Asset tracking is a location problem, but in practice, the data captured from assets is much more than the location.
A connected asset could be identified, tracked through locations, monitored for environmental conditions, and much more. The key is to bring disparate data points together to support an operational or analytical system.
A quick way to summarize a modern asset tracking stack is to think of it as:
Asset → Identification → Location → Sensor → Connectivity → Data Platform → Analytics → Action
and combine different technologies, such as:
RFID
GPS
BLE
IoT sensors
on the path of transforming physical objects into data.
1. RFID: The Identification Layer
RFID is often used in a scenario where the primary goal is to identify an asset.
An RFID system is made of:
Tags
Readers
Antennas
And a backend database to tie RFID tag data to assets.
In practice, the RFID tag is read by an RFID reader when an asset comes in proximity to it, allowing the reader to tie the RFID tag and its contents to the asset.
This makes RFID ideal for use cases such as:
Warehouses
Manufacturing facilities
Inventory spaces
Supply-chain oriented operations
Asset check-in/out processes
And similar situations where RFID tags and readers can be used to identify or inspect a tagged asset.
RFID, however, does not offer continuous geographic tracking of tagged or identified asset. RFID’s value is often tied to answering the question of what tagged assets were present at a specific location.
2. GPS: The Location Layer
GPS becomes more interesting and valuable when assets are not static and need geographic tracking.
A GPS tracking device can track the current GPS location of the asset and relay the data to some backend system over a cellular or other connectivity.
A simplified view would involve a GPS tracker:
GPS device → Connectivity → Backend Platform
This makes GPS trackers valuable in scenarios such as:
Tracking vehicles
Tracking containers
Tracking equipment
Tracking distributed assets
Tracking assets in outdoor conditions
GPS is relatively ineffective in indoor conditions due to interference, which is why a lot of the asset tracking stacks use a combination of different wireless communication methods. Engineers designing such trackers need to keep in mind that the data model needs to unify multiple data sources.
3. BLE: Indoor and Proximity Location Information
BLE tags can be another source of location information, and in particular, indoor location information.
BLE tags or beacons can be read by BLE gateways or BLE readers, or other BLE tags depending on the configuration, yielding additional sources of location information for assets that are equipped with BLE tags.
Such a system could be a combination of:
BLE tags for indoor location tracking
And GPS trackers for outdoor location tracking
But the engineering complexity comes from ensuring that the data from different systems are normalized properly.
4. IoT Sensors: Asset Conditions
Location is valuable, but a more complete picture of an asset’s state involves sensor data.
Depending on the use-case, sensors can provide information such as:
Temperature
Humidity
Vibrations
Movement
And more.
A connected asset, for instance, could provide a data stream involving:
Asset ID
+-- Location
+-- Temperature
+-- Humidity
+-- Motion
+-- Timestamp
5. Connectivity Is Part of the Stack
The data from the above technologies needs to be sent through some type of connectivity channel to a backend system. Here, the asset tracking system begins to converge with IoT data pipeline and data ingestion systems.
Connectivity options include:
Cellular
Wi-Fi
BLE
RFID
LoRaWAN
Satellite connectivity
And others, depending on the use-case.
The choice of a connectivity protocol for an IoT device is non-trivial and depends on such factors as:
Power
Connectivity range
Bandwidth
Latency
Power cycle requirements
Number of devices
Deployment environment
And much more.
For an asset tracker, the choice of connectivity also determines what data it can collect and how much information can be sent through the channel for processing and analytics.
6. Data Ingestion, Storage, and Processing
Once the data is received by the asset tracking system, the next step is to process it.
A simplified view of an IoT data pipeline could look like:
Physical Asset
↓
Tag / Tracker / Sensor
↓
Connectivity
↓
Gateway or Network
↓
Data Ingestion
↓
Storage / Processing
↓
Analytics
↓
Application / Dashboard
↓
Operational Action
Depending on the implementation, individual components could be replaced, modified, or omitted. For instance, an RFID reader may not require a gateway or network before it reaches the data ingestion step.
The data ingestion component should tie the data to the asset, preserving all important information, including timestamps, locations, sensor data, and any other data relevant to the context of the asset.
It also needs to normalize the data and deal with data quality issues. In cases where timestamps are incorrect, events are out of order, sensor data is missing, connectivity was lost, or the data was incorrectly associated with the asset, such data quality issues need to be addressed by the data ingestion and processing layers before analytics and presentation components consume the data.
7. Data Modeling: Representing Assets Consistently
An asset tracking system has to represent information about a physical asset consistently. This usually involves a data model that can store data about an asset in a consistent manner, enabling downstream systems to process tracking, analytics, and other use cases.
Let’s suppose that an asset has:
An RFID tag
A GPS tracker
A BLE tag
And various sensors:
Then, a good data model would tie this information together into one logical asset record:
Asset
├── RFID ID
├── GPS Device ID
├── BLE Tag ID
├── Sensor IDs
├── Current Location
├── Last Update
└── Historical Telemetry
By doing this, downstream applications can treat various technologies as attributes of a physical asset, enabling more valuable insights and applications.
8. From Analytics to Insights
Telemetry data is not useful until it is analyzed to create meaningful conclusions and insights about the asset.
Analytics can involve detecting patterns, detecting changes in behavior, or detecting anomalies in the data. With this information, an analytics layer could tie asset identification data, GPS location data, timestamps, and sensor data to detect meaningful patterns, such as the movement of an asset and its changing environmental conditions.
An AI-powered analytics layer can also help process such data to extract valuable patterns and insights.
Analytics also have to consider data quality, as discussed earlier. A high-fidelity analytics layer is only as good as the quality of data that it receives as input.
9. Choosing Based on Requirements
A practical guide to choosing what technology to use for asset tracking is to list requirements and map them onto appropriate technologies.
Requirement Technologies to Consider
Asset identification RFID, barcode, QR code
Outdoor location GPS, cellular
Indoor/proximity visibility BLE, Wi-Fi, RTLS
Environmental monitoring IoT sensors
Remote connectivity Cellular, LoRaWAN, satellite
Large-scale telemetry IoT platforms and data pipelines
As discussed in previous sections, each asset tracking project will involve choosing a combination of technologies. This may include RFID for identification, BLE readers for indoor location tracking, and IoTT sensors for environmental data, all for a warehouse use-case. In other scenarios, GPS devices may be paired with cellular connectivity and sensors for vehicle tracking. Or a combination of GPS, cellular, and satellites plus IoT sensors for remote asset tracking.
The choice of architecture depends on the data that needs to be collected about an asset and the operational environment of the asset.
For more information about the different approaches to asset tracking and technologies, including RFID, GPS, and BLE location tracking and IoT monitoring, refer to the Asset Track Pro blog site.
10. The Interesting Engineering Challenge
The interesting engineering challenge about asset tracking is not simply tagging an asset. It is about connecting events in the physical world to events in the digital world. A good asset tracking system needs to address several components:
Identification of the asset: What asset is being tracked?
Location of the asset: Where is the asset located?
Time: What time is the event taking place?
Condition: What is going on with the asset?
Connectivity: How is the information about the event being sent to the asset tracking system?
Processing: How is the event validated and stored?
Analytics: What can be done about the findings?
And finally:
Action: What should the asset tracking system or a human operator do with that information?
A well-architected system connects these components seamlessly, enabling more interesting and impactful use-cases: from simple location tracking of physical assets to creating a full IoT data pipeline connecting physical assets to the digital world.
Conclusion
Modern asset tracking solutions are usually a combination of different technologies. RFID tags can be used to identify an asset, GPS devices can be used to provide location data, BLE tags can be used to provide indoor or proximity location tracking, and IoT sensors can be used to monitor environmental conditions. Connectivity is required to send the data to the backend asset tracking system, and the data is then normalized and stored for processing, analytics, and operational actions.
For developers, the main challenge is ensuring that the technology used can gather the necessary data and that the data is formatted correctly so that it can be consumed by downstream systems.
Top comments (0)