What do cyber-physical systems have in store for us?

What is a cyber-physical system?

A cyber-physical system is an integration of systems with varying natures whose main purpose is to control a physical process and, through feedback, adapt itself to new conditions in real time. They’re created at the intersection of physical processes, networking, and computation.

The term ‘physical’ refers to the object as it is perceived by human senses, while the term ‘cyber’ refers to the virtual representation of the world in which the physical object belongs, while providing further details about the object – for example its make or model.

Being able to view things from both a physical and virtual perspective, cyber-physical systems – under human supervision – can autonomously evaluate operational conditions and support subsequent decision-making and operational adaptations.

The essential component of cyber-physical systems is the presence of interconnected objects which, through sensors, actuators and a network connection, generate and acquire data of various kinds.

With the help of integrated sensors, cyber-physical systems can autonomously determine their current operating status within the environment in which it is located, and the distances between objects. Actuators serve to carry out planned actions or to implement corrective decisions, for example, optimizing a process or solving a problem. Decisions are made by AI which evaluates information from its own internal sensors and information shared by other cyber-physical systems.

Communication is vital in cyber-physical systems, as they allow different objects to exchange information with each other and with humans, at any time and in any conditions. Information is shared at the cyber level, which is currently raising questions around whether or not it’s possible for physical objects, and their virtual representations, to create a ‘social network of things.’

Are we ready for them?

Network performance – in terms of latency, bandwidth, and reliability – largely impact the interactions between the different components of a cyber-physical system, and the ability to execute parallel tasks within a specific timeframe is critical.

Another key challenge for cyber-physical systems is identifying critical issues and faults. In a complex cyber-physical system, a minor fault can generate a cascade of alarms, making difficult to identify the problem and its root cause.

To ease this process, a massive amount of data must be able to flow between the physical and cyber levels to ensure to the system continues to operate within expected behavioral models. This approach also provides human supervisors with a clear analysis of any issues, their location, and the relevant actions that need to be taken.

There are many types of cyber-physical systems, each with different applications. However, the future network platform should provide the right level of connectivity for them all to deliver optimum performance. For example, latency is an issue for all cases where a complex AI must make real-time decisions, and this needs to be as low as possible.  

Each cyber-physical system has a specific architecture, so the future network platform needs to be flexible enough to adapt and comply with these architectures. This means specific ‘ad-hoc’ designs for indoor and outdoor coverage could be needed. In addition, network slicing will play a key role in enabling heterogeneous connectivity requirements on the same 5G network.  

Real-world examples of cyber-physical systems

In the future, cyber-physical systems will be present in all industry sectors. Here, we explore three examples that relate to industrial segments that many analysts consider highly important: smart manufacturing, port logistics, and automotive. In all cases, these complex systems are required to be self-configuring, self-adjusting, and self-optimizing; leading to greater agility, flexibility, and cost effectiveness.

Cyber-physical systems in Smart Manufacturing

The future factory will run on a cyber-physical system, or a set of interacting systems, where highly skilled workers will be provided with operational insights directly from coordinated intelligent machines controlled by a central entity.

Every functional aspect of production will be affected, from design, to manufacturing and supply chains, and later extending to customer service and support. The smart factory will be hyper-connected and data intensive, and rely on an industrial-grade 5G network that’s 100 percent secure. The 5G-enabled Digital Twin – which Ericsson is trialing with Comau is an excellent example of the ability of cyber-physical systems to create and combine the physical or digital aspects of products, systems, and processes. Here, physical machines in an automotive plant are fitted with massive sensors that send status data to a virtual reproduction on a constant basis. Engineers can work on potential problems and possible solutions directly in this cyber replica. In 2019, Ericsson and Comau demonstrated the 5G-enabled digital Twin in Hannover Messe 2019 and at Mobile World Congress Shanghai.

In the smart manufacturing scenario, Ericsson is also experimenting with Automated Guided Vehicles (AGV); trolleys that enable multidirectional layout of the production line instead of a linear layout based on conventional conveyors. AGVs ‘understands’ the surrounding environment, and ensure that various components are shuttled among the work cells in the plant and between the line and the warehouses/loading bays. Most of AGV control is happens in the cloud, and interaction with humans is facilitated by implementing visual navigation with collision avoidance functionalities.

Cyber-physical systems and the port of the future

Terminal port operations will increasingly become a mixture of physical machinery, robotics systems, automated vehicles, human-operated digital platforms, and AI-based software systems. This will transform future ports, and cyber-physical systems comprised of various ‘intelligent agents’ will be highly specialized in cargo loading/unloading and support the port logistics chain.

In the Livorno Port in Italy, Ericsson is experimenting with TIM key 5G-enabled AR applications using operational research algorithms, image recognition techniques, and AI technologies to support dock workers. Thanks to the abundance of 5G connected objects in the port, an AI operation management system determines the sequence of logistics tasks and activities, using real-time information captured from sensors, cameras, and other devices and vehicles in the area. Having such a detailed view of port operations allows the AI to feedback on processes and give live updates to its human supervisors.

Cyber-physical systems in Automotive

All new features in modern cars, like advanced driver assistance systems and connected vehicle services, are based on electronics and software rather than on mechanical engineering. These systems run on specific, compartmentalized in-car modules that interact with multiple sensors and actuators. In the automotive industry, cyber-physical systems could prevent accidents and casualties. For example, consider a highway fitted with different sensors that are able to detect objects or obstacles that could cause accidents in real time. A quick analysis of the data supplied by these sensors by a cyber-physical system could send an instant alert across a communication network to all potentially affected vehicles.

Providing this kind of cyber-physical system could run on a low-latency network, it could save many lives. Nowadays, navigators provide information about obstacles and queuing, but certainly not in real time.

In Torino, Ericsson is cooperating with Centro Rierche FIAT (CRF), the FCA research center, to experiment with Extended Virtual Sensing (EVS) techniques, blending enhanced onboard sensor measurements with network data to boost safety and maximize passenger comfort.

Ericsson and Comau’s Smart Manufacturing research findings

Caption: Orientation map of the physical elements in the smart manufacturing scenario where a complete set of use cases, relevant for 5G, have been deployed, experimented, and assessed by Ericsson and Comau in Italy.

Since 2016, Ericsson has been exploring smart manufacturing with Comau, a world leader in industrial automation. Together we’ve completed all the steps required to build the factory of the future in a real industrial context. More specifically, we’ve explored 4G/5G cellular networks and edge cloud for manufacturing plants.

Two parallel communicating layers worked together:

• The ‘physical’ layer is the factory floor in the production/assembly plant. It includes robots, systems and machines of various kinds, automated guided vehicles, physical controllers, sensors (including cameras) and tools
• The ‘cyber’ layer is constituted by an edge cloud, typically a ‘cloud on the premises’, where computing power, execution of applications, data elaboration, logging, storing, and alarm management are taken away from the shop floor, centralized, and virtualized. Edge cloud does not need contact with any centralized cloud, although it may interact with one

The 5G multipurpose network connects the physical layer with the cyber layer. New findings have emerged from our experiments, including:

• Robot and low-level systems control moved in cloud demands 1-2 mS latency
• Work cell control moved in cloud and virtualized (i.e. virtual PLCs replacing physical PLCs) requires 10-30 mS latency
• Replacement or supplementation of conveyors with automated guided vehicles (AGV) demands a maximum latency of 10 mS if the control runs in cloud, instead of running onboard the vehicle. A stable 70 Mbps rate is also a requirement for visual navigation
• Connected sensors can reach the number of one thousand in a plant. The use of 5G connectivity ensures smooth scalability and minimum energy consumption. For the specific Comau scenario sensors include cameras and pressure, temperature, vibration (IMU) sensors
• Tools like connected screwdrivers and drills can be moved from Wi-Fi or Bluetooth technology to 5G for better scalability and reduced battery drain

Moving data elaboration and intelligence into a cloud environment reduces the footprint of the robot, and allows for higher density production, as well as reducing the robot cost. Remotely located machines and portable systems can also be easily integrated in the chain. The cloud is where the digital twin replica of the plant can be constructed, completely mirroring the physical reality in a digital appearance.

This experimental deployment helped us establish the requirements needed to run cyber-physical systems in this kind of environment. It demonstrates, in a realistic smart manufacturing scenario, how 5G is vitally important to bridge the ‘physical’ world with the ‘cyber’ one.

Read the 2019 technology trends report in full.

Learn more about the industrial cloud and digitalization of industry.

Want the latest Ericsson Research articles straight to your inbox? Subscribe to the Ericsson Blog today.

Networking trends: A platform for next-level digitalization - Ericsson