Think of a commercial construction site with hundreds of workers, dozens of contractors, and big equipment and materials moving back and forth between site locations.
All activities create valuable data. RFID tags – supply data about the make-up of articles, the location of assets and tools on a GPS, and the output of anything that's connected and capable of sharing information.
But there's a caveat: gathering data doesn't necessarily translate to a smarter construction site.
The hard part is integrating these various data sources. What is difficult is ingesting the data and turning it into a usable form for project teams.
It's at this stage that the role of AIoT in technology and software development becomes truly exciting.
Here's how an AIoT data pipeline might work in a commercial construction setting, and what developers need to keep in mind to build it.
- Understanding the AIoT Architecture
AIoT is a combination of Internet of Things (IoT) and AI Data Analytics.
Construction-site architecture is built in the following layers:
- Device layer: Devices that generate data with RFID tags, BLE beacons, GPS trackers, environmental sensors, etc.
- Connectivity layer: Data is routed by gateways and communication networks.
- Data processing layer: Data is checked, organized, and standardized.
- Analytics layer: Rules and machine learning models recognize patterns and predict potential operational issues.
- Application layer: Construction teams use dashboard and reporting tools to provide valuable information.
All that's crucial is that every single layer has diverse responsibilities.
(For example, an RFID reader reads a tagged object, but another system has to read the ID, match that ID to a record in a database of inventory and make that ID available to the right users.)
- Connecting Different Sensor Technologies
It can also be difficult in construction technology because a variety of devices gather unique data.
Knowing the capacity of each of them before coming up with a combined device is always important.
RFID
Radio Frequency Identification (RFID) employs radio waves for Identification of tagged objects.
It can be useful for:
- Identifying construction materials.
- Recording inventory movement.
- Managing tool allocation.
- Following tagged equipment as it passes designated reader points.
RFID is especially helpful where you need to identify objects even though they don't have to be scanned one by one.
Bluetooth LE
BLE devices communicate using low-energy Bluetooth technology.
In Construction, BLE beacons can support: proximity applications, indoor positioning and visibility of assets.
But, tracking track varies depending on the location of beacons, the environment of the signal, and the tracking framework.
GPS
GPS provides location information using satellite signals.
Generally effective in tracking vehicles, heavy machinery and equipment used outside.
Indoor GPS can be restricted, and certain other positioning technologies might be necessary inside large structures.
Ultra-Wideband (UWB)
UWB employs wideband radio signals for accurate ranging and positioning within suitably designed systems.
It could enable applications with superior location data from indoors, such as monitoring equipment in selected work areas.
Selecting among the two technologies involves a comparison of site environment, accuracy required, costs and requirements for infrastructure and batteries.
- Building a Reliable Data Pipeline
Because connecting the devices is only the beginning. Developers also need a solid way to deal with the data generated by those devices.
Consider a simple equipment-tracking example.
A machine with GPS transmits the location data to a connected platform. The platform receives the event, confirms the event's structure and maps the equipment id to an existing asset record.
A reduced processing flow could be this:
Sensor Gateway Data Ingestion Validation Storage Analytics Dashboard
They all contribute to the robustness of the whole.
Data ingestion
The ingestion layer of the data platform handles information from various communication protocols and hardware.
Depending on where you plan to use this, developers can develop to APIs, or message based on MQTT, or whatever makes sense.
Data validation
Incoming events should be validated before using.
For example:
- Is the device identifier recognized?
- Is the timestamp valid?
- Are required fields present?
- Is the location data in an acceptable format?
- Is the event duplicated?
This monitoring can stop bad data from moving downstream.
Data normalization
Different devices use different formats to display information.
A normalization layer maps in events in such a way that this normalized input becomes easier to process by applications.
For instance, equipment details may be presented with common asset names and timestamps from GPS and BLE systems.
Storage and analytics
Processed events can be stored for operational reporting and historical statistics.
From there the analytics system can detect these trends including: repeated idle time of equipment, unwanted movement or transfer of assets, and inventory availability fluctuation.
The selection of storage technology should be based on data volume, query needs, retention policies and architecture.
- Where Artificial Intelligence Adds Value
Not every IoT application requires machine learning.
Certain aspects of construction workflows may be easily managed using the basic event rules and automated alerts and notifications.
Yet AI is effective if the company requires solving problems on a larger scale of data.
Consider three possible applications.
Analysis of equipment use
Past equipment activity can help identify multiple periods of idle time.
Project managers can also determine if there is a problem in allocating equipment, scheduling, or coordinating the operations.
Material availability assessment
Materials can be cross referenced against delivery information and project timelines.
These forecasts are made based on real time data.
Operational risk identification
Using past project data and operational metrics, an AI model can identify a pattern indicating an imminent delay.
Therefore, these should be aimed at assisting human decision making and not used to replace project managers or safety managers.
- Important Technical Challenges
There are some practical challenges when developing an AIoT solution on a construction site.
Device interoperability
Manufacturers may implement a different protocol, data format and communication method.
Developers shall provide standard device integrations and a standard data model.
Network connectivity
On building sites, the layout can be dynamic, with temporary structures and areas that do not have network coverage.
Systems need to be capable of handling transient failure with buffer, attempts and synchronization.
Data quality
Faulty timestamps, duplicated events and absent sensor readings can skew analysis.
Validation, monitoring and data-quality checks need to be built into the architecture from the outset.
Privacy and security
Workforce-related monitoring requires particular care.
Organizations must decide what data they are collecting, the purpose of that collection, who has access to it and how long they plan on keeping it.
Encryption, authentication, role-based controls, and retention policies can all help safeguard sensitive data.
- Start With a Small AIoT Pilot
Introducing the entire connected construction environment at one time might add needless complexity.
The more realistic way is to start from a single, well-defined operational problem.
Example: Equipment visibility in a specific work zone.
A straightforward pilot may include:
- Prioritizing a small subset of equipment assets.
- Choosing suitable tracking devices.
- Configuring the required connectivity.
- Creating a basic asset identification system.
- Creating a dashboard where the available information will be presented.
- Measuring tracking reliability and equipment visibility.
The post-implementation analysis is then used to determine if the program should be extended.
This will then be easier to evaluate technical limitations and what the real operational benefit is before scaling the system up.
Conclusion
AIoT in commercial construction isn't just about the sensors.
It is an experience of a connected designed system that converts data generated by a device into trustworthy solution insights.
Dev problems seem to be all about device interoperability, ingest, event processing, analytics, cybersecurity, and system integration.
https://commconai .com/
Top comments (0)