Through KIASH, integrators, installers, electrical engineers and tradesmen will be empowered to offer their customers security monitoring for smart homes.
This means that new value creation processes, including simple digital contracting, can be established without having to have expertise in AI or security themselves.
In this way, targeted manipulation attempts and malfunctions in smart homes can be avoided.
In a nutshell
The KIASH-Security-Box monitors the data streams of the Smart Home for anomalies and has the task to react to them intelligently to avoid drastic consequences for the Smart Home user. The KIASH-Security-Box is integrated in the Smart Home by the craftsman and implemented in a power-saving way. A setup wizard helps with the secure, DSGVO-compliant setup. The KIASH-Security-Box passes into the sovereignty of the user. The craftsman has no access to the processed data without explicit approval by the end user. The user controls the box via a simple app. One task of the cloud platform developed in the project is the management of the anomaly detection algorithm, which is used locally on each KIASH box and in particular employs machine learning methods for detection that constantly improve themselves in a “federated learning” approach.
The KIASH concept has two central, DSGVO-compliant components. The first component is the KIASH Security Box, a system consisting of software and hardware, in parts based on generic KOBIL technology, with secure Internet connection via a smart meter gateway (more precisely: its so-called CLS channel). The second core component is a cloud platform. This is to be operated in the GAIA-X sovereign data infrastructure. This offers all participants throughout Europe data-secure and authorized access and guarantees offers and options for action, especially for small and medium-sized companies. By using federated learning, the AI improvements achieved among users flow into the shared recognition models managed in the cloud, without the need to communicate the smart home data itself to the cloud.
Targeted manipulation attempts and malfunctions in smart homes are increasing due to the progressive digitalization and networking of originally analog measurement and control components. According to an AVAST study, 16.7% of networked German households have at least one insecure device that criminals can use to infiltrate the entire network. In addition to network storage devices, smart household appliances also represent significant vulnerabilities.
Accessing and encrypting sensitive files can cause financial as well as personal damage. KIASH enables integrators, installers, electrical engineers and craftsmen to offer their customers security monitoring for smart homes and thus to establish new value-added processes, including simple digital contracting, without themselves having to have expertise in AI or security. To date, companies have lacked resources such as budget, personnel and expertise. Security and safety are essential factors in the smart home: not only in the detection of (virtual) intrusions, but also in the protection of privacy and the functionality of devices.
16.7% of germ. households
at least one unsafe device
to date, no AI-based
IT security solution
The August-Wilhelm Scheer Institute develops the following components:
Development of AI procedures for anomaly detection
The data basis of the AI-based methods of intentional anomaly detection is the data collected by the KIASH security box per household. For the purpose of data sovereignty of each household, a federated learning approach is followed. The AI models trained per KIASH Security Box are then aggregated into an overall model in the intended cloud platform by the August-Wilhelm Scheer Institute.
Realization of the cloud platform
To ensure scalability, accessibility, data protection and sovereignty, the AI procedures for anomaly detection, and functionalities necessary to connect the KIASH security box, as well as the end-user app, will be realized by the August-Wilhelm Scheer Institute in the form of a GAIA-X compatible cloud platform. For this purpose, the existing, necessary and available GAIA-X functionalities are taken into account to ensure a possible compatibility of the overall solution.
Development of the end user app
The August-Wilhelm Scheer Institute focuses on the user-friendly integration of its components into the end-user app in order to make the developed AI algorithms accessible to all users.
The KIASH project is funded by the German Federal Ministry of Education and Research.
Funding code: 16KIS1614
Duration: 01.07.2022 – 30.06.2025