Prototype design of securing the internet of things in smart homes

The Internet of Things (IoT) technology has revolutionized how businesses operate and changed our daily lives. IoT devices are used in different areas, including smart cities, smart agriculture, smart healthcare, and smart homes. The number of IoT devices connected worldwide continues to rise, and 75 billion devices are expected to be connected by 2025. Even though IoT devices are rapidly spreading, they come with security and privacy challenges. Traditional methods for securing against cyber-attacks are inefficient and inadequate for securing IoT devices. This study aimed to design and implement a secure hub ecosystem prototype with an Intrusion Detection System (IDS), including Machine Learning (ML), to defend IoT devices in a smart home. After the literature about the security of IoT devices in smart homes was analyzed to identify current challenges and limitations, a secure IoT hub ecosystem prototype was implemented. Benign data and malicious data were generated in the IoT testbed. Data was collected from the IoT smart home testbed and implemented using a supervised IDS with ML. The results demonstrate the effectiveness of network segmentation using a hub in mitigating device detection, Denial-of-Service (DoS), and Man-In-The-Middle (MITM) attacks, which were successful in unsegmented networks. Additionally, supervised machine learning classifiers, such as Random Forest and J48, exhibited exceptional performance with precision, recall, and F-measure scores exceeding 97%, highlighting their potential for detecting malicious activities and classifying IoT devices accurately. These findings underscore the importance of combining network segmentation, advanced machine learning algorithms, and user education to strengthen IoT security in smart homes. The results contribute valuable insights to the development of resilient IoT security frameworks.