Toward Native ISAC Support in O-RAN Architectures for 6G

TL;DR

This paper proposes architectural extensions to O-RAN for integrating sensing capabilities in 6G networks, enabling environmental sensing through new service models, primitives, and interfaces, validated via a prototype implementation.

Contribution

It introduces novel extensions to O-RAN specifications for monostatic sensing, including service models, primitives, and metadata, facilitating near-real-time environmental sensing in 6G.

Findings

Prototype implementation demonstrates low-latency sensing control.

Extensions support monostatic, bistatic, and cooperative sensing scenarios.

Validated architecture enables near-real-time environmental sensing.

Abstract

ISAC is an emerging paradigm in 6G networks that enables environmental sensing using wireless communication infrastructure. Current O-RAN specifications lack the architectural primitives for sensing integration: no service models expose physical-layer observables, no execution frameworks support sub-millisecond sensing tasks, and fronthaul interfaces cannot correlate transmitted waveforms with their reflections. This article proposes three extensions to O-RAN for monostatic sensing, where transmission and reception are co-located at the base station. First, we specify sensing dApps at the O-DU that process IQ samples to extract delay, Doppler, and angular features. Second, we define E2SM-SENS, a service model enabling xApps to subscribe to sensing telemetry with configurable periodicity. Third, we identify required Open Fronthaul metadata for waveform-echo association. We validate the architecture through a prototype implementation using beamforming and Full-Duplex operation, demonstrating closed-loop control with median end-to-end latency suitable for near-real-time sensing applications. While focused on monostatic configurations, the proposed interfaces extend to bistatic and cooperative sensing scenarios.