Nondeterministic Evaluation Mechanism for User Recruitment in Mobile Crowd-Sensing

Abstract

Based on the Internet of Behavior (IoB), mobile crowd-sensing (MCS) utilizes the Internet of Things (IoT) to recruit users by analyzing behavioral patterns. MCS is widely used in numerous large-scale and complex monitoring services, but it cannot provide stable and high-quality services due to nondeterministic user mobility and behaviors , which has a vital impact on recruiting high-quality users. In this paper, a stochastic semi-algebraic hybrid system (SSAHS) model is constructed to characterize the user mobility and behaviors of the MCS systems. Based on the definition of probabilistic path and task execution rate, a nondeterministic evaluation mechanism is proposed to measure the nondeterministic user mobility and behaviors, and to give the probability of the user completing the MCS task under the specified time bound and space conditions. The greater the probability is , the higher the quality of the user. Furthermore, a user recruitment scheme based on a nondeterministic evaluation mechanism (NUR) is developed. The NUR uses historical user data to predict user mobility and behaviors ; therefore, high-quality users are recruited to quickly upload reliable sensing data. We conduct simulation experiments based on a real-world user traces dataset , Geolife. The results show that compared with competing recruitment strategies, NUR achieves a higher QoS for the same MCS sensing tasks.