6G-SENSES aims to develop and assess an integrated communication and sensing next-generation network architecture capable of supporting advanced services, such as positioning and tracking, environmental awareness, and digital twin (DT) applications. The system design also focuses on enhancing network intelligence and improving sensing accuracy, thereby enabling data-driven and adaptive network operation.

Building on the multi-layer, disaggregated 6G-SENSES architecture aligned with 3GPP and Open Radio Access Network – Open RAN or O-RAN – principles, which was defined in deliverable D2.2, this deliverable further refines the architecture to enhance support for Integrated Sensing and Communication (ISAC) and intelligent control. These refinements include the acquisition of sensing data from heterogeneous sensor types, the extension of the Open RAN (O-RAN) E2 interface for sensing data, and support for intent-based control at the Service Management and Orchestration (SMO) layer.

It also provides a systematic evaluation of key user plane, control plane, and end-to-end (E2E) components across multiple Wireless Access Technologies (WATs) and  control timescales. On the user plane, the report evaluates ISAC-enabling technologies, such as multi-Access Point (AP) localization and tracking in the millimeter wave (mmWave) band, Sub-8 GHz Wi-Fi-based human presence detection (HPD), and intelligent sensing-assisted Medium Access Control (MAC) scheduling. It further assesses coordination in Sub-6 GHz Cell-Free Multiple-Input Multiple-Output (CF-MIMO), Reconfigurable Intelligent Surfaces (RISs) capabilities across multiple RIS-enabled  scenarios and frequency bands, and Multiaccess Edge Computing (MEC)-assisted wireless edge caching (WEC). These evaluations quantify the impact of the proposed solutions on representative performance indicators including positioning, localization, and sensing accuracy, delay, and resource utilization.

On the control plane, the evaluation serves two complementary purposes. First, it defines and applies a structured assessment procedure to verify the compliance of the Non-Real-Time (Non-RT) and Near-RealTime (Near-RT) RAN Intelligent Controllers (RICs) with O-RAN standards, including functional validation of standardized interfaces and service models (SMs), as well as performance profiling of control-loop latency. Second, it independently evaluates advanced control mechanisms developed within 6G-SENSES, including feedback-based control at the SMO level for enforcing sensing delay and freshness requirements and nearreal-time eXtended application (xApp)-based solutions that combine non-3GPP sensing context with 3GPPnative Quality of Service (QoS) information. In particular, Wi-Fi-assisted beamforming and Channel Quality
Indicator (CQI) + 5G QoS Identifier (5QI)-aware scheduling are evaluated in a unified control loop, where compact sensing context is conveyed via a Wi-Fi Sensing Indicator (WSI) through the Physical Uplink Control Channel (PUCCH), and are shown to jointly improve channel quality, QoS satisfaction, and fairness.

At the E2E level, the deliverable evaluates how Artificial Intelligence (AI)/Machine Learning (ML)-driven automation enhances the orchestration and operation of ISAC services across the network. A Deep Reinforcement Learning (DRL)-based slice orchestration framework is assessed in terms of its ability to balance energy efficiency and service-level performance for concurrent communication and sensing slices, while a techno-economic analysis provides insights into the broader benefits and trade-offs in ISAC-enabled 6G network deployments.

Overall, this deliverable demonstrates the feasibility and effectiveness of the 6G-SENSES architecture through an extensive, multi-layer evaluation. The results confirm that integrating sensing into both the user plane and control plane, combined with AI/ML-driven automation and O-RAN-compliant control, can significantly enhance network adaptability, efficiency, and service support.