Autonomous spatial mapping and analysis of heterogeneous ambient sensor data in factories

The measurement and analysis of spatially distributed ambient data (e.g. temperature, noise, lighting, or emissions) is an important aspect in factories. Ambient data is directly linked to various factory subsystems and ongoing activities and impacting operational performance, employee wellbeing, environmental outcomes, and associated costs. Due to the limited number of ambient sensor data collection points, interpolation methods are necessary to estimate ambient values at unmeasured locations, thereby generating a complete environmental distribution map. However, existing methods for analyzing spatially distributed ambient data face limitations in terms of availability, scope, and necessary efforts. The paper proposes an autonomous spatial mapping and analysis approach with multiple ambient data sources. It was validated in a factory setting with five ambient sensor data maps, and the evaluation of 12 spatial mapping methods revealed optimal interpolation techniques for different variables. This approach enables innovative manufacturing use cases with reasonable data acquisition efforts.