KITT– Context‑Information Sharing for IoT Applications

By utilizing such information, applications may at each session be tailored to the user’s needs. The amount of CI is increasing as sensors are becoming cheaper and being integrated into more products. This enables even better services, but can also cause communication channels to become congested by CI-related traffic. It is therefore necessary to sort out and only distribute the relevant CI. KITT will create a fully distributed platform that can handle this and also provide the opportunity to use CI to create seamless services.

Technical Background

Programs and services that change their behaviour depending on the user context are called context-aware applications. This type of applications received a very large impact on the market with the introduction of smart phones. The reason is that smart phones are suitable platforms with high-speed connections, relatively large computing capacity, and the fact that they can convey information from sensors and actuators that are integrated or connected wirelessly.
 
Sensor-based services is not isolated to smart phones, they also come in a variety of other everyday objects that are connected to the Internet. It is through these everyday objects the Internet of Things (IoT) will become a reality. Therefore, this project focus on research into support for services based on the information from sensors and actuators on the IoT.

Context information (CI) is constituted by data from sensors and actuators that relate to the user or the user's environment. For context-aware applications must provide high user satisfaction (Quality of Experience, QoE), it is essential that the CI can be shared and accessed anywhere. The amount of available CI increases as the sensors are becoming cheaper and integrated into more and more products. This allows for even better services, but can also lead to the communication channels becoming overloaded by CI-related traffic. This can become a serious obstacle to the development of sensor-based services. It is therefore necessary to sort, and transfer only the CI that is relevant to users.
 
Another important aspect is the enabling of seamless services. For the service to provide high QoE users must be able to switch communication techniques to achieve the best possible performance at each time. By using CI, conditions can be predicted and hence facilitate communication. It becomes possible to predict the best communication and create effective and seamless transitions, ie. rapid transitions that do not result in interruption of service. The project builds on the results of the project Media Sense to create a fully distributed technical platform that can handle this.

The Project Goals

Project goals can be divided into general objectives and technical goals.

Overall Goals

• New products, jobs and businesses.

Technical Goals

• DCXP to be further developed to become a stable and mature platform for exchange of context information.
• DCXP to be operated as an international open-source projects so that partners and external stakeholders can participate and contribute to development.
• A prototype implementation that can learn to identify new states of entities and share this information with their peers so that they can adapt to the circumstances of the devices they are communicating with.
• The project will deliver two working demonstrators.

Implementation Plan

The project is divided into four work packages (WP) which are managed as sub-projects.
 
WP1 will create the basis for the remaining WPs to build on, improve DCXP, an overlay network for sharing of CI on the Internet developed in Media Sense. The work will create a stable and functioning storage structure for CI, as well as implement Reliable UDP for good scalability.
 
WP2 is based on previous work and focuses on the seamless exchange of information. A specification for CI-access will be created. CI should be possible to connect to various devices and a mechanism to adjust communication on IoT will be developed.
 
WP3 builds on previous work to allow for heterogeneous connections with real-time performance. By anticipating connection conditions, data can be prefetched at the break, or the session can be transferred to other communication technologies. The CI of a device is modelled as a finite state machine and a prototype is created to predict fluctuations.
 
WP4 handles integration, testing and demonstrators. The criteria and scenarios are to be defined for the verification of QoE, demonstrators will be created and empirical studies should be conducted.