fingerprint-based approach

Abstract Mobile crowdsensing (MCS) leverages the multi-sensory capabilities of mobile devices to collect diverse data efficiently. However, in MCS, inefficient task assignment strategies seriously affect the overall effectiveness, while efficient task assignment frequently requires the collection of sensitive information about users and tasks. In order to effectively trade-off task assignment efficiency and bilateral privacy security, we propose a privacy-preserving task assignment framework based on location fingerprinting. In this investigation, we propose a location fingerprinting-based for privacy preservation mechanism (LFPM) based on Monte Carlo stochastic algorithm to bidirectionally protect the location privacy of workers and tasks. Meanwhile, to overcome the challenge of utilizing location information while protecting privacy, a two-stage task allocation algorithm (TSTA) is proposed. This mechanism facilitates precise task assignment through segmental location fingerprinting, aiming to minimize the total cost of completing all tasks. We theoretically analyze its lightweight design and privacy features. Comparative experiments on real datasets show that this strategy achieves significant improvements in communication efficiency, computational performance, and task assignment accuracy compared to other methods.