Noise estimation method by cyclic redundancy check in multi-hop decode-and-forward systems

Abstract

The present invention relates to communication in multi-hop decode-and-forward communication systems and, in particular, to a noise estimation method using Cyclic Redundancy Check (CRC) for error correction.

Description

[0001] NOISE ESTIMATION METHOD BY CYCLIC REDUNDANCY CHECK IN MULTIHOP DECODE-AND-FORWARD SYSTEMS

[0002] Technical Field of the Invention

[0003] The present invention relates to communication in multi-hop decode-and-forward communication systems and, in particular, to a noise estimation method using Cyclic Redundancy Check (CRC) for error correction.

[0004] State of the Art

[0005] The use of various network nodes, such as Integrated Access Backhaul (IAB) and Network-Controlled Repeaters (NCR), is considered essential for enabling high- frequency communications in next-generation network technologies. From the perspective of mobile networks, these nodes are expected to be deployed as ultra- dense networks to address the increasing demand for capacity in future use and the challenges associated with communication in the higher part of the spectrum. Ultra- dense networks, in this context, also necessitate a widespread, high-capacity main carrier solution. It is known that distributing fibre optics to all base stations is generally very costly for network operators.

[0006] Multi-hop transmission is already a technique used in LTE-Advanced and 5G, and it is expected to become increasingly important in 6G networks. This technique not only optimises network performance but also enhances connectivity. The integration of multi-hop transmission into 6G networks is anticipated to further improve both network performance and connection quality. Additionally, it should be noted that Cyclic Redundancy Check (CRC) plays a crucial role in detecting errors in Integrated Access Backhaul (IAB) and Network-Controlled Repeaters (NCR).

[0007] Multi-hop decode-and-forward systems, which are expected to be used in future wireless networks, are exposed to critical levels of errors due to the high frequencies at which they will operate. It is known that, at millimetre-wave frequencies, communication signals experience greater path loss and are more prone to reflections rather than scattering. In multi-hop decode-and-forward systems, such challenging channel conditions may lead to problems such as reduced communication quality and coverage.

[0008] Although various proposals and implementations have been developed in the prior art for algorithms using Cyclic Redundancy Check (CRC) and noise estimation, these developments remain insufficient. Some applications related to inventions developed for this purpose are provided below.

[0009] Patent document no “US11381260B2” in the state of the art is reviewed. The invention subject to this application relates to the field of data decoding and, specifically, to decoding using noise estimation. It provides a solution to the problem of designing a decoder (GRAND) based on noise estimation over candidate error sequences for a noisy channel. In the decoding process, candidate error sequences are individually added to the received codeword from the channel, using a list ordered by decreasing likelihood, until a positive result is obtained from a codebook membership test. The error sequence that yields a positive result is added to the received codeword to produce the decoding output. This document specifies the design of GRAND and pertains solely to the design of the GRAND decoder; it does not include the application of GRAND / ORBGRAND in multi-hop decode-and-forward systems together with CRC.

[0010] Patent document no “CN102938686A” in the state of the art is reviewed. The invention subject to this application belongs to the field of wireless communication and relates to a high-diversity decode-and-forward cooperative communication method used for cooperative communication in a wireless network. The invention addresses the problem of improving communication robustness and reducing bit error probability in decode-and-forward systems. In solving this problem, it utilises a cooperative communication method that incorporates high-degree diversity. Two rotation symbol components are derived from two original symbols. The components are transmitted in four stages. Relay nodes adjust cooperative bit statuses, while the destination node applies maximum composite demodulation. Rotation symbol components are obtained from the original symbols, but high-degree diversity is not employed. Here, the gain of near-maximum likelihood decoding provided by GRAND / ORBGRAND is used to reduce errors. Patent document no “CN117560019A” in the state of the art is reviewed. The invention subject to this application belongs to the technical field of channel coding and decoding and relates to an ORB-GRAND optimisation method for PAC codes. In particular, it offers a solution to the problem of optimising the block error probability of high-rate PAC codes using ORB-GRAND. The problem is addressed by adding CRC to the code to improve the distance characteristics of the PAC code through secondary control. As in ORBGRAND, the binary noise sequences to be tested are ordered based on bit reliabilities. It solves the problem of reducing the block error probability of PAC codes using ORBGRAND. However, no optimisation related to PAC codes is targeted here, and the application of GRAND / ORBGRAND to multi-hop decode-and-forward systems is not designed.

[0011] There remain unmet needs in the prior art concerning algorithms that use Cyclic Redundancy Check (CRC) and noise estimation.

[0012] As a result, due to the drawbacks mentioned above and the inadequacy of current solutions regarding the subject matter, a development in the relevant technical field has become necessary.

[0013] The Aim of the Invention

[0014] The primary aim of the invention is to transform the error control mechanism provided by CRC into an error correction mechanism by integrating CRC with GRAND under various scenarios in multi-hop decode-and-forward systems, thereby improving error performance.

[0015] Another aim of the invention is to ensure that the signal reaches the destination intact by fully eliminating errors through the GRAND algorithm, thus avoiding the delay caused by retransmission.

[0016] A further aim of the invention is to leverage the error-reducing capability of GRAND in multi-hop decode-and-forward systems that utilise CRC.

[0017] Another aim of the invention is the use of a noise estimation algorithm, adapted to the defined scenario, in relay-based systems through efficient software implementations and specialised hardware solutions. An additional aim of the invention is to enable more effective error correction and communication with lower complexity through the use of GRAND.

[0018] The structural and characteristic features of the invention, along with all of its advantages, will become more clearly understood through the figures provided below and the detailed description referring to these figures. For this reason, the evaluation should be made by taking these figures and detailed description into consideration.

[0019] Description of Drawings

[0020] FIGURE 1 is a diagram illustrating the flowchart of a multi-hop system utilising the noise estimation algorithm subject to the invention.

[0021] FIGURE 2 is a diagram illustrating the flowchart of decode-and-forward using GRAND in conjunction with the noise estimation algorithm subject to the invention.

[0022] Reference Numbers

[0023] DM. Demodulation

[0024] G. GRAND decoding

[0025] M. Modulation

[0026] Description of the invention

[0027] The present invention relates to communication in multi-hop decode-and-forward communication systems and, in particular, to a noise estimation method using Cyclic Redundancy Check (CRC) for error correction. The method subject to the present invention is executed and operated on a communication information system. In addition, the method may also be executed on a controller / microcontroller or on any electronic device equipped with a controller.

[0028] While Cyclic Redundancy Check (CRC) is traditionally used solely for error detection in decode-and-forward systems, with the integration of the GRAND (Guessing Random Additive Noise Decoding) algorithm into this system, CRC can also be used for error correction. The invention proposes the use of the GRAND algorithm with CRC for error correction in communication systems, under different scenarios. In this way, it is used in multi-hop decode-and-forward systems to reduce errors and address the issue of improving connection quality and coverage.

[0029] While the GRAND algorithm offers an innovative solution for error correction in noisy communication channels based on noise estimation, GRAND algorithms that enable maximum likelihood decoding for both hard and soft decoding already exist. These algorithms are designed to improve performance over generalised fading channels. In multi-hop decode-and-forward systems that use CRC, the error-reducing feature of GRAND is turned into an advantage. Therefore, it involves the optimisation of multihop communication systems and the integration of noise estimation-based error correction algorithms into such systems.

[0030] There are two different transmission system models: GRAND-Based (Single-Hop) Transmission and GRAND-Based (Multi-Hop) Transmission. In the GRAND-Based (Single-Hop) Transmission model; when single-input single-output (SISO) wireless communication is considered and a single hop is assumed on top of it, the communication occurs over generalised fading channels. In this case, the relationship between the transmitted and received symbol blocks is expressed as:

[0031] The vector represents the transmitted block consisting of M modulated symbols. Similarly, represents the received block consisting of M noisy modulated symbols. The transmitted block x is subjected to the channel effect characterised by the diagonal matrix which represents point-to-point fading. The vector represents AWGN (additive white Gaussian noise). The transmitted symbols xm, where (1 < m < M), are selected with equal probability from a normalised complex constellation X, and it is assumed that It is assumed that each symbol carries bits. The bit representation of the symbol xmcan be expressed as follows: Therefore, it can be stated that each transmitted block carries N = qM bits, and it is represented as follows:

[0032] It is expressed in the form |c| = N. Without loss of generality, c is assumed to be protected by CRC. Here, CRC can be represented by a one-to-one function CRC with a code rate of "R = K / N", and the domain of this function is equivalent to a codebook. This codebook, contains all possible codewords. Here, b represents the uncoded bit vector. According to the conventional perspective, by conditioning on the received bits a machine learning decoder calculates the probability that each codeword was transmitted, and the codeword cMLwith the highest probability among the codewords in the set C is selected, resulting in the following equation:

[0033] Here, the received bits represented by "c" may contain errors denoted by the vector “e”. “e” arises due to various factors such as added noise and interference from different sources, channel fading, and hardware imperfections. These factors may distort the transmission and cause inconsistencies between the bit representations of the transmitted block and the received block Therefore, the received bits can be expressed as c e F" , the transmitted bits as "c, and the erroneous bits — also referred to as additive noise — as That is:

[0034] Here, ®, denotes the operation. Using the equation , the additive noise can be expressed in terms of the transmitted and received bits as Accordingly, the given machine learning decoder can be simplified.

[0035]

[0036] Here, in the transition from step (a) to (b), the denominator can be neglected using , since it is assumed that the symbols are selected with equal probability for transmission. GRAND, instead of searching codewords in machine learning-based decoding, seeks the most likely random additive noise, that is, e = which corrupts the transmitted bits, by leveraging the error detection capability of the decoder and CRC. For this purpose, all possible additive noise sequences are ordered in descending likelihood. This alternative decoding approach is known as GRAND. The GRAND solution is defined as follows:

[0037] Here, the GRAND algorithm actively attempts to find the most likely additive noise (error sequence), thereby maximising the probability that the CRC returns a successful result. In this way, the transmitted bits are recovered with maximum accuracy, that is:

[0038] The term on the right-hand side of the equation above represents the bits detected by GRAND.

[0039] Additionally, since the entropy of the additive noise is generally lower than that of the information bits, the use of GRAND allows decoding complexity to be reduced while maintaining optimal error bounds. This increase in efficiency relies on two main strategies: one is to rank the error sequences in descending order of their probabilities, and the other is to stop the error estimation after a predefined computational limit is reached.

[0040] The other of the two system models is known as GRAND-Based (Multi-Hop) Transmission.

[0041] Considering a multi-hop (L-hop) DF system model, in this model, information is transmitted from the source node (S) to the destination node (£)) via "L - 1" relay nodes, denoted as This system uses the DF methodology, in which each relay node decodes the received signal and then re-encodes and forwards it for the next hop. The successive transmissions between adjacent nodes take place over generalised fading channels, which are assumed to be subject to channel impairments such as path loss, shadowing, and multipath fading, and it is also assumed that AWGN is added to the signal.

[0042] To eliminate the error in the received bits using GRAND and to enhance the robustness of multi-hop transmission over generalised fading channels, two approaches are proposed. In the first scenario, the GRAND technique is used only at the destination node D to estimate and eliminate the error sequence added to the received bits from the source node S through multi-hop relaying. Here, "c" represents a code block. The source node S protects c with a CRC code, modulates it, and then obtains the symbol block "x". Then, "x" is transmitted to the destination node D, reaching S through a series of relay nodes. The relay node Rereceives the noisy symbol block y{from the previous relay node R{-1. Here, Rorepresents the source node S'. The noisy symbol block y{is expressed as follows:

[0043] Here denotes the symbol block transmitted by the relay node and denotes the symbol block transmitted by the source node Figure 1 visually summarises the transmission in the multi-hop system. The relay node decodes as , as explained below. represents the additive noise that occurs during the transmission from relay node Rt-rto relay node R{, and therefore c0denotes the bit representation of the symbol block c transmitted by the source node S. At the end of all transmissions throughout the entire relay node chain, the received bit representation (c) at the destination node D is found as: denotes the total additive noise that occurs in multi-hop decode-and-forward transmission.

[0044] Finally, to reduce the binary noise bits exadded by the destination node D, the GRAND technique is used in the form of.

[0045] By estimating the most likely error sequence, GRAND maximises the probability In this way, it ensures the best possible recovery of the transmitted bits. In other words, the transmitted bits satisfy the equation Accordingly, the BER (Bit Error Rate) calculation is performed as follows:

[0046] This equation gives the received bit error rate when GRAND is applied only at the final destination node (scenario 1 ).

[0047] In the second scenario, each relay node Reapplies GRAND to the bits received from the previous relay node Here, the relay ois the source node S. This noise reduction approach applies noise estimation and elimination processes throughout the transmission chain. At each relay node, the GRAND technique is used to accurately estimate and suppress the noise bits added before the received signal is forwarded to the next relay. This repeated noise suppression process at each relay node contributes to improved signal quality and enhanced protection against channel impairments. One of the limitations of CRC is its vulnerability to undetectable errors. Although CRC is designed to detect errors, it may fail to recognise certain types of errors. For example, when the number of corrupted bits in a message equals the degree of the CRC polynomial, the errors may cancel each other out, causing CRC to produce a false positive result. As a result, CRC may not always correctly detect when errors have occurred, which can lead to data corruption or communication issues. In accordance with these stages, since each relay node applies GRAND decoding, the received bit block at the destination node "D" is obtained as at the end of all transmissions through the relay node chain, here e is:

[0048] If the GRAND decoding method naturally fails to detect an additive noise that results in a codeword included in the CRC codebook C, then the sum of this noise and the transmitted codeword constitutes a valid codeword. These estimated noise bits, denoted by for each hop, go undetected and the error is not identified by the CRC. The codeword obtained as a result of GRAND at the destination node can be calculated as:

[0049] As transmitted bit block is BER can be written as follows. This equation gives the received bit error rate when GRAND is applied at each relay and destination node (scenario 2).

[0050] The two scenarios mentioned above involve transmission over generalised fading channels. The choice of scenario depends on the trade-off between computational complexity, delay, and channel variations. Scenario 1 simplifies the transmission but places a greater noise suppression burden on the destination node. In contrast, Scenario 2 distributes the noise suppression responsibility across all relay nodes, which increases system robustness while potentially also increasing the computational load.

[0051] As a result, this approach enables the elimination of errors by applying GRAND technology at relay nodes or the destination node depending on the scenario. In a classical decode-and-forward system, the operations applied to the signal by the destination node or relays are visible. In such systems, when CRC detects an error, the initially transmitted signal must be retransmitted.

[0052] In the multi-hop decode-and-forward system, the steps applied to the signal by the relay nodes using GRAND are shown in Figure 2. Figure 2 includes the steps performed by a processor equipped with an interface. The relay node receives the noisy symbol block from the previous relay node Within the relay node, after demodulation is performed in the DM section, the process continues to G as In G, using CRC, GRAND decoding is applied, and the result is forwarded to M for modulation as . In this way, the transmitted symbol block is sent from the relay node. The complete elimination of error by the GRAND algorithm means that the signal reaches the destination intact and avoids the delay caused by retransmission.

[0053] This technology aims to reduce block and bit errors while operating with lower complexity compared to other similar systems. Especially when used at the final relay, it enables error reduction while keeping system complexity to a minimum. Furthermore, when used at every relay, it can be applied without introducing additional complexity for error reduction. GRAND can be easily integrated into existing communication systems without requiring any changes to the communication protocol or hardware structure. Thanks to these features, it can be readily used in various communication systems, including 5G and potential 6G networks, thereby enabling more efficient and reliable communication. These advantages highlight the value and uniqueness of the invention, as GRAND provides more effective error correction and communication capability with reduced complexity. In the context of 5G NR and 6G systems, where ultra-reliable low-latency communication (URLLC) and energy efficiency are critical, the proposed use of GRAND supports these requirements by enhancing reliability without increasing overhead or power consumption. Its compatibility with existing standards also aligns with the need for scalable and backward-compatible solutions in next-generation wireless communication systems.

[0054] The invention includes: the use of GRAND with CRC in multi-hop decode-and-forward systems for error correction to reduce block and bit errors; the use of GRAND only at the final relay to reduce block and bit errors without significantly increasing complexity; the use of GRAND at every relay to greatly reduce block and bit errors at the expense of increased complexity; and the application of GRAND in 5G, potential 6G, and other existing communication systems without requiring any changes in the communication protocol or hardware structure.

Claims

CLAIMS1. A method for communication in multi-hop decode-and-forward communication systems and, in particular, for error correction using Cyclic Redundancy Check (CRC) and noise estimation, characterised in that the communication system performs the process steps of,- receiving the bits from the source node and performing estimation and suppression through the relay nodes,- eliminating the errors in the received bits using GRAND and increasing the robustness of multi-hop transmission over generalised fading channels,- using GRAND at each relay node to accurately estimate and suppress the noise bits added before the received signal is forwarded to the next relay,- applying the noise reduction process, including noise estimation and elimination, throughout the transmission chain, and- applying GRAND decoding at the final relay node and receiving the transmitted symbol block at the destination node "D" at the end of all transmissions through the relay node chain.

2. A Cyclic Redundancy Check (CRC) and noise estimation method according to Claim 1 , characterised in that, in the step of performing estimation and suppression of the bits received from the source node through the relay nodes, the communication system performs the steps of:- receiving, by the relay node, the noisy symbol block from the previous relay node,- performing demodulation within the relay node, followed by applying GRAND decoding to the bit block using CRC,- applying modulation to the bit block that has undergone GRAND decoding using CRC, andtransmitting the transmitted symbol block resulting from the modulation process from the relay node.

3. A Cyclic Redundancy Check (CRC) and noise estimation method according to Claim 1 , characterised in that, in the process step of eliminating the error in the received bits using GRAND and increasing the robustness of multi-hop transmission over generalised fading channels, the decoder comprises the process step of• ensuring the application of GRAND only at the final destination node by means of the equationand providing the received bit error rate,GRAND estimating and suppressing the error sequence added to the received bits from the source node S through multi-hop relaying only at the destination node D.

4. A Cyclic Redundancy Check (CRC) and noise estimation method according to Claim 1 , characterised in that, in the process step of eliminating the error in the received bits using GRAND and increasing the robustness of multi-hop transmission over generalised fading channels, the decoder comprises the process step of• ensuring the application of GRAND at each relay and destination node by means of the equationand providing the received bit error rate.each relay node{applying GRAND to the bits received from the previous relay node

Images