Incremental Redundancy

Description

Incremental Redundancy is a fundamental component of the Hybrid ARQ (HARQ) process in 3GPP systems like LTE and NR. It operates by initially transmitting a data block with a certain level of channel coding (e.g., a mother code rate). If the receiver fails to decode this initial transmission and sends a Negative Acknowledgement (NACK), the transmitter does not simply resend the identical data. Instead, it sends a different, complementary set of coded bits, known as redundancy versions (RVs). These RVs contain additional parity bits generated from the same original data block but punctured differently. The receiver stores the soft bits (log-likelihood ratios) from the failed transmission in a buffer. Upon receiving a retransmission, it combines these new soft bits with the stored ones, effectively creating a lower overall code rate and a more robust composite codeword. This process of Chase Combining or Incremental Redundancy continues until the data is successfully decoded or a maximum number of retransmissions is reached.

The architecture supporting IR is integrated within the physical layer's channel coding and HARQ management functions. Key components include the HARQ entity at both transmitter and receiver, which manages multiple parallel HARQ processes for continuous data flow; the redundancy version (RV) generator, which selects the specific pattern of systematic and parity bits to transmit; and the soft buffer at the receiver for storing and combining log-likelihood ratios. The eNodeB/gNB schedules transmissions and retransmissions, while the UE performs the decoding and combining. The specific coding scheme, such as Turbo codes in LTE or LDPC/Polar codes in NR, defines the mother code and the rules for generating different RVs.

IR's role is critical for link adaptation and robust data delivery over unreliable radio channels. By incrementally sending more parity information, it allows the system to adapt the effective code rate to the instantaneous channel conditions. In good conditions, the initial transmission might suffice, maximizing throughput. In poor conditions, the accumulated redundancy from multiple transmissions provides powerful error correction. This makes IR highly efficient, as it avoids wasting bandwidth on full retransmissions of already-received information. It is a key enabler for achieving high data rates and low latency with high reliability, especially for control channels and data channels in challenging radio environments.

Purpose & Motivation

Incremental Redundancy was created to address the fundamental challenge of efficient and reliable data transmission over error-prone wireless channels. Traditional ARQ schemes, which discard failed packets and request identical retransmissions, are inefficient in terms of spectral usage and delay. The purpose of IR is to enhance the Hybrid ARQ process by making each retransmission more valuable than a simple repeat. It solves the problem of wasted bandwidth by transmitting new, complementary information with each retransmission, increasing the probability of successful decoding with each attempt.

Historically, before sophisticated HARQ with IR, systems relied on simpler repetition coding or ARQ with Chase combining (which repeats the same bits). These methods were less spectrally efficient. IR, introduced as a core part of HARQ in 3GPP Rel-8 for LTE, represented a significant evolution. It was motivated by the need for higher data rates and better performance at cell edges in emerging OFDMA-based systems. By allowing the receiver to combine different pieces of the same encoded packet, IR effectively creates a stronger, lower-rate error-correcting code on the fly, adapting to channel conditions without requiring explicit signaling to change the initial modulation and coding scheme (MCS).

This technique is fundamental to meeting the Quality of Service (QoS) requirements for various services. It improves throughput and reduces latency for packet data services by minimizing the number of required retransmissions and making each one more effective. In later releases like NR (Rel-15), IR remains a cornerstone, now applied to new channel coding schemes like LDPC for data, ensuring continued gains in spectral efficiency and reliability for 5G's diverse use cases, from enhanced mobile broadband to ultra-reliable low-latency communications.