On the Design of Adaptive Robotic Systems Using Room Sensors, Anchoring, Semantic, and Low-Code Technologies

Enabling robotic solutions in our everyday lives implies an evolution in the ability of a robot to adapt to its context and how it performs a task there. Camera and sensor data streams, object recognition, semantic data representations, and low-code programming are examples of technology candidates to address the requirements of context adaptation. These technologies provide the means to increase the data available about a specific context of a robot and provide richer abstractions for programming the adaption of tasks to a context, forming a digital twin for this setting. In this paper, we present the design concepts of a robotic system for facility management tasks. We discuss these design concepts that enable the system to adapt to context changes and programmatic adaption of tasks. Based on the development process and several deployments of the robotic system, we summarize the lessons learned for software engineering of the system.