An Integrated System for Global Tracking

An Exploration of the New EPCIS 2.0 Standard and a Global Internet of Things and the exciting opportunities for integration, analysis and prediction that can be unlocked when the various participants “speak the same language"

July 22, 2022 7 mins. read Andrey Tagarev

The network of global transportation is an incredibly complex system made up of a huge number of moving parts. It includes multinational corporations; tracking of multi-modal globe-spanning delivery routes; details of consignments, shipments and their related parties; details on all kinds of delivery vehicles and their current loads and possibly much more besides that.

In this post, we will examine the opportunities for integration, analysis and prediction that will be unlocked when the various participants “speak the same language.” Specifically, we will look at the new EPCIS 2.0 extension of the GS1 EPCIS standard and how it can enable disparate applications and organizations to exchange and process semantic data in a global transportation context, thus starting a global Logistics Knowledge Graph using Linked Data and semantic technologies.

To illustrate the functionality of a fully-integrated system, we will use a somewhat tongue-in-cheek example. Nonetheless, it reflects the underlying reality of the global supply chain in which materials, new machines and replacement parts can come from anywhere in the world and the routes they take to reach their destination are very different and usually involve several modes of transportation and crossing multiple borders.

Logistics Semantic Master Data

The greatest innovation of EPCIS 2.0 is its tight integration with the GS1 Digital Link and the Web Vocabulary (GS1 Voc) standards. These allow describing product and company details, enabling the identification of certified producers and products meeting certain requirements.

The example will follow a consumer looking to purchase and track two related items: from identifying suppliers to actual delivery. Our fictional consumer has decided to open a tomato juice shop in Germany 30 days from today. To that end, they need to obtain a suitable industrial juicer and a large quantity of tomatoes from reputable sources, and ensure they are delivered in time for the grand opening.

Now that they know what they need, the next step is to discover proper suppliers. Here, we will leverage the integration of GS1 Digital Link and GS1 Voc to express both product type identity (as a Global Trade Item Number) and certification for production of various products.

For the juicer, the search is going to start by identifying a suitable GS1 Global Product Classification (GPC) code and then searching the database for companies certified to produce matching products with their respective GTIN. This leads to the producer’s Party Global Location Number (PGLN). That leads to data on the company itself, including information on how to order the machine and an estimate on how long it will take to be delivered.

For the tomatoes, we follow the exact same process making use of the fact that product information with associated GTIN, company information with associated PGLN and certificate information with associated GDTI are all presented according to the EPCIS standard.

Transportation Steps

After the products have been ordered, the next step is to receive identifiers for each item that will be used to track them while in transit. These identifiers are now a Serialized Global Trade Item Number (SGTIN) for the juicer: essentially the serial number uniquely identifying the specific machine, and a Lot Global Trade Item Number (LGTIN) for the tomatoes: a lot identifier for 200kg of tomatoes produced in the selected farm.

The system then collects the major steps in the transportation process in fundamentally the same manner as in EPCIS 1.0 – the products are put in crates with unique Global Returnable Asset Identifiers (GRAIs). These are represented as aggregation events and the crates are then allocated to containers (GRAI) and to logistic units (SSCC) for this specific delivery represented as an aggregation event. Then the logistic units travel from their initial to their final destination, being loaded on and off different modes of transport and sometimes temporarily aggregated into larger units such as containers that might present further levels of (more temporary) logistic units.

This is where the connections between various data sources become crucial. The SGTIN or LGTIN of the product itself will only appear at the very first and very last step of the process. At any time in between, the question “Where is my item?” requires following multiple steps: aggregating, disaggregating, adding and removing various objects to figure out which logistics identifiers are currently of interest.

Tracking With IoT Sensors

A major extension of EPCIS 2.0 over the earlier versions of the standard is including support for IoT sensors. This allows seamless extension from tracking specific major events during transportation to much finer-grained tracking capability. A sensor can tell us where the package is at any given time as well as record many other aspects of its handling during transportation, such as temperature, acceleration, presence and concentration of chemicals or micro-organisms, etc.

Now, the first step in using the sensor output is to associate it with the object in which it is installed. Presently, for practical reasons (such as sensor size and cost) we can expect them to be limited to the delivery vehicles themselves and larger returnable assets such as containers. The approach to data representation itself, however, can easily scale to a true physical internet, where in the near future each crate or package may have some form of sensor associated with it.

Sensor outputs are represented as the new “sensor_reporting” CBV BizStep. In their very basic application, they include the current location of the shipment and sensor in the epcis:readPoint property (a GPS geo location or GLN place), but they also support any other kind of measurements such as speed, acceleration, heading, temperature, luminosity and so on.

Decision Making Through Real-time ETA

One obvious use of the tracking capability of sensors is, of course, to enable more informed decisions and updated delivery time estimates based on real-time data. To demonstrate, we will consider two worrying events that occur during the delivery of our goods.

The first concerns the juicer’s train journey. As we initially said, the goods need to be delivered within 30 days for the grand opening and the juicer is initially estimated to take 24 days for delivery. However, about ten days into the journey, the train’s GPS sensor brings up a possible issue – the train has stopped at the China-Kazakhstan border in Alshankou for an inordinately long time and that has triggered a warning to be sent to the recipients of all packages about expected delays.

The recipient now has 20 days to track the situation and decide whether any action needs to be taken. Specifically, any delay of more than 5 days would require some action on their end for which they can plan now and actually do 15 days before the deadline thanks to this real-time tracking. Three days into the delay, the train passes customs and reaches Druzhba – the juicer will be delayed by 3 days but will still arrive in time so no action is actually required.

The second concerns the tomatoes’ trip over the Atlantic. The sensor in their container triggers a “serious bump” warning – it has detected a momentary acceleration of over 3g, which might break fragile items. In this case, the ship has encountered a serious storm and rough seas, which might cause a slight delay. The client decides that potentially bruised tomatoes and a day’s delay will not actually affect their plans, so this is not a concern.

Takeaway

In summary, this is a simple example of a fully-integrated system following the EPCIS 2.0 standard. Our aim is to illustrate the basic value to a recipient of tracking and tracing goods in the immensely complex network that is global transportation.

In a way, this initial integration only scratches the surface of what is possible. Once the data is accessible through a uniform standard, there are many opportunities for analytics and predictive systems that estimate times and costs of delivery, suggest optimal routes according to different criteria, do real-time rerouting, calculate environmental impact, issue early warnings to inform recipients of relevant events and so on.

Want to learn how a Logistics Knowledge Graph can hep your business?

 

New call-to-action

 

This work is carried out under the PLANET project that has received funding from the European Union’s Horizon 2020 research and innovation programme under the Grant Agreement No 860274.

Article's content

Researcher at Ontotext

Andrey Tagarev has a MSc degree in Computing Specialism (Machine Learning) from the Imperial College London. He joined Ontotext in 2015 and since then is working on the development of machine learning algorithms for document classification, sentiment analysis, rumor detection and claim identification. Andrey is also specialized in development of pipelines for natural language processing of documents in several European languages – English, French, Dutch.