Exploring a Hybrid Control Approach for enhanced User Experience of Interactive Lighting

Modern lighting systems allow for light settings that are more in tune with users’ activities, by going beyond mere functional illumination. These systems have a large amount of controllable parameters such as intensity and colour of individual light sources. Using an autonomous control system is therefore an attractive option, especially since such control systems may also lead to reduced energy consumption. From a user experience point of view however, there are certain drawbacks to this automation. This paper proposes a hybrid approach towards lighting control to create a dynamic balance between user control and system automation. Such a hybrid system has the ability to autonomously set the lighting according to its knowledge about the current context, while offering users the possibility to manually adapt the light settings. These manual adaptations can in turn be used by the system to learn about user preferences in various situations, and thereby to improve its future lighting suggestions. To explore and evaluate this approach, a smart lighting system was developed as an initial implementation, and installed in a real office environment. The system employs a machine learning algorithm to achieve intelligent behaviour and provides users with an interface to control the lights and give feedback to the system. In a six-week study, the user experience of this initial implementation is evaluated. The results provide an insight in design considerations when adopting this approach for the design of smart lighting control systems. The considerations regard the type of machine learning, the degrees of freedom offered to the user, the insight in the system’s decision making process, and the user interface.