Demos

00Days
00Hours
00Minutes
00Seconds

one6G Summit 2026

King's College London

London 6G Week

6G native AI

one6G SUMMIT 2026

The Exhibitors

The call for demos and posters will remain open until July 31, 2026!

Visit the call for demos page to learn more!

The one6G Summit 2026 provides a premier platform for innovators to showcase groundbreaking prototypes to a global audience. The demo session elevates visibility and fosters live discussions with top industry and academic experts. Don’t miss the chance to connect with exhibitors and explore these innovations during dedicated breaks!

Demo by

6G-Enabled Environmental Reconstruction for Digital Twin

This demo showcases real-time 3D joint perception by fusing ISAC, LiDAR, and camera data into a live 3D scene, continuously tracking humans and robots with high localization resolution.

This demo showcases real-time 3D joint perception enabled by ISAC and multi-modal sensor fusion. ISAC-generated point clouds from massive MIMO sensing with 3GPP-compatible waveform are fused with LiDAR and camera data to create a live 3D scene representation. The system continuously tracks moving objects, including humans and robots, while supporting resolution refinement for improved object localization, environment awareness, and dynamic scene understanding.

Demo by

6G-Enabled Remote Teleoperation and Autonomous Control for Humanoid Waste-Sorting Robots

TeknTrash’s ALPHA humanoid automates waste sorting via multi-camera vision and real-time control. See our live teleoperation demo showcasing the low-latency connectivity demands 6G will solve.

Material Recovery Facilities need robots that can be monitored, controlled, and updated remotely — but multi-camera vision, real-time inverse kinematics, and gripper feedback loops demand connectivity that today's networks struggle to deliver reliably at scale, especially across multiple facilities. TeknTrash's ALPHA is a legged humanoid robot for automated waste sorting, combining multi-camera YOLO-based object detection, real-time inverse kinematics, and dexterous gripper control. At the booth, we demonstrate live teleoperation and remote monitoring of ALPHA's dual-arm sorting system, streamed and controlled over a public tunnel/relay architecture — the same latency- and bandwidth-sensitive pipeline that 6G's URLLC and network slicing are designed to serve at production scale.

Demo by

6G for Medical Applications

Explore the potential of 6G networks for emerging medical applications: Our demonstrator showcases how MAF-based network resource optimization and distributed in-network computing can be leveraged to enhance the performance of medical applications.

Demo by

6G Robotics Services Enabled by Agentic AI-Based Mobile Core Network (A-CORE)

This demo showcases a dynamic network architecture where smart agents manage communication, computing, and AI resources, proven via real-time human-humanoid robot interaction.

The network is dynamically configured to run and support communication, computing and AI resources based on smart agents. A human interaction with a humanoid robot will demonstrate the new architectural concept.

Demo by

Agentic Networks for Autonomous Exploration with Connected Intelligence

CNIT’s booth showcases the “Agentic Network” concept for multi-robot disaster exploration, ahead of a planned 2027 real-world validation with robot fleets in Bologna, Italy.

Natural disasters often leave emergency teams with limited information about affected areas. Autonomous exploration with connected robot fleets can provide timely situational awareness, but requires intelligent coordination across robotics, communication networks, and semantic services. Existing approaches often lack the flexibility and scalability needed for such multi-domain challenges. At our booth, we demonstrate a toy example of the Agentic Network concept, showing how agent-based intelligence can improve multi-robot exploration efficiency while highlighting the challenges of extending autonomous coordination toward networking and semantic services. CNIT aims to validate this agentic approach through a demonstration with real robots equipped with communication technologies, planned for 2027 in Bologna, Italy, where a realistic post-disaster scenario will be replicated.

Demo by

Antenna Tilt Failure Estimation

ISAC uses 3D point clouds to autonomously detect antenna tilt failures with 0.4 RMSE accuracy. This hardware-free internal network maintenance framework drastically reduces CAPEX and OPEX.

Integrated Sensing and Communications (ISAC) is widely recognized as a pillar of next-generation networks, typically framed around external environmental use cases such as traffic monitoring or drone detection. This demo introduces a paradigm shift by positioning ISAC as an internal ”sensor-free sensor” dedicated to self-healing network infrastructure. Utilizing an experimental 10 GHz cmWave ISAC testbed featuring a virtual MIMO architecture, we demonstrate how physical antenna tilt failures (ATFs) can be autonomously detected and estimated. By processing 3D point cloud data through novel estimation frameworks, we achieve an accuracy of 0.4 RMSE. This operator- vendor collaborative approach opens new horizons for drastically re- ducing CAPEX and OPEX through predictive, hardware-free network maintenance.

Demo by

BEACON: Demonstrating AI-Enabled Intent-Driven Services Across the RAN, Core, and Edge for Automotive and Robotic Applications

BEACON is a University of Sussex, Honda, and SUTD collaboration building AI-driven, beyond-5G/6G network API frameworks to power real-time robotic and automotive applications.

A Platform for Beyond-5G Intent Driven Advanced Connectivity Services with End-to-End Network Exposure and Network APIs (BEACON) is a trilateral collaboration between the University of Sussex’s 6G Lab, Honda R&D Europe (UK) Ltd., and the Future Communications Connectivity Lab (FCCLab) at Singapore University of Technology and Design SUTD aims to develop an AI-enabled, intent-driven end-to-end service-based architecture (SBA) and Network Exposure Function (NEF) framework for real-time robotic and automotive applications over beyond 5G/emerging 6G networks. It builds on Sussex and Honda’s successful 5G-Advanced proof-of-concept for real-time vehicle video streaming via telecom APIs and leverages the Sussex–SUTD MoU established in 2023 to advance 6G and AI research.

Demo by

CHIRP: Distributed ISAC with Unsynchronized FMCW Sensors

CHIRP is an unsynchronized 3-node 77-81 GHz ISAC system. It exploits mutual interference to recover bistatic channels, enabling scalable 6G tracking and imaging without complex synchronization.

Distributed ISAC systems require spatially separated sensing nodes to cooperate, but tight synchronization between devices can increase complexity, cost, and deployment constraints. CHIRP demonstrates a distributed architecture based on three independent 77–81 GHz FMCW nodes operating without synchronization. Their mutual interference is exploited to recover not only monostatic channels, but also bistatic channels between node pairs. At the booth, visitors will observe the individual channels, the extraction of bistatic information from interference, and its use for environment reconstruction and people tracking. The demo showcases how unsynchronized multi-node systems can enable richer spatial diversity and scalable multi-static perception for future 6G sensing applications.

Demo by

Digital Twin for Pre-Deployment Validation of AI-Driven Safety-Critical Industrial Control Loops

This demo showcases a Digital Twin unifying physical, 5G network, and AI layers. Validated on a hazardous-liquid AMR, it accurately predicts spill alerts, RSRP, and RTT for safety-critical control.

As industrial systems become more complex, understanding the interactions between physical processes, communication networks, and intelligent applications is essential for design and validation. This demo presents a Digital Twin (DT) that jointly reproduces these layers for an industrial deployment. We validate the DT against a Proof-of-Concept featuring a 5G-connected Autonomous Mobile Robot (AMR) transporting hazardous liquids, remotely controlled by an AI application that anticipates imminent spill events. Integrating robot trajectory, network dynamics, and application logic, the DT reproduces the physical system's behavior, including the spill alert. Its predictions closely match field measurements of Reference Signal Received Power (RSRP) and application Round-Trip Time (RTT), demonstrating the potential of integrated DTs for designing and validating safety-critical control loops.

Demo by

Distributed Edge & Split Learning for Spatiotemporal Forecasting in 6G

DSIPTS-P is a K8s-native platform for 6G edge-cloud spatiotemporal forecasting. It supports split and federated learning, offering automated, resilient, and energy-efficient distributed AI.

We present DSIPTS-P, a Kubernetes-native platform for distributed spatiotemporal forecasting across edge–cloud infrastructures for AI-native 6G. DSIPTS-P supports split learning and federated execution while automating deployment, training, monitoring, and recovery. Built on the DSIPTS stack, it provides GPU- and edge-aware scheduling, YAML- or GUI-based configuration, and integrated telemetry through Prometheus, Grafana, Kepler, and AIM. In the live demo, attendees configure forecasting models, split-learning partition points, and edge/cloud placement, then inject failures and network impairments. Real-time dashboards reveal the impact on forecasting accuracy, convergence, communication overhead, latency, and energy efficiency, demonstrating resilient and resource-efficient distributed AI for future 6G edge–cloud systems.

Demo by

Energy-Efficient Radio Frequency Chain Activation for Cell-Free Massive MIMO

This work optimizes uplink mmWave cell-free mMIMO energy efficiency under imperfect CSI. Proposed graph-theory and DRL approaches for joint AP/RF activation yield high EE gains with minimal SE loss.

In this work, we investigate energy efficiency (EE) optimization subject to quality of service (QoS) requirements for uplink millimeter-wave (mmWave) cell-free massive multiple-input multiple-output (CF-mMIMO) systems under imperfect channel state information (CSI). To maximize the total EE in practical scenarios, we propose two energy-efficient joint access point (AP) and radio frequency (RF) chain activation approaches based on graph theory and deep reinforcement learning (DRL), respectively. Numerical results demonstrate that the proposed adaptive RF chain activation (ARFA) schemes both achieve a favorable trade-off between EE and SE, yielding significant EE improvement at the cost of negligible SE degradation, while maintaining a manageable computational complexity and enabling real-time operation.

Demo by

Explainable Deep Reinforcement Learning for O-RAN Slice Resource Control

Explore our live O-RAN demo using DRL for 5G slice adaptation. Perturb cell traffic, monitor real-time PRB reallocations, and query a local LLM to trace every decision back to its root cause.

Deep Reinforcement Learning (DRL) can adapt 5G slice resources to meet SLAs, but operators still struggle to see why a controller moved Physical Resource Blocks (PRBs) when traffic shifts. This demo connects visitors to a live remote O-RAN testbed: three UEs, one per eMBB, URLLC, and mMTC slice, and a constrained Soft Actor-Critic (SAC) near-RT xApp that reallocates downlink PRBs using E2SM-KPM telemetry and E2SM-RC control. Visitors watch KPMs and DRL actions on Grafana, then use a browser-based MGEN UI to set traffic direction, rate, packet size, and duration, perturb the cell, and observe how slice quotas and SLA flags respond in real time. They can then query a local LLM about selected events and decisions. The booth demonstrates measurable SLA-aware PRB adaptation and lets visitors trace each decision back to the KPMs and controller state that triggered it.

Demo by

From Locomotion and Manipulation to Loco-Manipulation: Planning and Safety

To address contact-rich loco-manipulation in humanoids, this demo introduces LLM-guided motion planning combined with safe predictive filters to enable safe planning and control.

Safe planning and control in locomotion and manipulation is extremely challenging, mainly due to their contact-rich nature. When combined, the loco-manipulation becomes arguably the most challenging problem in humanoid deployment. In this demo, we present our research on LLM-guided motion planning and safe predictive filters towards enabling safe loco-manipulation.

Demo by

Importance of Service-Level Energy Consumption Control in 6G

This demo highlights end-to-end energy metering for 6G services like video and AI bots, showing how workload placement, prompt complexity, and incentive models drive sustainable usage.

As sustainability and energy efficiency become vital to our society, they have also become key objectives for future 6G networks. Due to complex and interdependent infrastructures that enable provisioning of modern ICT services, certain combined systemic impacts can generate hidden environmental costs despite local efficiency gains. To uncover and avoid these impacts, metering the energy consumption of individual ICT services across the entire end-to-end provisioning chain is essential. Our interactive demo showcases this approach using two representative services: video streaming and chatting with an AI bot. Visitors will explore how service quality, AI prompt complexity, and workload placement influence energy consumption. Furthermore, the demo will show how this data, combined with innovative incentive mechanisms, can drive more sustainable service provisioning and user consumption.

Demo by

ISAC Real-Time Experimentation Platform

This real-time mmWave ISAC system uses a cross-layer HW/SW design to achieve continuous sensing over 5G communication, extracting CSI for live micro-Doppler target visualization.

ISAC is a key enabler for future wireless systems, yet realizing it in practice requires tightly coupled communication and sensing pipelines under strict real-time constraints. In this demonstration, we present a real-time Millimeter-Wave ISAC system which addresses these challenges through a cross-layer hardware/software design. Our approach preserves a standard 5G-like communication pipeline and enables sensing by appending bursts of reference signals for beam-sweeping-based channel probing. The system partitions the processing across the FPGA logic, an embedded ARM processor, and an external host, enabling continuous sensing without disrupting the communication link. We demonstrate an end-to-end system where Channel State Information is extracted on-device and streamed for real-time processing and visualization of micro-Doppler signatures from moving targets.

Demo by

MOVE: AI-Driven Smart Switching for Hybrid TN/NTN Connectivity

OrbitsIQ Global’s MOVE is an AI platform managing hybrid terrestrial/satellite networks. The demo showcases smart hardware and dashboards to ensure resilient connectivity for future 6G mobility.

As connected and autonomous systems increasingly depend on uninterrupted communications, intelligent management of hybrid terrestrial and non-terrestrial (TN/NTN) networks is becoming essential for resilient connectivity and future 6G applications. OrbitsIQ Global will showcase MOVE, its AI-driven smart-switching platform for hybrid connectivity. The demonstration features both the hardware and software components of the solution, including the physical deployment platform and pre-recorded dashboard visualisations highlighting intelligent network selection, real-time connectivity management, and AI-assisted switching across terrestrial and satellite networks. MOVE is designed to minimise connectivity disruptions while optimising network performance for connected mobility and other mission-critical applications.

Demo by

Powering Trusted AI with 5G in a Box: Kai and Aruva

AWTG’s Flex 5G private network enables rapid deployment of AI services like Kai (enterprise solutions) and Aruva (education), showcasing secure connectivity as a practical bridge toward AI-native 6G.

Organisations need secure, rapidly deployable connectivity and AI services without complex infrastructure. AWTG will demonstrate its Flex 5G Pop Up Mobile Private Network, a dis-aggregated 5G standalone platform that can go live in minutes and integrate Open RAN components. Devices connected to the private network will access two live AI applications. Kai will demonstrate multilingual customer service, enterprise knowledge retrieval and guided resolution. Aruva will demonstrate personalized learning, explainable feedback and institution controlled educational AI. Together, they show how portable private networks can securely connect trusted AI services and provide a practical bridge towards AI native 6G for enterprises, public services and universities.

Demo by

Real-time AI Decisions for Mobility Use Cases

The demonstration showcases agentic core mobility use-cases, by embedding LLM-based agents in UE association and hand-over decisions, while simultaneously compensating for high latency or potential lack of accuracy originating from LLMs.

Demo by

Real-time Generative Multicasting with On-Device Intent-Aware Semantic Decomposition

Generative semantic multicasting splits video into shared semantic maps and user-specific classes. Running in real-time on a Coral TPU, it reduces transmitted payload bits by 48%.

Demo by

Remote Teleoperation and Training of Robot through VR

Extend Robotics will demonstrate their AMAS remote robot monitoring and intervention capability for industrial use cases. Booth visitors can experience fully immersive robot control via standard 5G.

Extend Robotics will demonstrate their AMAS remote robot monitoring and intervention capability being deployed for industrial use cases. Visitors to the booth will be able to experience fully immersive robot control through standard 5G internet.

one6G Summit 2026

King's College London

London 6G Week

6G native AI