A method for controlling one or more lighting devices according to a user request
A hybrid generative AI approach for lighting control uses edge models for quick responses and transitions to cloud models for accurate settings, addressing response time and accuracy challenges in existing technologies.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- SIGNIFY HOLDING BV
- Filing Date
- 2025-12-18
- Publication Date
- 2026-07-09
AI Technical Summary
Existing generative AI models for lighting control face challenges in choosing between cloud-based and edge-based solutions, with cloud models requiring significant resources leading to longer response times and higher costs, while edge models lack accuracy and capabilities.
A hybrid approach combining edge and cloud generative models, where edge models provide fast responses and cloud models offer more accurate settings, transitioning smoothly from initial edge-based control to refined cloud-based settings.
Ensures fast reaction times without sacrificing accuracy by leveraging edge models for initial control and transitioning to more accurate cloud-based settings, enhancing user experience in lighting applications.
Smart Images

Figure EP2025088206_09072026_PF_FP_ABST