Publikation
ORION-RAN: RAN Intelligence and Observability for Near-Real-Time Open RAN Systems
Devanshu Anand; Michael Gundall; Hans Dieter Schotten
In: 31st European Wireless 2026. European Wireless (EW-2026), IEEE, 6/2026.
Zusammenfassung
The evolution towards sixth-generation (6G) net-
works and Open RAN architectures introduces significant chal-
lenges for radio access network (RAN) monitoring due to in-
creasing system complexity, heterogeneous traffic demands, and
dynamic radio conditions. Existing solutions are largely reactive
and rely on static threshold-based indicators, providing limited
visibility into emerging performance degradation. This paper
proposes ORION-RAN, an AI-assisted RAN intelligence and
observability framework for Open RAN systems. ORION-RAN
introduces a telemetry-driven intelligence layer that transforms
near-real-time RAN measurements into high-level indicators cap-
turing network stability and degradation risk. By decoupling
data-driven inference from control logic, the framework enables
reusable and system-level observability within the near-real-time
RAN Intelligent Controller (RIC). The proposed framework is
implemented as a lightweight xApp leveraging standardized Open
RAN interfaces and is evaluated using an end-to-end testbed inte-
grating ns-3, ns-O-RAN, and a near-real-time RIC under hetero-
geneous traffic conditions. Results demonstrate strong alignment
with network degradation, improved early detection capability,
and enhanced reliability compared to conventional threshold-
based and reactive monitoring schemes. These findings highlight
the role of AI-assisted RAN intelligence as a key enabler for
proactive and scalable Open RAN management beyond traditional
control-centric approaches.