Information support systems for coping with and mastering increasing process complexity are becoming a success factor for many companies. Growing functional scopes increase the demands on the user interface, which in practice are tailored to the users. They are either reduced in functionality or require special training, which is costly and impractical for some user groups. In this context, modern speech interaction interfaces have advantages over purely graphical user interfaces, as they both allow direct access to diverse and complex functions and reduce the cognitive load by using natural language.
The AI-Consult project aims to develop a system that provides the most intuitive access to contextual information possible. This shall be achieved by a natural and low-threshold communication in combination with optical recognition methods by a multifunctional 2D/3D scanner. At the same time, it is intended to provide experienced users with direct, fast and non-contact access to a wide range of functions. The communication system consists of a multimodal interface, in which the acquisition and output of natural language, the graphic representation of data as well as the three-dimensional representation of the conversational partner are fused. Personal image and speech data are to be processed by an integrated computing unit to ensure data privacy.
At its core, an adaptable off-the-shelf overall system is to be developed. First, the system requirements will be determined mainly on the basis of application scenarios in the logistics and construction industries, which differ in terms of user experience, process complexity, and environmental conditions. The applicability of the overall system in these scenarios is then evaluated through a user study. The hardware and software development consists of the development of different AI models and the creation of tools for the management of the training data. The training data will be extended through a data acquisition campaign and the system will be adapted to the application scenarios.
Partner
