Commercial construction projects entail the generation of operational data relating to work coordination, equipment use, project scheduling, inspections, and subcontractor work. When the same is distributed across disparate sets of information, the involved teams are likely to see an emerging need for greater insight into the state of their jobsite operations.
Artificial Intelligence of Things (AIoT) is an emerging concept of using connected things and artificial intelligence to gather and utilize data.
In this article, we examine what AIoT is, and how one can use an AIoT-driven approach to gain better understanding of the commercial construction jobsite and improve coordination between different elements of the project.
We will use examples specific to the commercial construction domain.
What Is AIoT?
AIoT is a combination of two technologies: the Internet of Things (IoT), and artificial intelligence (AI).
IoT refers to information generated by connected devices, sensors, and monitoring instruments. AI is used in data processing applications, where relevant insights are extracted from a given set of information.
The commercial construction use case involves bringing relevant information streams together in order to form an improved understanding of the construction jobsite, as opposed to using any given set of information in isolation.
- Building a More Connected Jobsite
The construction process may involve a range of different information sets, including general project scheduling information, work coordination and progress bulletins specific to each team, equipment logs and information, and inspections and site conditions.
While each such element represents an information vector, its use is not standardized and the involved parties update their records at different intervals.
A connected jobsite is one where all the relevant data is brought together into a single frame of reference.
It is useful, although not strictly necessary, that the underlying architecture reflects that these are the same underlying facts, just presented from different points of view.
There are several ways in which such a project can be structured, using anything from APIs and system integration tools, through middleware, cloud platforms, and edge devices, to systems-level integration at the level of ERP/MRP systems and other back-end tools.
We will return to this concept later, but it is important to note that any such initiative must begin with a discussion about what data is relevant, who keeps it, and how frequently it should be refreshed.
- Improving Coordination Between Trades
Commercial construction projects are characterizd by the interdependency of different works, with specific tasks being interdependent, or being part of a critical path that determines the overall project timeline. For example, HVAC installation works may be interdependent with electrical works, and both are preceding framing activities. An ability to share progress reports and updates can help the trades keep better visibility on interdependencies, as well as manage their own critical paths.
A deeper ability to process this information and analyze patterns can be useful, so long as the data has sufficient depth and is not misleading at any point – an analytical solution should not be considered an authoritative source of information on the jobsite, but rather a tool for contextualizing events.
Other rules that apply to construction-specific coordination tools apply, including ensuring that project supervisors have the ability to override AI-derived conclusions.
- Designing the Data Pipeline
An AIoT initiative may be summarized through an engineering pipeline, starting with raw data and proceeding through several stages of processing before reaching operational reporting and analytics tools.
In an abbreviated form, this can be expressed as:
Data sources -> Ingestion -> Validation & normalization -> Storage and integration -> Analytics -> Operational reporting
All intermediary stages may raise specific issues and require specific architectural considerations.
Ingestion: sources use different protocols and methods of operation, which may involve varying intervals of refresh, and degrees of precision.
Validation & normalization: timestamps and identifiers such as unique IDs and labels need to be standardized, with a specific definition of what constitutes an event, and methods for managing duplicates, normalization, and disambiguation of source data.
Storage and integration: databases should use appropriate schema and data organization techniques to ensure that relational integrity (project worksites) are respected, and that data that should not be shared remains confidential.
Analytics: rule-based analysis is useful in situations where a known standard must be followed, while machine learning systems should be implemented only when there is sufficient quantity and quality of training data to produce useful results.
Operational reporting: information should be visualized and reported in a manner that is relevant to the decisions being made, rather than presenting raw findings in meaningless detail.
- Security and Data Governance
Projects in the commercial construction space will almost always involve multiple stakeholders, each of whom may require specific degrees of access to the overall information sphere.
An AIoT implementation should address security and governance rules at each stage of the pipeline.
Such measures typically include granular access management and role-based access control (RBAC), end-to-end encryption, secure data lifecycle management, auditing capabilities, and secure onboarding of new elements into the system.
The question of who owns which data, including information on equipment, worksites, and workers, and what confidentiality rules apply, should also be addressed at the stage of implementation.
Finally, where personally identifiable information (PII) is concerned, additional measures may be required in order to reduce the risks of worker data being used for unauthorized purposes.
- Measuring Whether the System Works
When evaluating the utility of a commercial construction technology initiative, it is important to compare the goals of the implementation to its results.
This can include reduced time between updates being pushed to the field, reduced number of invalid or irrelevant entries, reduced time to understand and resolve interdependency issues, and increased amount of relevant data at the disposal of the management team on any given project.
A baseline understanding of the state of affairs prior to the implementation of an AIoT initiative is also necessary to establish whether it has had a positive effect or not.
However, one must always measure utility based on concrete results, rather than theoretical gains.
Conclusion
AIoT represents an opportunity to create a deeper understanding of commercial construction jobsites and the interactions between various works and worksites. It is important to keep the focus on jobssite understanding and not get bogged down in the technical aspects of interconnected data streams.
Some of the considerations related to an AIoT jobsite initiative are likely to apply more broadly, including a focus on data quality, integration, security, visualization, and coordination-specific rules.
The AIoT construction space is rapidly-evolving, and there is still room for innovation and research.
As far as commercial construction goes, "CommCon AI" is a representative example of an AIoT-driven construction space initiative.
The engineering challenge of this sort of endeavor boils down to connecting fragmented pieces of information into accurate and actionable insights.
Top comments (0)