IoTGuard: IoT Security in Smart Homes, Production, and Sensitive Infrastructure

The IoTGuard project is dedicated to leveraging advanced Artificial Intelligence (AI) technologies to establish robust, secure, and user-friendly mechanisms for large-scale smart home environments. In collaboration with esteemed industry partners such as Intel Corporation, Kobil GmbH, and IoT-Venture GmbH, along with the Systems Security Lab of TU Darmstadt serving as a university partner, the Secure Software Systems Group is actively engaged in developing a comprehensive framework tailored to enhance the security of IoT devices, with a particular emphasis on, but not confined to, smart home ecosystems.

Despite the considerable excitement surrounding the Internet of Things (IoT), its development remains in its early stages, with numerous challenges requiring attention before it can fully realize its potential. The pervasive diversity in the landscape of IoT devices and the associated heterogeneity present pose the greatest challenges for developers in this domain.

At present, the security of IoT is far from assured, as evidenced by numerous attacks on a spectrum of devices, ranging from smartphones to critical breaches in automotive systems and smart home devices. The IoTGuard project conscientiously investigates existing and potential future attack scenarios. It strives to formulate effective solutions that secure IoT devices and the broader network environments, particularly within smart homes, against adversarial exploits.

The research conducted by the Secure Software Systems Group is particularly concentrated on the security of distributed learning systems, which are integral to a federation of multiple IoT devices collaboratively training AI systems. Additionally, the project delves into the realm of context-based attack and failure detection, placing specific emphasis on text and voice-controlled devices prevalent in smart home settings.

The insights garnered from this initiative, primarily within the context of smart homes, are anticipated to have broader applications across various domains. These domains include critical environments such as smart cities, smart factories, smart grids, and autonomous driving systems, where the transferability of findings from the IoTGuard project promises to enhance overall security and resilience.

People involved: Prof. Dr. Alexandra Dmitrienko, Torsten Krauß, Jasper Stang.