How FEC and ARQ Work: Core Mechanisms Defined

Forward Error Correction (FEC) and Automatic Repeat Request (ARQ) solve the same problem — wireless transmission errors — through opposite strategies. FEC embeds mathematically redundant data within each transmitted packet, enabling the receiver to detect and correct errors entirely on its own, without any communication back to the sender. ARQ, by contrast, relies on the receiver detecting an error and sending a retransmission request to the sender, which then re-sends the affected data.

In an FEC system, the transmitter applies an error-correcting code — such as a Reed-Solomon code, a convolutional code, or a Low-Density Parity-Check (LDPC) code — to the original data before transmission. The receiver uses the redundant bits to reconstruct the original message even if some bits are corrupted in transit. The key property is that this entire process is one-way: no feedback channel is required, and no round-trip delay is incurred.

ARQ operates differently. The receiver uses an error-detection code (most commonly a Cyclic Redundancy Check, or CRC) to determine whether a received packet is intact. If an error is detected, the receiver discards the packet and sends a Negative Acknowledgement (NACK) to the transmitter, which then re-sends the data. Three classic ARQ variants exist: Stop-and-Wait ARQ (the sender waits for acknowledgement before sending the next packet), Go-Back-N ARQ (the sender retransmits from the errored packet onward), and Selective Repeat ARQ (only the specific errored packets are retransmitted).

Latency, Reliability, and Bandwidth: The Core Tradeoffs

FEC and ARQ impose different costs on the three resources that matter most in industrial wireless design: latency, reliability, and spectral bandwidth. FEC adds overhead to every packet regardless of whether errors occur, consuming bandwidth even on a clean channel. ARQ consumes bandwidth only when errors are detected, but at the cost of variable and potentially unbounded latency when channel conditions degrade.

For hard real-time industrial control — such as closed-loop motor control or robotic arm coordination — the latency budget is fixed and tight. A missed deadline in these systems can cause physical damage or safety incidents. FEC’s deterministic latency profile, determined solely by the code block size and the decoder’s processing time, makes it the natural choice for such applications. The latency introduced by FEC encoding and decoding is fixed and known at design time, regardless of channel conditions.

“In hard real-time industrial control, a missed deadline can cause physical damage. FEC’s deterministic, bounded latency makes it the natural choice where ARQ’s variable retransmission delay is unacceptable.”

ARQ’s variable latency arises from the round-trip propagation delay plus processing time multiplied by the number of retransmissions. In a lightly loaded, low-noise industrial environment, ARQ can be highly efficient — packets are rarely retransmitted, and bandwidth is used only when needed. However, in environments with bursty radio-frequency interference (common in factory floors with variable-speed drives, arc welders, and dense Wi-Fi deployments), error bursts can trigger cascading retransmissions that push latency well beyond acceptable limits for control applications, according to standards bodies including IEEE.

Hybrid ARQ: Combining Both Approaches for Industrial Links

Hybrid ARQ (HARQ) resolves the FEC–ARQ tradeoff by combining both mechanisms in a single protocol layer. The transmitter applies an FEC code to the data before sending; the receiver first attempts to correct any errors using the FEC component. Only if FEC-based correction fails — indicating an error burst severe enough to exceed the code’s correction capacity — does the receiver issue a NACK and request retransmission. This means HARQ retransmissions are rarer than in pure ARQ, and when they do occur, the receiver can combine the new transmission with the previously received (but corrupted) copy to improve decoding success.

Two principal HARQ combining strategies are used in practice. Chase Combining retransmits an identical copy of the original packet; the receiver combines both copies using maximum-ratio combining to improve the signal-to-noise ratio before decoding. Incremental Redundancy (IR) retransmits additional parity bits not included in the original transmission, progressively increasing the effective code rate with each retransmission attempt until decoding succeeds. Incremental Redundancy is more spectrally efficient because it avoids re-sending bits the receiver already has, a property particularly valuable in bandwidth-constrained industrial wireless bands.

HARQ is the error-control mechanism underpinning 5G New Radio (NR), including the Ultra-Reliable Low-Latency Communication (URLLC) service class designed specifically for industrial automation. According to 3GPP specifications, 5G URLLC targets a one-way latency of 1 millisecond with a packet error rate of 10⁻⁵, requirements that would be impossible to meet with pure ARQ alone given the retransmission round-trip overhead at typical industrial cell sizes.

Industrial Wireless Standards and Their Error-Control Choices

The major industrial wireless standards have made distinct choices about FEC and ARQ based on the latency and reliability profiles of their target applications. Understanding these choices illuminates the practical tradeoffs each standard’s designers prioritised.

WirelessHART (IEC 62591) and ISA-100.11a

WirelessHART and ISA-100.11a are the dominant standards for wireless process automation — monitoring pressure, temperature, and flow in oil refineries, chemical plants, and water treatment facilities. Both standards are built on IEEE 802.15.4 physical and MAC layers and employ mesh network topologies with multi-hop routing. Their primary error-control mechanism is ARQ-based retransmission: if a packet fails its CRC check at any hop, the receiving node requests retransmission from the previous hop. The relatively relaxed latency requirements of process monitoring (update rates of seconds to tens of seconds) make ARQ’s variable latency entirely acceptable, while ARQ’s bandwidth efficiency suits the low-data-rate, battery-powered sensor nodes common in process plants. Standards documentation is maintained by ISA.

IEEE 802.15.4 and Optional FEC

The IEEE 802.15.4 standard — the physical and MAC foundation for WirelessHART, ISA-100.11a, Zigbee, and Thread — includes an optional FEC mode in its O-QPSK physical layer. When enabled, a rate-1/2 convolutional code is applied to the data stream before transmission, doubling the bandwidth overhead but significantly improving link margin in noisy environments. Most industrial deployments do not enable this optional FEC mode, preferring to rely on ARQ retransmission and mesh path diversity for reliability.

5G URLLC for Factory Automation

5G NR’s URLLC service class represents the state of the art in industrial wireless error control. It uses LDPC codes for data channels and Polar codes for control channels — both chosen for their near-Shannon-limit performance at the short block lengths characteristic of industrial control messages. HARQ with Incremental Redundancy provides the retransmission layer. The combination achieves the 10⁻⁵ packet error rate and sub-millisecond latency required for closed-loop motion control and collaborative robotics, as documented by ETSI in its 5G standards series.

Selecting the Right Protocol for Your Industrial Application

Choosing between FEC, ARQ, and HARQ for an industrial wireless deployment requires a structured evaluation of three application-level parameters: the latency budget, the required packet error rate, and the expected channel conditions. No single protocol is universally superior — the right choice depends on the specific combination of these factors in the target environment.

Applications with hard real-time latency requirements — closed-loop motion control, safety-instrumented systems, and synchronised multi-axis robotics — should favour FEC or HARQ with aggressive FEC coding rates. The bounded latency of FEC decoding is a non-negotiable property for these use cases. Applications with soft real-time or non-real-time requirements — process monitoring, asset tracking, and condition monitoring — can tolerate ARQ’s variable latency and benefit from its superior bandwidth efficiency, particularly where sensor nodes are battery-powered and spectrum is shared.

Channel conditions also drive the choice. In environments with predominantly random bit errors (Gaussian noise channels), even a modest FEC code can eliminate nearly all errors without retransmission. In environments with bursty errors caused by impulsive interference — common near industrial power electronics — burst-error-correcting codes such as Reed-Solomon, or HARQ with Incremental Redundancy, perform significantly better than either pure FEC or pure ARQ alone. The ITU provides channel modelling guidelines relevant to industrial radio environments that engineers can use to characterise their specific deployment conditions before selecting an error-control strategy.