Double-set online fountain code of underwater acoustic network based on state feedback and encoding and decoding method

Abstract

The invention relates to the technical field of underwater acoustic networks, in particular to a state feedback-based double-set online fountain code and encoding and decoding method for an underwater acoustic network, which comprises an encoding and decoding process under a lossless channel condition and an encoding and decoding process under a lossy channel condition. The double-set online fountain codes adopt a double-set coding strategy in the establishment stage and the completion stage so as to quickly form a maximum connected component; and under the condition of a lossy channel, a state feedback strategy is adopted, so that the coding and decoding efficiency is improved, the number of feedback packets is effectively reduced, and the channel utilization rate is improved. The method shows better performance in the aspects of overhead and the number of feedback packets, and is more suitable for the UANs.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of underwater acoustic networks, in particular to a state feedback-based double-set online fountain code and encoding and decoding method for underwater acoustic networks. BACKGROUND

[0002] As a bridge connecting the underwater world and the information system on land, underwater acoustic networks (UANs) play an irreplaceable role. UANs are widely used in many fields such as ocean monitoring, ocean exploration, and underwater target tracking. With the continuous development of the exploitation of marine resources by countries around the world, the development of UANs has become particularly urgent. However, UANs have characteristics such as long propagation delay, high error rate, energy limitation, and half-duplex communication, which makes the underwater nodes have a high error or loss in the data forwarding process. Therefore, improving the transmission reliability of data has become a research hotspot for UANs.

[0003] In recent years, many scholars have extensively studied UANs reliable transmission technology based on fountain codes. However, due to the lack of control and supervision of the sender in the decoding process of traditional fountain codes, fountain codes with online encoding and decoding capabilities have gradually attracted widespread attention. Some scholars have proposed online fountain codes (OFC) with low overhead, which can calculate the optimal encoding strategy according to any given decoding state. Compared with fountain codes, OFC has lower encoding and decoding overhead and good online performance. Some scholars have compared OFC-related literature in the past ten years and analyzed the overhead and feedback performance of different encoding mechanisms. It is concluded that OFC and improved OFC have poor performance when applied to UANs. In view of the above shortcomings, two encoding mechanisms, recursive online fountain code with limited feedback (ROFC-LF) and sequential recursive online fountain code with limited feedback (SROFC-LF), are proposed, which have low overhead, low energy consumption, and less feedback. Theoretical analysis and simulation experiments prove that the proposed encoding mechanism has good performance in terms of overhead, feedback packet quantity, encoding efficiency, and computational complexity. The performance of SROFC-LF is better than that of ROFC-LF, and it is more suitable for UANs.

[0004] The characteristics of narrow bandwidth, high error rate, and long delay of UANs require a coding mechanism with lower overhead and less feedback. Therefore, a coding mechanism with lower encoding and decoding overhead and fewer feedback packets is needed. SUMMARY

[0005] The present application provides a kind of underwater acoustic network Double-SetOnline Fountain Code Based on State Feedback (DSOFC-SF) and coding and decoding method, which can overcome certain or some defects of prior art.

[0006] The underwater acoustic network Double-SetOnline Fountain Code Based on State Feedback according to the present application coding and decoding method includes coding and decoding process under lossless channel condition and coding and decoding process under lossy channel condition; Coding and decoding process under lossless channel condition is specifically: Establishment stage: first, the sender according to Double-Set coding strategy, all original packages are divided into two sets, unselected set And selected set ; First send package, sender randomly selects original package in set Generation 2 code package is sent, and the receiver updates decoding graph Also update two sets, selected original package is placed in set ; Then every time select package, sender randomly selects an original package from set And set XOR generation 2 code package is sent, until the maximum connected component reaches Wherein, Represent all original package quantity; Represent the proportion of maximum connected component in establishment stage, and the receiver sends feedback package containing information to the sender: maximum connected component has been formed;Then, randomly generate degree 1 code package in set And send until the maximum connected component decoding is successful, and the receiver then sends feedback package containing information: maximum component decoding is successful to the sender, at this time, establishment stage ends, enters completion stage; Completion stage: same as establishment stage, sender also adopts Double-Set coding strategy and divides the remaining undecoded original package into two sets, unselected set And selected set , Selected set after deleting code package in completion stage, Selected original package quantity in completion stage, initial state , Sequence number of original package in set , when , sender first from set Two original packets are randomly selected to generate degree-2 coded packets, and two sets are also updated, and the selected original packets are put into the set The receiver updates the decoding graph Then, the sender randomly selects one original packet from each of the two sets to form a degree-2 coded packet to be sent until the maximum connected component reaches , is a ratio representing the maximum connected component in the completion phase, the receiver sends two feedback packets and the process of decoding the maximum connected component by using the degree-1 coded packet is the same as that in the establishment phase; The process of constructing the maximum connected component is repeated in the completion phase, and finally, when the undecoded original packets are less than or equal to 4, the sender directly sends the degree-1 coded packet until all original packets are successfully decoded; The encoding and decoding process under the lossy channel condition is as follows: Establishment phase: if the entire set is not selected empty, the encoding process in the establishment phase is the same as that under the lossless channel condition; however, the feedback is different from that under the lossless channel condition, when the maximum connected component in the establishment phase reaches , the receiver sends the type-1 feedback packet to the sender; then, the minimum ID of the feedback maximum connected component is generated to form a degree-1 coded packet and is sent until the maximum connected component is successfully decoded, the receiver then sends the type-2 feedback packet to the sender to inform that the maximum component is successfully decoded, at this time, the establishment phase ends and enters the completion phase; Completion phase: when entering the completion phase, the set is the same as that in the completion phase under the lossless channel condition, a double-set encoding strategy is adopted, and the remaining undecoded original packets are divided into two sets, the unselected set and the selected set , wherein represents the selected set after the original packets to be decoded in the completion phase are deleted, represents the number of original packets selected in the completion phase, and in the initial state, , represents the number of original packets contained in the maximum connected component, represents the sequence number of the original packets in the set , and the sender randomly selects one original packet from each of the set and the set to form a degree-2 coded packet to be sent, and the receiver updates the decoding graph until the maximum connected component reaches , the receiver sends a type 1 feedback packet to the sender; then, the sender generates and sends a degree 1 coded packet with the smallest ID of the maximum connected component, until the decoding of the maximum connected component succeeds, the receiver then sends a type 2 feedback packet to the sender to inform the sender of the decoding success of the maximum connected component, and the above process of constructing the maximum connected component is repeated, finally, when the number of undecoded original packets is less than or equal to 4, the sender directly sends a degree 1 coded packet until all original packets are decoded successfully. When the set is empty, the sender encodes the original packets in the set using the improved double-set encoding strategy, that is, the sender randomly selects one original packet in the maximum connected component and randomly selects one original packet in the set to generate and send a degree 2 coded packet, and updates the decoding graph ; the process is repeated until a type 1 feedback packet containing the maximum connected component is received, and the subsequent encoding and decoding process is exactly the same as that when the set is not empty.

[0007] As a preferred embodiment, in the encoding and decoding process under the condition of a lossy channel, when the case occurs in the encoding process of the establishment stage, the double-set encoding strategy is no longer used, and the sender randomly selects two original packets from all original packets to generate and send a degree 2 coded packet, and the process is repeated until the maximum connected component in the decoding graph reaches , the receiver sends a type 1 feedback packet to the sender; then, the sender generates and sends a degree 1 coded packet with the smallest ID of the maximum connected component, until the decoding of the maximum connected component succeeds, the receiver then sends a type 2 feedback packet to the sender to inform the sender of the decoding success of the maximum connected component, and the above process of constructing the maximum connected component is repeated, finally, when the number of undecoded original packets is less than or equal to 4, the sender directly sends a degree 1 coded packet until all original packets are decoded successfully.

[0008] As a preferred embodiment, the type 1 feedback packet includes the information: the maximum connected component has been formed, and the smallest ID in the maximum connected component.

[0009] As a preferred embodiment, the type 2 feedback packet includes the information: the decoding of the maximum connected component succeeds, the formation of the maximum connected component, and the set of the smallest IDs of all connected components except the maximum connected component.

[0010] As a preferred embodiment, the state feedback strategy is as follows: (1) The information that the maximum connected component contained in the type 1 feedback packet has been formed is set as a 1-bit flag, when the flag is 1, it indicates that the maximum connected component has reached the proportion; the purpose of setting the 1-bit flag in the feedback packet is to avoid the case that the feedback packet occupies multiple bytes; (2) The connected state of the second maximum connected component in the type 2 feedback packet is represented by formula (1): (1) represents the number of original packets contained in the second maximum connected component, represents the serial number of an original packet in the set ; the connected state of the original packet in the second maximum connected component is encoded in ascending order of the serial number, and the connected state of the second maximum connected component is represented as: (2) In formula (2), there are bits in total, each bit represents an original packet, wherein 0 indicates that the original packet is not in the second maximum connected component, and 1 indicates that the original packet is in the second maximum connected component; (3) The state of the minimum ID set of all connected components except the second maximum connected component is similar to the connected state , and the state of the minimum ID set of the connected component is represented as: (3) In equation (3), bits represent the total number of original packets, wherein each bit represents an original packet; the value of 0 indicates that the corresponding original packet is the minimum original packet in the undecoded component in the serial number, and the values of the remaining bit positions are all 1; in addition, represents the total number of undecoded connected components, that is, there are 0s and 1s in the decoding state of the minimum ID set of the connected component; finally, the candidate minimum ID set is set according to the state of the minimum ID set of the connected component , wherein only contains the original packet with the minimum serial number in the undecoded connected component.

[0011] The application provides an online fountain code based on state feedback of an underwater acoustic network, which adopts the above-mentioned double-set online fountain code encoding and decoding method based on state feedback of an underwater acoustic network.

[0012] The beneficial effects of the application are as follows: ​This invention improves upon SROFC-LF by employing a dual-set coding strategy and a state feedback strategy, proposing the DSOFC-SF (Double-Set Online Fountain Code Based on State Feedback) coding mechanism. This mechanism introduces a dual-set coding strategy during the setup phase to accelerate the construction of the maximum connected component; during the completion phase, it uses both a dual-set coding strategy and a state feedback strategy to achieve fast encoding and decoding, and assigns new... The value is used to limit feedback and reduce the number of feedback packets. Experimental results show that, compared with the existing OFC, the DSOFC-SF coding mechanism has better performance in terms of both the number of required coding packets and the number of feedback packets, and is more suitable for UANs. Attached Figure Description

[0013] Figure 1 This is a schematic diagram illustrating the construction of the maximum connected component in the embodiment; Figure 2 This is a schematic diagram of the encoding and decoding process in the DSOFC-SF completion stage of the embodiment; Figure 3(a) is a schematic diagram illustrating the impact of the original number of packets on the overhead in the embodiment; Figure 3(b) is a schematic diagram illustrating the influence of the original number of packets on the number of feedback packets in the embodiment; Figure 4 For the example , and A diagram illustrating the relationship between the number of coded packets sent and the number of original packets recovered. Figure 5 For the example , , and A diagram illustrating the relationship between the number of coded packets sent and the number of original packets recovered. Figure 6(a) shows the implementation example. A diagram illustrating the impact on the number of required encoding packets; Figure 6(b) shows the implementation example. A diagram illustrating the impact on the number of feedback packets. Detailed Implementation

[0014] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0015] Example This embodiment provides a dual-set online fountain code and encoding / decoding method for underwater acoustic networks based on state feedback, which includes lossless channel conditions (i.e., channel erasure rate). The encoding and decoding process and the conditions under lossy channel conditions (i.e., channel erasure rate) The encoding and decoding process of Double-Set Online Fountain Code Based on State Feedback (DSOFC-SF) consists of two phases: the setup phase and the completion phase. Assume the sender needs to transmit... The receiver maintains a decoding graph of the original packet's block data. Initial stage of decoding , The decoded image contains A white dot, without a border.

[0016] The encoding and decoding process under lossless channel conditions is as follows: Setup Phase: First, the sender divides all raw packets into two sets according to a biset coding strategy: the unselected set and the... and the selected set The first time a packet is sent, the sender is in the set. The receiver randomly selects an encoded packet with a degree of 2 from the original packet and sends it; the receiver then updates the decoding graph. At the same time, update both sets, adding the selected original packages to the set. middle; Subsequently, each time a packet is selected, the sender selects from the set. and set Each packet is randomly selected from the original packets and XORed with a generated degree of 2 to send an encoded packet, until the maximum connected component is reached. ( ),in, Represents the total number of all original packets; Representing the proportion of the largest connected components during the establishment phase, the receiver sends a feedback packet to the sender containing information: the largest connected components have been formed; then, in the set The receiver randomly generates encoded packets of degree 1 and sends them until the largest connected component is successfully decoded. The receiver then sends back a feedback packet containing the information that the largest component has been successfully decoded. At this point, the establishment phase ends and the completion phase begins.

[0017] Completion Phase: Similar to the setup phase, the sender also employs a dual-set encoding strategy to divide the remaining undecoded raw packets into two sets: the unselected set and the... and the selected set , This represents the selected set after the decoded encoded packets have been removed during the completion phase. This indicates the number of original packets selected during the completion phase, in the initial state. , Represents a set The serial number of the original packet, when At that time, the sender first selects from the set The system randomly selects two raw packets and generates encoded packets with a degree of 2 to send. Simultaneously, it updates two sets by adding the selected raw packets to the set. In the middle; the receiver updates the decoding graph. Subsequently, the sender randomly selects one original packet with a forming degree of 2 from each of the two sets and sends it until the maximum connected component is reached. , It is a proportion of the largest connected component in the completion phase. The process of the receiver sending two feedback packets and an encoded packet with a generation degree of 1 to decode the largest connected component is exactly the same as in the establishment phase. The process of building the maximum connected component is repeated throughout the completion phase. Finally, when the number of undecoded raw packets is less than or equal to 4, the sender directly sends encoded packets with a degree of 1 until all raw packets are successfully decoded. The encoding and decoding process under lossy channel conditions is as follows: Establishment phase: If the entire establishment phase set The coding process during the setup phase is exactly the same as under lossless channel conditions, except that the feedback differs from that under lossless channel conditions. When the maximum connected component reaches [a certain value] during the setup phase... At this point, the receiver sends a type 1 feedback packet to the sender; then it sends an encoded packet of the minimum ID generation degree 1 of the maximum connected component until the maximum connected component is successfully decoded. The receiver then sends a type 2 feedback packet back to the sender to inform the sender that the maximum component has been successfully decoded. At this point, the establishment phase ends and the completion phase begins.

[0018] When the phase coding process occurs In such cases, the dual-set encoding strategy is no longer used. Instead, similar to the OFC setup phase, random encoding is performed. The sender randomly selects two raw packets from all raw packets to generate degree-2 encoded packets and sends them. This process is repeated until the maximum connected component in the decoded graph is reached. The receiver sends a type 1 feedback packet to the sender; then it sends an encoded packet with a minimum ID generation degree of 1 for the maximum connected component until the maximum connected component is successfully decoded. The receiver then sends a type 2 feedback packet back to the sender to inform the sender that the maximum component has been successfully decoded, and enters the completion phase.

[0019] Completion Phase: When entering the completion phase, the set At the same time, similar to the completion phase under lossless channel conditions, a dual-set coding strategy is adopted to divide the remaining undecoded raw packets into two sets: the unselected set and the unselected set. and the selected set ,in, This represents the selected set after the original decoded packets have been removed during the completion phase. This indicates the number of original packets selected during the completion phase, in the initial state. , This indicates the number of original packets contained in the largest connected component. Represents a set The sequence number of the original packet; the sender from the set and set The receiver randomly selects one original packet with a degree of 2 and sends it, and updates the decoding graph. until the largest connected component reaches The receiver sends a type 1 feedback packet to the sender; then, it generates an encoded packet of degree 1 for the minimum ID of the maximum connected component and sends it, until the maximum connected component of the completed phase is successfully decoded. The receiver then sends a type 2 feedback packet back to the sender to inform that the maximum component has been successfully decoded. The process of building the maximum connected component is repeated. Finally, when there are 4 or fewer undecoded original packets, the sender directly sends an encoded packet of degree 1, until all original packets are successfully decoded.

[0020] When set During the completion phase, the sender uses a status feedback strategy to update the minimum ID set. The original packets in the first and second-maximum connected components are encoded using an improved binary encoding strategy. Specifically, the sender randomly selects one original packet from the second-maximum connected component and simultaneously encodes the original packets from the second-maximum connected component. Randomly select one original packet with a generation degree of 2 to generate an encoded packet and send it, then update the decoding graph. Repeat this process until a component containing the maximum connectivity has been received. For type 1 feedback packets, besides selecting the set of changes from the original packet, the subsequent encoding and decoding processes are similar to the set. The encoding and decoding processes are completely identical.

[0021] Specifically, the biset encoding strategy is described in detail below: All raw data packets are divided into two sets: the unselected set. and the selected set (in, This represents the number of original packets that were not selected, in the initial state. , and These represent the sequence numbers of the original packets in the two sets, respectively. During the first packet selection, the sets... If empty, it will be placed in the set. The receiver randomly selects two raw packets, performs an XOR operation to generate an encoded packet with degree 2, and sends it to the receiver. The sequence number of the selected raw packets is then removed from the set. Remove from the middle and put into the collection In the next step, each time the original package index that has been selected will be removed from the set. Remove from the middle and put into the collection In the middle; except for the first time, each time a package is selected, it is from the set. Randomly select one original packet from the set and simultaneously select one original packet from the set. One original packet is randomly selected from the array and XORed to generate an encoded packet of degree 2, which is then sent to the receiver. Without considering channel packet loss, each subsequent encoded packet will always have one original packet connected to the maximum connected component. Therefore, the dual-set packet selection strategy can quickly achieve the maximum connected component. The process of forming the maximum connected component is as follows: Figure 1 As shown. Figure 1 In (a), the set Contains circles , , ;gather Includes The circles represent the original packets in the set, and the edges between the circles represent the encoded packets. The blue circles form the maximally connected component of the decoding graph. Continue by randomly selecting three degree-2 encoded packets from the two sets of original packets and sending them. , and Then, the formation process of the largest connected component is as follows: Figure 1 As shown in (b)(c)(d).

[0022] When the sender uses a biset encoding strategy, it continuously updates the selected set as soon as an encoded packet is generated. and unselected sets Because the underwater acoustic channel has a high bit error rate, some encoded packets may be corrupted or even lost. Therefore, there will be unselected sets. All the original packages in the file were selected (i.e.) In this case, besides the existence of a maximum connected component, the receiver will also have other connected components. However, the probability of this situation occurring during the establishment phase is very small. In this case, the binary encoding strategy is no longer used. Instead, the sender randomly selects two encoded packets with a generation degree of 2 from the original packets and sends them until the maximum connected component is reached. Therefore, when the set The coding process for the completion stage will be described in detail.

[0023] After the largest connected component is successfully decoded during the setup phase, the receiver sends the sender information including "the composition of the second largest connected component" and "the minimum set of IDs for all connected components". (in, This represents the total number of undecoded connected components. The feedback packet represents the smallest sequence number in a connected component. Therefore, during the completion phase, one original packet is randomly selected from the second largest connected component in the feedback, and then the smallest ID set is selected. One original packet is randomly selected from the data, and an encoded packet with a degree of 2 is generated and sent to the receiver.

[0024] The specific status feedback strategy is as follows: Whether in the setup or completion phase, the receiver needs to provide feedback when the maximum connectivity component is reached and when decoding of the maximum connectivity component is successful. In a lossless channel, the feedback packet only needs to report whether the maximum connectivity component has been reached and whether decoding was successful; the composition of the feedback packet is simple and will not be described in detail. Because the bit error rate of underwater acoustic channels is high, this embodiment focuses on the feedback packet under lossy channel conditions. When the maximum connectivity component is reached, the receiver sends a type 1 feedback packet to the sender; after successful decoding of the maximum connectivity component, the receiver sends a type 2 feedback packet to the sender. The definitions of the two types of feedback packets are as follows: Type 1: This feedback packet contains the information "The largest connected component has been formed + the smallest ID in the largest connected component".

[0025] Type 2: This feedback packet contains information such as "the largest connected component was successfully decoded + the composition of the second largest connected component + the set of the smallest IDs of all connected components except the second largest connected component".

[0026] The primary function of Type 1 feedback packets is to inform the sender of the minimum ID in the largest connected component, enabling the sender to decode the largest connected component as quickly as possible. Type 2 feedback packets report the composition of the second largest connected component, aiming to achieve the largest connected component as quickly as possible. Feedback of the minimum IDs of all other connected components is intended to reduce the number of loops within the same component. When sending feedback packets, due to varying numbers of original packets, there may be cases where the original packet sequence numbers are very large and occupy multiple bytes. If the feedback packet is filled with the original packet sequence numbers, the length of the feedback packet may be very long. Since the bit error rate of underwater acoustic channels is high, the longer the feedback packet, the more prone it is to errors. Therefore, to reduce the length of the feedback packet, a state feedback strategy is proposed for the above feedback packet types, mainly consisting of three parts, with the strategy settings as follows: (1) Set the information "maximum connected components have been formed" in the Type 1 feedback packet to occupy 1 flag bit. When the flag bit is 1, it means that the maximum connected components have reached the ratio. The purpose of setting 1 flag bit in the feedback packet is to avoid the feedback packet occupying multiple bytes.

[0027] (2) The connectivity state of the second-largest connected component in the Type 2 feedback packet is represented by formula (1): (1) This indicates the number of original packets contained in the second-largest connected component. Represents a set The sequence number of a certain original packet; to reduce the length of the feedback packet, the connectivity states of the original packets in the second-largest connected component are encoded in ascending order of their sequence numbers. The connectivity state of the second-largest connected component is then represented as: (2) There are a total of in equation (2) Each bit represents an original packet, where 0 indicates that the original packet is not in the second-largest connected component, and 1 indicates that the original packet is in the second-largest connected component; for example, there are 10 original packets, with the sequence number set as follows: The information of the receiver forming the second largest connected component is The connectivity state of the second-largest connected component is represented as follows:

[0028] (3) The state of the minimum ID set of all connected components except the second largest connected component. With connected state Similarly, Figure 2(a) shows the decoding graph just entering the completion stage, where there are 3 undecoded connected components, namely... , and .like Figure 2 As shown, the receiver returns the smallest ID in each undecoded component, which are 2, 1, and 12 respectively. Then, based on the set of smallest IDs, it randomly selects the original packet encoding and sends the encoded packets to the receiver sequentially. The decoding diagram is shown below. Figure 2 As shown in (b), (c), and (d), in this case, the state of the minimum set of IDs of the connected components. Represented as: (3) In equation (3), `bits` represents the total number of raw packets, where each bit represents one raw packet; a value of 0 indicates that the corresponding raw packet is the smallest sequence number in its undecoded component, while the values ​​of the remaining bits are all 1; furthermore, This represents the total number of undecoded connected components, that is, the number of connected components that exist in the decoding state of the smallest set of connected component IDs. 0 and One by one; finally, based on the state of the minimum ID set of connected components. Set the minimum set of candidate IDs ,in It contains only the original packet with the smallest sequence number in the undecoded connected components.

[0029] In summary, this feedback strategy addresses the issue of excessive packet loss in the underwater acoustic channel, where the ensemble... The selection of empty packets avoids the situation where subsequent packet selection results in many useless packets. This effectively prevents the transmission of too many redundant packets during the completion phase, reduces the energy consumption of UANs, decreases the length of feedback packets, reduces the probability of errors in feedback packets, and improves channel utilization.

[0030] This embodiment provides an online fountain code based on state feedback for an underwater acoustic network, which adopts the above-described dual-set online fountain code encoding and decoding method based on state feedback for an underwater acoustic network.

[0031] Simulation Experiment To verify the performance of DSOFC-SF, it was compared with existing technologies such as SROFC-LF, OFI-OFCNB (Optimized Feedback Iterative Online Fountain Code without Build-up Phase), P-ZROFC-LF (Priority-Based and Zigzag-Decodable online fountain codes for underwateracoustic networks), ICM-OFC (Online Fountain Code with An Improved Caching Mechanism), ROFC-LF, OFC, IOFC-SLDD (Improved online fountain codes based on shaping for left degree distribution), IOFC (Improvement of online fountain codes), ZDOF-BDM (Zigzag Decodable Online Fountain with Buffer Decoding Method), WOFC (Weighted online fountain codes), and OFCNB-EBDS (Estimation Based Degree Selection scheme for Online Fountain Codes with No Build-up Phase). To ensure the reliability of the experimental results, the average of multiple repeated experiments was taken as the final experimental result. Unless otherwise specified in the experiment, the channel erasure rate will be assumed. Since the simulation parameter settings differ between the comparison techniques, the DSOFC-SF proposed in this embodiment is plotted and compared with the above techniques for analysis.

[0032] Simulation experiments revealed that the number of feedback cycles during the completion phase was related to... The value is inversely proportional to the number of encoded packets sent. The values ​​are basically proportional (see Table 1 for specific parameter values). Because the bit error rate of underwater acoustic channels is relatively high, the channel erasure rate is generally... To optimize both the number of feedback packets and the number of encoding packets, then... The value is set to 0.8.

[0033] Table 1. Statistics on the number of encoded packets and feedback packets under different parameter combinations. ; Figures 3(a) and 3(b) are The impact of OFC, IFOC, SROFC-LF, P-ROFC-LF, and DSOFC-SF on the number of feedback packets and overhead is investigated. Based on the simulation parameters of these four mechanisms, DSOFC-SF adopts... , , The parameters are as follows. When recovering the same number of original packets, IFOC and OFC consistently have higher overhead, which remains relatively constant. The overhead of DSOFC-SF remains relatively constant at around 0.1, which is significantly lower than the overhead of the above coding mechanisms. Compared with SROFC-LF, P-ZOFC-LF, IOFC, and OFC, DSOFC-SF reduces overhead by an average of 13.7%, 14.7%, 23.9%, and 32.3%, respectively.

[0034] Regarding the number of feedback packets, see Figure 3(b). The number of feedback packets for IFOC, OFC, SROFC-LF, and P-ROFC-LF all increase with... The number of feedback packets increases significantly with the increase of the value, while the number of feedback packets in DSOFC-SF remains stable, consistently around 11. Specifically, the number of feedback packets in DSOFC-SF is reduced by an average of 7.97 compared to P-ZROFC-LF; 18.7 compared to SROFC-LF; 32.37 compared to IOFC; and 41.05 compared to OFC. This result indicates that DSOFC-SF has a significant advantage in feedback efficiency. For UANs, excessive feedback packets consume valuable channel resources and increase communication latency; at the same time, higher overhead means the generation of more invalid coded packets, thereby increasing the energy consumption of underwater acoustic communication equipment and reducing channel bandwidth efficiency. Comprehensive performance analysis shows that DSOFC-SF is more suitable for UANs.

[0035] Figure 4 Showing , and The relationship between the number of encoded packets sent and the number of original packets recovered is analyzed. During the setup phase, OFI-OFCNB first sends degree 1 encoded packets, thus exhibiting good intermediate decoding performance from the outset. In contrast, DSOFC-SF employs a dual-set coding strategy for constructing the maximum connected component in both the setup and completion phases. At the beginning of both phases, only degree 2 encoded packets are sent; degree 1 encoded packets are only sent and the original packets are recovered after the maximum connected component is successfully constructed. Therefore, DSOFC-SF exhibits a more pronounced broken curve compared to other coding mechanisms. However, due to the dual-set coding and state feedback strategies employed by the DSOFC-SF mechanism, the dual-set updates are timely, resulting in fewer useless encoded packets. Therefore, it demonstrates better intermediate decoding capabilities in the later stages of decoding compared to OFI-OFCNB and OFC, and requires the fewest encoded packets to recover all original packets. Compared to P-ZROFC-LF, SROFC-LF, and IOFC, DSOFC-SF recovers all original packets much faster, regardless of the channel erasure rate.

[0036] Figure 5 set up , , and The relationship between the number of encoded packets sent and the number of original packets recovered for six encoding mechanisms (OFC, IFOC-SLDD, SROFC-LF, ZDOF-BDM, P-ZROFC-LF, and DSOFC-SF) was investigated. Before sending 651 encoded packets, ZDOF-BDM showed stronger intermediate decoding performance, while after sending 915 encoded packets, DSOFC-SF demonstrated superior intermediate decoding performance. Furthermore, when recovering all original packets, DSOFC-SF required fewer encoded packets than P-ZROFC-LF, SROFC-LF, IOFC, and OFC, making it more suitable for UANs.

[0037] To compare with the five mechanisms mentioned above, the parameters of DSOFC-SF are set according to the parameters of the five encoding mechanisms: IOF-OFCNB, ROFC-LF, OFC, IOFC, and WOFC. , , As can be observed from Figure 6(a), the number of encoding packets required by the six encoding mechanisms all increases with... The value increases significantly. The DSOFC-SF and OFI-OFCNB encoding mechanisms require a minimum number of encoded data packets to recover the same amount of original packets. However, as shown in the figure, the DSOFC-SF proposed in this embodiment requires the fewest encoded data packets. In Figure 6(b), The impact of each of the six encoding mechanisms (OFC, IFOC, WOFC, ROFC-LF, OFI-OFCNB, and DSOFC-SF) on the number of feedback packets required shows that DSOFC-SF requires the fewest feedback packets to recover the same number of original data packets.

[0038] Table 2 gives , , and The experiment compared the number of required coding packets and feedback packets between DSOFC-SF and other online fountain code coding mechanisms. The results show that DSOFC-SF reduces the number of coding packets by 27.27, 42.87, 46.15, and 319.87 respectively compared to P-ZROFC-LF, ROFC-LF, SROFC-LF, and OFCNB-EBDS; and reduces the number of feedback packets by 4.42, 6.18, and 5.36 respectively compared to P-ZROFC-LF, ROFC-LF, and SROFC-LF. These data clearly demonstrate that DSOFC-SF exhibits significant advantages in coding efficiency and feedback overhead, making it more suitable for UANs with high channel erasure rates.

[0039] Table 2. Number of encoding packets and feedback packets required for five encoding mechanisms ( , , , ) ; Based on the experimental data analysis in Table 3, it can be concluded that different coding mechanisms have different effects on various aspects. Performance varies significantly under certain conditions. At that time, the DSOFC-SF coding mechanism required 144.44, 240.99, and 174.5 fewer coding packets than ICM-OFC, OFC, and IOFC, respectively; and correspondingly fewer feedback packets by 10.8, 21.62, and 14.26. With... When the value increases to 1500: the performance difference widens further, with DSOFC-SF showing a more significant advantage in the number of encoded packets, reducing them by 219.2, 369.19, and 265.78 packets respectively compared to the comparison mechanism; the number of feedback packets also shows a similar trend, with the largest reduction reaching 17.53 packets. When the value is increased to 2000, the advantages of DSOFC-SF continue to strengthen, with the reduction in the number of encoded packets expanding to 309.35-511.32, and the reduction in the number of feedback packets reaching 13.04-26.18. At that time, compared with other mechanisms, DSOFC-SF maintained the best performance, with the number of encoded packets continuously decreasing, reaching a maximum of 649.27, and the number of feedback packets decreasing by a maximum of 28.06. Experimental data fully demonstrates that DSOFC-SF outperforms other comparative mechanisms in both encoding efficiency and feedback, and that as... As the value increases, its performance advantages become more prominent; therefore, this indicates that DSOFC-SF is more suitable for UANs.

[0040] Table 3 Required number of encoding packets and feedback packets ( , , , ) ; This embodiment proposes a dual-set OFC (DSOFC-SF) coding mechanism with low overhead, low energy consumption, and minimal feedback. Secondly, DSOFC-SF employs a dual-set coding strategy in both the setup and completion phases to quickly form the maximum connected component (UAN). Furthermore, under lossy channel conditions, a state feedback strategy is used to accelerate encoding and decoding efficiency, effectively reduce the number of feedback packets, and improve channel utilization. Finally, experimental results show that compared with existing online fountain codes, DSOFC-SF exhibits better performance in terms of both overhead and the number of feedback packets; therefore, DSOFC-SF is more suitable for UANs.

[0041] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A dual-set online fountain code encoding and decoding method based on state feedback in underwater acoustic networks, characterized in that: The coding process under the lossless channel condition and the coding process under the lossy channel condition are included; The coding process under the lossless channel condition is specifically: Establishment phase: first, the sender according to the double set encoding strategy, all the original package is divided into two sets, not selected over the set And the selected set ; The first time, the sender selects a degree-2 encoding packet from the set and sends it, and the receiver updates the decoding graph and also updates the two sets, putting the selected original packet into the set ​ Then, the sender randomly selects one original packet from the set and one original packet from the set to generate a degree-2 encoded packet by XOR operation, and sends it until the maximum connected component reaches , where represents the number of all original packets; represents the proportion of the maximum connected component in the establishment stage, and the receiver sends a feedback packet to the sender containing the information that the maximum connected component has been formed; then, the receiver randomly generates a degree-1 encoded packet from the set and sends it until the maximum connected component is successfully decoded, and the receiver then sends a feedback packet to the sender containing the information that the maximum connected component is successfully decoded, at which time the establishment stage ends and the completion stage begins. Completion phase: same as the setup phase, the sender divides the remaining undecoded original packets into two sets, the unselected set and the selected set , denotes the selected set after the sender removes the decoded coded packets in the completion phase, denotes the number of original packets selected in the completion phase, initially , denotes the sequence number of the original packets in the set When , the sender randomly selects two original packets from the set to generate a degree-2 coded packet and sends it, while also updating the two sets by placing the selected original packets into the set ; the receiver updates the decoding graph ; then the sender randomly selects one original packet from each of the two sets to form a degree-2 coded packet and sends it until the maximum connected component reaches , is a ratio that represents the maximum connected component in the completion phase, and the receiver sends two feedback packets and decodes the maximum connected component using degree-1 coded packets in the same way as in the setup phase; The process of repeatedly constructing the maximum connected component is completed, and finally, when the undecoded original packet is less than or equal to 4, the sender directly sends the coding packet with a degree of 1 until all original packets are successfully decoded; The coding process under the lossy channel condition is specifically: Establishment phase: if the entire establishment phase set is not selected empty, the encoding process of the establishment phase is exactly the same as that under the lossless channel condition; but the feedback is different from that under the lossless channel condition, when the maximum connected component of the establishment phase reaches , the receiving party sends the type 1 feedback packet to the sending party; then the minimum ID of the maximum connected component of the feedback is generated to the type 1 encoding packet and sent, until the maximum connected component is decoded successfully, the receiving party then sends the type 2 feedback packet to the sending party to inform that the maximum component is decoded successfully, at this time, the establishment phase ends and enters the completion phase; Completion phase: when entering the completion phase, the set is divided into two sets, the unselected set and the selected set , where represents the selected set after the completion phase deletes the decoded original packets, represents the number of original packets selected in the completion phase, and in the initial state , represents the number of original packets contained in the maximum connected component, represents the sequence number of the original packets in the set ; the sender randomly selects 1 original packet from the set and the set to form a degree-2 encoded packet for transmission, and the receiver updates the decoding graph until the maximum connected component reaches , and the receiver sends a type-1 feedback packet to the sender; then, the minimum ID of the maximum connected component generates a degree-1 encoded packet and is sent until the decoding of the maximum connected component in the completion phase is successful, and the receiver subsequently sends a type-2 feedback packet to the sender to inform that the decoding of the maximum component is successful, and the above process of building the maximum connected component is repeated, and finally, when the number of undecoded original packets is less than or equal to 4, the sender directly sends a degree-1 encoded packet until all original packets are successfully decoded. When set , the completion phase sends the original packets in the minimum ID set and the original packets in the second largest connected component by state feedback strategy. The sender encodes the original packets in the second largest connected component and the original packets in the set using the improved double-set encoding strategy, that is, the sender randomly selects one original packet in the second largest connected component, and randomly selects one original packet in the set to generate a coded packet with a degree of 2 and sends it, and updates the decoding graph ; repeat the process until a feedback packet containing type 1 that has reached the largest connected component is received. Except for the change in the set of selecting original packets, the subsequent encoding and decoding process is exactly the same as that when the set 2. The state feedback based two-set online fountain code encoding and decoding method of underwater acoustic network according to claim 1, characterized in that: In the case of , the double-set encoding strategy is not used any more, and the sender randomly selects two original packets from all original packets to generate a degree-2 encoded packet and sends it, repeating the process until the largest connected component in the decoding graph reaches , the receiver sends a feedback packet of type 1 to the sender. The coding packet with a degree of 1 is generated and sent according to the minimum ID of the maximum connected component, and the decoding of the maximum connected component is successful, and the receiver sends the feedback packet of type 2 to the sender to inform the decoding success of the maximum connected component, and enters the completion stage.

3. The state feedback based two-set online fountain code encoding and decoding method of the underwater acoustic network according to claim 2, characterized in that: The feedback packet of type 1 includes information: the maximum connected component has been formed, and the minimum ID of the maximum connected component.

4. The method of claim 3, wherein the state feedback-based two-set online fountain code encoding and decoding method for underwater acoustic networks is characterized in that: The feedback packet of type 2 includes information: the decoding of the maximum connected component is successful, the formation of the second maximum connected component, and the set of minimum IDs of all connected components except the second maximum connected component.

5. The method of claim 4, wherein the state feedback-based two-set online fountain code encoding and decoding method for underwater acoustic networks is characterized in that: The state feedback strategy is specifically: (1) The information that the maximum connected component has been formed in the feedback packet of type 1 includes a 1-bit flag, and when the flag is 1, it means that the maximum connected component has reached the proportion; The purpose of setting a 1-bit flag in the feedback packet is to avoid the case that the feedback packet occupies multiple bytes; (2) The connected state of the second maximum connected component in the feedback packet of type 2 is represented by formula (1): (1); the number of original packets contained in the second largest connected component, the set the serial number of the original packet in the second largest connected component; the connected state of the second largest connected component is represented as: （2）; There are totally bits in formula (2), each bit represents an original packet, wherein 0 represents that the original packet is not in the second largest connected component, and 1 represents that the original packet is in the second largest connected component. (3) The state of the minimum ID set of all connected components except the next largest connected component With the connected state Similarly, the state of the minimum ID set of connected components Is represented as: （3）; In equation (3), bits represent the total number of original packets, wherein each bit represents an original packet; a value of 0 indicates that the corresponding original packet is the original packet with the smallest sequence number in the undecoded component, and the values of the remaining bit positions are all 1; in addition, represents the total number of undecoded connected components, i.e., there are 0s and 1s in the decoding state of the connected component minimum ID set; finally, the candidate minimum ID set is set according to the state of the connected component minimum ID set , wherein only contains the original packet with the smallest sequence number in the undecoded connected component.

6. A state feedback based two-set online fountain code for underwater acoustic networks, characterized in that: It adopts the underwater acoustic network double-set online fountain code coding and decoding method based on state feedback according to any one of claims 1-5.

Images