HanKA: Enriched Knowledge Used by an Adaptive Cooking Assistant

Nils Neumann, Sven Wachsmuth

Abstract: Cooking a meal is a challenging and recurring task that requires the consideration of various environmental influences and constraints as well as significant domain knowledge. One way to reduce the complexity is to follow recipes that provide an ordered set of tasks for the preparation of a dish. This concept was already transferred to cooking assistants which present the recipe to the cook while adding further assistance like device control or step-by-step visualization. Although recipes and assistants simplify the cooking process itself, other factors like the available devices or differences in cooking skills are ignored. Aside from that, current assistants are often limited to one recipe at a time, ignoring the regularly occurring requirement to prepare a meal of multiple components. Considering these challenges, we propose our adaptive cooking assistant HanKA that considers the individual user skill and environmental kitchen setup, while assisting the cook at the preparation of their freely combined and synchronized recipes. This is achieved through a modular approach consisting of the automatic detection and control of the available devices and user interfaces, the scheduling of multiple recipes based on the distributed knowledge representation, and a deviation management that considers the user experience. Hereby, we created an adaptive cooking assistant that considers various influences that occur in a cooking scenario, resulting in a better assistance for the user.