Introducing the Yard Management System (YMS): a Proof of Concept

At Sweet Mustard, we are always looking for ways to help improve the efficiency of our clients’ processes. Recently, we have been working on a proof of concept for a new Yard Management System (YMS). This system could help automate the delivery process for our clients by guiding drivers through their yard. It is built on top of several pieces of software and hardware, including IoT devices, and uses integrations to streamline communication between its components. In this blog post, we will introduce the new system, give a high-level technical overview, and discuss the next steps.

What is the Yard Management System?

The Yard Management System uses IoT and integrations to improve the delivery process at large companies. It automates several time-consuming manual steps and guides truck drivers toward the right delivery area. Its automation and symbol-based signalling also helps to deal with the communication issues caused by a language barrier.

When drivers arrive, they scan a QR code that informs the system of their license plate, load and destination. The YMS uses this info to show truck drivers where they need to deliver their goods. It does this by integrating the cameras, OLED screens, and other pieces of hardware that are used to navigate the yard.

Where did the idea come from?

One of our founders, Mike Seghers, has been wanting to develop something along these lines for a while. He got the idea from a large client, who ran into issues with their delivery system. The large amount of traffic on their yard caused a lot of traffic jams, unintended blockages, and tons of manual administrative work for their employees. Automation could help solve these issues, but the project remained a far-off ambition due to time constraints.

With the What’s NXT exhibition on the horizon on May 14th, we decided to finally tackle this interesting challenge for real. We started two months in advance, and our employees really went above and beyond when developing this solution. We used the Innovation Day initiative to further fine-tune and debug our initial design, so it was ready to be presented on the exhibition.

Who helped develop the Yard Management System?

We approached this project as a company-wide team effort. We used our broad range of expertise and our subsidiaries to ensure a holistic approach.

The backend is hosted entirely in our private cloud using Linode. We used several hardening methods to make sure that the system was iron-clad and production-ready. We developed the frontend in React, together with our colleagues at Octoo. No-wire helped us to develop the backend with their expertise in hardware, machine learning, and IoT integrations.

We kept everyone in this relatively large project team in sync by sending ad-hoc updates. The team also hosted regular stand-up meetings: weekly at first, and daily as the project neared its deadline. We also realised that we had underestimated the time that was needed, especially when it came to testing the hardware. Thanks to our self-sustaining teams and their collegial can-do attitude, everything was concluded successfully.

What is the technical backbone of the YMS?

At Sweet Mustard, one of our pillars is to build an inspiring community. That is why we want to be transparent in how we built the technical backbone of our Yard Management System. We hope to inspire others in the development community to come up with their own related solutions. That being said, the list and image below are just a high-level introduction to how the system works. Fully explaining its intricacies would take many more paragraphs.

The system is made up of the following components:

  • Seven USB cameras, each connected to one of two mini-computers which converts their feed to a RTSP stream by re-encoding it.
  • An Intel NUC PC, which uses a TensorFlow & OpenCV machine learning model to recognise cars and trucks, takes a picture, and sends it to our cloud server.
  • A cloud server, which uses the MQTT messaging protocol to execute the following logic:
  • Analysis of the licence plates in the pictures
  • Calculating GPS coordinates to compare a vehicle’s current location with its destination
  • Communicating with Linux IoT devices placed throughout the yard to control OLED screens and gates
  • A frontend user interface, built in React, which shows the yard in real time.

What are the next steps of the project?

At the moment, the Yard Management System is still a proof of concept. We have already fully implemented the technical backend. Everything in the front- and backend was developed with scalability in mind from the get-go. The system currently handles seven cameras, but is designed to handle hundreds. We also made a fully-functioning scale model completely by hand to demonstrate its potential.

We debuted the model on the What’s NXT exhibition, which has already led to several leads. We are currently in talks to see when and how the YMS could be implemented at our clients’ sites. It has also lead to some interesting suggestions, such as using traffic lights instead of OLED screens, and integrating the system with GPS through a mobile app.

In the meantime, the YMS will serve as a project for our future interns. Because it includes every aspect of our company and our cluster Cronos aan de Leie, and it represents our development philosophy, it is the ideal onboarding project. Stay tuned for more updates on this exciting concept!

Want to know more about the Yard Management System? Interested in using it to automate deliveries at your own organisation? Contact us, and we’ll gladly discuss the possibilities together!