CET-AoTM: Cloud-Edge-Terminal Collaborative Trust Evaluation Scheme for AIoT Networks

With the emergence of 5G (5th Generation mobile communication technology), the integration of AI (Artificial Intelligence) and IoT (Internet of Things) has gained momentum, facilitating the rapid development of AIoT (Artificial Intelligence of Things). Through sensor-enabled data collection, smart terminals are able to analyze, forecast, and make decisions based on data using AI technology. However, smart terminals may inadvertently contribute corrupted and forged data, or malicious terminals may intentionally spread false data, which poses a significant threat to the credibility of AIoT services. Therefore, evaluating the trustworthiness of smart terminals plays a crucial role in ensuring high-quality sensing data and reducing the risk of AIoT. To address this issue, we propose a novel cloud-edge-terminal collaborative AIoT trust model (CET-AoTM). CET-AoTM aggregates the cumulative experience attribute of smart terminals in AIoT and evaluates their credibility by leveraging the collaborative architecture of cloud-edge-terminal. In order to solve the challenge that a large number of new smart terminals lack historical interaction due to the high dynamic nature of AIoT, CET-AoTM evaluates the credibility of the terminals based on the fuzzy attributes such as location attribute, propagation attribute and communication attribute of the smart terminals as a supplement to the trust framework. And a demand-driven cloud-edge-terminal collaboration mechanism is designed to flexibly adapt to different service requirements. The experimental results show that the proposed method has high detection rate under low historical interaction scenario, which is not inferior to popular approaches at prensent.