Autonomous Uplink Downlink Feedback Indication

Description

Autonomous Uplink Downlink Feedback Indication (AUL-DFI) is a key feature of 5G New Radio (NR) uplink transmission schemes, defined within the context of configured grant (Type 1) operation. It operates within the physical layer framework specified in 3GPP TS 36.212 (multiplexing and channel coding) and TS 37.213 (physical layer procedures). The architecture involves the User Equipment (UE), the gNodeB (gNB), and specific physical layer signaling channels. AUL-DFI enables a paradigm shift from purely grant-based uplink access to a hybrid model where a UE can transmit autonomously on pre-configured resources.

At its core, AUL-DFI works by separating the uplink data transmission trigger from the downlink feedback for that transmission. The network pre-configures the UE with a set of periodic time-frequency resources (a Configured Grant) for potential uplink transmission. The UE, based on its data arrival and internal logic, decides autonomously when to use these resources to send a Transport Block (TB). Crucially, the UE does not send a Scheduling Request (SR) or wait for a dynamic Uplink Grant (UL Grant) from the gNB for this specific transmission instance. This eliminates the scheduling request and grant exchange latency.

The 'Downlink Feedback Indication' component is the gNB's response to this autonomous transmission. After decoding the UE's transmission on the configured grant resources, the gNB must inform the UE whether the transmission was successful (ACK) or needs retransmission (NACK). This feedback is not sent via the conventional Physical HARQ Indicator Channel (PHICH) as in LTE, but is instead carried on the Physical Downlink Control Channel (PDCCH) using a specific DCI format, namely DCI format 0_1 or 0_2 with the DFI flag set. This DCI contains the Hybrid Automatic Repeat Request (HARQ) process ID and the New Data Indicator (NDI) corresponding to the autonomous transmission, allowing the UE to unambiguously associate the feedback with its specific TB.

Key components in the AUL-DFI procedure include the pre-configured grant resources (periodicity, time/frequency allocation, MCS), the HARQ process ID associated with the configured grant, the UE's autonomous transmission decision logic, and the gNB's DFI transmission on PDCCH. Its role in the network is to provide ultra-reliable low-latency communication (URLLC) for uplink-centric traffic, support industrial IoT applications with deterministic traffic patterns, and improve spectral efficiency by reducing control signaling overhead for predictable uplink flows. The mechanism requires tight time alignment; the gNB must be prepared to receive and decode on the configured resources, and the UE must monitor for DFI in a specific time window after its autonomous transmission.

Purpose & Motivation

AUL-DFI was created to address the fundamental latency and efficiency limitations of dynamic grant-based scheduling for certain 5G use cases. In traditional dynamic scheduling, a UE with data to send must first transmit a Scheduling Request (SR), wait for the gNB to respond with an Uplink Grant, and then transmit the data. This multi-step process introduces significant latency (often several milliseconds), which is unacceptable for mission-critical URLLC applications like factory automation, remote control, and augmented reality, where uplink latency is a key performance indicator.

The technology was motivated by the need to support uplink traffic with sporadic, periodic, or predictable characteristics more efficiently. Prior to AUL-DFI, LTE and early NR offered semi-persistent scheduling (SPS) for downlink and configured grants for uplink, but the feedback mechanism for uplink configured grants was less flexible. AUL-DFI specifically solves the problem of providing efficient, low-latency HARQ feedback for autonomous transmissions without resorting to a always-on, dedicated feedback channel like PHICH, which is inefficient for sporadic traffic. It allows the network to retain control over the HARQ process (via feedback) while granting the UE autonomy for the initial transmission, striking a balance between UE autonomy and network management.

Historically, the limitations of previous approaches included the fixed timing of PHICH (limiting flexibility), the latency of the SR/Grant cycle, and the inefficiency of always allocating resources for potential feedback. AUL-DFI, introduced in Rel-15 as part of the foundational NR URLLC toolkit, provided a more dynamic and efficient feedback channel (PDCCH-based) that could be shared among multiple UEs and configured with different periodicities, directly addressing these shortcomings and enabling new low-latency uplink services.

Key Features

• Enables UE autonomous uplink transmission on pre-configured grants without dynamic scheduling request
• Provides downlink HARQ feedback (ACK/NACK) via PDCCH using DCI format 0_1 or 0_2 with DFI flag
• Reduces uplink latency by eliminating the SR and UL Grant exchange delay
• Decreases control signaling overhead for predictable or periodic uplink traffic patterns
• Supports multiple HARQ processes for configured grant operation
• Facilitates reliable communication for URLLC services by maintaining HARQ retransmission capability

Evolution Across Releases

Defining Specifications