3GPP Release 20 Standardizes Voice over NB-IoT NTN for Satellite Connectivity

·2026.10.12 10:31

Key point

The standard targets geostationary satellites with a new ultra-low bitrate codec operating at 1.0 and 1.4 kbps.

1 / 2

Details

3GPP Release 20 is extending NB-IoT NTN (Non-Terrestrial Networks) to support voice communications, addressing the need for satellite-based emergency services and connectivity for wearables and vehicles. This work builds on the NB-IoT NTN foundation established in Release 17, which was the first widely available form of satellite-based 3GPP cellular connectivity. The initiative aims to leverage existing commercial deployments of NB-IoT NTN infrastructure rather than waiting for NR NTN maturity.

Technical Constraints and Design Decisions

The primary use case focuses on Geostationary (GEO) satellite access, where propagation delays exceed 250 milliseconds and data rates are constrained to below 10 kbps. To accommodate voice within these limits, 3GPP selected the User Plane approach over the Control Plane, aligning with conventional 4G voice procedures to reduce implementation complexity. Key technical additions include Semi-Persistent Scheduling (SPS) to minimize signaling overhead for voice frames and Robust Header Compression (RoHC) to reduce protocol overhead. Additionally, Release 20 permits device transmit power up to 37 dBm to close the link budget gap for GEO uplinks.

Voice Codec and Open Issues

A critical component of this standardization is the development of an Ultra Low Bitrate Codec (ULBC) designed to operate at 1.0 kbps and 1.4 kbps. This codec is expected to be AI-based, marking a first for 3GPP voice codec standardization, and aims to deliver speech quality comparable to terrestrial HD Voice despite the extreme bandwidth constraints. Ongoing discussions address the integration of voice procedures with terrestrial networks, the necessity of specific access barring mechanisms, and the computational complexity limits for neural models in constrained devices like wearables. First published standards for these features are targeted for early 2027.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.