Description
The Uplink Sounding Reference Signal (UL-SRS) is a known sequence transmitted by the UE that allows the gNodeB (gNB) to perform uplink channel estimation. Unlike demodulation reference signals (DM-RS) which are tied to a specific PUSCH/PUCCH transmission, SRS is transmitted independently, often on a configured, periodic, semi-persistent, or aperiodic basis. The gNB configures the SRS transmission parameters via RRC signaling and can trigger aperiodic SRS via DCI. Key parameters include bandwidth, time/frequency position, transmission comb, cyclic shift, and sequence ID, allowing for flexible and UE-specific sounding.
The primary function is channel sounding. By analyzing the received SRS, the gNB can estimate the uplink channel state information (CSI), including the channel frequency response across the sounded bandwidth. This estimation is crucial for several network functions. First, it enables frequency-selective scheduling for the uplink UL-SCH. The scheduler can allocate PUSCH resources in frequency regions where the UE has a strong channel, thereby maximizing throughput and spectral efficiency. Second, it facilitates uplink link adaptation by providing the information needed to select the appropriate modulation and coding scheme (MCS) for a given UE's PUSCH transmission, balancing data rate and error probability.
In Time Division Duplex (TDD) systems, where uplink and downlink share the same frequency band, the channel reciprocity principle can be applied. The uplink channel estimate derived from SRS can be used to infer the downlink channel conditions. This is a cornerstone for massive MIMO and beamforming. The gNB can use the SRS-based channel knowledge to calculate precoding weights for downlink transmissions, effectively steering energy towards the UE and creating narrow, focused beams. This improves downlink signal strength, reduces interference to other users, and increases overall network capacity. Furthermore, SRS can be configured with antenna switching, allowing a UE with multiple transmit antennas to sound the channel from different antennas sequentially, enabling the gNB to acquire channel knowledge for all UE transmit antennas, which is vital for uplink MIMO and downlink beamforming based on reciprocity.
Purpose & Motivation
UL-SRS was introduced to provide the network with a dedicated and flexible mechanism for obtaining uplink channel state information, a capability that became increasingly critical with the evolution of LTE and the advent of 5G NR. Early LTE releases focused on downlink CSI feedback (CQI, PMI, RI) reported by the UE. However, for optimal uplink performance and to fully exploit the potential of advanced antenna systems, the network needed direct, wideband knowledge of the uplink radio channel.
The motivation stems from several limitations of relying solely on downlink measurements or demodulation references. Demodulation reference signals (DM-RS) are only present during a PUSCH transmission and cover only the scheduled bandwidth, providing no information about channel quality on other frequencies. SRS solves this by allowing the network to proactively sound the channel over a configurable, potentially system-wide, bandwidth even when the UE has no data to send. This enables proactive scheduling decisions.
With the push towards massive MIMO and beamforming, especially in TDD deployments, channel reciprocity became a highly efficient alternative to explicit CSI feedback. UL-SRS is the enabling signal for this approach. By sounding the uplink, the gNB, equipped with a large antenna array, can estimate the spatial characteristics of the channel and derive accurate beamforming vectors for downlink transmission. This is far more scalable and provides more timely information than having each UE measure and report downlink CSI for dozens of antenna ports. Thus, UL-SRS is a key enabler for the high spectral efficiency and capacity targets of 5G networks.
Key Features
- Wideband or partial bandwidth channel sounding for uplink CSI acquisition
- Multiple transmission configurations: periodic, semi-persistent, and aperiodic (triggered by DCI)
- Support for antenna switching, enabling channel sounding from multiple UE transmit antennas
- Configurable parameters for bandwidth, comb structure, cyclic shift, and time-domain position for multi-user multiplexing
- Enables uplink frequency-selective scheduling and link adaptation
- Critical for channel reciprocity-based downlink beamforming in TDD systems
Evolution Across Releases
Introduced in NR with enhanced flexibility compared to LTE SRS. Key initial features included support for wider bandwidths aligned with NR carrier bandwidths, flexible placement within a slot, and robust support for aperiodic SRS triggering. It was designed from the ground up to support massive MIMO and beamforming, with explicit configurations for spatial relation info and usage parameters (e.g., 'codebook', 'nonCodebook', 'antennaSwitching', 'beamManagement').
Enhanced SRS capabilities for improved performance and efficiency. Introduced SRS switching gaps to allow a single RF chain to sound multiple panels/antennas. Enhanced support for multi-panel and multi-TRP (Transmission Reception Point) operation, allowing more sophisticated beam management and coordination between multiple network transmission points.
Defining Specifications
| Specification | Title |
|---|---|
| TS 37.571 | 3GPP TR 37.571 |
| TS 38.305 | 3GPP TR 38.305 |
| TS 38.455 | 3GPP TR 38.455 |
| TS 38.473 | 3GPP TR 38.473 |