Link Adaptation

The most fundamental link adaptation technique, AMC varies the modulation order (e.g., QPSK to 64-QAM) and forward error correction (FEC) code rate on a per-frame basis. When the Signal-to-Noise Ratio (SNR) is high, the transmitter switches to a higher-order modulation and a weaker code to maximize throughput. As the channel fades, it falls back to a robust, low-data-rate combination like BPSK with a strong convolutional code to preserve the link. This is the primary mechanism behind the variable data rates seen in 4G LTE and 5G NR standards.

TPC dynamically adjusts the radio's output power to the minimum level required for the receiver to achieve a target Bit Error Rate (BER). This is a critical interference management tool in CDMA and OFDMA networks.

• Near-Far Problem Mitigation: Prevents a terminal close to the base station from drowning out a distant terminal.
• Battery Conservation: Reduces energy consumption in mobile handsets.
• Co-Channel Interference Reduction: In dense cellular deployments, TPC limits the noise floor rise caused by adjacent cells, directly increasing overall network capacity.

HARQ is a time-domain link adaptation technique that combines Forward Error Correction (FEC) with Automatic Repeat reQuest (ARQ). Instead of discarding a corrupted packet, the receiver stores the soft information in a buffer and requests a retransmission. The receiver then combines the multiple transmissions using Chase Combining or Incremental Redundancy to effectively increase the coding gain. This allows the system to operate at a higher average BLER (Block Error Rate) target, squeezing more spectral efficiency out of the channel by relying on rapid retransmissions to correct rare errors.

In multi-antenna systems, link adaptation extends to selecting the optimal Multiple-Input Multiple-Output (MIMO) transmission mode. When the channel is highly correlated (e.g., a static rooftop link), the system may switch from Spatial Multiplexing (sending independent data streams) to Transmit Diversity (sending the same data over multiple antennas) or Beamforming (focusing energy in a specific direction). The choice is driven by the Rank Indicator (RI) and Channel Quality Indicator (CQI) reported back from the receiver, adapting the spatial structure of the transmission to the instantaneous multipath environment.

Link adaptation relies on a closed-loop feedback mechanism. The receiver estimates the downlink channel conditions and reports a CQI value back to the transmitter. This CQI is a quantized recommendation that maps directly to a specific modulation scheme, code rate, and transport block size that the receiver believes it can decode with a BLER below 10%. The latency of this feedback loop is critical; a stale CQI causes the transmitter to select parameters optimized for a past channel state, leading to either wasted capacity or a decoding failure. 5G NR uses ultra-fast sub-millisecond feedback cycles to track vehicular mobility.

The CQI reports from the receiver are inherently imperfect due to estimation errors and quantization. Outer Loop Link Adaptation (OLLA) corrects this by monitoring the actual HARQ ACK/NACK statistics. If the block error rate is too high, OLLA applies a negative back-off offset to the reported CQI, forcing the inner loop to select a more robust MCS. If the BLER is too low, it applies a positive offset to increase throughput. This dual-loop structure ensures the link adaptation converges to the target BLER even when the receiver's CQI estimates are systematically biased.

The primary mechanism for link adaptation at the physical layer. AMC dynamically varies the modulation order (e.g., QPSK, 16-QAM, 64-QAM) and the forward error correction (FEC) code rate on a per-frame basis.

• Goal: Maximize throughput under a target Block Error Rate (BLER).
• Mechanism: When the Signal-to-Noise Ratio (SNR) is high, the system switches to a higher-order modulation and a higher code rate to increase spectral efficiency. When the SNR drops, it falls back to a more robust, lower-order scheme.
• Example: A 5G NR base station uses Channel Quality Indicator (CQI) reports from the User Equipment (UE) to select the optimal Modulation and Coding Scheme (MCS) index from a predefined table.

A closed-loop or open-loop mechanism that dynamically adjusts the transmitter's output power to maintain a target received signal quality while minimizing interference to co-located networks.

• Objective: Achieve a target Signal-to-Interference-plus-Noise Ratio (SINR) at the receiver using the minimum necessary transmit power.
• Near-Far Problem: TPC is critical in CDMA and OFDMA systems to prevent a nearby transmitter from drowning out a distant one.
• Implementation: The receiver sends Transmit Power Control (TPC) commands back to the transmitter, instructing it to increase or decrease power by a fixed step size (e.g., ±1 dB).

A metric reported by the receiver back to the transmitter that quantifies the current state of the wireless channel. It is the fundamental feedback mechanism that drives link adaptation decisions.

• Content: A CQI report typically maps to a specific combination of modulation, code rate, and transport block size that the receiver estimates it can decode with a BLER not exceeding 10%.
• Reporting: Can be wideband (a single value for the entire bandwidth) or sub-band (per frequency sub-band) to enable frequency-selective scheduling.
• Latency: The delay between CQI measurement and its application must be minimized to prevent the link adaptation from being outdated in fast-fading channels.

A retransmission protocol that combines physical-layer Forward Error Correction (FEC) with link-layer Automatic Repeat Request (ARQ) to create a highly robust link adaptation mechanism.

• Chase Combining: The simplest HARQ method where the receiver stores a failed packet and soft-combines it with the retransmitted copy to increase the effective SNR.
• Incremental Redundancy: A more advanced method where each retransmission sends a different set of coded bits (parity bits), progressively lowering the effective code rate until decoding succeeds.
• Link Adaptation Synergy: HARQ allows the AMC to operate at a higher target BLER (e.g., 10% instead of 1%), relying on fast retransmissions to correct the occasional errors, thus maximizing spectral efficiency.

A reinforcement learning framework used to select the optimal transmission rate when the channel statistics are unknown or non-stationary. The cognitive radio treats each available MCS as an 'arm' of a slot machine.

• Exploration: The radio occasionally tries a higher MCS to see if the channel can support a faster rate.
• Exploitation: The radio uses the MCS that has historically yielded the highest throughput.
• Thompson Sampling: A specific MAB algorithm that maintains a probability distribution over the success rate of each MCS and selects an arm by sampling from these distributions, naturally balancing the exploration-exploitation trade-off.

A design paradigm that violates the strict OSI model layering to enable superior link adaptation. The physical layer shares real-time channel state information directly with higher layers.

• PHY-MAC Interaction: The physical layer provides instantaneous SNR and Doppler spread estimates to the MAC layer, which uses this data to select the MCS and decide on packet scheduling.
• PHY-Network Interaction: Channel state information is shared with the network layer to enable joint optimization of routing and transmission parameters, avoiding routes through deeply faded links.
• Benefit: Prevents the 'layer isolation' problem where a TCP congestion control algorithm might misinterpret a physical-layer packet loss as network congestion, triggering an unnecessary and counterproductive rate reduction.