6G and Robotics Experts Align at ERF 2026

one6G and euRobotics have released a new highlight video capturing the discussions from Workshop 47, “6G Empowering Future Multi-Purpose Robotics,” held at the European Robotics Forum (ERF) 2026 in Stavanger. The workshop brought together specialists from telecommunications, robotics, academia, research and industry to explore how future communication systems could support the next generation of connected robots.

One of the strongest messages from the video is that the robotics and telecommunications communities need to work much more closely together. The participants described a “communication chasm” between the two sectors. Robotics experts understand autonomy, sensing, control and physical interaction, but may not always appreciate the complexities and limitations of wireless networks. Telecom experts, meanwhile, understand connectivity but may not fully grasp the hardware, safety and operational constraints faced by robots.

Closing this knowledge gap is essential if 6G is to support practical robotic systems rather than become an engineering exercise looking for applications. The workshop participants stressed that new network capabilities should be driven by clear operational requirements and real-world problems.

There are also lessons to be learned from 5G. Even where advanced connectivity features are available, integrating them into industrial systems can be difficult. Robots, networks, edge platforms and applications may all come from different suppliers, while each deployment may require substantial configuration and systems integration. The discussions suggested that 6G should be easier to implement, potentially using AI-assisted configuration and orchestration to reduce this integration burden.

Resilience is equally important. In search and rescue, industrial inspection, public safety and other critical applications, robots cannot simply stop functioning whenever communication conditions deteriorate. They must recognise when connectivity is impaired, adjust their behaviour and continue performing their mission safely wherever possible.

This does not necessarily mean placing every robotic function in the network. Immediate safety functions may still need to remain on the robot, particularly where losing a connection could put people at risk. The challenge is determining how intelligence, perception, control and safety should be distributed across the robot, the network edge and the cloud.

The workshop included two particularly interesting provocation talks. Professor Davide Scaramuzza from the University of Zurich examined whether future networks could meet the requirements of high-speed aerial robots operating in rapidly changing environments. Professor Tamás Haidegger discussed critical teleoperation in healthcare and industrial robotics, where the barriers extend beyond average latency to include latency variation, instability, integration complexity, certification and regulation.

A major part of the workshop was the launch of the second joint one6G and euRobotics whitepaper, “6G Architectural Foundations and AI-Native Solutions for Future Connected Robotics.” Developed with contributions from more than 20 organisations, the paper attempts to provide a common technical foundation for the telecommunications and robotics communities.

The central idea is that 6G should not be treated simply as a faster connection between a robot and a remote application. Instead, it could become a common platform providing communication, sensing, positioning, computing and intelligence as services.

In this model, the network evolves from a transport mechanism into an active participant in robotic perception, cognition and control. The whitepaper describes concepts such as the network acting as a sensor, the network acting as an actuator and the network responding dynamically to the intent of a robotic application.

For example, a robot might tell the network that it is entering a phase of its mission requiring highly reliable, low-latency control. At another point, it might require high uplink capacity to send perception data, precise positioning support or access to an AI model running at the edge. Intent-based interfaces would allow the robotic application to describe the required outcome without needing to configure every underlying network parameter.

The paper organises the main technology requirements around five connected-robotics pillars:

  • Integrated Sensing and Communication
  • AI-native architecture
  • Real-time digital twins
  • Fleet management and multi-robot control
  • Cyber-physical systems for immersive human-robot interaction

Integrated Sensing and Communication, generally known as ISAC, could allow communication infrastructure to contribute to environmental sensing. Base stations, access points and distributed antennas could analyse reflected radio signals to detect objects, movement and changes in the surrounding environment.

This could extend a robot’s perception beyond its cameras, LiDAR and other onboard sensors. It may be particularly useful in cluttered environments, factories, warehouses, disaster areas and locations where satellite positioning is unavailable or unreliable. The network could effectively become an additional distributed sensor for the robotic system.

AI-native networking is another important element. The whitepaper considers AI not merely as an application running over the network, but as a capability distributed across the robot, RAN, core, edge and cloud.

A robot could offload demanding perception, reasoning or planning functions to nearby edge infrastructure. Split inference could divide an AI model between the robot and the edge, allowing early processing to take place locally while more computationally demanding layers run in the network.

Federated learning could also allow fleets of robots to improve shared models without sending all their raw sensor data to a central location. Robots could exchange model parameters or compact representations instead, helping preserve privacy while reducing the volume of information transported over the network.

The paper goes further by considering interactions between AI agents residing on robots and AI agents operating within the network. A network-based agent might optimise coverage, allocate computing resources, protect a safety-critical data flow or determine when a fleet should participate in a federated learning update.

Real-time digital twins would provide virtual representations of individual robots, complete fleets and their operating environments. By synchronising these twins with data from robots and network-based sensing, operators could monitor behaviour, detect anomalies, predict failures and test alternative operating strategies.

Digital twins could also provide safer environments for training robotic control systems. New policies could be tested against unusual network conditions, equipment failures and potentially dangerous scenarios before being introduced into a live robotic system.

For multi-robot operations, the whitepaper considers both centralised and decentralised fleet management. A central planner may be able to optimise tasks, routes, charging and maintenance across an entire fleet. Decentralised communication, including direct robot-to-robot links, can allow robots to coordinate locally when central connectivity is unavailable or when an immediate reaction is required.

In practice, future systems are likely to combine both approaches. Longer-term planning could take place at the edge or in the cloud, while robots retain sufficient local intelligence to avoid hazards and continue operating when communication is disrupted.

To bring these elements together, the whitepaper proposes a four-plane architecture for connected robotics.

The Robotic Service Plane contains the robots, robotic middleware and applications. It is where physical sensing, movement, control and task execution take place.

The Intelligent Service Plane provides distributed AI, reasoning, planning, semantic communication and coordination. It could host AI agents that manage robot fleets, optimise control policies or determine where models should run.

The Network Service Plane provides communication, positioning, sensing, time synchronisation and computing support. This is where capabilities such as ISAC, edge computing, reliable connectivity and task-oriented network services become available to robotic applications.

The Data Governance Plane manages data and AI models throughout their lifecycle. It supports access control, privacy, security, traceability, auditing and regulatory compliance. This is especially important when robots operate in healthcare, public environments or across organisational boundaries.

The architecture is validated against several representative use cases. These include dynamic safety zones for collaborative robots in factories, fleet management, remote driving and teleoperation of mobile robots, search-and-rescue robotics, energy-aware robot coordination, construction and logistics, and robotic systems for stroke treatment and rehabilitation.

Dynamic safety zones provide a useful example of why communications and robotics must be designed together. Instead of relying entirely on fixed barriers or a robot’s onboard sensors, a connected system could combine cameras, radio sensing, positioning and predictions of human movement.

The permitted operating area around a robot could then expand, contract or change shape according to the position and expected movement of nearby workers. The system could slow the robot, modify its route or stop it when the risk became unacceptable.

However, such a system cannot be judged only by radio-interface latency. The complete loop from detecting a person to changing the robot’s behaviour must be dependable, predictable and ultimately certifiable. The robot, network, edge platform, AI model and safety system must therefore be considered as one end-to-end cyber-physical system.

This highlights one of the biggest challenges for 6G-enabled robotics. Impressive peak data rates and theoretical latency figures will not be sufficient. Real deployments will require dependable performance, predictable behaviour during failures, interoperable interfaces and a clear division of responsibility between local and network-based functions.

The whitepaper identifies several gaps that still need to be addressed. These include the absence of widely accepted certified wireless safety loops, limited support for integrated sensing, insufficient data-exposure interfaces and immature communication profiles for medical-grade remote robotics. There are also unresolved questions around security, privacy, ethics, liability and regulatory compliance.

Standards will therefore be central to future progress. The ERF workshop provided input to the IEEE P1955 standardisation activity and the euRobotics Connected Robotics Topic Group, helping identify hardware, implementation and interoperability requirements that should be considered as 6G develops.

The roadmap described in the paper begins with 5G-Advanced platforms, controlled trials and early 6G research. These are expected to evolve towards interoperable, multi-site and increasingly AI-native deployments around 2030 and beyond.

The broader lesson from the video and whitepaper is that robotics provides a valuable reality check for 6G research. A robot does not benefit simply because a network offers more capacity or lower theoretical latency. It requires the right performance at the right time, together with accurate sensing, appropriately located computing, trustworthy intelligence and safe behaviour when something fails.

Achieving this will require the telecommunications and robotics communities to develop common terminology, architectures, interfaces, test environments and standards. The discussions at ERF 2026 and the accompanying whitepaper represent an important step towards that shared understanding.

The one6G highlight video is embedded below, and the full whitepaper is available through the one6G website.

Finally, for anyone interested in the wider development of 6G, registration is now open for the sixth one6G Summit, taking place in London on 10–11 September 2026. The annual event brings together experts from industry, academia, research and policy to discuss the technical progress, practical applications and cross-sector collaboration needed to turn the 6G vision into reality. Further details and registration are available on the one6G Summit website. 

Related Posts: 

Posted by on

Comments

Post a Comment