Channel-coded optical decoding method and decoding device
By combining optical phase shifters and optical interferometers with the quantum-inspired algorithm dSB, the power consumption and signal delay problems of electronic decoders in the high-speed direction are solved, and the performance of LDPC decoding is improved by achieving high efficiency.
Patent Information
- Application Number
- CN202510622096.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-14
- Publication Date
- 2026-11-17
AI Technical Summary
Electronic decoders face power consumption and signal delay bottlenecks in their development towards high speeds, which are difficult to solve effectively with existing technologies.
An optical decoding method is adopted, which uses an optical phase shifter and an optical interferometer to realize LDPC decoding. Combined with the quantum-inspired dSB algorithm, iterative decoding is performed through optical computation, making full use of the sparsity of the parity-check matrix to reduce hardware complexity and power consumption.
It significantly improves decoding performance, reduces power consumption, increases decoding speed and system throughput, and adapts to the needs of high-speed and low-latency scenarios.
Smart Images

Figure CN123122736A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of communications, and more specifically, to a channel-coded optical decoding method and decoding apparatus. Background Technology
[0002] LDPC (Laser-Dependent Proof-of-Stake) is a linear block code based on a sparse parity-check matrix (SPC) and has become a core coding technology in modern wireless communication systems such as mobile communication systems and Wi-Fi systems. The core advantage of LDPC codes lies in their ability to approach the Shannon limit through iterative decoding algorithms, significantly improving spectral efficiency. Simultaneously, the sparse parity-check matrix structure supports low-complexity, high-parallelism hardware implementation, adapting to the demands of high-speed, low-latency scenarios. The flexibility and scalability of LDPC codes allow for dynamic adjustment of code length and code rate to cope with channel interference, and combined with interleaver design, it can effectively resist sudden errors in the wireless environment. In the 5G NR standard, it was selected by 3GPP as the eMBB data channel coding scheme, supporting a peak rate of 20Gbps, and is widely used in Wi-Fi 6's massive MIMO transmission and low signal-to-noise ratio (SNR) links in satellite communication. With its low power consumption and simplified architecture, it has become a key technological foundation for IoT terminals and future terahertz communication.
[0003] As electronic decoders evolve towards higher speeds, they face numerous challenges. First, the limitations of Moore's Law mean that while chip size continues to shrink to improve performance, power consumption becomes increasingly prominent. Furthermore, the wiring within the internal networks of electronic chips has become more complex, leading to signal delay and increased power consumption, becoming a significant bottleneck for improving decoding speed. Summary of the Invention
[0004] This application provides a channel-coded optical decoding method and decoding apparatus that can improve decoding performance and reduce decoding power consumption.
[0005] Firstly, an optical decoding method is provided. This method can be executed by a communication device, which may be a communication equipment or a unit / module / component (such as a chip, chip system, logic circuit, or software) configurable in (or usable in) a communication equipment. For example, the communication equipment may be a terminal or a network device; the following description uses the execution of this method by a communication device as an example.
[0006] The method includes: In the t-th iteration, the communication device inputs a light source to at least one optical decoding unit to obtain the output result of the at least one optical decoding unit. The optical decoding unit includes at least one multiplication link, and each multiplication link includes multiple optical phase shifters. The number of multiplication links and the number of optical phase shifters in each multiplication link are determined based on the parity-check matrix of the channel coding, where t is a positive integer. Based on the output result of the at least one optical decoding unit, the communication device obtains the iteration result of the t-th iteration. This iteration result is a sequence containing N elements, where N is the code length of the channel coding, and N is a positive integer. Based on the iteration result of the t-th iteration, the communication device assigns values to the optical phase shifters included in each multiplication link of the at least one optical decoding unit. The communication device performs the (t+1)-th iteration, and repeats this process multiple times until the iteration result converges, resulting in a decoding sequence.
[0007] According to the above scheme, decoding of channel coding is realized based on optical decoding unit, which makes full use of the sparsity of parity check matrix to reduce hardware implementation complexity. Compared with decoding based on electrical energy, iterative calculation using optical energy can significantly reduce power consumption and significantly improve decoding performance.
[0008] In conjunction with the first aspect, in some implementations of the first aspect, the Hamiltonian obtained by solving the LDPC decoding problem based on the discrete simulation bifurcation algorithm is... satisfy:
[0009]
[0010] Where σ = 1 - 2x, x is a coded sequence of code length N, and r i It is the demodulated value corresponding to the i-th symbol in x obtained after demodulation of the received signal by the receiving end, where K is the number of parity bits and N is the number of parity bits. k Let k be the number of non-zero elements in the k-th row of the parity check matrix. m Let λ be the column index of the m-th non-zero element in the k-th row, λ be the hyperparameter of positive real numbers in the LDPC decoding problem, and K and N be the column indices of the m-th non-zero element in the k-th row. k m is less than N, k is less than or equal to K, k m Less than or equal to N k ;
[0011] The Hamiltonian Compared to the i-th symbol x in the decoded sequence i gradient for:
[0012]
[0013] Among them, W i Is with x i The corresponding number of non-zero elements in the i-th column of the parity check matrix. It is the k-th non-zero element in the i-th column of the parity check matrix. w The number of non-zero elements in the row. It is the kth w OK Dividing x from non-zero elements i The value of the symbol corresponding to the m-th non-zero element (excluding the corresponding non-zero element) in the t-th iteration, and the at least one chain link is used to implement the gradient. Multiplication terms in
[0014] According to the above scheme, the quantum-inspired dSB algorithm can significantly improve decoding performance compared to current decoding algorithms (such as sum-product decoding). It hardens the LDPC decoding problem to a photonic decoder, fully utilizing the sparsity of the parity-check matrix to reduce hardware implementation complexity. Using light energy for iterative computation significantly reduces power consumption and improves decoding performance compared to decoding based on electrical energy.
[0015] In conjunction with the first aspect, in some implementations of the first aspect, the optical decoding unit further includes a bias link and / or a coherent summation structure, the bias link including an optical interferometer, and the coherent summation structure being used to coherently sum the light beams of the light source passing through the at least one optical decoding unit and the bias link.
[0016] For example, the bias link is used to implement the following bias term in the Hamiltonian relative to the gradient of the i-th symbol in the decoded sequence:
[0017]
[0018] The Hamiltonian is obtained by solving the LDPC decoding problem using the discrete simulation bifurcation algorithm.
[0019] Optionally, the link may also include an optical interferometer.
[0020] In conjunction with the first aspect, in some implementations of the first aspect, the optical interferometer is located at the end of the chain link, and the optical interferometer is used to realize the chain accumulation term in the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence. of Scaling.
[0021] Based on the above scheme, implementing other computational tasks in the dSB decoding algorithm in the optical domain can further reduce decoding power consumption and improve decoding efficiency.
[0022] In conjunction with the first aspect, in some implementations of the first aspect, the at least one optical decoding unit includes N optical decoding units, each of which corresponds one-to-one with one of the N symbols in the decoding sequence. The output result includes the beam output by each of the N optical decoding units corresponding to the corresponding symbol, where N is an integer greater than 1. Specifically, the i-th symbol in the decoding sequence corresponds to the first optical decoding unit, where i is a positive integer less than or equal to N, and the first optical decoding unit outputs the beam corresponding to the i-th symbol.
[0023] The i-th symbol corresponds to the i-th column of the parity check matrix, which includes W. i The first optical decoding unit includes W non-zero elements. i A multiplication link, the W i A link and the W i W of the parity matrix containing the non-zero elements i Each row corresponds one-to-one with the other, and the w-th non-zero element in the i-th column is located at the k-th column of the parity matrix. w Okay, the kth w Lines include The k-th non-zero element w The corresponding chain links include An optical phase shifter, A light phase shifter and Each code element corresponds one-to-one. The code element is this Non-zero elements divided by x i Other than the corresponding non-zero elements The code corresponding to each non-zero element Each optical phase shifter is assigned a value based on the element corresponding to the corresponding symbol in the previous iteration result.
[0024] According to the above scheme, ultra-high decoding throughput can be provided through the parallel expansion of multiple decoding units. The connection relationship of multiple optical decoding units can be freely configured, and different code lengths and code rates can be mapped to parity check matrices, supporting large-scale multi-scenario application expansion.
[0025] In conjunction with the first aspect, in some implementations of the first aspect, the at least one optical decoding unit includes an optical decoding unit comprising K chain links, the K chain links corresponding one-to-one with the K rows of the parity-check matrix, and the k-th row of the parity-check matrix comprising N k The k-th row contains N non-zero elements, and the chain link corresponding to this chain includes N. k A light phase shifter, the N k One optical phase shifter and N k Each code element corresponds one-to-one with the N code. k Each code element is the decoded sequence and the N. kThe code corresponding to each of the N non-zero elements k Each optical phase shifter is assigned a value based on the element corresponding to the corresponding symbol in the previous iteration result.
[0026] According to the above scheme, for LDPC decoding, a dedicated decoding module based on the quantum-inspired algorithm dSB was designed based on the parity check matrix. This fully utilizes the sparsity of the parity check matrix to realize a simplified optical decoding unit, which can improve decoding performance while reducing hardware overhead.
[0027] In conjunction with the first aspect, in some implementations of the first aspect, the optical decoding unit further includes a broadcast multiplexing route and N beam combining nodes, wherein the N beam combining nodes correspond one-to-one with the N symbols in the decoding sequence, and the broadcast multiplexing route routes the output beam of the multiplication link corresponding to the k-th row to the N beam combining nodes. k N corresponding to each code element k There are N beam combining nodes. The output includes N beams, each corresponding to one of the N symbols, output by the N beam combining nodes.
[0028] According to the above scheme, the characteristics of the parity check matrix are used to realize the broadcast multiplexing of photons, thereby realizing a simplified optical decoding unit. This improves decoding performance while reducing hardware overhead.
[0029] In conjunction with the first aspect, in some implementations of the first aspect, the communication device obtains the iteration result of the t-th iteration based on the output result of the at least one optical decoding unit, including: the communication device converts the optical domain output result of the at least one optical decoding unit into an electrical domain output result using a detector. Based on the electrical domain output result, the communication device obtains the gradient of the Hamiltonian relative to the i-th symbol in the decoding sequence. And, based on this gradient The iteration result x of the t-th iteration of the i-th symbol is obtained. i (t+1). Wherein, this gradient... and the iteration result x i (t+1) satisfies:
[0030] x i (t+1)=x i (t)+a0y i (t+1)Δ t ,
[0031]
[0032] Where a0 is the scaling factor for each iteration, Δ t This represents the step size for each iteration update.
[0033] Secondly, an optical decoding method is provided. This method can be executed by a communication device, which may be a communication equipment or a unit / module / component (such as a chip, chip system, logic circuit, or software) configurable in (or usable in) a communication equipment. For example, the communication equipment may be a terminal or a network device; the following description uses the execution of this method by a communication device as an example.
[0034] The method includes: In the t-th iteration, the communication device randomly selects m variables from the sequence x(t), changes the values of these m variables, and obtains the sequence x′(t). The communication device modulates the optical signal based on the sequence x′(t), and inputs the modulated optical signal into the optical matrix part. The optical calculation result is used to determine the Hamiltonian of the t-th iteration. Based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration, the communication device obtains the iteration result x(t+1) of the t-th iteration. This iteration result x(t+1) is used for the (t+1)-th iteration, and the iterations are repeated until the sequence iteration result converges. The converged iteration result is the decoding sequence.
[0035] According to the above scheme, the Hamiltonian is calculated through an optical decoding unit, thus completing the solution of the quantum-inspired algorithm. This offers significant decoding performance advantages, reducing decoding power consumption and improving decoding efficiency. In conjunction with the second aspect, in some implementations of the second aspect, the HUBO-form Hamiltonian obtained by solving the LDPC decoding problem based on the discrete simulation bifurcation algorithm is used. satisfy:
[0036]
[0037] Where σ = 1 - 2x, x is a coded sequence of code length N, and r i It is the demodulated value corresponding to the i-th symbol in x obtained after demodulation of the received signal by the receiving end, where K is the number of parity bits and N is the number of parity bits. k Let k be the number of non-zero elements in the k-th row of the parity check matrix. m Let λ be the column index of the m-th non-zero element in the k-th row, λ be the hyperparameter of positive real numbers in the LDPC decoding problem, and K and N be the column indices of the m-th non-zero element in the k-th row. k m is less than N, k is less than or equal to K, k m Less than or equal to N k ;
[0038] Hamiltonian based on this HUBO form The Hamiltonian in QUBO form obtained after adding auxiliary bits satisfies:
[0039]
[0040] The optical matrix section is configured as follows: When the input to the optical matrix is a sequence x′(t), the amplitude of the output beam is E. out =Ux′(t), the calculated result of this light beam is the light intensity I. out =|E out | 2 =(Ux′(t)) T (Ux′(t)).
[0041] In conjunction with the second aspect, in some implementations of the second aspect, the communication device obtains the sequence x(t+1) based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration, including: the communication device acquiring the change ΔH of the Hamiltonian of the t-th iteration compared to the Hamiltonian of the (t-1)-th iteration. The communication device receives the sequence x′(t) with probability min[1,exp(-ΔH / T)] to obtain the sequence x(t+1), where T is the annealing temperature in the simulated annealing algorithm.
[0042] Thirdly, a communication device is provided. In one design, the device may include modules corresponding to the methods / operations / steps / actions described in the first aspect or any embodiment of the first aspect. These modules may be hardware circuits, software, or a combination of hardware circuits and software. In one design, the device includes at least one optical decoding unit, a processing unit, and a detection unit. The optical decoding unit includes at least one multiplication link, which includes multiple optical phase shifters. The number of multiplication links and the number of optical phase shifters in the multiplication link are determined based on the parity-check matrix of the channel coding, where t is a positive integer. The processing unit is used to control the light source input to the at least one optical decoding unit in the t-th iteration to obtain the electrical domain output result of the at least one optical decoding unit. The detection unit is used to convert the electrical domain output result into an electrical domain output result. The processing unit is also used to obtain the iteration result of the t-th iteration based on the electrical domain output result. The iteration result is a sequence containing N elements, where N is the code length of the channel coding, and N is a positive integer. The processing unit is also used to assign values to the optical phase shifters contained in each multiplication link of the at least one optical decoding unit based on the iteration result of the t-th iteration, and then use them for the (t+1)-th iteration until the iteration result converges and the obtained iteration result is the decoding sequence.
[0043] Fourthly, a communication device is provided. In one design, the device may include modules corresponding to the methods / operations / steps / actions described in the second aspect or any of the embodiments of the second aspect. These modules may be hardware circuits, software, or a combination of hardware circuits and software. In one design, the device includes a processing unit, an optical modulation unit, and an optical matrix unit. The processing unit is used to randomly select m variables from a sequence x(t) in the t-th iteration, change the values of the m variables, and obtain a sequence x′(t). The optical modulation unit is used to modulate an optical signal based on the sequence x′(t) to obtain a modulated optical signal. The optical matrix unit is used to output an optical calculation result based on the input modulated optical signal, and this optical calculation result is used to determine the Hamiltonian of the t-th iteration. The processing unit is also used to obtain a sequence x(t+1) based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration. This sequence x(t+1) is used in the (t+1)-th iteration, and the iterations are repeated until the sequence x converges. The converged sequence x is the decoding sequence.
[0044] Fifthly, a communication device is provided, including a processor. The processor can implement the methods of the first to second aspects and any possible implementations thereof. Optionally, the communication device further includes a memory, and the processor is coupled to the memory and can be used to execute instructions in the memory to implement the methods of the first to second aspects and any possible implementations thereof. Optionally, the communication device further includes a communication interface, and the processor is coupled to the communication interface. In the embodiments of this application, the communication interface may be a transceiver, a pin, a circuit, a bus, a module, or other types of communication interface, and is not limited thereto.
[0045] In one implementation, the communication device is a communication equipment (such as a terminal device or access network equipment). When the communication device is a communication equipment, the communication interface can be a transceiver, or an input / output interface.
[0046] In another implementation, the communication device is a chip configured within a communication device. When the communication device is a chip configured within a communication device, the communication interface can be an input / output interface.
[0047] Optionally, the transceiver can be a transceiver circuit. Optionally, the input / output interface can be an input / output circuit.
[0048] A sixth aspect provides a processor, comprising: an input circuit, an output circuit, and a processing circuit. The processing circuit is configured to receive signals through the input circuit and transmit signals through the output circuit, causing the processor to execute the methods described in the first to second aspects and any possible implementation thereof.
[0049] In specific implementation, the processor can be one or more chips, the input circuit can be input pins, the output circuit can be output pins, and the processing circuit can be transistors, gate circuits, flip-flops, and various logic circuits. The input signal received by the input circuit can be received and input by, for example, but not limited to, a receiver, and the signal output by the output circuit can be, for example, but not limited to, output to and transmitted by a transmitter. Furthermore, the input circuit and the output circuit can be the same circuit, which is used as both the input circuit and the output circuit at different times. This application does not limit the specific implementation of the processor and various circuits.
[0050] In a seventh aspect, a computer program product is provided, comprising: a computer program (also referred to as code or instructions) that, when run, causes a computer to perform the methods described in the first to second aspects and any possible implementation thereof.
[0051] Eighthly, a computer-readable storage medium is provided that stores a computer program (also referred to as code or instructions) that, when executed on a computer, causes the computer to perform the methods described in the first to second aspects and any possible implementation thereof.
[0052] Ninthly, a communication system is provided, including a decoding device and an encoding device for any of the methods provided in the first or second aspect.
[0053] It should be understood that the beneficial effects of the features corresponding to the first or second aspect in the third to ninth aspects can be referred to the relevant descriptions of the first or second aspect above, and will not be repeated here. Attached Figure Description
[0054] Figure 1 This is a schematic diagram of the communication system architecture applicable to the embodiments of this application;
[0055] Figure 2 This is a schematic diagram of the structure of an optical decoding unit in a general optical decoding module provided in this application embodiment;
[0056] Figure 3 , Figure 4 This is a schematic diagram of different optical decoding units in the same decoding module provided in the embodiments of this application;
[0057] Figure 5 This is a schematic diagram of the structure of the optical decoding unit of the dedicated optical decoding module provided in the embodiments of this application;
[0058] Figure 6This is a schematic flowchart of an optical decoding method for channel coding provided in an embodiment of this application;
[0059] Figure 7 This is a schematic diagram of the structure of the QUBO-based optical decoding module provided in the embodiments of this application;
[0060] Figure 8 This is another schematic flowchart of the optical decoding method for channel coding provided in the embodiments of this application;
[0061] Figure 9 This is a schematic block diagram of an example of a communication device provided in an embodiment of this application;
[0062] Figure 10 This is a schematic structural diagram of another example of the communication device provided in the embodiments of this application;
[0063] Figure 11 This is a schematic structural diagram of another example of the communication device provided in the embodiments of this application. Detailed Implementation
[0064] To facilitate understanding of the embodiments of this application, the following description is provided first:
[0065] In this application, "instruction" can include direct instruction, indirect instruction, explicit instruction, and implicit instruction. When describing a certain instruction information for the purpose of instructing A, it can be understood that the instruction information carries A, directly instructs A, or indirectly instructs A.
[0066] In this application, " / " can indicate that the objects before and after are in an "or" relationship. For example, A / B can mean A or B. "And / or" can be used to describe three relationships between the related objects. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. A and B can be singular or plural.
[0067] In this application, "at least one" means one or more, and "more than one" means two or more, such as three, four, or more. Similar expressions (such as at least one, at least one, etc.) are used in the same way. "At least one of the following," "one or more of the following," or similar expressions refer to any combination of these items, which may include only a single item or a combination of multiple items. For example, at least one of a, b, or c can mean: a, or b, or c; a and b; or a and c; or b and c; or a, b, and c. Where a, b, and c can be single or multiple.
[0068] In this application, for the convenience of describing the technical solutions of the embodiments of this application, the terms "first" and "second" may be used to distinguish them. The terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.
[0069] In this application, the words "exemplary," "example," or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary," "example," or "for example" should not be construed as being more preferred or advantageous than other embodiments or designs. The use of the words "exemplary," "example," or "for example" is intended to present the relevant concepts in a specific manner to facilitate understanding.
[0070] In this application, "sending information / data" only indicates the direction of information / data transmission, including direct transmission via the device's communication interface (such as an air interface, or simply air interface). "Sending" can also be understood as the "output" of a module interface. "Sending" can include indirect transmission by the processing unit through the communication interface, meaning that after the processing unit outputs information / data through the module interface, it is transmitted to the device's communication interface and then sent out. "Receiving information / data" only indicates the direction of information / data transmission, including direct reception via the communication interface. "Receiving" can also be understood as the "input" of a module interface. "Receiving information / data" can include indirect reception by the processing unit through the communication interface, meaning that after the communication interface receives information / data, it is transmitted to the processing unit's module interface and then input to the processing unit. "Sending information / data to… (such as a terminal)" can be understood as the destination of the information being the terminal. It can include sending information / data directly or indirectly to the terminal. "Receiving information / data from… (such as a terminal)" can be understood as the source of the information being the terminal, and can include receiving information / data directly or indirectly from the terminal. Information / data may undergo necessary processing, such as format changes, between the source and destination, but the destination can understand the valid information / data from the source. Similar statements in this application can be understood in a similar way, and will not be repeated here.
[0071] The technical solutions of this application can be applied to various communication systems, such as Long Term Evolution (LTE) systems, 5th Generation (5G) communication systems, satellite communication systems, Wireless Fidelity (WiFi) systems, and the solutions provided in this application can also be applied to future communication systems or other communication systems. This application does not limit these applications.
[0072] Figure 1This is a schematic diagram illustrating yet another possible, non-limiting system. For example... Figure 1 As shown, the communication system 10 includes a radio access network (RAN) 100, a core network (CN) 200, and a data network (DN) 300. RAN 100 includes at least one RAN node (e.g., Figure 1 110a and 110b (collectively referred to as 110) and at least one terminal (such as Figure 1 RAN 100, denoted as RAN 120a-120j, is collectively referred to as RAN 120. RAN 100 may also include other RAN nodes, such as wireless relay equipment and / or wireless backhaul equipment. Figure 1 (Not shown in the image). Terminal 120 is connected to RAN node 110 wirelessly. Access network node (or RAN node) 110 is connected to core network 200 wirelessly or via wired connection. The core network equipment in core network 200 and access network node 110 in RAN 100 can be different physical devices, or they can be the same physical device integrating core network logical functions and wireless access network logical functions.
[0073] RAN 100 can be a cellular system related to the 3rd Generation Partnership Project (3GPP), such as 4G, 5G mobile communication systems, or future evolution systems. RAN 100 can also be an open access network (openRAN, O-RAN, or ORAN), a cloud radio access network (CRAN), or a wireless fidelity (WiFi) system. RAN 100 can also be a communication system that integrates two or more of the above systems.
[0074] Access network node 110, sometimes also referred to as access network equipment, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple access network nodes 110 in communication system 10 can be of the same type or different types. In some scenarios, the roles of access network node 110 and terminal 120 are relative, for example... Figure 1 Network element 120i can be a helicopter or a drone, and it can be configured as a mobile base station. For terminals 120j that access RAN 100 through network element 120i, network element 120i is a base station; however, for base station 110a, network element 120i is a terminal. Access network node 110 and terminal 120 are sometimes referred to as communication devices, for example... Figure 1Network elements 110a and 110b can be understood as communication devices with base station functions, while network elements 120a-120j can be understood as communication devices with terminal functions.
[0075] In one possible scenario, the access network node can be a base station, such as an evolved NodeB (eNodeB), a next-generation NodeB (gNB), or a base station in a future mobile communication system. The access network node can also be a macro base station (e.g., Figure 1 110a), micro base stations or indoor stations (such as Figure 1 The access network node can be a relay node or donor node (as described in 110b), or a wireless controller in a CRAN scenario. Alternatively, the access network node can be an access point (AP), a transmission reception point (TRP), or an access node in a WiFi system. Optionally, the access network node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network device in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). All or part of the functions of the access network node in this application can also be implemented through software functions running on hardware, or through virtualization functions instantiated on a platform (e.g., a cloud platform). The access network node in this application can also be a logical node, logical module, or software capable of implementing all or part of the access network node functions.
[0076] In another possible scenario, multiple access network nodes collaborate to assist the terminal in achieving wireless access, with each access network node performing a portion of the base station's functions. For example, access network nodes can be central units (CUs), distributed units (DUs), CU-control plane (CPs), CU-user plane (UPs), or radio units (RUs), etc. CUs and DUs can be set up separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).
[0077] A terminal can also be called a terminal device, user equipment (UE), mobile station, mobile terminal, etc. Terminals can be widely used for communication in various scenarios. These scenarios include, but are not limited to, at least one of the following: enhanced mobile broadband (eMBB), ultra-reliable low-latency communication (URLLC), massive machine-type communications (mMTC), device-to-device (D2D) communication, V2X, machine-type communication (MTC), Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, intelligent transportation, sensing terminals, terminals integrating communication and sensing, or smart cities, etc. A terminal can be a mobile phone (e.g., Figure 1 120a, 120j and 120e), tablet computers, computers with wireless transceiver capabilities (such as...) Figure 1 120g), customer-premises equipment (CPE), point-of-sale (POS) machines, wearable devices, vehicles (such as...) Figure 1 120b), drones, helicopters, aircraft (such as Figure 1 120i), ships, robots, robotic arms, sensors, detectors, or smart home devices (such as Figure 1 (e.g., 120h).
[0078] It should be understood that the decoding device provided in the embodiments of this application can be a communication device (such as...) Figure 1 The RAN node or terminal shown can also be a device / module (such as a chip, chip system, logic circuit or software) configured in (or usable in) a communication device, and this application does not limit it in this regard.
[0079] The technical solutions in this application will now be described with reference to the accompanying drawings.
[0080] 1. Low-density parity-check codes (LDPC)
[0081] LDPC is a linear block code based on a sparse parity-check matrix. Its core advantage lies in its ability to approach the Shannon limit through iterative decoding algorithms, significantly improving spectral efficiency. Furthermore, the sparse parity-check matrix structure supports low-complexity, high-parallelism hardware implementation, adapting to high-speed, low-latency scenarios. As the name suggests, the parity-check matrix of an LDPC contains only a small number of 1s, with the majority being 0s. Utilizing the sparsity of the parity-check matrix, the decoding complexity is linearly related only to the code length, ensuring that the decoding process remains relatively simple even with long code lengths.
[0082] In the 5G NR standard, it was selected by 3GPP as the eMBB data channel coding scheme, supporting a peak rate of 20Gbps. It is also widely used in Wi-Fi 6's massive MIMO transmission and satellite communication low signal-to-noise ratio links. With its low power consumption and simplified architecture, it has become a key technological foundation for IoT terminals and future terahertz communication.
[0083] 2. Quantum Annealing-Inspired Algorithm (QAIA)
[0084] QAIA is a class of classical optimization algorithms inspired by the quantum annealing process. Quantum annealing is an optimization technique based on quantum mechanics, aiming to find the optimal solution to a problem by simulating the evolution of a quantum system, physically achieved by adjusting the Hamiltonian to realize the evolution of the ground state. Quantum annealing-inspired algorithms simulate this process, using classical methods similar to simulated annealing to solve complex combinatorial optimization problems, such as graph coloring, the Traveling Salesman Problem (TSP), and the Max-Cut problem. These algorithms gradually converge to the global optimum by simulating the low-energy states of the system, possessing strong global search capabilities and effectively avoiding getting trapped in local optima. Although quantum annealing itself relies on quantum computing resources, the classical algorithms inspired by it can be implemented on conventional computers, demonstrating good practicality and broad application prospects.
[0085] 3. Discrete simulated bifurcation (dSB)
[0086] dSB is a heuristic optimization algorithm that combines simulated annealing and the bifurcation phenomenon in physics to improve the efficiency of the search process and the quality of the solution. The algorithm guides the search process step-by-step by simulating the transition of the system from a steady state to multiple possible solutions, using control parameters (similar to temperature). During iterations, the system avoids local optima by "jumping" to new states, enabling more efficient exploration of the solution space. The dSB algorithm demonstrates good performance and fast convergence when solving complex combinatorial optimization problems (such as graph optimization and constrained optimization).
[0087] 4. Simulated Annealing (SA) Algorithm
[0088] Stochastic Algorithm (SA) is a general-purpose global optimization algorithm that simulates the solid annealing process to find the global optimum in the search space. Its basic principle is to simulate the annealing process in physics, controlling the probability of accepting a suboptimal solution during the search by using a temperature parameter, thereby avoiding getting trapped in local optima.
[0089] Currently, the wiring of internal networks in electronic chips has become more complex, leading to increased signal delay and power consumption, which has become a significant bottleneck in improving decoding speed. Optical computing, as an emerging computing paradigm, is gradually gaining popularity. Compared to traditional electronic computing architectures, optical computing offers higher energy efficiency and greater computing power. Optical signal transmission is unaffected by electromagnetic interference, enabling lower latency and higher bandwidth. Based on this, this application proposes using optical computing to improve the decoding efficiency of channel coding. Specifically, it proposes a scheme using the discrete analog bifurcation algorithm (dSB) from quantum heuristics to solve LDPC decoding, designs an optical computing architecture based on the quantum heuristic algorithm, and fully leverages the performance advantages of optical computing to accelerate LDPC decoding, thereby improving decoding performance and increasing system throughput.
[0090] The technical solutions provided in the embodiments of this application will be described in detail below.
[0091] The LDPC decoding optimization problem can be described as follows:
[0092]
[0093] Where, x * Let x be the decoded sequence carrying the checksum, x be the possible encoded sequence carrying the checksum (i.e., the decoded result to be solved), r be the signal obtained by the signal receiver through multiple input multiple output (MIMO) detection, H be the binary checksum matrix, and λ be a positive real hyperparameter. In the above relation (1), the first term ‖rx‖ 2 The first term, mod2(Hx), means that the decoded sequence is as close as possible to the received signal. The second term, mod2(Hx), means that the decoded sequence x satisfies the parity check matrix H. The decoding based on the parity check matrix described above can be performed using a Hamiltonian of the HUBO form. For example, the Hamiltonian satisfy:
[0094]
[0095] Where σ = 1 - 2x, K is the number of parity bits, i.e., the number of rows in the parity check matrix H, and N kTo check the number of non-zero elements in the k-th row of the matrix, k m It is the column index of the m-th non-zero element in the k-th row.
[0096] This application proposes using the quantum-inspired dSB algorithm to solve the above-mentioned LDPC decoding optimization problem. The dSB algorithm requires solving for the gradients of variables x and y for iterative updates. The relationship between y and x satisfies the following equation:
[0097] x i (t+1)=x i (t)+a0y i (t+1)Δ t (3)
[0098] Where, x i (t), x i (t+1) represent the i-th symbol in the t-th iteration result and the i-th symbol in the (t+1)-th iteration result, respectively. The length of sequence x is N, meaning sequence x includes N symbols, where i is a positive integer less than or equal to N. x in the (t+1)-th iteration... i (t+1) is based on y i The value is obtained by updating at (t+1), where a0 is the scaling factor for each iteration, and Δ t Let be the step size for each iteration. In each iteration, y... i Based on y i gradient Update, y i (t+1) satisfies:
[0099]
[0100] gradient Through Hamiltonian For x i Taking the partial derivative, we get the value in the t-th iteration. satisfy:
[0101]
[0102] Among them, W i Is with x i The number of non-zero elements in the i-th column of the corresponding parity check matrix. It is the k-th element of the parity check matrix containing the w-th non-zero element in the i-th column. w The number of non-zero elements in the row. It is the kth w N of the row kw Dividing x from non-zero elements i The value of the symbol corresponding to the m-th non-zero element other than the corresponding non-zero element in the t-th iteration.
[0103] In this dSB algorithm, gradient calculation is performed iteratively until all y i gradient All convergent, meaning the gradient corresponding to each of the N symbols in sequence x is convergent. All convergences occur. At this point, the obtained x is the decoding result. Based on this, embodiments of this application propose an optical decoding module based on the dSB algorithm, used to quickly solve for the gradient through the optical decoding unit.
[0104] The following first introduces a general-purpose photonic decoder (PDEC) based on the dSB algorithm proposed in the embodiments of this application. The photonic decoder can be configured in a decoding device, and the decoding device uses the photonic decoder to implement LDPC decoding.
[0105] This universal optical decoding module includes a decoding unit corresponding to each of the N symbols in the sequence x. The following section uses the symbol x in the sequence x as an example. i The corresponding decoding unit will be used as an example for detailed explanation. Other decoding units can be found by referring to code element x. i The corresponding decoding unit is implemented, which will not be elaborated further.
[0106] From relation (5), we can see that the gradient Including the cumulative terms of multiplication This application embodiment proposes that it can be achieved through W i A high-speed multiplication link implements the multiplication term, where W i The decoder x is obtained from the parity check matrix H. i The number of non-zero elements in the corresponding i-th column, i.e., the code element x in the parity check matrix. i The corresponding column weight of the i-th column. That is, the i-th column of the verification matrix includes W. i If there are 1 non-zero element, then the code element x i The corresponding decoding unit includes W i A high-speed interconnection link, this W i A high-speed multiplication link and the W in the i-th column i W of the parity matrix containing the non-zero elements i Each row corresponds one-to-one. The w-th non-zero element in the i-th column is located in the k-th column of the parity matrix. w Okay, the kth w Lines include A non-zero element, in the W i In the k-th high-speed link w The corresponding high-speed connecting links include A high-speed optical phase shifter, A high-speed optical phase shifter and Each code element corresponds one-to-one. The k-th code element is the w In the line Among non-zero elements, x i Other than the corresponding non-zero elements The code corresponding to each non-zero element Each high-speed optical phase shifter assigns a value to the element corresponding to the corresponding symbol in the previous iteration result (such as the (t-1)th iteration).
[0107] Because high-speed interconnect links require high-speed iterative updates, high-speed optical phase shifters are used. Exemplary high-speed optical phase shifters may include, but are not limited to, LNOI electro-optic phase shifters, BTO electro-optic phase shifters, carrier injection phase shifters, carrier dissipation phase shifters, slot waveguide phase shifters, MEMS phase shifters, etc. High-speed optical phase shifters can support encoding bandwidths exceeding 100 GHz.
[0108] The optical decoding unit may also include a beam-splitting node for splitting the input light source into multiple beams, including W. i Each beam is input to W. i A high-speed multiplication link to achieve gradient The multiplicative terms in the formula.
[0109] Optionally, gradient The cumulative term of the product The accumulation term in the optical domain can also be implemented. The optical decoding unit may also include a beam combiner node, which can combine W... i The output beams (i.e., the product results) of each high-speed multiplication link are coherently summed to achieve the product accumulation term.
[0110] For example, the beam combining node can be a low-power beam combining node, including but not limited to a multimode interferometer (MMI), a directional coupler (DC), a Y-branch λsplitter, etc.
[0111] Optionally, gradient The cumulative term of the product The scaling factor can also be implemented in the optical domain.
[0112] In one implementation, the W i Each high-speed multiplication link in the series may terminate at a low-speed optical interferometer, through which λ / 2 information is modulated for scaling of each multiplication result. This embodiment is primarily used as an example in the description of this application; however, it should be understood that this application is not limited thereto.
[0113] In another implementation, the W i The output beams of the high-speed multiplication link are combined at the beam combining node and then passed through a low-speed optical interferometer to obtain a multiplication accumulation term containing a scaling factor λ / 2.
[0114] Optionally, the gradient in relation (5) Also includes bias terms This bias term can be implemented in the optical domain. The optical decoding unit may also include a bias link for implementing this bias term, which may include a low-speed optical interferometer and a low-speed optical phase shifter. The low-speed optical interferometer can implement… By shifting the term by -π using a low-speed optical phase shifter, the bias term can be achieved.
[0115] For example, low-speed optical phase shifters may include, but are not limited to, thermo-optic modulators, phase change material modulators, MEMS phase shifters, acousto-optic phase shifters, etc.
[0116] This optical decoding unit can be as follows: Figure 2 As shown, the beam splitter node can divide the input light source into W equal parts. i +1 beam is input to W respectively. i Each high-speed interconnecting link includes a high-speed interconnecting link and an offset link. Each high-speed interconnecting link includes... The bias link includes a high-speed optical phase shifter and a low-speed optical interferometer. i +1 beam of light passes through the W i After adding one more link, W is combined by the bundle node (also known as the coherent summation structure). i The product results of each link in the high-speed multiplication link and the bias link are coherently summed to achieve accumulation. The output of the coherent summation structure is x in the dSB algorithm. i corresponding gradient
[0117] It should be understood that Figure 2 For illustrative purposes only, in actual implementation, at least one of the following: the summation of the product result, the scaling factor, and the bias term, can be implemented in the electrical domain. For example, one or more of these can be determined to be implemented in the electrical domain based on the precision and rate of the optical decoding unit. The following will use... Figure 2 The structure of the optical decoding unit shown is illustrated by way of example, and it should be understood that the embodiments of this application are not limited thereto.
[0118] For example, if the number of parity bits K is 5, the number of code elements N in the encoded sequence x carrying the parity bits is 10, and the parity check matrix H is:
[0119]
[0120] The 10 code elements in sequence x correspond one-to-one with the 10 columns of the parity-check matrix. This optical decoding module includes N optical decoding units, each corresponding to one of the N code elements. Based on the preceding description, the optical decoding unit corresponding to code element x1 can be as follows: Figure 3 As shown, x1 corresponds to the first column of the parity check matrix (i.e., i = 1), and the first column contains two non-zero elements. Therefore, as follows... Figure 3 As shown, the optical decoding unit corresponding to symbol x1 includes two high-speed multiplication links (i.e., W1), and the row containing the first non-zero element in the first column (i.e., the first row) contains four non-zero elements (i.e., w). i =1, R1=4), one of the high-speed multiplication links contains 3 high-speed optical phase shifters, and each corresponds one-to-one with the 3 non-zero elements in the first row except for the non-zero element corresponding to symbol x1. These 3 high-speed optical phase shifters are assigned values based on symbols x2(t), x3(t), and x4(t) in the sequence x(t) obtained in the previous iteration (such as the (t-1)th iteration). By passing the light beam through this high-speed multiplication link, the multiplication result of these 3 symbols in the optical domain can be obtained. The other high-speed multiplication link contains 3 high-speed optical phase shifters, and each of the 3 non-zero elements in the second row except for the non-zero element corresponding to symbol x1 is assigned values based on symbols x5(t), x6(t), and x7(t) in the sequence x(t) obtained in the previous iteration (such as the (t-1)th iteration). By passing the light beam through this high-speed multiplication link, the multiplication result of these 3 symbols in the optical domain can be obtained. The coherent summation structure coherently sums the three output beams from the two high-speed multiplication links and the bias link, performing an accumulation operation. The output of the coherent summation structure is the gradient corresponding to x1 in the dSB algorithm. The decoding device can use a detector to transfer the calculation results in the optical domain to the electrical domain, and obtain the gradient. Decoding devices can be based on gradients Update y1, such as y1(t+1) based on the above relation (4), and then update x1 based on the above relation (3) to obtain x1(t+1), complete the t-th iteration of x1, and assign values to the high-speed optical phase shifters corresponding to x1 in the optical decoding units corresponding to each symbol based on x1(t+1) to carry out the next iteration (i.e. the t+1-th iteration).
[0121] Similarly, the optical decoding unit corresponding to symbol x2 can be as follows: Figure 4As shown, x2 corresponds to the second column of the parity-check matrix (i.e., i=2). The two non-zero elements in the second column are located in the first and third rows, respectively. Therefore, one high-speed multiplication link in the optical decoding unit corresponding to x2 includes three high-speed optical phase shifters that correspond one-to-one with the three non-zero elements in the first row (excluding the non-zero element corresponding to symbol x2). These three high-speed optical phase shifters are assigned values based on symbols x1(t), x3(t), and x4(t) in the sequence x(t) obtained in the previous iteration (e.g., the (t-1)th iteration). By passing the light beam through this high-speed multiplication link, the multiplication result of these three symbols in the optical domain can be obtained. Another high-speed multiplication link includes three high-speed optical phase shifters that correspond one-to-one with the three non-zero elements in the second row (excluding the non-zero element corresponding to symbol x2). These three high-speed optical phase shifters are assigned values based on symbols x5(t), x8(t), and x9(t) in the sequence x(t) obtained in the previous iteration (e.g., the (t-1)th iteration). By passing the light beam through this high-speed multiplication link, the multiplication result of the three symbols in the optical domain can be obtained. The coherent summation structure coherently sums the three output beams from the two high-speed multiplication links and the bias link, performing an accumulation operation. The output of the coherent summation structure is the gradient corresponding to x2 in the dSB algorithm. The decoding device can use a detector to transfer the calculation results in the optical domain to the electrical domain, and obtain the gradient. Decoding devices can be based on gradients Update y2, such as y2(t+1) based on the above relation (4), and then update x2 based on the above relation (3) to obtain x2(t+1), complete the t-th iteration of x2, and assign values to the high-speed optical phase shifters corresponding to x2 in the optical decoding units corresponding to each symbol based on x2(t+1) to carry out the next iteration (i.e. the t+1-th iteration).
[0122] Similarly, for the optical decoding unit corresponding to each symbol in the optical decoding module, please refer to the previous introduction of the optical decoding units corresponding to symbols x1 and x2 for implementation details, which will not be repeated here. Iterative gradient calculation continues until the y-value converges, i.e., the gradient. The sequence x obtained at this point is the decoding result.
[0123] It should be noted that the parity check matrix example provided by equation (6) is a regular parity check matrix, that is, each symbol is associated with two parity bits, and each parity bit is obtained based on four symbols. It should be understood that the embodiments of this application are not limited to this, and the schemes provided by the embodiments of this application can also be applied to irregular parity check matrices.
[0124] Based on the above scheme, a general optical decoding module based on the quantum-inspired algorithm dSB was designed for LDPC decoding. The dSB algorithm significantly improves decoding performance compared to current decoding algorithms (such as sum-product decoding), hardening the LDPC decoding problem to a photonic decoder and fully utilizing the sparsity of the parity-check matrix to reduce hardware implementation complexity. Utilizing optical energy for iterative computation significantly reduces power consumption compared to decoding based on electrical energy, and parallel expansion of multiple modules can provide ultra-high decoding throughput. The connection relationships of multiple optical decoding units can be freely configured, enabling mapping of parity-check matrices with different code lengths and rates, supporting large-scale, multi-scenario application expansion.
[0125] As can be seen from the above description of the general optical decoding module proposed in the embodiments of this application, for multiple symbols corresponding to non-zero elements in the same row of the parity check matrix, the optical decoding unit corresponding to each symbol includes a high-speed multiplication link corresponding to that row of the parity check matrix. For x1, x2, x3, and x4 corresponding to the non-zero elements in the first row of the parity check matrix, the optical decoding unit corresponding to these four symbols includes a high-speed multiplication link corresponding to the first row, such as... Figure 3 , Figure 4 The first high-speed multiplication link shown differs from the high-speed multiplication link corresponding to the first row of the optical decoding units for these four symbols in that x i The high-speed multiplication link corresponding to the first row in the optical decoding unit corresponding to (i = 1, 2, 3, 4) does not include x. i The corresponding high-speed optical phase shifter.
[0126] To reduce hardware overhead and achieve link multiplexing, this application also proposes a dedicated optical decoding module based on dSB, which can also be called a photonic hardware accelerator (PHAC). The dedicated optical decoding module will be described in detail below.
[0127] The dedicated optical decoding module includes an optical decoding unit, which comprises K high-speed multiplication links. Each of these K links corresponds one-to-one with K rows in the parity check matrix. Each high-speed multiplication link includes multiple high-speed optical phase shifters, which correspond one-to-one with multiple non-zero elements in a row of the parity check matrix corresponding to that link, or in other words, with multiple code elements corresponding to the non-zero elements in that row. Taking the parity check matrix in equation (6) as an example, the number of parity bits K (i.e., the number of rows in the parity check matrix) is 5. Figure 5As shown, the optical decoding unit includes five high-speed multiplication links, each corresponding one-to-one with one of the five rows in the parity-check matrix. For example, the high-speed multiplication link corresponding to the first row of the parity-check matrix includes four high-speed optical phase shifters. These four high-speed optical phase shifters correspond one-to-one with the four symbols x1, x2, x3, and x4 corresponding to the non-zero elements in the first row of the parity-check matrix. These four high-speed optical phase shifters are assigned values based on the symbols x1(t), x2(t), x3(t), and x4(t) in the sequence x(t) obtained in the previous iteration (e.g., the (t-1)th iteration), and are used for the calculation in the tth iteration. As another example, the high-speed multiplication link corresponding to the fourth row of the parity-check matrix includes four high-speed optical phase shifters, corresponding to the four symbols x3, x6, x7, and x4 corresponding to the non-zero elements in the fourth row. 10 One-to-one correspondence, these four high-speed optical phase shifters are based on the code elements x3(t), x6(t), x7(t), and x in the sequence x(t) obtained in the previous iteration (e.g., the (t-1)th iteration). 10 (t) is assigned a value for the t-th iteration calculation. Similarly, the correspondence between other high-speed multiplication links and the rows of the parity-check matrix can be obtained, which will not be elaborated here. By mapping the connection relationship described by the parity-check matrix of the LDPC code (the position of the non-zero element (i.e., the 1 element) in the matrix to the optical decoding unit, the code element corresponding to the 1 element in the matrix corresponds to the high-speed optical phase shifter of the optical decoding unit. In the optical domain, the 0 and 1 bit flipping in the digital domain is simulated by the phase shift of 0 and π, thus realizing multiplication.
[0128] The optical decoding unit also has N low-loss combining nodes and a broadcast multiplexing route corresponding to N symbols. The broadcast multiplexing route is used to distribute the multiplication results of K high-speed multiplication links to N low-loss combining nodes. The multiplication result of a high-speed multiplication link will be distributed to the low-loss combining nodes corresponding to multiple symbols of the high-speed optical phase shifter in the high-speed multiplication link. Each low-loss combining node accumulates the multiplication results distributed to it.
[0129] For example Figure 5 As shown, the third high-speed multiplication link corresponds to the third row of the parity check matrix. This third row includes non-zero elements that correspond one-to-one with symbols x2, x5, x8, and x9. Therefore, this high-speed multiplication link includes four high-speed optical phase shifters that correspond one-to-one with symbols x2, x5, x8, and x9. The multiplication result of this high-speed multiplication link will be distributed to the low-loss combining nodes corresponding to x2, x5, x8, and x9. The same applies to other high-speed multiplication links.
[0130] A coherent summation structure composed of low-loss bundle combining nodes enables coherent summation of the product results distributed to each low-power bundle combining node. However, as can be seen from relation (5), the actual x i corresponding gradient The product terms in the formula do not include x. iAnd x i The corresponding multiplication term W i Each link contains x i The corresponding high-speed phase shifter makes x i The chained multiplication term implemented by the corresponding bundle node includes a redundancy coefficient x. i This term can be eliminated in the electric domain, that is, in x i The electric domain result obtained after photoelectric conversion of the beam output from the corresponding beam combiner node is based on x corresponding to the current iteration (e.g., the t-th iteration). i (t) is used for scaling.
[0131] The above can complete the multiplication and accumulation operation for gradient update in the dSB algorithm, so that based on the output of each low-loss bundle node, the value used to update each x can be obtained. i gradient The term of the product accumulation in the middle
[0132] In one implementation, the bias term and / or coefficients This can be implemented in the optical domain. For example, you can refer to the optical domain implementation method in the general optical decoding module introduced earlier, such as adding it to each high-speed interconnect link. The corresponding low-speed optical interferometer, and / or, adding a bias term. The biased links, the product of the biased links, are distributed to each low-loss bundle node in the broadcast multiplexing route.
[0133] In another implementation, the bias term and / or coefficients It can be implemented in the electrical domain.
[0134] Optionally, the choice of which of the two implementation methods described above to adopt can be determined based on the accuracy and speed of the optical decoder.
[0135] After adding coefficients and bias terms to the cumulative terms of the product, we can obtain the terms used to update each x. i gradient The decoding device can be based on each gradient Update y i For example, based on the above relation (4), we can obtain y i (t+1), and then update x based on the above relation (3). i Get x i (t+1), completing the t-th iteration of x, assigning values to the high-speed optical phase shifter in the optical decoding unit based on x(t+1) for each symbol, to proceed to the next iteration (i.e., the (t+1)-th iteration). Iterative gradient calculation continues until the y-value converges, i.e., each gradient... The obtained x is the decoding result.
[0136] Based on the above scheme, a dedicated decoding module based on the quantum-inspired algorithm dSB was designed for LDPC decoding. The quantum-inspired algorithm dSB has significantly improved decoding performance compared to current decoding algorithms (such as sum-product decoding algorithms). Using light energy to achieve iterative calculation can significantly reduce power consumption compared to decoding based on electrical energy. The LDPC decoding problem is hardened to a photonic decoder, which makes full use of the sparsity of the parity-check matrix and the broadcast multiplexing characteristics of photons to realize a minimalist optical decoder and reduce hardware overhead.
[0137] Based on the above introduction, this application provides an optical decoding method based on the dSB algorithm. The following is a detailed explanation... Figure 6 The optical decoding method 600 provided in the embodiments of this application will be described, and the method 600 includes, but is not limited to, the following steps:
[0138] S601, in the t-th iteration, the decoding device inputs the light source to at least one optical decoding unit and obtains the output result of at least one optical decoding unit. The optical decoding unit includes at least one multiplication link, and the multiplication link includes multiple optical phase shifters. The number of multiplication links in the optical decoding unit and the number of optical phase shifters in the multiplication link are determined according to the parity check matrix of the channel coding, where t is a positive integer.
[0139] Optionally, the chain link in the optical decoding unit is used to implement the gradient shown in equation (5). Multiplication terms in
[0140] Optionally, the optical decoding unit may also include an offset link and / or a coherent summation structure.
[0141] For example, the bias link includes an optical interferometer; optionally, the bias link further includes an optical interferometer. This bias link can realize the gradient shown in equation (5). Mid-bias term The coherent summation structure is used to coherently sum a beam that has passed through at least one optical decoding unit, or a beam that has passed through at least one decoding unit and a bias link, and can realize the gradient shown in equation (5). The summation operation in [the context of the text].
[0142] In one embodiment, the at least one optical decoding unit includes N optical decoding units, where N is the code length of the channel coding and is a positive integer greater than 1. The N optical decoding units correspond one-to-one with the N symbols in the decoding sequence. Each of the N optical decoding units outputs a beam corresponding to its corresponding symbol, and each beam includes the gradient multiplication result corresponding to that symbol. Specifically, the i-th symbol in the decoding sequence corresponds to the first optical decoding unit, where i is a positive integer less than or equal to N. The first optical decoding unit outputs the beam corresponding to the i-th symbol, and the i-th symbol corresponds to the i-th column of the parity-check matrix. The i-th column of the parity-check matrix includes W... i The first optical decoding unit includes W non-zero elements. i A multiplication link, W i A link and W i W of the parity matrix containing the non-zero elements i Each row corresponds one-to-one with the other, and the w-th non-zero element in the i-th column is located at the k-th position of the parity matrix. w Line, k w Lines include The kth non-zero element w The corresponding chain links include One optical phase shifter, A light phase shifter and Each code element corresponds one-to-one. Each code element is Non-zero elements divided by x i Other than the corresponding non-zero elements The code elements corresponding to each non-zero element Each optical phase shifter is assigned a value based on the element corresponding to the symbol in the previous iteration (i.e., the (t-1)th iteration). Each iteration results in a sequence of N elements, each corresponding one-to-one with one of the N symbols in the decoding sequence. One element is used to assign a value to the optical phase shifter corresponding to the given symbol.
[0143] If t = 1, that is, the t-th iteration is the first iteration, then... Each optical phase shifter can be randomly assigned a value by the decoding device or assigned a value based on a preset value; this application embodiment does not limit this.
[0144] In this embodiment, the decoding device can use the general optical decoding module described above for decoding. For the specific structure, please refer to the description of the general optical decoding module described above, which will not be repeated here.
[0145] In another embodiment, the at least one optical decoding unit includes one optical decoding unit, which includes K chain links, each of which corresponds one-to-one with one of the K rows of the parity-check matrix. The k-th row of the parity-check matrix includes N kThere are N non-zero elements, and the chain link corresponding to the k-th row includes N. k One optical phase shifter, N k One optical phase shifter and N k Each code element corresponds one-to-one, N k Each code element is a decoded sequence and N k The code elements corresponding to N non-zero elements, N k Each optical phase shifter assigns a value based on the element corresponding to the symbol in the previous iteration result. If t = 1, that is, the t-th iteration is the first iteration, then... Each optical phase shifter can be randomly assigned a value by the decoding device or assigned a value based on a preset value; this application embodiment does not limit this.
[0146] Optionally, in this embodiment, the optical decoding unit further includes a broadcast multiplexing route and N beam combining nodes, where each of the N beam combining nodes corresponds one-to-one with one of the N symbols in the decoding sequence. The broadcast multiplexing route routes the output beam of the chain link corresponding to the k-th row to the N beam combining nodes. k N corresponding to each code element k The output consists of N beam combining nodes, each corresponding to one of the N symbols.
[0147] In this embodiment, the decoding device can use the dedicated optical decoding module described above for decoding. For the specific structure, please refer to the description of the dedicated optical decoding module described above, which will not be repeated here.
[0148] S602, the decoding device obtains the iteration result of the t-th iteration based on the output result of at least one optical decoding unit.
[0149] S603, the decoding device assigns values to the optical phase shifters contained in each multiplication link of the at least one optical decoding unit based on the iteration result of the t-th iteration.
[0150] S604, the decoding device performs the (t+1)th iteration, and the iteration result obtained when the iteration result converges is the decoding sequence.
[0151] According to the above scheme, a decoding method based on optical decoding units can be realized through quantum-inspired algorithms, which can significantly improve decoding performance and reduce hardware overhead and decoding power consumption.
[0152] As can be seen from equation (2), the Hamiltonian of the HUBO form There are higher-order terms Transforming it into a quadratic unconstrained binary optimization (QUBO) form allows for the optical matrix mapping of the Hamiltonian. QUBO is an optimization model where the objective function is quadratic and the variables are binary (0 / 1 or ±1), with no explicit constraints. By introducing synthetic bits... Dimensionality reduction of higher-order terms can... Written in QUBO format:
[0153]
[0154] in, These are additional auxiliary bits. Furthermore, the Hamiltonian of the LDPC can be mapped to QUBO form:
[0155]
[0156] The matrix Q can be decomposed using singular value decomposition (SVD):
[0157] Q = V T ∑V (9)
[0158] Hamiltonian in QUBO form It can be written as:
[0159]
[0160] Based on this relation (10), embodiments of this application propose that can On-chip coding is performed using an optical matrix, with x as the input vector and E as the output amplitude. out =Ux, the light intensity detected by the decoder through the detector is I. out =|E out | 2 =(Ux) T (Ux), thus the Hamiltonian in the QUBO form of LDPC can be directly calculated through optical computation decoding, completing the solution of the quantum-inspired algorithm.
[0161] This application also provides an optical decoding module based on quadratic unconstrained binary optimization (QUBO) for accelerating LDPC decoding. For example... Figure 7 The QUBO-based optical decoding module shown consists of the following five parts.
[0162] 1) Beam splitting node: The fractional structure is used to divide the input light source from the laser into N+L beams, where N is the number of columns in the parity check matrix H, and L is the number of auxiliary bits, L = 2N(K-1).
[0163] 2) High-speed vector input unit, which drives N+L high-speed optical modulators to perform high-speed 0 / 1 modulation on the input optical signal through serializer / deserializer (SerDes) to represent the encoding in the annealing process.
[0164] In each iteration, m variables (where m∈[1,2,…,(NK)]) are randomly selected from the input vector x(t), and these variables are changed between 0 and 1 to obtain the vector x′(t). N+L high-speed optical modulators are driven by a serializer / deserializer (SerDes) to perform high-speed 0 / 1 modulation on the input optical signal based on the vector x′(t), and then input to the optical matrix.
[0165] 3) For the optical matrix section, configure the optical matrix as follows: This matrix corresponds to the weight matrix Q, realizing the Hamiltonian mapping in the LDPC decoding problem. The digital-to-analog converter (DAC) is used to encode the weights of the matrix for analog mapping. As mentioned earlier, x′(t) serves as the input to the optical matrix part, and the output amplitude is E. out =Ux′(t), the light intensity detected by the decoder through the detector is I. out =|E out | 2 =(Ux′(t)) T (Ux′(t)) can be directly calculated using optical computation decoding, allowing the variable in the LDPC QUBO form Hamiltonian that varies with vector x. Adding a constant C yields the QUBO form Hamiltonian, completing the solution using the quantum-inspired algorithm. Specifically, this constant C can be added in the electrical or optical domain; this application does not limit this. The high-speed coding optical matrix is linearly related to the code length, and the matrix can be implemented using low-power static maintenance techniques.
[0166] For example, optical matrices include, but are not limited to, Mach-Zehnder interferometer (MZI) network architectures such as triangular (e.g., Reck-type), rectangular (e.g., Clements-type / Bell-type), hexagonal loop networks, micro-loop networks, and cross-bar networks.
[0167] 4) Detection and transimpedance amplifier (TIA) are used to convert the results of optical calculations into electrical signals.
[0168] 5) The comparator performs a min[1,exp(-ΔH / T)] operation on the calculation result, meaning the state vector x′(t) is accepted with probability min[1,exp(-ΔH / T)], where T is the annealing temperature and ΔH is the Hamiltonian. The change relative to the previous iteration. That is, if 1 ≤ exp(-ΔH / T), then x(t+1) = x″(t+1); if exp(-ΔH / T) < 1, then x″(t+1) is accepted with probability exp(-ΔH / T), resulting in the state variable x(t+1), which is used for the next iteration update until the convergent output decoded value is reached. The calculation result of the comparator input is the Hamiltonian obtained by converting the analog electrical signal output from the TIA using an analog-to-digital converter (ADC). Digital electrical signals.
[0169] According to the above scheme, by transforming the higher-order Hamiltonian into a quadratic unconstrained binary optimized QUBO form, the QUBO form Hamiltonian can be calculated through the optical decoding module, thus completing the solution of the quantum-inspired algorithm. With short code lengths, the performance of SA-QUBO approaches the BP bound, exhibiting significant decoding performance advantages, reducing decoding power consumption, and improving decoding efficiency.
[0170] Based on the above description, this application provides another optical decoding method, which is described below in conjunction with... Figure 8 The optical decoding method 800 provided in the embodiments of this application will be described, and the method 800 includes, but is not limited to, the following steps:
[0171] S801, the decoding device randomly selects m variables in the sequence x(t) in the t-th iteration, changes the values of the m variables, and obtains the sequence x′(t).
[0172] S802, the decoding device modulates the optical signal based on the sequence x′(t), and then inputs the modulated signal into the optical matrix unit to obtain the optical calculation result output by the optical matrix unit. This optical calculation result is used to determine the Hamiltonian of the t-th iteration.
[0173] As mentioned earlier, the decoding device can control the SerDes to drive N+L high-speed optical modulators to modulate the input optical signal based on the vector x′(t) and then input it to the optical matrix unit.
[0174] The optical matrix unit can be the optical matrix section described earlier. As mentioned earlier, the optical matrix unit is configured as follows: When the input to the optical matrix is a sequence x′(t), the amplitude of the beam output by the optical matrix unit is E. out =Ux′(t), the optical calculation result is the light intensity I of the output beam. out =|E out | 2 =(Ux′(t)) T (Ux′(t)) can be obtained through optical matrix units, which are the variables in the Hamiltonian of LDPC in QUBO form that change with vector x. After adding a constant C, the Hamiltonian in QUBO form can be obtained, thus completing the solution of the quantum-inspired algorithm. Specifically, this constant C can be added in the electric domain or the optical domain, but this application does not limit this.
[0175] S803, the decoding device obtains the iteration result x(t+1) based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration. This vector is used for the (t+1)-th iteration, and the iteration continues until the iteration result converges. The converged iteration result is the decoding sequence.
[0176] In one embodiment, the decoding device obtains the sequence x(t+1) based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration, including: the decoding device obtains the change ΔH of the Hamiltonian of the t-th iteration compared to the Hamiltonian of the (t-1)-th iteration, and accepts the sequence x′(t) with probability min[1,exp(-ΔH / T)] to obtain the sequence x(t+1), where T is the annealing temperature in the simulated annealing algorithm.
[0177] Figure 8 For specific implementation details of the embodiments shown, please refer to the preceding text. Figure 7 The description in the illustrated embodiment will not be repeated here.
[0178] According to the above scheme, by transforming the higher-order Hamiltonian into a quadratic unconstrained binary optimized QUBO form, the QUBO form Hamiltonian can be calculated through the optical decoding module, thus completing the solution of the quantum-inspired algorithm. With short code lengths, the performance of SA-QUBO approaches the BP bound, exhibiting significant decoding performance advantages, reducing decoding power consumption, and improving decoding efficiency.
[0179] It is understood that, in order to achieve the functions in the above embodiments, the communication device may include hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and method steps of the various examples described in conjunction with the embodiments disclosed in this application, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.
[0180] Figure 9 , Figure 10 and Figure 11 The diagram illustrates the possible structures of communication devices provided in embodiments of this application. These communication devices can be used to implement the functions of the decoding devices in the above-described method embodiments, and thus also achieve the beneficial effects of the above-described method embodiments. In embodiments of this application, the communication device can be as follows: Figure 1 One of the terminals 120a-120j shown can also be as follows: Figure 1 The network devices 110a or 110b shown can also be modules (such as chips or chip systems) applied to terminals or network devices.
[0181] like Figure 9 The communication device 900 shown includes at least one decoding unit 910, a detection unit 920, and a processing unit 930. This communication device 900 can be used to implement the above-described... Figure 6 The method shown is used when the communication device 900 is used to implement the above. Figure 6 In the illustrated method embodiment: the optical decoding unit 910 includes at least one multiplication link, which includes multiple optical phase shifters. The number of multiplication links and the number of optical phase shifters in the multiplication links of the decoding unit 910 are determined according to the parity check matrix of the channel coding, where t is a positive integer. The processing unit 930 is used to control the light source input to the at least one optical decoding unit in the t-th iteration to obtain the electrical domain output result of the at least one optical decoding unit. The detection unit 920 is used to convert the electrical domain output result into an electrical domain output result. The processing unit 930 is also used to obtain the iteration result of the t-th iteration based on the electrical domain output result. The iteration result is a sequence containing N elements, where N is the code length of the channel coding, and N is a positive integer. The processing unit 930 is also used to assign values to the optical phase shifters included in each multiplication link in the at least one optical decoding unit 910 based on the iteration result of the t-th iteration, and then use this value for the (t+1)-th iteration until the iteration result converges, at which point the obtained iteration result is the decoding sequence.
[0182] like Figure 10 The communication device 1000 shown includes a processing unit 1030, an optical modulation unit 1010, and an optical matrix unit 1020. This communication device 1000 can be used to implement the above-mentioned... Figure 8 The method shown is used when the communication device 1000 is used to implement the above. Figure 8In the illustrated method embodiment: the processing unit 1010 randomly selects m variables from the sequence x(t) in the t-th iteration, changes the values of the m variables, and obtains the sequence x′(t). The optical modulation unit 1010 modulates the optical signal based on the sequence x′(t) to obtain the modulated optical signal. The optical matrix unit 1020 outputs an optical calculation result based on the input modulated optical signal, which is used to determine the Hamiltonian of the t-th iteration. The processing unit 1030 further obtains the sequence x(t+1) based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration. This sequence x(t+1) is used in the (t+1)-th iteration, and the iterations are repeated until the sequence x converges. The converged sequence x is the decoding sequence.
[0183] It should be understood that Figure 9 , Figure 10 The communication device shown may be a chip configured in (or used in) a communication device. The communication device may also include a transceiver unit, which may be an input / output interface or circuit of the chip. The processing unit in the communication device may be a processor in the chip.
[0184] Optionally, such as Figure 9 , Figure 10 The communication device shown may also include a storage unit, which can be used to store instructions or data. The processing unit can execute the instructions or data stored in the storage unit to enable the communication device to perform corresponding operations.
[0185] For a more detailed description of each unit in the above communication device, please refer to the relevant descriptions in the method embodiments above, which will not be repeated here.
[0186] It should be understood that the processing unit in the communication device can be implemented by at least one processor, and the processing unit in the communication device can also be implemented by at least one logic circuit. Optionally, the communication device further includes a storage unit, which can be implemented by a memory.
[0187] like Figure 11 As shown, the communication device 1200 includes a processor 1210 and an interface circuit 1220. The processor 1210 and the interface circuit 1220 are coupled to each other. It is understood that the interface circuit 1220 can be a transceiver or an input / output interface. Optionally, the communication device 1200 may also include a memory 1230 for storing instructions executed by the processor 1210, or storing input data required by the processor 1210 to execute instructions, or storing data generated after the processor 1210 executes instructions.
[0188] In one implementation, the memory 1230 may be integrated into the processor 1210 or independent of the processor 1210.
[0189] When the communication device 1200 is used to implement Figure 6 , Figure 8 In the method shown, processor 1210 is used to implement the functions of the above-mentioned processing unit, and interface circuit 1220 is used to implement the transmission of control instructions to realize the decoding function described above.
[0190] When the aforementioned communication device is a chip applied to a terminal device, the terminal device chip can implement the functions of the terminal (such as a test terminal or a first terminal) in the above method embodiments. The terminal device chip receives information from other modules (such as a radio frequency module or antenna) in the terminal device, information sent by the network device to the terminal device; or, the terminal device chip sends information to other modules (such as a radio frequency module or antenna) in the terminal device, information sent by the terminal device to the network device.
[0191] When the aforementioned communication device is a module applied to a network device, the network device module can implement the functions of the network device (such as a base station, a network device implementing the functions of a network management node or a terminal management node) in the above method embodiments. The network device module receives information from other modules (such as a radio frequency module or antenna) in the network device, which is information sent from the terminal device to the network device; or, the network device module sends information to other modules (such as a radio frequency module or antenna) in the network device, which is information sent from the network device to the terminal device. Here, the network device module can be the baseband chip of the network device, or it can be a DU or other modules. The DU here can be a DU under an open radioaccess network (O-RAN) architecture.
[0192] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), microprocessor units (MPUs), microcontroller units (MCUs), graphics processing units (GPUs), artificial intelligence processors (AI processors), neural processing units (NPUs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.
[0193] The method steps in the embodiments of this application can be implemented in hardware or in software instructions executable by a processor. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. The storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in an access network device or a terminal device. The processor and storage medium can also exist as discrete components in the access network device or terminal device.
[0194] According to the method provided in the application embodiments, this application embodiment also provides a computer program product, the computer program product comprising: computer program code, which, when executed by one or more processors, causes a device including the processor to perform as described above. Figure 6 , Figure 8 The method shown.
[0195] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. This computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of this application are performed, in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, network equipment, user equipment, or other programmable device.
[0196] According to the method provided in the embodiments of this application, the embodiments of this application also provide a computer-readable storage medium that stores the aforementioned computer program or instructions. When the computer program or instructions are executed by one or more processors, they cause a device including the processor to perform actions such as... Figure 6 , Figure 8 The method shown.
[0197] As described above, computer programs or instructions can be stored in or transferred from one computer-readable storage medium to another. For example, the computer programs or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video optical disc; or it can be a semiconductor medium, such as a solid-state drive. The computer-readable storage medium can be a volatile or non-volatile storage medium, or it can include both volatile and non-volatile types of storage media.
[0198] According to the method provided in the embodiments of this application, this application also provides a communication system, including the aforementioned decoding device and encoding device. The encoding device encodes information bits to obtain an encoded sequence, which is then transmitted to the decoding device via a channel. The decoding device demodulates the encoded sequence to obtain the demodulated sequence r described above, thereby realizing the optical decoding method described above.
[0199] In the various embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus described above is merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0200] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this solution according to actual needs.
[0201] In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions of different embodiments are consistent and can be referenced by each other. The technical features of different embodiments can be combined to form new embodiments according to their inherent logical relationship.
[0202] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A channel-coded optical decoding method, characterized in that, include: In the t-th iteration, the light source is input to at least one optical decoding unit to obtain the output result of the at least one optical decoding unit. The optical decoding unit includes at least one multiplication link, and the multiplication link includes multiple optical phase shifters. The number of multiplication links in the optical decoding unit and the number of optical phase shifters in the multiplication link are determined according to the parity check matrix of the channel coding, where t is a positive integer. Based on the output of the at least one optical decoding unit, the iteration result of the t-th iteration is obtained. The iteration result is a sequence containing N elements, where N is the code length of the channel coding and N is a positive integer. Based on the iteration result of the t-th iteration, values are assigned to the optical phase shifters contained in each multiplication link of the at least one optical decoding unit; Perform the (t+1)th iteration, and repeat the iterations multiple times until the iteration results converge. The resulting iteration result is the decoding sequence.
2. The method according to claim 1, characterized in that, Hamiltonian obtained by solving LDPC decoding problem based on discrete simulated bifurcation algorithm satisfy: wherein σ = 1 - 2x, x is a coded sequence with a code length of N, r i is a demodulation quantity corresponding to the ith symbol in x obtained after a receiving end demodulates a received signal, K is a number of check bits, N k is a number of non-zero elements of a kth row of the check matrix, k m is a column index of an mth non-zero element of the kth row, λ is a hyperparameter of a positive real number in an LDPC decoding problem, K, N k , m are less than N, k is less than or equal to K, k m is less than or equal to N k ; The Hamiltonian Compared to the i-th symbol x of the decoded sequence i gradient for: Among them, W i Is with x i The corresponding number of non-zero elements in the i-th column of the verification matrix. It is the k-th non-zero element in the check matrix of the i-th column. w The number of non-zero elements in the row. It is the kth w OK Dividing x from non-zero elements i The value of the symbol corresponding to the m-th non-zero element (excluding the corresponding non-zero element) in the t-th iteration, wherein the at least one chain link is used to implement the gradient. Multiplication terms in 3. The method according to claim 1 or 2, characterized in that, The optical decoding unit further includes a bias link and / or a coherent summation structure, the bias link including an optical interferometer, and the coherent summation structure being used to coherently sum the light beams of the light source passing through the at least one optical decoding unit and the bias link.
4. The method according to claim 3, characterized in that, The bias link is used to implement the following bias term in the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence: The Hamiltonian is obtained by solving the LDPC decoding problem based on the discrete simulation bifurcation algorithm.
5. The method according to any one of claims 1 to 4, characterized in that, The link also includes an optical interferometer.
6. The method according to claim 5, characterized in that, The optical interferometer is located at the end of the chain link, and is used to realize the addition of the chain term in the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence. of Scaling.
7. The method according to any one of claims 1 to 6, characterized in that, The at least one optical decoding unit comprises N optical decoding units, each of which corresponds one-to-one with one of the N code elements in the decoding sequence. The output result includes the light beam output by each of the N optical decoding units corresponding to the corresponding code element, where N is an integer greater than 1. In this decoding sequence, the i-th symbol corresponds to the first optical decoding unit, where i is a positive integer less than or equal to N. The first optical decoding unit outputs the light beam corresponding to the i-th symbol. The ith symbol corresponds to an ith column in the check matrix, the ith column in the check matrix includes W i nonzero elements, the first optical decoding unit includes W i serial multiplication links, the W i serial multiplication links correspond to W i rows of the check matrix where the W i nonzero elements are located, respectively, The w-th non-zero element in the i-th column is located at the k-th position of the check matrix. w Line, the kth w Lines include The kth non-zero element, the kth element w The corresponding chain links include A light phase shifter, the A light phase shifter and Each code element corresponds one-to-one with the aforementioned The code element is the aforementioned Non-zero elements divided by x i Other than the corresponding non-zero elements The code elements corresponding to the non-zero elements, the Each optical phase shifter is assigned a value based on the element corresponding to the corresponding symbol in the previous iteration result.
8. The method according to any one of claims 1 to 6, characterized in that, The at least one optical decoding unit includes one optical decoding unit. The optical decoding unit includes K chain links, each of which corresponds one-to-one with one of the K rows of the parity-check matrix. The k-th row of the parity-check matrix includes N... k There are N non-zero elements, and the chain link corresponding to the k-th row includes N. k A light phase shifter, the N k One optical phase shifter and N k Each code element corresponds one-to-one, the N k Each code element is the decoded sequence and the N. k The code elements corresponding to the N non-zero elements, the N k Each optical phase shifter is assigned a value based on the element corresponding to the corresponding symbol in the previous iteration result.
9. The method according to claim 8, characterized in that, The optical decoding unit further includes a broadcast multiplexing route and N beam combining nodes, wherein each of the N beam combining nodes corresponds one-to-one with one of the N code elements in the decoding sequence. The broadcast multiplexing route routes the output beam of the chain link corresponding to the k-th row to the N-th row. k N corresponding to each code element k A bundle node, The output results include N beams output by the N beam combining nodes, each corresponding to one of the N symbols.
10. The method according to any one of claims 1 to 9, characterized in that, The step of obtaining the iteration result of the t-th iteration based on the output result of the at least one optical decoding unit includes: The detector converts the optical domain output of the at least one optical decoding unit into the electrical domain output. Based on the electrical domain output, the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence is obtained. Based on the gradient The iteration result x of the t-th iteration of the i-th symbol is obtained. i (t+1), Wherein, the gradient and the iteration result x i (t+1) satisfies: x i (t+1)=x i (t)+a0y i (t+1)Δ t , Where a0 is the scaling factor for each iteration, Δ t This represents the step size for each iteration update.
11. A decoding method for channel coding, characterized in that, include: In the t-th iteration, m variables in the sequence x(t) are randomly selected, and the values of the m variables are changed to obtain the sequence x′(t), where m is less than N, N is the code length of the channel coding, and the sequence x(t) includes N elements; The optical signal is modulated based on the sequence x′(t), and the modulated optical signal is input to the optical matrix unit to obtain the optical calculation result output by the optical matrix unit. The optical calculation result is used to determine the Hamiltonian of the t-th iteration. Based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration, the iteration result x(t+1) of the t-th iteration is obtained. The iteration result x(t+1) is used for the (t+1)-th iteration. Multiple iterations are performed until the sequence iteration result converges. The converged iteration result is the decoding sequence.
12. The method according to claim 11, characterized in that, The HUBO-form Hamiltonian obtained by solving the LDPC decoding problem based on the discrete simulation bifurcation algorithm. satisfy: Where σ = 1 - 2x, x is a coded sequence of code length N, and r i It is the demodulated value corresponding to the i-th symbol in x obtained after demodulation of the received signal by the receiving end, where K is the number of parity bits and N is the number of parity bits. k Let k be the number of non-zero elements in the k-th row of the parity check matrix. m Let λ be the column index of the m-th non-zero element in the k-th row, λ be the hyperparameter of positive real numbers in the LDPC decoding problem, and K and N be the column indices of the m-th non-zero element in the k-th row. k m is less than N, k is less than or equal to K, k m Less than or equal to N k ; Hamiltonian based on the HUBO form The Hamiltonian in QUBO form obtained after adding auxiliary bits satisfies: The optical matrix unit is configured as follows: When the input to the optical matrix is a sequence x′(t), the amplitude of the beam output by the optical matrix is E. out =Ux′(t), where the calculated light intensity I is the light intensity of the light beam. out =|E out | 2 =(Ux′(t)) T (Ux′(t)).
13. The method according to claim 11 or 12, characterized in that, The sequence x(t+1) obtained based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration includes: Obtain the change ΔH of the Hamiltonian in the t-th iteration compared to the Hamiltonian in the (t-1)-th iteration; The sequence x′(t) is accepted with probability min[1,exp(-ΔH / T)] to obtain the sequence x(t+1), where T is the annealing temperature in the simulated annealing algorithm.
14. A decoding device, characterized in that, It includes at least one optical decoding unit, a processing unit, and a detection unit. The optical decoding unit includes at least one multiplication link, and the multiplication link includes multiple optical phase shifters. The number of multiplication links in the decoding unit and the number of optical phase shifters in the multiplication link are determined according to the parity check matrix of the channel coding, where t is a positive integer. The processing unit is used to control the light source input to the at least one optical decoding unit in the t-th iteration to obtain the electrical domain output result of the at least one optical decoding unit; The detection unit is used to convert the electrical domain output result into an electrical domain output result. The processing unit is further configured to obtain the iteration result of the t-th iteration based on the electrical domain output result, wherein the iteration result is a sequence containing N elements, where N is the code length of the channel coding and N is a positive integer; The processing unit is further configured to assign values to the optical phase shifters contained in each multiplication link of the at least one optical decoding unit based on the iteration result of the t-th iteration, and then use them for the (t+1)-th iteration until the iteration result converges, and the resulting iteration result is the decoding sequence.
15. The apparatus according to claim 14, characterized in that, Hamiltonian obtained by solving LDPC decoding problem based on discrete simulated bifurcation algorithm satisfy: Where σ = 1 - 2x, x is a coded sequence of code length N, and r i It is the demodulated value corresponding to the i-th symbol in x obtained after demodulation of the received signal by the receiving end, where K is the number of parity bits and N is the number of parity bits. k Let k be the number of non-zero elements in the k-th row of the parity check matrix. m Let λ be the column index of the m-th non-zero element in the k-th row, λ be the hyperparameter of positive real numbers in the LDPC decoding problem, and K and N be the column indices of the m-th non-zero element in the k-th row. k m is less than N, k is less than or equal to K, k m Less than or equal to N k ; The Hamiltonian Compared to the i-th symbol x of the decoded sequence i gradient for: Among them, W i Is with x i The corresponding number of non-zero elements in the i-th column of the verification matrix. It is the k-th non-zero element in the check matrix of the i-th column. w The number of non-zero elements in the row. It is the kth w OK Dividing x from non-zero elements i The value of the symbol corresponding to the m-th non-zero element (excluding the corresponding non-zero element) in the t-th iteration, wherein the at least one chain link is used to implement the gradient. Multiplication terms in 16. The apparatus according to claim 14 or 15, characterized in that, The optical decoding unit further includes a bias link and a coherent summation structure. The bias link includes an optical interferometer, and the coherent summation structure is used to coherently sum the light beams of the light source that have passed through the N optical decoding units and the bias link.
17. The apparatus according to claim 16, characterized in that, The bias link is used to implement the following bias term in the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence: The Hamiltonian is obtained by solving the LDPC decoding problem based on the discrete simulation bifurcation algorithm.
18. The apparatus according to any one of claims 14 to 17, characterized in that, The link also includes an optical interferometer.
19. The apparatus according to claim 18, characterized in that, The optical interferometer is located at the end of the chain link, and is used to realize the addition of the chain term in the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence. of Scaling.
20. The apparatus according to any one of claims 14 to 19, characterized in that, The at least one optical decoding unit includes N optical decoding units, each of which corresponds one-to-one with one of the N code elements in the decoding sequence. The output result includes the light beam output by each of the N optical decoding units, corresponding to the corresponding code element. In this decoding sequence, the i-th symbol corresponds to the first optical decoding unit, where i is a positive integer less than or equal to N. The first optical decoding unit outputs the light beam corresponding to the i-th symbol. The i-th symbol corresponds to the i-th column of the parity check matrix, and the i-th column of the parity check matrix includes W. i The first optical decoding unit includes W non-zero elements. i A multiplication link, the W i Each link is connected to the W i W of the parity matrix containing the non-zero elements i Each row corresponds one-to-one. The w-th non-zero element in the i-th column is located at the k-th position of the check matrix. w Line, the kth w Lines include The kth non-zero element, the kth element w The corresponding chain links include A light phase shifter, the A light phase shifter and Each code element corresponds one-to-one with the aforementioned One codeword is the Non-zero elements divided by x i Other than the corresponding non-zero elements The code elements corresponding to the non-zero elements, the Each optical phase shifter is assigned a value based on the element corresponding to the corresponding symbol in the previous iteration result.
21. The apparatus according to any one of claims 14 to 19, characterized in that, The at least one optical decoding unit includes one optical decoding unit. The optical decoding unit includes K chain links, each of which corresponds one-to-one with one of the K rows of the parity-check matrix. The k-th row of the parity-check matrix includes N... k There are N non-zero elements, and the chain link corresponding to the k-th row includes N. k A light phase shifter, the N k One optical phase shifter and N k Each code element corresponds one-to-one, the N k Each code element is the decoded sequence and the N. k The code elements corresponding to the N non-zero elements, the N k Each optical phase shifter is assigned a value based on the element corresponding to the corresponding symbol in the previous iteration result.
22. The apparatus according to claim 21, characterized in that, The optical decoding unit further includes a broadcast multiplexing route and N beam combining nodes, wherein each of the N beam combining nodes corresponds one-to-one with one of the N code elements in the decoding sequence. The broadcast multiplexing route routes the output beam of the chain link corresponding to the k-th row to the N-th row. k N corresponding to each code element k A bundle node, The output results include N beams output by the N beam combining nodes, each corresponding to one of the N symbols.
23. The apparatus according to any one of claims 14 to 22, characterized in that, The step of obtaining the iteration result of the t-th iteration based on the output result of the at least one optical decoding unit includes: The detector converts the optical domain output of the at least one optical decoding unit into the electrical domain output. Based on the electrical domain output, the gradient of the Hamiltonian relative to the i-th symbol in the decoded sequence is obtained. Based on the gradient The iteration result x of the t-th iteration of the i-th symbol is obtained. i (t+1), Wherein, the gradient and the iteration result x i (t+1) satisfies: x i (t+1)=x i (t)+a0y i (t+1)Δ t , Where a0 is the scaling factor for each iteration, Δ t This represents the step size for each iteration update.
24. A decoding device, characterized in that, It includes a processing unit, an optical modulation unit, and an optical matrix unit. The processing unit is used to randomly select m variables in the sequence x(t) in the t-th iteration, change the values of the m variables to obtain the sequence x′(t), and control the decoding unit to input the sequence x′(t) into the optical matrix unit; The optical modulation unit is used to modulate the optical signal based on the sequence x′(t) to obtain the modulated optical signal; The optical matrix unit is used to output optical calculation results based on the input modulated optical signal, and the optical calculation results are used to determine the Hamiltonian of the t-th iteration. The processing unit is further configured to obtain a sequence x(t+1) based on the Hamiltonian of the t-th iteration and the Hamiltonian of the (t-1)-th iteration. The sequence x(t+1) is used for the (t+1)-th iteration, and the iteration is repeated until the sequence x converges. The converged sequence x is the decoding sequence.
25. The apparatus according to claim 24, characterized in that, The HUBO-form Hamiltonian obtained by solving the LDPC decoding problem based on the discrete simulation bifurcation algorithm. satisfy: Where σ = 1 - 2x, x is a coded sequence of code length N, and r i It is the demodulated value corresponding to the i-th symbol in x obtained after demodulation of the received signal by the receiving end, where K is the number of parity bits and N is the number of parity bits. k Let k be the number of non-zero elements in the k-th row of the parity check matrix. m Let λ be the column index of the m-th non-zero element in the k-th row, λ be the hyperparameter of positive real numbers in the LDPC decoding problem, and K and N be the column indices of the m-th non-zero element in the k-th row. k m is less than N, k is less than or equal to K, k m Less than or equal to N k ; Hamiltonian based on the HUBO form The Hamiltonian in QUBO form obtained after adding auxiliary bits satisfies: The optical matrix unit is configured as follows: When the input to the optical matrix is a sequence x′(t), the amplitude of the beam output by the optical matrix is E. out =Ux′(t), where the calculated light intensity I is the light intensity of the light beam. out =|E out | 2 =(Ux′(t)) T (Ux′(t)).
26. The apparatus according to claim 24 or 25, characterized in that, The processing unit is specifically used for: Obtain the change ΔH of the Hamiltonian in the t-th iteration compared to the Hamiltonian in the (t-1)-th iteration; The sequence x′(t) is accepted with probability min[1,exp(-ΔH / T)] to obtain the sequence x(t+1), where T is the annealing temperature in the simulated annealing algorithm.
27. A communication device, characterized in that, The device includes a processor coupled to a memory for storing a computer program, the processor for executing the computer program stored in the memory to cause the communication device to perform the method as described in any one of claims 1 to 10.
28. A communication device, characterized in that, It includes a processor and a communication interface, the processor being used to control the communication interface to implement the method as described in any one of claims 1 to 13.
29. A computer-readable storage medium, characterized in that, The computer stores instructions that, when executed on the computer, cause the computer to perform the method as described in any one of claims 1 to 13.
30. A computer program product, characterized in that, The computer program product includes: a computer program that, when run, causes a computer to perform the method of any one of claims 1 to 13.
