
O-RAN ALLIANCE’s Industry Engagement Focus Group (IEFG) and next Generation Research Group (nGRG) are jointly hosting a one‑day workshop on next-G research and on Integrated Sensing & Communications (ISAC).
It's scheduled on June 4, 2026 from 8:30am – 7:30pm PDT as an interactive online event, broadcast from the O-RAN F2F meeting in Seattle, USA.
O-RAN members and participants are welcome to attend in person, as part of the O-RAN F2F meeting Seattle-Bellevue, 1-5 June 2026 - register through the wiki (requires O-RAN membership login).
Organized by O-RAN ALLIANCE's next Generation Research Group (nGRG).
Pacific Daylight Time zone (PDT)
______________
O-RAN has built the open RAN. Agentic AI systems reason over goals, plan across steps, and act autonomously. This talk argues that the O-RAN community should proactively define how agentic intelligence integrates into the RAN architecture, align that work with parallel efforts in the broader standards ecosystem, and establish clear boundaries before fragmented vendor implementations close the cognitive layer that open interfaces were meant to protect.
______________
Exploring the role of, and expectations on, AI agents in the management and automation systems being studied for 6G in both O-RAN and 3GPP SA5. Autonomously acting management and automation functions can be realized using AI and agentic techniques, and the management systems need to support them and manage their agentic capabilities.
______________
6G is evolving beyond “enhanced connectivity” toward native integration of communication, computing, and intelligence at the access edge. The radio access network is shifting from conventional wireless access nodes to comm–compute–intelligence convergence anchors that host the radio protocol stack, edge compute, and intelligent closed-loop control. Network services are also moving from human- or application-driven connectivity to Agentic Communication—intent-driven, goal-negotiable, and partially autonomous sessions in which terminal-side and network-side agents use tasks, semantics, and policies as inputs to autonomously perform discovery, negotiation, resource requests, and quality-of-service assurance, upgrading the communication system from a bit pipe to a collaborative agent network. This talk outlines the base-station evolution path from 5G-Advanced to 6G, presents a reference architecture for comm–compute–intelligence converged gNBs, and clarifies core capability boundaries at the base station around hardware/software disaggregation, edge compute pooling and placement, coordinated scheduling of communication and compute resources, and unified orchestration of communication, compute, and intelligence. On the intelligence plane, we propose co-design of lightweight Agentic AI with Agentic Communication. Drawing on industry practice, the session offers reference perspectives for the coordinated evolution of 6G comm–compute–intelligence converged base stations and Agentic Communication.
______________
Multivendor Cloud RAN faults routinely span RAN software, the cloud platform, and underlying hardware, making Day-2 triage slow and brittle. Agentic systems integrated through an Agentic Gateway over MCP have begun to address the diagnosis half of the problem by coordinating domain-specific agents over controlled interfaces while preserving vendor-private knowledge. PTP synchronization is a particularly demanding test, with root causes ranging from hardware timestamp inconsistency to host-level misconfiguration interfering with PTP traffic.
This talk asks what it takes to extend such diagnostic substrates into closed-loop Day-2 operations. We propose an agent harness pattern that bridges the diagnostic layer to O-Cloud via the O-RAN O2 IMS API and bare-metal lifecycle management, using Red Hat's open-source O-Cloud Manager as the reference O2 IMS implementation. The pattern layers three contributions on top of generic agent frameworks
We walk host-platform PTP scenarios (a fw-lldp-agent misconfiguration and an outdated NIC driver) through the harness, from linuxptp state changes published as O-RAN CloudEvents by Cloud Event Proxy, to fixes routing into the O-Cloud as MachineConfig and KMM Module artifacts delivered via O2 IMS rather than escalating to the SMO. We close with open questions on human-machine teaming for closed-loop remediation (agent governance, digital-twin sandboxing prior to live action, and EvalOps for agentic confidence scoring), alongside the role of MCP relative to O-RAN-native interfaces.
This talk argues that 6G should be designed as an AI‑native system, where an Edge AI stack and distributed AI inference infrastructure come first, and the RAN is built on top of this foundation rather than augmented by AI after the fact. From this perspective, the 6G rollout becomes the buildout of a global, distributed AI‑inference substrate, leveraging the telco’s massive cellular and edge footprint to host low‑latency, privacy‑preserving AI services. We examine the architectural implications of co‑designing RAN and AI workloads on shared edge infrastructure: programmable edge platforms, distributed AI runtimes, and tight edge–cloud orchestration under real‑time constraints. This reframing reshapes our 6G research priorities and enables new classes of applications—such as physical AI, sensing, robotics, and ultra‑low‑latency inference—positioning 6G as a convergence point of connectivity and distributed intelligence.
______________
This presentation showcases a comprehensive OpenAirInterface (OAI) implementation for 5G O-RAN systems featuring 7.2 fronthaul split with end-to-end emulation capabilities and advanced compression techniques. The presentation demonstrates the baseline OAI 7.2 fronthaul implementation where O-DU and O-RU are connected via 7.2 split architecture, with UE connectivity achieved through vrtsim channel emulator enabling realistic end-to-end testing scenarios that support multi-UE configurations with heterogeneous antenna setups, GPU-accelerated channel convolution, and angle-of-arrival modeling for spatial effects evaluation. Furthermore, this presentation covers the integration of compression techniques within the OAI framework, presenting a full-stack fronthaul simulation approach that enables comprehensive analysis across complete UE-RU-DU deployments, capturing interactions between the wireless stack and fronthaul transport to evaluate the impact of IQ compression on system-level performance metrics including throughput, signal quality, and fronthaul bandwidth usage, with experimental results highlighting tradeoffs between fronthaul efficiency and wireless performance for scalable O-RAN deployments.
______________
______________
The ISAC workshop, organized by O-RAN ALLIANCE's Industry Engagement Focus Group (IEFG), dives deep into various aspects of Integrated Sensing and Communications (ISAC) – standards, architecture, implementation, role of open interfaces, and path to 6G native sensing. Discussions will span from current state of ISAC implementations, performance, and validating testing to path to commercialization. We will examine how RAN sensing can leverage E3 (Edge‑Enhanced Exposure) interface and dApps to support ISAC use cases such as drone swarm sensing and detection.
O-RAN IEFG is co-chaired by: AT&T - Pavan Gupta, NEC - Marko Babovic, Radisys - Ganesh Shenbagaraman.
Pacific Daylight Time zone (PDT)
______________
______________
______________
______________
______________
______________
______________
______________
This topic will highlight the recent acceleration of Open Source, Open RAN with OCUDU as a platform to rapidly build out AI-RAN capabilities and specifically ISAC as a key NextG capability. We will highlight our work in this area, alongside the OCUDU project and its partners, leveraging Open RAN as a basis for accelerating pre-standardization activities in AI-RAN, 5G, 6G, and ISAC - and highlight key ISAC data collection, prototyping, interfaces, and paths towards standardization and broader adoption that we are focused on alongside this ecosystem, partners, and our efforts in this area.
______________
______________
Integrated Sensing and Communication (ISAC) is becoming a foundational technology for both O‑RAN evolution and future 6G systems. ISAC enables the RAN to perform joint communication and radio‑based sensing, allowing the network to detect, localize, and characterize passive, non‑connected objects. By embedding sensing as a native RAN function, ISAC allows O‑RU, O‑DU, and O‑CU nodes to reuse communication waveforms, reference signals, and spectrum resources for environmental sensing. Depending on the waveform and processing chain, ISAC can estimate object range, angle, Doppler shift, and in some cases infer size or shape through micro‑doppler and scattering signatures. These capabilities rely on radio signals transmitted and received by both network infrastructure and user equipment.
Radisys has integrated ISAC signal processing and control functions into its O‑RAN compliant DU/CU software. The pre‑E3 O‑RAN interface enables coordination between ISAC sensing modules and RAN control logic. The E3 interface—recently introduced within O‑RAN—provides a standardized mechanism for real‑time, AI‑driven RAN control via distributed applications (dApps). Through E3, dApps can interact directly with O‑DU and O‑CU nodes, enabling low‑latency, closed‑loop control for functions such as dynamic spectrum allocation, positioning, and sensing‑assisted mobility or interference management in 5G and future 6G deployments.
The increasing geopolitical and public‑safety focus on detecting UAVs, drones, and other low‑altitude or small‑RCS objects has intensified the need for wide‑area sensing in both urban and rural environments. Deploying dedicated radar or sensing infrastructure for these use cases requires substantial capital and operational expenditure. Leveraging existing 5G—and future 6G—RAN deployments for ISAC provides a cost‑efficient alternative by reusing existing radio infrastructure, fronthaul/backhaul transport, and compute resources. This approach enables large‑scale sensing coverage without the need for standalone radar systems, positioning the mobile network as a unified communications‑and‑sensing platform.
______________
______________
______________

O-RAN ALLIANCE’s Industry Engagement Focus Group (IEFG) and next Generation Research Group (nGRG) are jointly hosting a one‑day workshop on next-G research and on Integrated Sensing & Communications (ISAC).
It's scheduled on June 4, 2026 from 8:30am – 7:30pm PDT as an interactive online event, broadcast from the O-RAN F2F meeting in Seattle, USA.
O-RAN members and participants are welcome to attend in person, as part of the O-RAN F2F meeting Seattle-Bellevue, 1-5 June 2026 - register through the wiki (requires O-RAN membership login).
Organized by O-RAN ALLIANCE's next Generation Research Group (nGRG).
Pacific Daylight Time zone (PDT)
______________
O-RAN has built the open RAN. Agentic AI systems reason over goals, plan across steps, and act autonomously. This talk argues that the O-RAN community should proactively define how agentic intelligence integrates into the RAN architecture, align that work with parallel efforts in the broader standards ecosystem, and establish clear boundaries before fragmented vendor implementations close the cognitive layer that open interfaces were meant to protect.
______________
Exploring the role of, and expectations on, AI agents in the management and automation systems being studied for 6G in both O-RAN and 3GPP SA5. Autonomously acting management and automation functions can be realized using AI and agentic techniques, and the management systems need to support them and manage their agentic capabilities.
______________
6G is evolving beyond “enhanced connectivity” toward native integration of communication, computing, and intelligence at the access edge. The radio access network is shifting from conventional wireless access nodes to comm–compute–intelligence convergence anchors that host the radio protocol stack, edge compute, and intelligent closed-loop control. Network services are also moving from human- or application-driven connectivity to Agentic Communication—intent-driven, goal-negotiable, and partially autonomous sessions in which terminal-side and network-side agents use tasks, semantics, and policies as inputs to autonomously perform discovery, negotiation, resource requests, and quality-of-service assurance, upgrading the communication system from a bit pipe to a collaborative agent network. This talk outlines the base-station evolution path from 5G-Advanced to 6G, presents a reference architecture for comm–compute–intelligence converged gNBs, and clarifies core capability boundaries at the base station around hardware/software disaggregation, edge compute pooling and placement, coordinated scheduling of communication and compute resources, and unified orchestration of communication, compute, and intelligence. On the intelligence plane, we propose co-design of lightweight Agentic AI with Agentic Communication. Drawing on industry practice, the session offers reference perspectives for the coordinated evolution of 6G comm–compute–intelligence converged base stations and Agentic Communication.
______________
Multivendor Cloud RAN faults routinely span RAN software, the cloud platform, and underlying hardware, making Day-2 triage slow and brittle. Agentic systems integrated through an Agentic Gateway over MCP have begun to address the diagnosis half of the problem by coordinating domain-specific agents over controlled interfaces while preserving vendor-private knowledge. PTP synchronization is a particularly demanding test, with root causes ranging from hardware timestamp inconsistency to host-level misconfiguration interfering with PTP traffic.
This talk asks what it takes to extend such diagnostic substrates into closed-loop Day-2 operations. We propose an agent harness pattern that bridges the diagnostic layer to O-Cloud via the O-RAN O2 IMS API and bare-metal lifecycle management, using Red Hat's open-source O-Cloud Manager as the reference O2 IMS implementation. The pattern layers three contributions on top of generic agent frameworks
We walk host-platform PTP scenarios (a fw-lldp-agent misconfiguration and an outdated NIC driver) through the harness, from linuxptp state changes published as O-RAN CloudEvents by Cloud Event Proxy, to fixes routing into the O-Cloud as MachineConfig and KMM Module artifacts delivered via O2 IMS rather than escalating to the SMO. We close with open questions on human-machine teaming for closed-loop remediation (agent governance, digital-twin sandboxing prior to live action, and EvalOps for agentic confidence scoring), alongside the role of MCP relative to O-RAN-native interfaces.
This talk argues that 6G should be designed as an AI‑native system, where an Edge AI stack and distributed AI inference infrastructure come first, and the RAN is built on top of this foundation rather than augmented by AI after the fact. From this perspective, the 6G rollout becomes the buildout of a global, distributed AI‑inference substrate, leveraging the telco’s massive cellular and edge footprint to host low‑latency, privacy‑preserving AI services. We examine the architectural implications of co‑designing RAN and AI workloads on shared edge infrastructure: programmable edge platforms, distributed AI runtimes, and tight edge–cloud orchestration under real‑time constraints. This reframing reshapes our 6G research priorities and enables new classes of applications—such as physical AI, sensing, robotics, and ultra‑low‑latency inference—positioning 6G as a convergence point of connectivity and distributed intelligence.
______________
This presentation showcases a comprehensive OpenAirInterface (OAI) implementation for 5G O-RAN systems featuring 7.2 fronthaul split with end-to-end emulation capabilities and advanced compression techniques. The presentation demonstrates the baseline OAI 7.2 fronthaul implementation where O-DU and O-RU are connected via 7.2 split architecture, with UE connectivity achieved through vrtsim channel emulator enabling realistic end-to-end testing scenarios that support multi-UE configurations with heterogeneous antenna setups, GPU-accelerated channel convolution, and angle-of-arrival modeling for spatial effects evaluation. Furthermore, this presentation covers the integration of compression techniques within the OAI framework, presenting a full-stack fronthaul simulation approach that enables comprehensive analysis across complete UE-RU-DU deployments, capturing interactions between the wireless stack and fronthaul transport to evaluate the impact of IQ compression on system-level performance metrics including throughput, signal quality, and fronthaul bandwidth usage, with experimental results highlighting tradeoffs between fronthaul efficiency and wireless performance for scalable O-RAN deployments.
______________
______________
The ISAC workshop, organized by O-RAN ALLIANCE's Industry Engagement Focus Group (IEFG), dives deep into various aspects of Integrated Sensing and Communications (ISAC) – standards, architecture, implementation, role of open interfaces, and path to 6G native sensing. Discussions will span from current state of ISAC implementations, performance, and validating testing to path to commercialization. We will examine how RAN sensing can leverage E3 (Edge‑Enhanced Exposure) interface and dApps to support ISAC use cases such as drone swarm sensing and detection.
O-RAN IEFG is co-chaired by: AT&T - Pavan Gupta, NEC - Marko Babovic, Radisys - Ganesh Shenbagaraman.
Pacific Daylight Time zone (PDT)
______________
______________
______________
______________
______________
______________
______________
______________
This topic will highlight the recent acceleration of Open Source, Open RAN with OCUDU as a platform to rapidly build out AI-RAN capabilities and specifically ISAC as a key NextG capability. We will highlight our work in this area, alongside the OCUDU project and its partners, leveraging Open RAN as a basis for accelerating pre-standardization activities in AI-RAN, 5G, 6G, and ISAC - and highlight key ISAC data collection, prototyping, interfaces, and paths towards standardization and broader adoption that we are focused on alongside this ecosystem, partners, and our efforts in this area.
______________
______________
Integrated Sensing and Communication (ISAC) is becoming a foundational technology for both O‑RAN evolution and future 6G systems. ISAC enables the RAN to perform joint communication and radio‑based sensing, allowing the network to detect, localize, and characterize passive, non‑connected objects. By embedding sensing as a native RAN function, ISAC allows O‑RU, O‑DU, and O‑CU nodes to reuse communication waveforms, reference signals, and spectrum resources for environmental sensing. Depending on the waveform and processing chain, ISAC can estimate object range, angle, Doppler shift, and in some cases infer size or shape through micro‑doppler and scattering signatures. These capabilities rely on radio signals transmitted and received by both network infrastructure and user equipment.
Radisys has integrated ISAC signal processing and control functions into its O‑RAN compliant DU/CU software. The pre‑E3 O‑RAN interface enables coordination between ISAC sensing modules and RAN control logic. The E3 interface—recently introduced within O‑RAN—provides a standardized mechanism for real‑time, AI‑driven RAN control via distributed applications (dApps). Through E3, dApps can interact directly with O‑DU and O‑CU nodes, enabling low‑latency, closed‑loop control for functions such as dynamic spectrum allocation, positioning, and sensing‑assisted mobility or interference management in 5G and future 6G deployments.
The increasing geopolitical and public‑safety focus on detecting UAVs, drones, and other low‑altitude or small‑RCS objects has intensified the need for wide‑area sensing in both urban and rural environments. Deploying dedicated radar or sensing infrastructure for these use cases requires substantial capital and operational expenditure. Leveraging existing 5G—and future 6G—RAN deployments for ISAC provides a cost‑efficient alternative by reusing existing radio infrastructure, fronthaul/backhaul transport, and compute resources. This approach enables large‑scale sensing coverage without the need for standalone radar systems, positioning the mobile network as a unified communications‑and‑sensing platform.
______________
______________
______________