Gábor Nemes (AX SYSTEM): „AI enables to make forklift traffic safer without complex hardware”

Gábor Nemes, founder of the Hungarian startup AX SYSTEM, discovered Log!Ville during the 2022 Logiconomi event in Antwerp. The concept appealed to him so much that he decided to become a Start2Scale-up partner. In Log!Ville, his company shows its system that, thanks to artificial intelligence, makes forklift traffic in a logistics or industrial environment much safer. He explains how this smart solution works and shares with us his vision of safety in the warehouse and production plants.

How do you describe AX SYSTEM?

The AX SYSTEM is designed to protect pedestrians from stepping in front of a forklift and to safeguard forklift drivers from hitting a pedestrian or colliding with another vehicle. Accidents at crossing points or in the aisles of warehouses mostly occur because people do not see the vehicle or pedestrian coming.

Thanks to Artificial Intelligence, the camera-based system detects and recognizes moving machines and pedestrians, follows their movements, and reacts to potentially dangerous situations or controls traffic. The system relies on what it sees, not on signals from different sensors. This has many advantages: for example, we can determine how it should react to a traffic event, prioritize between pedestrians or forklifts, only signal in actual emergency situations, and so forth.

Depending on design and function, two systems are distinguished: the compact system and the modular system. The compact device, with built-in LED-panels and cameras, is hung above potentially dangerous crossings, passages, gateways and pedestrian routes. The LED panel on each side gives feedback in two colours – green or red - to the traffic participants.

In this video, you can discover how the AX SYSTEM works. And in this one, you can see some practical examples.

How was the idea born?

I was a forklift driver and witnessed several accidents as well as numerous near misses. I thought that we needed a solution that warns about a collision risk in time, thus avoiding emergency braking, which can result in human and cargo injuries. I wanted a system that is not only intelligent but also convenient for the employees, because I also observed that the use of several sensors on the vehicle, the infrastructure or clothes generates too much and sometimes confusing alarms signals. It is also very expensive.

That's how the idea was born of a detection system that would be AI-based and wouldn’t need devices that people have to wear or that have to be installed on forklifts.

When was AX SYSTEM developed?

I started to define the key features of the system in 2016. It took some time to find the proper technology and the team. A big part of the research – which took quite a long time – was related to the implementation of AI. This development was made in collaboration with the University of Györ. The key members from that time are now my co-founders. Some investors also joined in to provide financing, since every key development of the hardware and software was done in-house.

Our first prototype was ready in 2019 and the first system was sold in 2021, in Hungary. Since then, we installed systems in Romania, Slovakia, Italy, Switzerland, UK and Serbia. Currently, we have projects in three additional countries where there is a good chance that customers will choose us. These are Germany, Sweden and Poland.

Our compact device is the most popular because it is very easy to install and can be useful at symmetrical intersections and pedestrian crossings. It is the best choice for most traffic situations. However, clients have all specific requirements and infrastructure. Every solution is thus tailormade.

Are traffic lights or projectors sufficient to prevent human errors?

The answer is simple: they are the best solution thanks to their simplicity. Compare it to a public environment. Would you like to drive in a busy street crisscrossed by many pedestrian paths - some unmarked - with no traffic lights and sometimes questionable rules? Of course, it’s up to you the respect the rules, but traffic lights clearly indicate you what you have to do.

In our recommended solution, we always make sure that signals are easily noticeable and detectable by all traffic participants. Besides clear visual signals, we can also use sound signals to trigger more than one sense. Well-chosen signals are sufficient enough to prevent accidents.

What is the role of AI in AX SYSTEM?

The role of AI is to detect and distinguish objects, be they forklifts, vehicles or persons. Through the images of the cameras, the AX SYSTEM can react to events. For example, if a person and a vehicle are moving towards each other from an unseen direction, we warn them. If, on the other hand, there are just two people walking, then the warning signal is not necessary. My experience is that if a warning signal is triggered many times without reason, we do not pay attention to it after a while, so it loses its function. That's why we thought about developing a system that intelligently responds to traffic situations.

With the help of AI, we can also handle many specific traffic situations. We can determine how the system responds to a variety of traffic events. Other systems are unable to do this.

A few years ago, people weren’t so familiar with the possibilities of artificial intelligence. So it wasn't easy for us to make clients understand that a system can react to an event based on what it sees, like a human, without sensors. Now, they are more aware that AI enables to combine efficiency, accuracy and flexibility.

Couldn’t you use AI to generate automatic stops instead of warning the drivers or pedestrians?

We chose not to, for several reasons. The most important is that automatic speed limiting or braking devices and/or systems have to be installed on every forklift or worn by all the pedestrians. They have to be interconnected, which is quite complex and expensive. Our customers are usually large multinational companies with extensive fleets of machinery, where the cost of installing sensors on every machine would significantly inflate the investment.

One of the advantages of our system is that the investment cost is independent of the size of the machinery fleet or the number of employees working there. And our experience indicates that our customers find clearly perceptible signals sufficient.

Nevertheless, we began in 2021 to develop an onboard unit for forklifts, which would have facilitated communication between our system and the forklift. However, due to the high cost and reliability concerns associated with establishing a wireless connection, we have currently halted this development. Nonetheless, it is planned that our systems will be able to communicate with other manufacturers' UWB systems in the future.

You are a Hungarian company. How did you discover Log!Ville?

We were invited to showcase our solution at the 2022 Logiconomi-event in Antwerp, organized by Toyota Material Handling, as an interesting startup. The event featured not only keynotes and debates, but also an interactive exhibition in Log!Ville, where we displayed our solutions. This opportunity was invaluable for gathering feedback.  

The AX SYSTEM tackles a problem that is global. Feedback from end users is thus very useful, whatever the country they operate in, to improving our product.

What attracted you in Log!Ville?

This is not only an excellent environment to gather feedback, since visitors are end users involved in logistics operations, but also a good place to promote our product and showcase the new developments. It also offers opportunities to exchange information and discuss about innovations.

As I said, the safety problem is worldwide. It doesn’t really matter if the feedback comes from a Belgian user or one from another country. The essence is the quality of the feedback. Face to face contacts are essential in this respect. I wish we could have such a demonstration and inspirational centre in Hungary.

If you had to make one proposal to improve Log!Ville, what would it be?

Work more on the international exposure of Log!Ville. It deserves more international notoriety. It is not known enough, for example, that the interesting content on the website also exists in English. I would therefore promote these articles, for example by posting a lot more in English on LinkedIn.