Steam engines powered the first industrial revolution, electricity enabled the second, and the third revolution was digital. Today, we are at the verge of the fourth Industrial Revolution. What are the technologies that are driving the factories of tomorrow? In this post, we’ll discuss the key innovations and how they can impact your business processes today.

Neural Networks & Deep Learning

Artificial Intelligence (AI) isn’t what it used to be. Before the past decade, AI was, in all honesty, a glorified set of if-else statements or statistical rules. However, in the 2010s a new approach changed the AI game entirely: neural networks, or deep learning. Loosely simulating the workings of neurons in the human brain, artificial neural networks are designed to recognise patterns from training data. Adding more layers to a neural network makes them ‘deeper’, enabling them to tackle more complex patterns. These deep-neural networks form the basis of deep learning.

Artificial Intelligence has evolved from if-else statements
Artificial Intelligence (AI) has far evolved from a set of if-else statements or statistical rules. 

Since then, deep learning has made the impossible possible: it has taken the driver’s seat in fully autonomous self-driving cars, it can translate up to more than 100 languages in human-like quality and below and behold: it even extends the battery life in your mobile phone. AI has made our machines smarter than ever.

But deep learning is so much more than a mere academic curiosity. It has already proven its utility for businesses. Today’s smart systems, robots, and applications are fully capable of automating complex operations on the work floor. At Robovision, we have already deployed AI-enabled applications that can autonomously spot material defects, sort different foods, plant seeds or recognise the items in your shopping basket. Furthermore, the hardware, connectivity and applications that underpin AI have also drastically improved thanks to concurrent advances in 5G, chip technology and the Internet of Things (IoT). It should come as no surprise that AI investments are skyrocketing across sectors while even nations are pursuing AI as a core strategic focus. The potential is just that great.

Robovision's AI-enabled defect detection system used in manufacturing production process
An example of Robovision’s AI-enabled application that can autonomously spot defects using deep learning. 

5G Cellular Technology

While 5G may sound like just a slight improvement over its predecessor 4G, in reality it is a total game-changer. Think of it this way: 4G makes it possible to stream your Netflix in HD, while 5G unlocks connected driving. The increase in speed and stability, its ultra-low latency, higher bandwidth and larger connection density make it possible for cars to communicate with each other in real time at speed and intelligently make life-saving decisions faster than any human could. It’s also a true game changer for manufacturing.

Why? Because of IoT and edge devices. As these devices are fundamentally dependent on connectivity, 5G improves their performance by orders of magnitudes. Manufacturers in turn benefit from faster, more reliable and more secure devices that can be connected to one another, opening up new business possibilities for smart real-time applications and data-driven production.

A good example of this is the pilot drone with an AI-based weed control system that Robovision built together with Flanders Research Institute for Agriculture, Fisheries and Food (ILVO). As part of a precision weed control system, its task is to fly over fields and identify patches of weed. The drone’s 5G antenna enables it to transmit field images with GPS data to the cloud in real time, something that would not have been feasible with 4G. After these heavy images are processed in the cloud, a tractor on the ground receives GPS coordinates of the detected weed and moves to spray in that area only. By cutting down on pesticide usage, the farmer protects the environment while reducing costs in the process.

The autonomous drone captures and sends images to the Robovision’s AI-based system to analyse the field via 5G.

Industrial Internet-of-Things (IIoT) &
Artificial Intelligence-of-Things (AIoT)

Simply put, an Internet of Things (IoT) device is a physical object that can connect wirelessly to the Internet.This allows it to generate and transmit data while it can be controlled remotely. Common examples include smart home appliances such as remotely operable toasters, fridges or surveillance systems. But there is more to IoT technology than simply tweaking your thermostat on your smartphone.

IoT’s true promise lies in industrial applications. Industrial IoT (IIoT) can vastly increase productivity, cut down on waste, and improve both safety and quality. Artificial Intelligence of Things (AIoT) further complements this by adding AI in the mix. After all, since IIoT devices generate data, they make for a perfect marriage with AI. AI models can analyse and learn from vast amounts of IoT data that can be used to make intelligent decisions on the workfloor.

The use cases listed below are just some of the many possibilities enabled by AIoT:

  • Predictive maintenance: operators can remotely monitor and be notified when a piece of equipment is on the verge of breaking down, a liquid in a tank is running dangerously low or when a conveyor belt stops working. Proactively preventing a disaster is much cheaper than cleaning up afterwards.
  • Real-time data gathering and analytics: sensors embedded within systems can capture important information and reveal key parameters to operators. Manufacturers can access these quality metrics and performance indicators in real time to optimise workflow efficiency and cut down on any waste in production processes.
  • Supply chain and logistics: as painfully demonstrated by the pandemic, our global supply chains are quite vulnerable to disruptions. AIoT can provide actionable, real-time insights to all parties involved in the logistics chain by monitoring and keeping track of assets and inventory. For instance, Robovision partnered with Peripass and H.Essers to develop a real-time AI-powered container locator of heavy-goods vehicles (HGVs). This intelligent yard management system beats GPS coordinates, pinpointing a trailer anywhere on the yard with a 1-metre accuracy.
 
Combining AI vision learning and the latest generation of smart cameras, Robovision provides real-time insights that can monitor and keep track of assets with AIoT. 

Edge Computing: Real-time Intelligence for Manufacturing Production Process

For consumers, high latency can be frustrating. For manufacturers, it effectively chokes their productivity. In some applications such as telemedicine (remote surgery) or automated vehicle guidance, poor latency can even be life-threatening. The factories of tomorrow will only become reality if we can achieve real-time intelligence with nearly non-existent latency. This is where edge computing comes in.

With edge computing, data processing and storage happen locally via an edge device. These edge devices are located at the ‘edge’ of a network, close to the source where the input data is generated. When we integrate deep learning into these devices, we get intelligent edge devices—the cornerstone of tomorrow’s smart factories.

In manufacturing, edge devices play the role of orchestrators or coordinators, gathering information from machines and system sensors/cameras then making autonomous, immediate decisions based on on-site field data. A real-life example is AI-based safety waste removal, where dangerous goods must be removed from the conveyor and safely disposed of. Here, Edge devices gather real-time data from cameras, use deep learning to check images for flammable items such as gas cylinders, and have them removed if any are found.

The real magic happens when you combine the best of both worlds: the speed of Edge devices with the learning possibilities of big data in the cloud. Our own proprietary intelligent edge device, Robovision Edge, enables operators to steadily improve application performance over time: input data is stored in the device and uploaded to the cloud, where it is used to retrain AI models on. These iterative training cycles offer operators an easy and streamlined way to maintain and improve the AI models of edge devices, making the applications that run these models ever more accurate and effective at their tasks.

Robovision Edge devices are used in production environment to provide real-time and actionable insights with AI
Our own proprietary intelligent edge device, Robovision Edge, enables manufacturers to maintain and improve real-time applications. 

Unlocking Industry 4.0 for Your Factory

All these exciting technologies are gradually painting a picture of what the smart factories of tomorrow will look like.

IIoT infrastructure turns the factory into an interconnected network of systems where machines can communicate with one another in real time. 5G and edge computing make real-time automation possible for these systems and machines. And all of these processes will be driven and coordinated by human-level intelligence through deep-learning AI, often embedded within edge devices or combined with the connectivity of IIoT.

While we certainly live in exciting times for project or innovation managers, businesses are yet hesitant to start surfing the 4th industrial wave. Still, the question is not if, but when these technologies will arrive on the factory floors of their competitors.

Leading AI companies such as Robovision can help translate this huge potential into concrete productivity gains for your factories today. With our Robovision Platform and Robovision Edge device, we can already set up cutting-edge deep-learning applications with 5G for your industrial needs and deploy them at the edge or in the cloud. You can safely join the fourth Industrial Revolution and stay ahead of your competition.

Learn more about Robovision AI in Manufacturing and our technology: Robovision Platform and Robovision Edge