The Internet of Things (IoT) growth is extremely fast and it now has found its way to healthcare applications too. Many smart health gadgets and devices are helping practitioners in collecting medical information and monitoring patients. In this distributed system, information or service is sometimes shared and used by other devices. Considering the importance of health-related information and the decisions made based on it, there should be some sort of assurance on the security and quality of the services or information provided. Trust management is an efficient means of promoting application security and reliability in these cases. However, due to some limitations that are specific to IoT, traditional trust evaluation algorithms cannot be employed or do not yield satisfactory results. In this paper, evidence theory is exploited to design a decentralized service-oriented trust management model for healthcare IoT. A measure of evidence distance is used to reward well-behaving healthcare service/information providers as well as referrers and punish malicious entities. In this context-aware model, trust is estimated based on direct experiences and indirect feedbacks of recommenders. The process runs in two contexts; trust to healthcare service and trust to recommendation. When personal direct experience does not exist, trust to a source or service is estimated by applying the combinatorial laws of evidence theory and integrating indirect trust values. The proposed model is secure against bad-mouthing, good-mouthing, and on-off attacks due to its dynamic parameters and using the concept of evidence distance. Our results confirm the robustness and efficiency of this scheme.

中文翻译：

---

智能健康系统安全、去中心化的信任管理方案


物联网 (IoT) 的增长速度非常快，现在也已进入医疗保健应用领域。许多智能健康小工具和设备正在帮助从业者收集医疗信息和监测患者。在这种分布式系统中，信息或服务有时被其他设备共享和使用。考虑到与健康相关的信息以及基于其做出的决策的重要性，应该对所提供的服务或信息的安全性和质量有某种保证。在这些情况下，信任管理是提高应用程序安全性和可靠性的有效手段。然而，由于物联网特有的一些限制，传统的信任评估算法无法使用或无法产生令人满意的结果。本文利用证据理论为医疗物联网设计了一种去中心化的面向服务的信任管理模型。证据距离的衡量标准用于奖励行为良好的医疗服务/信息提供者以及推荐者并惩罚恶意实体。在这种上下文感知模型中，信任是根据推荐者的直接经验和间接反馈来估计的。该流程在两个上下文中运行；对医疗服务的信任和对推荐的信任。当个人直接经验不存在时，对来源或服务的信任可以通过应用证据理论的组合定律并整合间接信任值来估计。由于其动态参数和使用证据距离的概念，所提出的模型可以安全地抵御恶意攻击、恶意攻击和开关攻击。我们的结果证实了该方案的稳健性和效率。