LivingPackets

Sensing the environment, analyzing information, transmitting to the outside world: just like the human body, LivingPackets intelligent packaging performs these 3 functions thanks to :

numerous sensors and triggers to sense, measure and react to the environment
algorithms to analyze and interpret these measurements
interfaces to transmit and communicate information to users

The sensors = the senses

Just like the human body, LivingPackets technology has its own senses to feel and interpret its environment. Here are some of the sensors featured in LivingPackets products:

Indoor and outdoor temperature
Indoor and outdoor humidity
Pressure
Geolokalisierung
Erkennung von Eindringlingen
Shock
Camera

These sensors make it possible to know in which environment the packaging is circulating, what its condition is and in which circumstances the products is being transported are located. In this way, they ensure the full safety of shipments.

To guarantee this security, the packages also have a rigid structure and an electronic locking system. This system guarantees the receiver that the package has not been opened during transport.

Algorithms = the mind

The information gathered by the sensors is precious but must be assimilated and processed to be exploited. To do this, LivingPackets packagings have their own mind and the analogy is complete: the intelligence of our ecosystem is divided into 2 entities, real hemispheres:

The software embedded in the tablet (or edge computing)
The backend stored in the servers (or cloud computing)

As soon as a data processing must be carried out, these 2 entities work and cooperate according to our software architecture designs, which are based on the following basic principles:

All processing that must be done in real time can only be done locally. This is the case for the processing of the raw signal provided by some sensors or for an alert calculation (during a shock for example).

All non-fundamental processing for the proper functioning of the product must be done in the cloud. Indeed, the communication between the product and the cloud server is not 100% guaranteed nor guaranteed to be operated in real time.

The amount of information exchanged between the product and the servers must be limited in order to avoid transferring unnecessary data and limit energy consumption.

The local processing must therefore provide an aggregated view of the package, while the cloud processing is responsible for the aggregated view of the entire fleet. Thus, the packaging itself is able to perform simple calculations from its environment. The cloud can host more complex calculations and, above all, it can use the mass of information from the entire packaging fleet to perform much more complex cross-analyses by exploiting Big Data. The learning done on this mass of data can then be transmitted to all the packages to increase their local intelligence.

The cost of processing is generally lower on the cloud side, which enables mass storage. Complex calculations or those requiring a large amount of data will therefore be performed there.

Finally, the flexibility offered by the cloud allows us to experiment with algorithms and adjust them before implementing them locally once they have been validated.

In our ecosystem, the intelligence embedded in the packaging is responsible for data processing (conversion into physical units, filtering of measurement noise, calculation of hysteresis, etc.) and transmits to the backend the processed and useful physical data, as well as the associated alerts.

Temperature alerts chart — *Temperature alerts*

The intelligence of the backend processes this data to be used on our digital interfaces. For example, in the temperature alert shown above, the embedded intelligence converts the voltage transmitted by the sensor into temperature. Then, it compares this temperature to the minimum and maximum thresholds defined by the user. An alert start event is sent when these thresholds are exceeded (Alert #1 and Alert #3 in our example). An end of alert event is also sent when the temperature returns to the acceptable zone (Alert #2 and Alert #4 in our example). Of course, a hysteresis is set up between the triggering of the alert and the return to normal conditions to avoid triggering multiple alerts in case of low temperature oscillations.

Then, the backend will create a richer object with, in particular, the location of the product at the time of the alert, the maximum or minimum temperature reached, any other alerts raised, etc.

The backend can also be interfaced with external data: logistics flows, transport modes, weather, etc., to enhance a Big Data analysis.

User interfaces = the communication

In the end, these sensors and algorithms should allow us to offer services to our users. LivingPackets interfaces (on the product or by application) communicate with our users to transmit these messages, with several levels of abstraction:

The information alone from the sensor, for example the current temperature inside a product
The formatting of this information, for example the position of the product on a map
Functionalities based on sensor analysis, e.g. alerts on the product and on the digital interfaces
Functionalities based on learning, for example to detect that the product is not on a normal route.

Again, these interfaces are located both on the packaging and on digital media. Locally, the packages have an E Ink screen, interaction buttons and a buzzer. Remotely, LivingPackets offers its customers a web interface and a mobile interface.

Beyond these first functionalities which are already available, the technological ambition of LivingPackets is to provide an ever more intelligent packaging fleet.

The communication and exchange between LivingPackets and the cloud will provide a detailed view of the entire logistic flow and therefore the possibility for everyone to optimize it. A true technological innovation within the logistics ecosystems to offer absolute control over the goods shipped while allowing the reuse of resources and the reduction of logistics waste.

Vincent Le Lay, Chief technical officer, Software Architect at LivingPackets.