Blockchain technology has recently inspired remarkable attention due to its unique features, such as privacy, accountability, immutability, and anonymity, to name of the few. In contrast, core functionalities of most Internet of Things (IoT) resources make them vulnerable to security threats. The IoT devices, such as smartphones and tablets, have limited capacity in terms of network, computing, and storage, which make them easier for vulnerable threats. Furthermore, a massive amount of data produced by the IoT devices, which is still an open challenge for the existing platforms to process, analyze, and unearth underlying patterns to provide convenience environment. Therefore, a new solution is required to ensure data accountability, improve data privacy and accessibility, and extract hidden patterns and useful knowledge to provide adequate services. In this paper, we present a secure fitness framework that is based on an IoT-enabled blockchain network integrated with machine learning approaches. The proposed framework consists of two modules: a blockchain-based IoT network to provide security and integrity to sensing data as well as an enhanced smart contract enabled relationship and inference engine to discover hidden insights and useful knowledge from IoT and user device network data. The enhanced smart contract aims to support users with a practical application that provides real-time monitoring, control, easy access, and immutable logs of multiple devices that are deployed in several domains. The inference engine module aims to unearth underlying patterns and useful knowledge from IoT environment data, which helps in effective decision making to provide convenient services. For experimental analysis, we implement an intelligent fitness service that is based on an enhanced smart contract enabled relationship and inference engine as a case study where several IoT fitness devices are used to securely acquire user personalized fitness data. Furthermore, a real-time inference engine investigates user personalized data to discover useful knowledge and hidden insights. Based on inference engine knowledge, a recommendation model is developed to recommend a daily and monthly diet, as well as a workout plan for better and improved body shape. The recommendation model aims to facilitate a trainer formulating effective future decisions of trainee’s health in terms of a diet and workout plan. Lastly, for performance analysis, we have used Hyperledger Caliper to access the system performance in terms of latency, throughput, resource utilization, and varying orderer and peers nodes. The analysis results imply that the design architecture is applicable for resource-constrained IoT blockchain platform and it is extensible for different IoT scenarios.

中文翻译：

---

owards基于IoT的区块链网络结合机器学习算法的安全健身框架

区块链技术最近因其独特的功能而引起了人们的极大关注，例如隐私，问责制，不变性和匿名性等。相反，大多数物联网（IoT）资源的核心功能使它们容易受到安全威胁的攻击。IoT设备（例如智能手机和平板电脑）在网络，计算和存储方面的容量有限，这使得它们更容易受到易受攻击的威胁的影响。此外，物联网设备产生的大量数据，对于现有平台进行处理，分析和挖掘底层模式以提供便利的环境，仍然是一个开放的挑战。因此，需要一种新的解决方案来确保数据负责，改善数据隐私和可访问性，并提取隐藏的模式和有用的知识，以提供适当的服务。在本文中，我们提出了一个安全的适应性框架，该框架基于集成了IoT的区块链网络和机器学习方法。拟议的框架包括两个模块：一个基于区块链的IoT网络，可为传感数据提供安全性和完整性；一个增强的智能合约启用的关系和推理引擎，可从IoT和用户设备网络数据中发现隐藏的见解和有用的知识。增强的智能合约旨在通过实际应用程序为用户提供支持，该应用程序提供了对部署在多个域中的多个设备的实时监视，控制，轻松访问和不可变的日志。推理引擎模块旨在从物联网环境数据中挖掘潜在的模式和有用的知识，这有助于做出有效的决策，以提供便捷的服务。为了进行实验分析，我们基于增强的智能合约启用关系和推理引擎实施了一个智能健身服务，该案例研究使用了多个IoT健身设备来安全地获取用户个性化健身数据。此外，实时推理引擎会调查用户的个性化数据，以发现有用的知识和隐藏的见解。基于推理引擎的知识，开发了推荐模型以推荐每日和每月饮食以及用于改善和改善身体形态的锻炼计划。该推荐模型旨在帮助培训师根据饮食和锻炼计划制定受训者健康的有效未来决策。最后，为了进行性能分析，我们已经使用Hyperledger Caliper来访问系统性能，包括延迟，吞吐量，资源利用率以及不同的订购者和对等节点。分析结果表明，该设计架构适用于资源受限的IoT区块链平台，并且可以针对不同的IoT场景进行扩展。