1 Client Requirements

SECO’s client is a globally recognised supplier of vending machines of various types, looking to deploy new vending concepts. A future business is seen in vending machines for juices or soft drinks in canteens, among other things.

SECO was asked to provide a smart human-machine interface (HMI) to be incorporated into the new vending machine. The desired HMI  requirements are:

• The touch display must be large enough to easily select the desired beverage from various choices.

• The device has sufficient processing capability to run an application which performs the following functions:
  • Render video for the display
  • Accept user inputs via touchscreen
  • Communicate with an electromechanical control board within the vending machine
  • Internet of Things (IoT) functions both internal to the vending machine and communicating with the cloud

• IoT capability which provides insight into vending machine status and leads to improved business operational efficiency
  • Beverage dispense count and other usage metrics
  • Status of the vending machine and its internal components
  • Advertisement placement on the display
  • Remote (including over the air, or OTA) software and configuration updates

• Flexibility in terms of network connectivity: Wi-Fi, cellular, Ethernet

• Communicate to the cloud via a cellular gateway device that may already be nearby the vending machine

• Option to integrate a payment module

• Device and data security

The IoT cloud solution requires the following functions:

• Remote telemetry monitoring, including drink counts, fill levels, and measurements of critical machine component operating parameters

• Clear visualization of telemetry via graphs and colored text

• Automated messaging in case of malfunction or early detection of a maintenance need

• Remotely update vending machines for software and advertising

• Remote management of machines at various levels of granularity, including individually, within a geographic area, for a certain client, or fleet-wide

• Data security

In the future, it must be possible to implement value-added artificial intelligence (AI) applications which may further improve operational efficiency, enhance profitability, or increase revenue. These AI applications may be developed in-house by the client, by SECO or SECO partners, or be obtained as standard applications from an AI app store.

Because the vending machine will be rolled out on a large scale with tens of thousands of machines in the field, the IoT solution must easily scale to reliably accommodate geographically dispersed IoT infrastructure – and further expandability must be guaranteed!

2 Challenges

On the hardware side, the client expects a modern HMI touch solution with crisp performance. The HMI must be easily integrated into the vending machine mechanically, interface with internal circuitry, and have sufficient processing capability to manage all its tasks – all within aggressive unit price targets.

With the new line of self-serve beverage dispensers being deployed near other networked equipment, it became clear that maximum flexibility in terms of cloud connectivity was also required. In addition, with a vending machine fleet numbering in the tens of thousands, the system-wide solution must handle an explosion of data within a complex fleet.

The client already uses an IoT platform for another product line and is considering using its cloud side to manage this new beverage product line as well. IoT software must be implemented within the beverage vending machine’s HMI which communicates with this IoT cloud platform. They are not 100% satisfied with their current IoT solution and would have to make compromises. There is also a risk that the IoT platform will reach a point where it can no longer scale to manage the entire vending fleet. It is currently unclear when this might occur.

Although not specifically mentioned by the client, security software that can cope with the increasing number of cyber-attacks is a must.

3 Our Solution

3.1 Hardware

Thanks to our decades years of expertise delivering highly customised end solutions, we quickly defined the HMI component. Our client places great emphasis on the scalability of its new business and expects this from the custom embedded solution supplied, in terms of subject domain expertise, technical performance and manufacturing capability.

The proposed HMI device is a 15.6-inch display from SECO’s Modular Vision product line, equipped with NXP’s i.MX 8M Plus processor. In addition, the processor performance has enough capacity to provide AI at the edge in the future. The Modular Vision opens the possibility to easily replace the installed Computer-on-Module with an even more powerful processor when future requirements grow. Ensuring maximum flexibility in terms of connectivity, the HMI includes Wi-Fi, Bluetooth, Ethernet, RS-232 and an optional cellular connection. This means that all devices can  connect to the IoT cloud in the best possible way, depending on where they are used and the connections available.. The HMI also has the serial and I/O interfaces required to communicate with other circuitry within the vending machine.

3.2 Clea IoT Software

Clea, which is SECO’s IoT software solution, is created from an open source core and, thanks to its modularity, offers flexibility in its use and can be integrated into any IoT scenario. Especially for newly developed IoT systems, where  future growth is anticipated but uncertain, the solution requires an open structure that can be quickly and easily adapted.

When it comes to building IoT infrastructure, Clea is a complete solution. The big advantage: Clea can easily be used for early development and small scale deployments for manageable, low-budget, and exploratory projects. This established infrastructure can then be easily scaled up to a high count fleet of devices. The open source approach and modular design facilitate implementation of new scenarios and features.

Clea is at its strongest when all the building blocks are combined in such a way that they integrate seamlessly into complex ecosystems when building the IoT platform with Clea.

3.3 Introducing Clea

Clea is cloud agnostic. It is not dependent on any specific cloud host and is therefore freely selectable. This strengthens control over your data that much more. Thanks to the open source structure, Clea can easily be integrated into the client’s own server infrastructure.

Astarte is a simple framework for collecting, orchestrating, and analysing data from any device. Its API openness allows integration with other frameworks. As Astarte handles the data transfer, security features are built in to protect the data.

Edgehog runs in the cloud of the new vending machine IoT infrastructure. All the desired functions of an appliance management system are bundled here. Management ranges from individual appliances and groups to the entire fleet. In addition, the Edgehog module is not necessarily tied to SECO hardware, which is particularly interesting for the customer’s other product line and reinforced the decision.

Portal is the Clea IoT front end, where a user logs in to manage their fleet and access its data. Hosted in the cloud, this is where a user logs into the system, accesses reports, data, and graphs, communicates with remote vending machines, update software, configurations, and advertising, and more. Portal enables multiple levels of fleet granularity, accessing individual, groups, or all vending machines. Additionally, the Clea App Store and StudioX AI applications can be seamlessly integrated into its structure – extracting more value from the data to enhance operations and deploy new revenue streams.

Matching the cloud IoT functionality, Edgehog OS is installed within each beverage vending machine HMI. This operating system integrates the most relevant features of Astarte and Edgehog with a Yocto Linux distribution to form a single, efficient, ready-to-use platform. The vending machine application-level software accesses EdgehogOS and its IoT functions via APIs. Because Astarte utilizes edge-cloud IoT standards, it is able to communicate with the client’s current IoT cloud platform., which is separately configured to communicate with these new vending machines.  Built for cybersecurity, which includes real-time scanning and anomaly detection to halt any form of data theft or cyber-attack, the operating system is maintained with security patches and updatable via the cloud.

With this system architecture, SECO’s client achieves the following advantages:

• Utilize their existing cloud-based IoT platform with the new SECO-supplied HMI enabled for IoT operation.

• Manage the new beverage vending product line on the same platform as other vending product lines

• Utilize their current IoT gateway device with the beverage vending machines, leveraging existing infrastructure and cellular data plans

3.4 Clea as complementary upgrade on existing IoT Infrastructures

While Clea, via its Astarte and Edgehog components being integrated within the Edgehog OS, is deployed within the smart HMI incorporated into the beverage vending machines, the initial cloud platform utilizes our client’s current software solution. This solution has certain limitations, including constraints in its configurabiility to match all the features desired within the beverage vending machines, its device count scalability, and how granular the fleet of devices can be managed. This closed platform cannot be customised and is not able to implement value-added AI algorithms.

To overcome these issues, there is a longer term plan to replace the existing cloud software with Clea. Within this implementation, Clea’s Portal module is used to implement the main user interface. The user interface will provide access to all data (whether collected by the previous platform or newly acquired once Clea is deployed) and machines across the vending fleet – whether the beverage machines that incorporate the SECO HMI or other machines built by other vendors. Value-added services, such as advertising and information display are incorporated, as well as AI-based algorithms that utilize data from across the entire fleet to improve maintenance and operational efficiency.

To communicate with and manage individual vending machines, the Edgehog and Astarte modules are also integrated with the new cloud application – facilitating intercommunication with and operation of the actual vending fleet – whether with SECO or non-SECO devices. None of the vending devices will require a software or hardware upgrade. The SECO devices, including the Modular Vision HMIs, will continue to operate via Edgehog OS. Non-SECO vending IoT devices will continue to operate with their already installed software. Within the Clea cloud implementation, Edgehog and Astarte instances are developed to specifically match the operational, data collection, and communication capabilities of these vending devices. Due to their flexible implementations and open source architecture, developing instantiations to communicate with other vendor hardware is relatively easy, when provided with the IoT capabilities of this hardware.

This deployment of Clea seamlessly replaces the current IoT cloud software. It utilizes all historic IoT data, communicates with all existing vending devices (using both SECO and non-SECO hardware), can easily be adapted to add additional hardware, and has the ability to scale to very large number of devices (potentially milliions).

This complex case shows the versatility of Clea, as it forms the IoT interface for the self-service refreshment stations, and is intended to complement and interface with the existing IoT platform used with the client’s other vending product lines.

4 Deploy Clea as Your IoT Solution – Big or Small

IoT systems like these are not only complex, but also highly fine-tuned to meet verify specific technical and business requirements. SECO regularly implements embedded solutions to be client- and application-specific, based on an established degree of standardisation. The Clea software suite approach follows the same concept as our hardware solutions – based on standards, but can be tuned to specific requirements. There is a good reason we have chosen to make Clea’s structure as open as possible, so that Clea can be implemented anywhere easily and quickly, while unleashing the power of IoT with simplicity.

Thanks to our vast expertise in architecting and deploying small and large scale IoT systems, SECO is a partner of choice for the structured development of IoT infrastructures.

Take the opportunity to discuss your IoT challenges with our IoT experts and learn how Clea will work for you.