Building an AIoT Architecture for Commercial Construction

In commercial construction, a multitude of connected devices generate data that can provide valuable insights. However, connecting and processing this data to derive actionable intelligence poses a significant challenge. This is where AIoT (Artificial Intelligence of Things) architecture becomes crucial in bridging the gap between physical operations on-site and digital applications.

The proposed AIoT architecture consists of several layers, with devices at the bottom and applications at the top. The left side of the architecture represents the physical world, while the right side depicts the digital twin. The devices can be diverse, ranging from simple on/off switches to complex systems, depending on their intended use. The choice of hardware for devices depends on various factors such as context, range, platform requirements, and existing infrastructure.

Connectivity is essential for transmitting data from devices to the edge layer, databases, or applications. The chosen connectivity method should consider range, platform requirements, existing infrastructure, data size, and format. In large-scale projects, multiple technologies may be employed, and the project's context will determine the most suitable solutions.

Edge processing allows for in-place data manipulation, aggregation, summarization, or transformation before sending it further. This step can help optimize data transmission and reduce latency. Once the data reaches the edge, it needs to be stored, normalized, made accessible to consuming applications, and converted between different formats if necessary. A data layer plays a pivotal role in providing normalization code and facilitating integration between edge and application components.

Analytics is the next step, where data is analyzed to uncover patterns and relationships. This analysis can be applied to various aspects of construction, such as equipment performance, operations modeling, environmental conditions, and more. Artificial Intelligence (AI) can be employed to solve specific problems, but it should be used judiciously, ensuring that the right problem is addressed with the appropriate tool.

Integration is a critical aspect of AIoT architecture in large-scale construction projects. It involves connecting various software solutions, including BIM applications, project management systems, telematics, tracking systems, field applications, and enterprise applications. An AIoT architecture can serve as an intermediary layer between field operations and consuming applications, ensuring seamless data flow and preventing information silos.

Developers play a vital role in the AIoT ecosystem, as they are responsible for integrating systems and processes from numerous sources. Their tasks may include defining what and how devices and events are processed, designing APIs, managing data, handling message processing, and overseeing edge-to-cloud communication. The system must also account for storage, management, analytics, authorization, interoperability, system interaction, reliability, scalability, and other considerations.

When designing an AIoT architecture for commercial construction, it is essential to consider potential challenges and design the system to accommodate them. A laboratory setting, although useful for initial development, does not fully replicate the unpredictable conditions of a construction site. Devices may move, stop functioning, or fail, requiring the system to be resilient and adaptable to real-world constraints.