Random network coding based centerless broadband ad hoc network multipath anti-interference method

Abstract

This invention relates to the field of wireless networking technology, specifically to a multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding. It includes: relay nodes scanning the wireless spectrum to generate a network channel quality spectrum; generating network compensation commands based on forward link loss rate; responding to the commands by performing network packet reassembly on the basic network payload using the local network topology vector to generate link transmission compensation packets; executing a fluid routing distribution strategy based on channel quality to push the aggregated wireless relay data streams in parallel to the multipath next-hop nodes; and intercepting automatic retransmission requests based on the feedback effective arrival rank. This invention, through deep coupling of network layer topology reassembly and multipath routing scheduling, achieves proactive compensation for link loss and low-latency transmission in complex electromagnetic environments, significantly improving the throughput and reliability of ad hoc networks.

Description

Technical Field

[0001] This invention relates to the field of wireless networking technology, specifically to a multipath anti-interference method for decentralized broadband self-organizing networks based on random network coding. Background Technology

[0002] In modern communication technologies, decentralized broadband wireless networking technology is widely used in complex scenarios such as emergency rescue, military tactical communications, and maritime unmanned swarm collaboration due to its independence from fixed infrastructure, self-organization, self-healing, and rapid deployment capabilities. This type of wireless networking typically requires the ability to provide high-bandwidth, low-latency data transmission services even in dynamically changing topologies and complex electromagnetic environments. Existing wireless networking architectures primarily utilize peer-to-peer networks, multi-hop relay technology to extend coverage, and adaptive modulation and coding at the physical layer to improve link quality, in order to meet the growing demands for multimedia and real-time control data transmission.

[0003] However, existing wireless networking anti-interference technologies have significant limitations. Taking the closest existing technology as an example, traditional wireless networking communication typically uses a single-path routing mechanism (such as AODV, OLSR, etc.), that is, only one "optimal" path is selected between the source node and the destination node for data transmission. When this path encounters sudden strong interference or deep fading, two passive coping methods are often adopted: one is to rely on automatic retransmission requests at the link layer or error correction coding at the physical layer (such as LDPC), which will lead to a large amount of retransmission overhead when the interference is severe, seriously squeezing the bandwidth resources of the wireless network and causing severe latency jitter; the other is to adopt route switching or avoidance strategies (such as through physical movement or switching to alternative paths), which will introduce a long route rediscovery and handshake time, resulting in data flow interruption. This "reconnect after disconnection" or "retransmit after packet loss" mechanism makes existing wireless networks extremely prone to network congestion or even paralysis when facing high dynamic interference, and cannot guarantee the continuity and reliability of broadband services. Specifically, there is a need in this field for a wireless networking data transmission method that can utilize the diverse resources of network topology and achieve a shift from "passively avoiding interference" to "actively resisting interference" through multi-path concurrency and coding redundancy techniques.

[0004] To address this issue, a multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding is proposed. This method aims to solve the technical problems of high packet loss, high latency, and low throughput in data transmission under strong interference scenarios. Summary of the Invention

[0005] The purpose of this invention is to provide a multipath anti-interference method for decentralized broadband self-organizing networks based on random network coding. By constructing a closed-loop control mechanism for dynamic multipath routing, reliable data transmission, elimination of retransmission delay, and improvement of throughput are achieved in environments with strong interference.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A multipath interference mitigation method for decentralized broadband ad hoc networks based on random network coding includes:

[0008] During the routing and forwarding process of relay nodes in a wireless ad hoc network, the surrounding wireless network spectrum environment is scanned to generate a network channel quality spectrum; the link evaluation algorithm is invoked to calculate the forward link loss rate of the next hop node corresponding to the current node based on the network channel quality spectrum;

[0009] The forward link loss rate is compared with the preset network loss threshold to generate a network compensation command; the relay node receives the wireless heterogeneous data packets from the upstream node and responds to the network compensation command by extracting the basic network payload from the relay buffer queue.

[0010] Using locally generated network topology vectors, network packet reassembly calculations are performed on the basic network payload to generate link transmission compensation packets controlled by network compensation commands;

[0011] Read the wireless network topology information and identify the network neighbor list; aggregate the link transmission compensation packet with the basic network payload to construct the wireless relay data stream; based on the network channel quality spectrum, adopt a fluid routing distribution strategy to push the wireless relay data stream in parallel to the next hop node in the network neighbor list;

[0012] Receive link status feedback from the next-hop node, obtain the effective arrival rank, intercept automatic retransmission requests based on the effective arrival rank, and update network compensation instructions.

[0013] Preferably, the specific process of generating the network channel quality spectrum includes: capturing wireless probe pilot signals sent by adjacent nodes within a preset wireless listening time window; measuring the received power intensity of the wireless probe pilot signals and simultaneously detecting the background noise level of the current frequency band; calculating the channel interference congestion index of the wireless link by combining the received power intensity and the background noise level; and constructing a network channel quality spectrum that includes the channel interference congestion index, the received power intensity, and the background noise level.

[0014] Preferably, the specific process for calculating the forward link loss rate includes: running a link evaluation algorithm to perform feature analysis on the network channel quality spectrum, extracting the channel interference congestion index, and calculating the frequency band signal-to-noise ratio (SNR) value based on the received power intensity and background noise level recorded in the network channel quality spectrum; using the link quality mapping relationship built into the link evaluation algorithm to match the basic transmission packet error probability corresponding to the frequency band SNR value; executing the interference weighting correction logic of the link evaluation algorithm to adjust the basic transmission packet error probability using the channel interference congestion index, generating an instantaneous link loss index; performing moving average filtering to smooth the instantaneous link loss index within a continuous time period, and outputting the forward link loss rate.

[0015] Preferably, the specific process of generating the network compensation instruction includes: calculating the link loss deviation value where the forward link loss rate exceeds a preset network loss threshold; querying a preset transmission compensation mapping table based on the link loss deviation value to obtain a topology mapping redundancy coefficient that adapts to the current link state; calculating the compensation packet count value to be supplemented based on the topology mapping redundancy coefficient, defining the compensation packet count value as a reassembly parameter, and outputting a network compensation instruction containing the reassembly parameter.

[0016] Preferably, the specific process of extracting the basic network payload includes: parsing the network compensation instruction to obtain the target generation identifier; traversing the relay buffer queue to filter out the wireless heterogeneous data packets carrying the target generation identifier; parsing the preceding topology feature vector of the wireless heterogeneous data packets and performing vector space differential comparison; retaining the wireless heterogeneous data packets with differential topology features and outputting them as the basic network payload.

[0017] Preferably, the specific process of generating the link transmission compensation packet includes: parsing the network compensation instruction to obtain reassembly parameters; generating a network topology vector based on the reassembly parameters; performing topology mapping processing on the basic network payload using the network topology vector to complete the network packet reassembly operation and generate topology mapping data; encapsulating the network topology vector and the topology mapping data to output the link transmission compensation packet.

[0018] Preferably, the specific process of constructing the wireless relay data stream includes: monitoring the wireless control channel, parsing the periodically interacting topology discovery signaling, and updating the local dynamic routing table; using the local dynamic routing table as wireless network topology information, identifying the network neighbor list based on the wireless network topology information, parsing the network neighbor list, and extracting the next-hop route identifier; performing frame-level multiplexing of the basic network payload and the link transmission compensation packet to generate mixed payload data; encapsulating the next-hop route identifier into the mixed payload data to construct a network transmission frame; and outputting a wireless relay data stream containing consecutive network transmission frames.

[0019] Preferably, the specific process of pushing the wireless relay data stream to the next-hop node includes: parsing the network channel quality spectrum and extracting the link throughput of each node in the network neighbor list; calculating the dynamic traffic splitting ratio of different nodes based on the link throughput; executing a fluid routing distribution strategy according to the dynamic traffic splitting ratio, performing traffic adaptation mapping on the wireless relay data stream, and constructing a split transmission queue; triggering a multi-channel transmission mechanism to deliver the wireless relay data stream temporarily stored in the split transmission queue to the next-hop node in the network neighbor list in parallel.

[0020] Preferably, the process of intercepting the automatic retransmission request and updating the network compensation instruction includes: monitoring the reverse wireless channel and capturing the link status feedback sent by the next-hop node; parsing the protocol control field of the link status feedback and extracting the receive rank indicator; converting the receive rank indicator into a numerical form of the effective arrival rank; calculating the difference between the effective arrival rank and a preset payload recovery threshold to generate a rank deviation; when the effective arrival rank is not lower than the payload recovery threshold, generating a virtual acknowledgment signal and blocking the automatic retransmission request; adjusting the reassembly parameters according to the rank deviation and generating an updated network compensation instruction.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0022] 1. This invention achieves a digital panoramic mapping of the radio electromagnetic environment by constructing a network channel quality spectrum that includes channel interference and congestion indicators, received power intensity, and background noise level. Compared to traditional evaluation methods that rely solely on signal strength, this scheme uses a topology mapping redundancy coefficient to finely offset link losses, ensuring that coding redundancy can accurately adapt to real-time channel quality in complex and variable electromagnetic interference environments. This guarantees transmission reliability while avoiding bandwidth waste caused by fixed redundancy.

[0023] 2. This invention introduces a "rank-based packet replacement" feedback mechanism based on the received rank indicator. By analyzing the relationship between the effective arrival rank and the payload recovery threshold, a virtual acknowledgment signal is proactively generated to shield the underlying automatic retransmission request when the data is confirmed to be decodeable. This "retransmission" mechanism eliminates the delay caused by multiple round-trip retransmissions due to packet loss in traditional protocols, achieving near real-time data delivery and greatly improving the service experience of broadband ad hoc networks in environments with long latency and high packet loss.

[0024] 3. This invention dynamically decomposes a single high-bandwidth data stream into sub-streams adapted to multiple differentiated links. By combining vector space differentiation comparison of heterogeneous wireless data packets, linearly correlated invalid duplicate information is eliminated before reassembly, ensuring that each transmitted link compensation packet possesses algebraic innovation. This mechanism fully utilizes fragmented spectrum and multipath spatial resources in ad hoc networks, significantly enhancing overall network throughput.

[0025] 4. When performing network packet reassembly operations, this invention analyzes the reassembly parameters and dynamically adjusts them in conjunction with the local computing power status. It uses randomly generated network topology vectors to linearly combine the basic network payload, ensuring the linear degrees of freedom of the data flow to cope with deep fading interference. Furthermore, the optimized matrix operation logic reduces the computational complexity and energy consumption of relay nodes when performing network coding, enabling its widespread application in resource-constrained portable or unmanned decentralized networking devices. Attached Figure Description

[0026] Figure 1 This is a flowchart of the multipath anti-interference method for decentralized broadband self-organizing networks based on random network coding according to the present invention;

[0027] Figure 2 This is a schematic diagram of the fluid routing and distribution mechanism according to an embodiment of the present invention;

[0028] Figure 3 This is a schematic diagram of the closed-loop feedback and zero-retransmission logic in an embodiment of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Please see Figures 1 to 3 This invention provides a multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding, and the technical solution is as follows:

[0031] Example 1

[0032] Reference Figure 1 This invention presents a flowchart of a multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding. This embodiment provides a specific application scenario for the method, specifically an application in a decentralized broadband ad hoc network scenario constructed by a drone swarm or emergency communication vehicle. The specific steps include:

[0033] During the routing and forwarding process of relay nodes in a wireless ad hoc network, the surrounding wireless network spectrum environment is scanned to generate a network channel quality spectrum; the link evaluation algorithm is invoked to calculate the forward link loss rate of the next hop node corresponding to the current node based on the network channel quality spectrum;

[0034] The forward link loss rate is compared with the preset network loss threshold to generate a network compensation command; the relay node receives the wireless heterogeneous data packets from the upstream node and responds to the network compensation command by extracting the basic network payload from the relay buffer queue.

[0035] Using locally generated network topology vectors, network packet reassembly calculations are performed on the basic network payload to generate link transmission compensation packets controlled by network compensation commands;

[0036] Read the wireless network topology information and identify the network neighbor list; aggregate the link transmission compensation packet with the basic network payload to construct the wireless relay data stream; based on the network channel quality spectrum, adopt a fluid routing distribution strategy to push the wireless relay data stream in parallel to the next hop node in the network neighbor list;

[0037] Receive link status feedback from the next-hop node, obtain the effective arrival rank, intercept automatic retransmission requests based on the effective arrival rank, and update network compensation instructions.

[0038] Furthermore, the specific process of generating the network channel quality spectrum includes: capturing the wireless probe pilot signals sent by adjacent nodes within a preset wireless listening time window; measuring the received power intensity of the wireless probe pilot signals and simultaneously detecting the background noise level of the current frequency band; calculating the channel interference congestion index of the wireless link by combining the received power intensity and the background noise level; and constructing a network channel quality spectrum that includes the channel interference congestion index, the received power intensity, and the background noise level.

[0039] Specifically, the relay node reserves a physical layer listening time slot of 1ms to 10ms within the communication cycle as a preset wireless listening window. Within this window, the node captures a wireless probe pilot signal via radio frequency. This signal is a reference signal broadcast by neighboring nodes at the physical layer, carrying a Zadoff-Chu sequence of preset length, specifically used for channel estimation and neighbor discovery. The node performs energy detection on the captured signal and quantizes it to obtain the received power intensity. This intensity is the absolute power value (in dBm) of the probe signal when it reaches the receiving antenna port, characterizing the remaining energy of the signal after spatial attenuation. Simultaneously, during idle time slots within the window, the node measures the ambient electromagnetic noise floor to obtain the background noise level. This level includes the energy reference value formed by the superposition of thermal noise and external environmental interference, excluding the useful signal. Subsequently, the processor calculates the difference between the received power intensity and the background noise level to obtain the current signal-to-noise ratio (SNR), and counts the percentage of sampling points where the signal energy exceeds a preset threshold within the time window to obtain the frequency band occupancy rate. The SNR and frequency band occupancy rate are then weighted and mapped to calculate a normalized channel interference congestion index (CIC). A higher CIC value indicates poorer communication quality for the current link. Finally, the measured CIC, received power intensity, and background noise level are indexed and associated by node ID, and stored as a structured network channel quality spectrum, serving as the underlying data set for subsequent link evaluation algorithms to perform precise calculations.

[0040] This invention constructs a network channel quality spectrum containing multi-dimensional physical layer parameters by precisely measuring pilot signal energy and environmental noise floor, realizing a digital panoramic mapping of the radio electromagnetic environment, and providing reliable data support for subsequent accurate calculation of link loss and formulation of anti-interference strategies.

[0041] Further, the specific process for calculating the forward link loss rate includes: running a link evaluation algorithm to perform feature analysis on the network channel quality spectrum, extracting the channel interference congestion index, and calculating the frequency band signal-to-noise ratio (SNR) value based on the received power intensity and background noise level recorded in the network channel quality spectrum; using the link quality mapping relationship built into the link evaluation algorithm to match the basic transmission packet error probability corresponding to the frequency band SNR value; executing the interference weighting correction logic of the link evaluation algorithm to adjust the basic transmission packet error probability using the channel interference congestion index, generating an instantaneous link loss index; performing moving average filtering to smooth the instantaneous link loss index within a continuous time period, and outputting the forward link loss rate.

[0042] Specifically, the steps for the relay node's processor to load and execute the link evaluation algorithm include: First, the algorithm reads the network channel quality spectrum from memory, parses the recorded channel interference and congestion indicators, and subtracts the background noise level in dBm from the received power intensity to calculate the band signal-to-noise ratio (SNR) in dB. Then, the algorithm queries a pre-set link quality mapping relationship, which is a "SNR-packet error rate" lookup table pre-stored in memory, obtained through statistical modeling of a large amount of historical wireless channel test data. Specifically, in an interference-free experimental environment, the SNR is adjusted in 0.5 dB steps, and the packet error rate under different modulation and coding strategies is recorded to construct the lookup table. For example, when the SNR is 10 dB and QPSK modulation is used, the corresponding basic transmission packet error probability is set to... The base packet error probability is obtained by matching the calculated signal-to-noise ratio (SNR). In real-world environments, non-Gaussian white noise interference exists. The algorithm executes interference-weighted correction logic, using the channel interference congestion index as an adaptive weighting factor applied to the base packet error probability, through a linear compensation formula: Calculate and generate instantaneous link loss metrics ,in Based on the probability of packet errors in transmission, This is an indicator of channel interference and congestion. The preset interference sensitivity coefficient is used to adjust the contribution weight of congestion index to link loss. To avoid drastic fluctuations in values ​​caused by rapid fading of the wireless channel, the algorithm performs a moving average filtering on the instantaneous link loss index within the most recent N transmission cycles (N ranges from 5 to 10), calculates its arithmetic mean, and outputs the final stable forward link loss rate.

[0043] This invention combines signal-to-noise ratio mapping with interference weighting correction and introduces a moving average filtering mechanism to eliminate the impact of instantaneous channel jitter on the evaluation results, thereby achieving accurate quantification of the actual packet loss risk of wireless links and providing a reliable basis for the subsequent formulation of precise compensation strategies.

[0044] Furthermore, the specific process of generating the network compensation instruction includes: calculating the link loss deviation value where the forward link loss rate exceeds a preset network loss threshold; querying a preset transmission compensation mapping table based on the link loss deviation value to obtain a topology mapping redundancy coefficient that adapts to the current link state; calculating the compensation packet count value to be supplemented based on the topology mapping redundancy coefficient, defining the compensation packet count value as a reassembly parameter, and outputting a network compensation instruction containing the reassembly parameter.

[0045] Specifically, the relay node processor first performs an interpolation operation. This operation involves subtracting a preset network loss threshold from the currently calculated forward link loss rate. The resulting difference is the link loss deviation value, which quantifies the degree to which the current link quality deviates from the reliability requirements. The preset network loss threshold is configured according to the service reliability level. For voice or video streaming services with high real-time requirements, this threshold is set between 3% and 8%; for command-based reliable transmission services, the threshold is set to 0.5%. This threshold serves as the boundary for initiating anti-interference compensation. Subsequently, the processor accesses a preset transmission compensation mapping table in memory. This table is pre-set based on the decoding recovery probability curve of the network coding, recording the coding redundancy requirements corresponding to different gradient deviation values, and is recorded in a piecewise function form.

[0046] First interval: When the link loss deviation value is in the first interval At that time, the corresponding transmission compensation redundancy coefficient is set to 1.1-1.2;

[0047] Second interval: When the deviation value is in the second interval At that time, the redundancy coefficient increases to 1.3-1.6;

[0048] Third interval: When the link loss deviation value meets the requirement At that time, the redundancy coefficient was set to 1.7-2.2 to cope with the algebraic rank loss caused by deep fading;

[0049] Saturation range: When the link loss deviation value At that time, the redundancy coefficient is fixed at the preset maximum threshold of 2.5, and the multipath route reselection logic is triggered synchronously.

[0050] Through the aforementioned segmented mapping, the processor obtains a topology mapping redundancy coefficient adapted to the current link state. This topology mapping redundancy coefficient is a value greater than 1, reflecting the necessary data redundancy ratio to ensure successful decoding by the receiver using a random network coding algorithm under the current network topology path. For example, when the link loss deviation is 10%, the topology mapping redundancy coefficient obtained from the table is 1.2, indicating that 1.2 times the original data volume of encoded packets needs to be sent. Next, the processor calculates the count of compensation packets to be added based on the topology mapping redundancy coefficient, defines it as a coding matrix reassembly parameter, and outputs a network compensation instruction containing the reassembly parameter to drive the random network coding engine to generate redundant packets.

[0051] This invention achieves precise matching between coding redundancy and real-time link quality by calculating link loss deviation and matching topology mapping redundancy coefficients by looking up a table; by generating network compensation instructions containing specific reassembly parameters, it ensures that bandwidth resources are not wasted due to over-coding while guaranteeing transmission reliability.

[0052] Furthermore, the specific process of extracting the basic network payload includes: parsing the network compensation instruction to obtain the target generation identifier; traversing the relay buffer queue to filter out the wireless heterogeneous data packets carrying the target generation identifier; parsing the preceding topology feature vector of the wireless heterogeneous data packets and performing vector space differential comparison; retaining the wireless heterogeneous data packets with differential topology features and outputting them as the basic network payload.

[0053] Specifically, the relay node's processor first parses the received network compensation instruction and reads the specified target generation identifier. This identifier is used to define the generational affiliation of the data packets involved in the current reassembly operation within the time axis or data stream. Since random network coding only performs linear combinations between data packets within the same generation, the target generation identifier establishes the dimensional boundary of the decoding matrix, ensuring that the relay node extracts only valid components belonging to the same generation from the buffer, avoiding decoding failures caused by inconsistencies in vector spaces between different generations. The processor then traverses the relay buffer queue in memory, filtering all temporarily stored heterogeneous wireless data packets based on the target generation identifier. The heterogeneous wireless data packets refer to radio electromagnetic signals encapsulated with specific coding coefficients, forwarded from nodes on different physical paths in the ad hoc network to the current node. For each filtered data packet, the processor extracts the preceding topological feature vector carried in its header. This preceding topological feature vector represents the orientation of the heterogeneous wireless data packet in the algebraic space. Subsequently, a vector space difference comparison is performed on the preceding topological feature vector.

[0054] The specific process of performing vector space differential comparison includes: constructing a corresponding spatial basis matrix space for the target generation identifier; performing linear combination elimination operation between the currently extracted pre-order topological feature vector and the existing basis vectors in the spatial basis matrix space; determining whether the residual vector after elimination is greater than a preset null space determination threshold; if the residual vector is greater than the null space determination threshold, then determining that the wireless heterogeneous data packet has differential topological features and incrementally updating its corresponding residual vector to the spatial basis matrix space. Specifically, the processor uses incremental Gaussian elimination to calculate the feature vector projection residual; if the residual vector magnitude is greater than... If a packet is found to be a novel data packet, its base matrix is ​​updated accordingly. The processor then aggregates these selected, valid data packets with differentiated topological characteristics to form the basic network payload, which serves as effective input material for subsequent secondary network packet reassembly operations. This scheme effectively reduces the computational burden caused by redundant packets and improves real-time performance through generational filtering and incremental spatial comparison.

[0055] This invention eliminates linearly correlated (i.e. invalid duplicate information) data packets before reassembly by filtering based on generational identifiers and vector space differentiation comparison, ensuring the purity and validity of the basic network payload, thereby avoiding invalid data from participating in subsequent encoding operations, significantly reducing computational overhead and improving network bandwidth utilization efficiency.

[0056] Furthermore, the specific process of generating the link transmission compensation packet includes: parsing the network compensation instruction to obtain reassembly parameters; generating a network topology vector based on the reassembly parameters; performing topology mapping processing on the basic network payload using the network topology vector to complete the network packet reassembly operation and generate topology mapping data; encapsulating the network topology vector and the topology mapping data to output the link transmission compensation packet.

[0057] Specifically, the relay node's encoding processor parses the reassembly parameters carried in the network compensation instruction. These parameters define the redundant packet count to be generated and the finite field order used in the operation. The processor then activates a random number generator to generate a network topology vector, which is a row vector composed of random coefficients selected within a finite field, equal to the number of basic network payload packets. Subsequently, the network topology vector is used to perform topology mapping processing on the basic network payload to complete the network packet reassembly operation. The specific process of performing the network packet reassembly operation includes: obtaining the current node's computing load status and the priority of the tasks to be sent; selecting a matching finite field order and computational sparsity based on the priority; within the selected finite field, using the non-zero coefficients in the network topology vector to perform weighted accumulation on the basic network payload, and synchronously updating the global encoding vector corresponding to the heterogeneous wireless data packets to generate the topology mapping data that meets the preset sparsity requirements. Specifically, the processor dynamically selects the computation domain based on the computing load and reduces the number of basic payloads participating in the multiply-accumulate operation through sparse operators, thereby reducing computational latency. The processor uses the network topology vector to perform topology mapping processing on the basic network payload, completing the network packet reassembly operation: the data content of each basic network payload is multiplied by the corresponding non-zero coefficient in the vector, and the product is XORed bit-by-bit to generate new topology mapping data. Finally, the network topology vector and topology mapping data are encapsulated to form a link transmission compensation packet. This scheme, through a computing power-aware dynamic reassembly mechanism, ensures efficient data compensation capabilities even in complex interference environments.

[0058] This invention generates link transmission compensation packets carrying entirely new linear relationships by performing network packet reassembly operations on the basic payload based on locally randomly generated network topology vectors. This mechanism not only increases the linear degrees of freedom of data flow in the network, but also effectively compensates for link loss during transmission, enabling the receiving end to recover the original information through the remaining compensation packets even if some data packets are lost.

[0059] Furthermore, the specific process of constructing the wireless relay data stream includes: monitoring the wireless control channel, parsing the periodically interacting topology discovery signaling, and updating the local dynamic routing table; using the local dynamic routing table as wireless network topology information, identifying the network neighbor list based on the wireless network topology information, parsing the network neighbor list, and extracting the next-hop route identifier; performing frame-level multiplexing of the basic network payload and the link transmission compensation packet to generate mixed payload data; encapsulating the next-hop route identifier into the mixed payload data to construct a network transmission frame; and outputting a wireless relay data stream containing consecutive network transmission frames.

[0060] Specifically, relay nodes monitor the wireless control channel, parse topology discovery signaling, and update their local dynamic routing tables in real time. The multidimensional dataset recorded in the local dynamic routing table, containing real-time link bandwidth and historical packet loss rates, is defined as wireless network topology information. Link stability indicators are extracted from this wireless network topology information, and anti-interference potential scores for each candidate neighbor node are calculated. These anti-interference potential scores are compared to a preset forwarding reliability threshold, and candidate nodes with scores below the threshold are eliminated. The processor traverses the wireless network topology information, filters out active nodes that meet the quality threshold to construct a network neighbor list, and extracts the physical address of the target forwarding node as the next-hop route identifier. Subsequently, the data link layer processor performs frame-level multiplexing.

[0061] The frame-level multiplexing process includes: real-time acquisition of burst interference intensity of the wireless link and calculation of anti-interference priority weights, and dynamic adjustment of the insertion interval of link transmission compensation packets in the mixed load data. Specifically, the processor changes the interleaving ratio of the basic network payload and compensation packets according to the instantaneous packet loss risk of the channel, so that the anti-interference protection density and interference intensity are adaptively matched. The processor performs frame-level multiplexing of the basic network payload to be forwarded and the newly generated link transmission compensation packets according to a determined density ratio to form mixed load data. Then, the processor encapsulates the next-hop routing identifier in the header of the mixed load data to construct a network transmission frame. Finally, the output is a wireless relay data stream containing continuous network transmission frames. This scheme ensures the reliability and efficiency of the data stream in multipath ad hoc networks through real-time topology mapping and dynamic mixed frame construction.

[0062] This invention uses a neighbor identification and frame-level multiplexing mechanism based on real-time topology information to dynamically integrate the basic payload and compensation packet into mixed payload data and encapsulate it into a network transmission frame. This construction method ensures that the wireless relay data stream can carry effective information and redundancy protection in a single transmission stream while adapting to dynamic changes in network topology, providing structured support for multipath reception and decoding of downstream nodes.

[0063] Furthermore, the specific process of pushing the wireless relay data stream to the next-hop node includes: parsing the network channel quality spectrum and extracting the link throughput of each node in the network neighbor list; calculating the dynamic traffic splitting ratio of different nodes based on the link throughput; executing a fluid routing distribution strategy according to the dynamic traffic splitting ratio, performing traffic adaptation mapping on the wireless relay data stream, and constructing a split transmission queue; triggering a multi-channel transmission mechanism to deliver the wireless relay data stream temporarily stored in the split transmission queue to the next-hop node in the network neighbor list in parallel. (Refer to...) Figure 2 This is a schematic diagram of the fluid routing and distribution mechanism according to an embodiment of the present invention.

[0064] Specifically, the relay node first accesses the network channel quality spectrum, reads the historical transmission rate and signal-to-noise ratio of each neighboring node, and estimates the link throughput of each neighboring node. Based on the link throughput, the processor calculates the dynamic traffic splitting ratio, i.e., the weight ratio of each neighboring path in the current total bandwidth. Subsequently, a fluid routing distribution strategy is executed, treating the wireless relay data stream as a dynamically divisible scheduling set, and performing traffic adaptation mapping on the data stream according to the dynamic traffic splitting ratio. Specifically, the fluid routing distribution strategy uses a round-robin mechanism, specifically a weighted round-robin scheduling algorithm, to normalize the dynamic traffic splitting ratio into a scheduling weight vector. The scheduler maintains a set of token buckets, and each time slot is allocated according to... Data packets are distributed to the sending queue of the corresponding next-hop interface to ensure that the traffic load of each path is linearly related to the link throughput. The data packet sequence is assigned to the corresponding logical path, and a split transmission queue is constructed. The mapped data packets are temporarily stored in a buffer independently maintained for each next-hop node. When the queue triggers the transmission threshold, a multi-channel transmission mechanism is activated. This multi-channel transmission mechanism calls the node's multiple physical radio frequency interfaces to simultaneously perform signal modulation and power amplification on orthogonal frequency bands or spatial streams, delivering data from each split transmission queue in parallel to the corresponding next-hop node. Through fluidized splitting and concurrent multi-channel delivery, full utilization of spatial multipath resources is achieved, improving anti-interference capability and overall throughput. The multi-channel transmission mechanism performs differentiated parallel delivery based on the radio frequency hardware configuration of the relay node: for devices with multiple radio frequency links or supporting multiple spatial stream transmission, the split transmission queues of different paths are synchronously mapped to orthogonal physical channels or spatial stream indices to achieve real-time concurrent transmission at the physical layer; for single-radio frequency architecture devices, high-speed time division multiplexing is used to poll and distribute data packets of each queue in different time slots, or orthogonal frequency division multiple access is used to map different queues to non-overlapping subcarrier clusters to achieve logical parallel delivery within a single MAC scheduling cycle.

[0065] This invention utilizes a fluid routing distribution strategy based on link throughput and a multi-channel transmission mechanism to dynamically decompose and adapt high-bandwidth data streams to multiple differentiated links, making full use of fragmented spectrum and spatial path resources in the network, and significantly improving the concurrent transmission capability and overall throughput of broadband self-organizing networks.

[0066] Further, the process of intercepting automatic retransmission requests and updating network compensation instructions includes: monitoring the reverse wireless channel and capturing the link state feedback sent by the next-hop node; parsing the protocol control field of the link state feedback and extracting the receive rank indicator; converting the receive rank indicator into a numerical form of the effective arrival rank; calculating the difference between the effective arrival rank and a preset payload recovery threshold to generate a rank deviation; when the effective arrival rank is not lower than the payload recovery threshold, generating a virtual acknowledgment signal and blocking the automatic retransmission request; adjusting the reassembly parameters according to the rank deviation and generating an updated network compensation instruction. (Refer to...) Figure 3 This is a schematic diagram of closed-loop feedback and zero retransmission logic in an embodiment of the present invention.

[0067] Specifically, in this embodiment, while transmitting data, the relay node captures the link state feedback sent by the next-hop node via the reverse wireless channel. This link state feedback is a short control frame, encapsulated with the current decoding matrix rank information, transmitted back by the receiver based on a generational cycle. The processor parses the frame header control field of this feedback frame, extracts the received rank indicator, and converts it into a numerical form of the effective arrival rank. This value represents the total number of linearly independent vectors currently held by the receiver. The processor reads the payload recovery threshold, which is set to the total number of packets of the original basic network payload within the current generation, representing the minimum generational rank required for successful decoding. Subsequently, the effective arrival rank is subtracted from the payload recovery threshold to generate a rank deviation characterizing the generational redundancy margin. The automatic retransmission request blocking logic is based on rank deviation: When the valid arrival rank is not lower than the payload recovery threshold (i.e., rank deviation is greater than or equal to 0), it indicates that the receiver has met the decoding conditions. At this time, the processor generates a virtual acknowledgment signal in the driver layer, specifically based on the SoftMAC architecture (such as the mac80211 subsystem). In the transmit completion interrupt handling function of the driver layer, it intercepts the transmission failure status reported by the hardware and forges a TX_OK status for the upper layer protocol stack. At the same time, it clears the corresponding cache descriptor in the driver software queue, thereby forcibly resetting the underlying retransmission counter and blocking the automatic retransmission request of the physical layer to eliminate retransmission latency. Conversely, when the rank deviation is less than 0, it is determined that there is insufficient redundancy. The processor releases the blocking restriction on ARQ, allowing the underlying protocol stack to initiate incremental retransmission requests for specific generations. When the rank deviation is less than 0 for N consecutive cycles, preemptive reassembly compensation is triggered through network compensation instructions. By actively increasing the redundancy ratio in the reassembly parameters, a dynamic balance between algebraic error correction and link retransmission is achieved. Finally, the processor dynamically adjusts the reassembly parameters for the next transmission cycle based on the sign and magnitude of the rank deviation, and generates an updated network compensation instruction.

[0068] This invention analyzes the relationship between the effective arrival rank and the payload recovery threshold, and actively generates a virtual confirmation signal in the confirmed decodeable state to shield the underlying automatic retransmission request, thus realizing a "zero retransmission" mechanism and completely eliminating the round-trip delay caused by retransmission. At the same time, it dynamically adjusts the subsequent network compensation instructions based on the rank deviation, establishing an adaptive closed-loop balance between anti-interference redundancy and channel quality.

[0069] Example 2

[0070] This embodiment provides an application scenario for a decentralized broadband ad hoc network multipath anti-interference method based on random network coding. Specifically, it is a decentralized broadband ad hoc network scenario constructed by a drone swarm or emergency communication vehicle, which ensures stable transmission of high-definition video streams or control commands in a complex electromagnetic interference environment.

[0071] In actual operation, the UAV communication equipment, acting as a relay node, uses its built-in radio frequency to continuously poll and scan the surrounding spectrum. For example, during breaks in data forwarding tasks, the radio frequency front-end quickly switches to sensing mode to detect the background noise floor and burst interference signal strength in the 2.4GHz or 5.8GHz operating frequency band and its adjacent bands, generating a network channel quality spectrum that maps the electromagnetic characteristics of the current spatial and frequency dimensions. The relay node's processor then calls the link evaluation algorithm in real time, using this quality spectrum as input parameters and combining it with historical transmission statistics to calculate the probability of data packet loss when the current UAV reaches its next-hop neighbor UAV, i.e., the forward link loss rate.

[0072] Assuming a preset network loss threshold of 10%, when the real-time loss rate rises to 20% due to environmental shielding or increased interference, the link is determined to be in a sub-healthy state, and an adaptive compensation mechanism is triggered to generate a network compensation command containing the target redundancy. At this time, the relay node extracts the heterogeneous wireless data packets to be forwarded (such as video clips sent by upstream drones) from the relay buffer queue in the memory buffer and parses their target generation identifier. This identifier ensures that the relay node only extracts valid components belonging to the same generation space from the buffer, avoiding encoding conflicts caused by inconsistencies in vector spaces between different generations.

[0073] The processor uses a random number generator to generate a set of random numbers in a finite field such as GF(2). 8 Linearly independent random coefficients within the vector form the network topology vector. During this process, the processor needs to determine the encoding generation size K, which is the number of basic network payloads participating in a single linear reassembly operation. The value of K is dynamically adjusted according to the service type; for high real-time instruction services, K ranges from 8 to 16, and for high-bandwidth video services, K ranges from 32 to 64. Next, the encoding engine performs a linear reassembly operation (i.e., network packet reassembly) on the extracted K basic network payloads, generating M algebraically innovative link transmission compensation packets through weighted modulo-2 addition. The specific value of M is obtained by multiplying the topology mapping redundancy coefficient by K and rounding up. For example, if the instruction requires a redundancy of 1.4 and K is set to 10, then while sending 10 original data packets, 4 additional compensation packets are generated. These compensation packets do not carry copies of specific original packets, but rather carry linear combination information of the original packet set, thus enabling the receiving end to have "packet loss-independent" recovery capabilities.

[0074] Subsequently, the node acquires the latest wireless network topology information by listening to the control channel, accurately identifies a list of network neighbors with forwarding capabilities in the dynamically changing flight formation, and performs frame-level multiplexing of the original payload and the generated compensation packet, aggregating them into a wireless relay data stream containing the next-hop MAC address identifier. Based on the throughput of each branch link estimated by the network channel quality spectrum, a fluid routing distribution strategy is executed. The processor simulates the data stream as a finely segmentable set of traffic, constructs a 1:2 dynamic traffic splitting ratio based on the capacity ratio of neighbor node A (throughput 10Mbps) and neighbor node B (throughput 20Mbps), and interleaves the data packet sequences to distribute them to different radio frequency physical channels, achieving multi-path concurrent parallel delivery.

[0075] Finally, the relay node continuously monitors the reverse channel, receiving link status feedback from downstream nodes and extracting the effective arrival rank representing the receiver's matrix state. If the effective arrival rank has reached the payload recovery threshold (i.e., the number of original packets), it proves that the receiver can recover all the original high-definition video information through matrix inversion using the received partial data packets and compensation packets. At this point, the relay node directly generates a virtual acknowledgment signal at the link layer driver layer, forcibly resetting the underlying retransmission counter. This intercepts and blocks automatic retransmission requests that might otherwise be triggered by the physical layer, avoiding bandwidth waste caused by repeated retransmission requests in interference environments. Simultaneously, the processor dynamically fine-tunes the redundancy scale for the next cycle based on the rank deviation (the difference between the effective rank and the threshold). If the difference is too large, the redundancy is reduced to release bandwidth; if the difference is too small, the redundancy is increased to enhance robustness, thus completing the closed-loop optimization of the entire process.

[0076] This invention introduces link evaluation-driven random network coding and multipath flow distribution mechanism into wireless ad hoc networks, achieving proactive adaptation to interference environments and real-time compensation for link loss. Combined with closed-loop feedback control based on effective arrival rank to intercept automatic retransmission, it effectively reduces network transmission latency and improves data transmission reliability and throughput.

[0077] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding, characterized in that, include: During the routing and forwarding process of relay nodes in a wireless ad hoc network, the surrounding wireless network spectrum environment is scanned to generate a network channel quality spectrum. The link evaluation algorithm is invoked to calculate the forward link loss rate of the next hop node corresponding to the current node based on the network channel quality spectrum. The forward link loss rate is compared with the preset network loss threshold to generate a network compensation command; the relay node receives the wireless heterogeneous data packets from the upstream node and responds to the network compensation command by extracting the basic network payload from the relay buffer queue. Using locally generated network topology vectors, network packet reassembly calculations are performed on the basic network payload to generate link transmission compensation packets controlled by network compensation commands; Read the wireless network topology information and identify the network neighbor list; aggregate the link transmission compensation packet with the basic network payload to construct the wireless relay data stream; Based on the network channel quality spectrum, a fluid routing distribution strategy is adopted to push the wireless relay data stream in parallel to the next-hop node in the network neighbor list; Receive link status feedback from the next-hop node, obtain the effective arrival rank, intercept automatic retransmission requests based on the effective arrival rank, and update network compensation instructions.

2. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process of generating the network channel quality spectrum includes: capturing the wireless probe pilot signals sent by adjacent nodes within a preset wireless listening time window; measuring the received power intensity of the wireless probe pilot signals and simultaneously detecting the background noise level of the current frequency band; calculating the channel interference congestion index of the wireless link by combining the received power intensity and the background noise level; and constructing a network channel quality spectrum that includes the channel interference congestion index, the received power intensity, and the background noise level.

3. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process for calculating the forward link loss rate includes: running a link evaluation algorithm to perform feature analysis on the network channel quality spectrum, extracting channel interference and congestion indicators, and calculating the frequency band signal-to-noise ratio (SNR) value based on the received power intensity and background noise level recorded in the network channel quality spectrum; using the link quality mapping relationship built into the link evaluation algorithm to match the basic transmission packet error probability corresponding to the frequency band SNR value; executing the interference weighting correction logic of the link evaluation algorithm to adjust the basic transmission packet error probability using the channel interference and congestion indicators, generating an instantaneous link loss indicator; performing moving average filtering to smooth the instantaneous link loss indicator within a continuous time period, and outputting the forward link loss rate.

4. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process of generating the network compensation command includes: calculating the link loss deviation value where the forward link loss rate exceeds the preset network loss threshold; querying the preset transmission compensation mapping table based on the link loss deviation value to obtain the topology mapping redundancy coefficient adapted to the current link state; calculating the compensation packet count value to be supplemented based on the topology mapping redundancy coefficient, defining the compensation packet count value as the reassembly parameter, and outputting the network compensation command containing the reassembly parameter.

5. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process of extracting the basic network payload includes: parsing the network compensation command to obtain the target generation identifier; traversing the relay buffer queue to filter out the wireless heterogeneous data packets carrying the target generation identifier; parsing the preceding topology feature vector of the wireless heterogeneous data packets and performing vector space differential comparison; retaining the wireless heterogeneous data packets with differential topology features and outputting them as the basic network payload.

6. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process of generating the link transmission compensation packet includes: parsing the network compensation instruction to obtain reassembly parameters; generating a network topology vector based on the reassembly parameters; performing topology mapping processing on the basic network payload using the network topology vector to complete the network packet reassembly operation and generate topology mapping data; encapsulating the network topology vector and the topology mapping data to output the link transmission compensation packet.

7. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process of constructing the wireless relay data stream includes: monitoring the wireless control channel, parsing the periodically interacting topology discovery signaling, and updating the local dynamic routing table; using the local dynamic routing table as wireless network topology information, identifying the network neighbor list based on the wireless network topology information, parsing the network neighbor list, and extracting the next-hop route identifier; performing frame-level multiplexing of the basic network payload and the link transmission compensation packet to generate mixed payload data; encapsulating the next-hop route identifier into the mixed payload data to construct a network transmission frame; and outputting a wireless relay data stream containing consecutive network transmission frames.

8. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The specific process of pushing the wireless relay data stream to the next-hop node includes: parsing the network channel quality spectrum and extracting the link throughput of each node in the network neighbor list; calculating the dynamic traffic splitting ratio of different nodes based on the link throughput; executing a fluid routing distribution strategy according to the dynamic traffic splitting ratio, performing traffic adaptation mapping on the wireless relay data stream, and constructing a split transmission queue; triggering a multi-channel transmission mechanism to deliver the wireless relay data stream temporarily stored in the split transmission queue to the next-hop node in the network neighbor list in parallel.

9. The multipath anti-interference method for decentralized broadband ad hoc networks based on random network coding according to claim 1, characterized in that, The process of intercepting automatic retransmission requests and updating network compensation instructions includes: monitoring the reverse wireless channel and capturing the link status feedback sent by the next-hop node; parsing the protocol control field of the link status feedback and extracting the receive rank indicator; converting the receive rank indicator into a numerical form of the effective arrival rank; calculating the difference between the effective arrival rank and a preset payload recovery threshold to generate a rank deviation; when the effective arrival rank is not lower than the payload recovery threshold, generating a virtual acknowledgment signal and blocking the automatic retransmission request; adjusting the reassembly parameters according to the rank deviation and generating an updated network compensation instruction.

Images