On AI and Intelligent Lighting Control

Recently, I saw a report that American entrepreneur Elon Musk stated on September 18th during a conversation with Israeli Prime Minister Benjamin Netanyahu that China will be one of the top countries in the field of artificial intelligence (AI) and believes that China has the potential to occupy the top spot in this field; I have been working in the lighting industry for over a decade and want to apply AI to the field of intelligent lighting control as soon as possible.

However, currently, there are some difficulties in the application of AI in the field of intelligent lighting control:

1) Industry digitization level

The digital level of the intelligent lighting industry should be said to be still in its early stages, mainly manifested as: the penetration rate of intelligent lighting in the entire lighting industry is very low: currently around 5%; The adoption ratio of digital dimming technology is relatively low: thyristor and 0-10V are still mainstream; The communication protocols are diverse and dispersed: wired is the main method, while wireless is growing rapidly with its own advantages and disadvantages; The object model is not unified: each manufacturer has its own definition on the edge side and equipment side;

2) User habits

Firstly, there are many issues that users may not adapt to during the transition from manual control to automation. Previously, users have used tape to attach constant illumination sensors because they are not accustomed to automatically adjusting the brightness of the light according to the lighting environment. The development of human-computer interaction has not made scene switching easier, and many users have returned to the primitive switch panel operation. Secondly, users' awareness and acceptance of the energy conservation, health, and comfort brought by intelligent lighting are still at a relatively low level.

3) Cost

The lighting industry has always been an extremely fragmented industry, with even the largest lighting company having a market share of no more than 10%. Many small and medium-sized enterprises find it difficult to concentrate resources on developing large models. From the perspective of input-output ratio, polishing the application of small models and AI toolsets is more efficient.

4) Device side

Although intelligent lighting is a typical AIOT application, the technical system cannot be separated from cloud, pipe, edge, and end. But at present, the edge gateway technology that is most likely to bear the edge computing function is not mature, and the demand for computing power is not very clear. In addition, under competitive pressure, equipment prices are getting lower and non essential functions are gradually being abandoned. It is also necessary for users to clearly perceive the true benefits brought by intelligent lighting and accept high value-added products and solutions.

5) Data security

Data security and privacy are issues of great concern to users. For example, in the process of factory lighting renovation, some users have proposed that competitors can derive production and operation status based on factory lighting data, so the process of data cloud application is cautious and cautious.

So, how can AI be applied in the field of intelligent lighting control?

For large lighting enterprises, if there is sufficient budget, they can fully consider the overall layout more, but they should avoid being greedy and do a good job of calculation. For many small and medium-sized lighting enterprises, they must embrace pragmatism and form true productivity. It is recommended to consider the application of AI from several perspectives:

1) Rethinking the application of cloud, management, edge, and end

To achieve unified management in large-scale projects and cross building projects, priority can be given to cloud platform based models and applications. In small and medium-sized projects, edge computing can be considered. First, the computing power of the edge gateway can be improved to achieve edge computing on the gateway side. Secondly, we can consider to decentralize the edge computing capability to sensors and intelligent drivers to achieve end-to-end computing.

2) Lighting scheme design

Nowadays, whether it is home or commercial lighting, selling single lamps is no longer an effective method. High quality home lighting retail stores have also begun to attract customers by designing overall lighting effects and packaging and selling complete sets of lighting fixtures, such as non main lights. Now there are AI design tools that can quickly output lighting design renderings. After customers are satisfied, they can output equipment lists and quotations. Companies such as Kujiale and Xingchuang have corresponding products to achieve similar goals.

3) Equipment installation

Equip the correct team of technical personnel with the correct tools and materials at the required location and time. Mainly, equipment lists, point maps, wiring diagrams, etc. can be converted from traditional drawing methods to VR three-dimensional methods, providing clear guidance and technical notes for installation personnel in complex projects.

4) System installation and debugging

AI can be used to shorten the configuration time of intelligent lighting scenes. For example, during the scene configuration process, Baidu attempted to use voice based methods to automatically configure the scene, especially in the process of automated and logical configuration, without the need to re-enter the code or manually configure it. When configuring intelligent T8 light tubes in the garage scene, Feifan Shi achieved automatic configuration of scenes and groups solely by the movement of people and cars, greatly simplifying the configuration process.

5) User/Tenant

Artificial control and AI intelligent control, human factor lighting.

AI intelligent assistants can achieve more intelligent voice interaction with users and learn user habits. Human factor lighting requires different illuminance and color temperature outputs based on different times and changes in the surrounding environment, while maintaining excellent spectral and color rendering light quality. AI can use algorithms and models for automatic management in the background, and automatically adjust according to sensors and clocks to achieve more natural switching.

6) Owners and Facility Management

Equipment operation optimization, energy management, and fault prediction.

Utilize AI for platform based device management, such as device fault alarms and accurate indication of fault causes and locations; Energy management is carried out based on the peaks, valleys, and electricity prices of electricity. The optimal lighting energy-saving scenario is configured based on space, time, and sensors. Equipment replacement time is predicted based on equipment lifespan, early warning is given to avoid equipment failures, and asset management is achieved based on positioning.

7) Equipment manufacturer

AI can be used for marketing content, marketing plans, promotional posters, product introductions, internal and external customer training, recruitment robots, and customer service robots to improve content quality and explain logic, and achieve better communication and conversion.

Microsoft has announced that it will directly integrate the newly upgraded Copilot into the entire range of Windows 11, Office 365 products, and Edge. In the future, drawing, photos, as well as Word, Excel, PowerPoint, Outlook, Teams, and more, Microsoft 365 Chat, a new AI assistant, can access email, meetings, chats, documents, and networks to solve complex problems.

Similar AI applications to smart lighting and smart homes in the future should not be far away, and we are prepared.