Designing a Trust-Aware Reputation Model for Enhanced Data Quality in Human-Centric Sensing Systems

Human-Centric Sensing (HCS) systems rely heavily on the participation of individuals using their mobile devices to gather data. Ensuring the quality of the data collected in such systems is critical, as poor or malicious data can significantly degrade system performance. To address this, we propose a trust-aware reputation model that evaluates the quality of user submissions and dynamically adjusts participant reputations. Firstly, our model distinguishes between intentional and unintentional low-quality contributions, applying stricter penalties for deliberate misbehavior while mitigating harsh consequences for accidental errors. Additionally, the model incorporates multi-source feedback, including witness reports, to provide a more accurate assessment of participant behavior. A forgiving mechanism is implemented, allowing participants who have previously misbehaved to regain their reputation through consistent high-quality contributions. Simulation results show that the proposed model effectively reduces the influence of malicious users and improves the overall trustworthiness of the system by promoting high-quality data submissions. Finally, we have compared our approach to the PRBTD method, and the results show that our model offers faster reputation growth for honest users, stronger penalties for malicious users, and better handling of unintentional contributions through adaptive penalties.