Radio interphone voice transmission MESH ad hoc network method

Abstract

The invention relates to a radio interphone voice transmission MESH ad hoc network method. The method comprises the following steps: determining an initial networking network architecture corresponding to sending an initial voice data packet according to voice bandwidth information in the initial voice data packet output by radio equipment; controlling the initial voice data packet to be sent to a relay layer from the radio equipment through the initial networking network architecture to obtain a relay voice data packet; fusing the relay voice data packet and the relay voice adding packet to obtain a relay voice fusion packet; according to the relay voice fusion packet, adjusting the initial networking network architecture to obtain an adjusted networking network architecture; taking the relay voice fusion packet as an initial voice data packet, taking the adjusted networking network architecture as an initial networking network architecture, and returning to execute the step of obtaining the relay voice data packet; and obtaining an output voice data packet until the relay layer receives a voice output instruction for the relay voice data packet. Consumption of transmission resources can be effectively optimized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of radio control, in particular to a radio intercom voice transmission MESH self-organizing network method. BACKGROUND

[0002] In the prior art, the radio intercom voice transmission first uses a voice codec (such as AMBE) to compress and quantize the voice, and then performs channel coding and interleaving to enhance the error code resistance, and is transmitted on a 12.5 kHz or narrower bandwidth by FDMA or TDMA (such as DMR, P25, TETRA, etc.), supplemented by frame synchronization, error correction and retransmission strategies to cope with noise, fading and multipath; meanwhile, encryption, authentication and group call / individual call / emergency call and other service controls are introduced. However, the conventional radio intercom voice transmission has not been effectively optimized in terms of transmission resource consumption. SUMMARY

[0003] Therefore, it is necessary to provide a radio intercom voice transmission MESH self-organizing network method, device and computer equipment which can effectively optimize the consumption of transmission resources in view of the above technical problems.

[0004] In a first aspect, the present application provides a radio intercom voice transmission MESH self-organizing network method, comprising: determining an initial networking network architecture corresponding to sending an initial voice data packet according to voice bandwidth information in the initial voice data packet output by a radio device; controlling the initial voice data packet to be sent from the radio device to a relay layer through the initial networking network architecture to obtain a relay voice data packet; fusing the relay voice data packet and a relay voice addition packet obtained by the relay layer to obtain a relay voice fusion packet; adjusting the initial networking network architecture according to the relay voice fusion packet to obtain an adjusted networking network architecture; taking the relay voice fusion packet as the initial voice data packet and taking the adjusted networking network architecture as the initial networking network architecture, and returning to execute the step of controlling the initial voice data packet to be sent from the radio device to the relay layer through the initial networking network architecture to obtain the relay voice data packet; until a voice output instruction for the relay voice data packet is received by the relay layer to obtain an output voice data packet.

[0005] In a second aspect, the present application further provides a radio intercom voice transmission MESH self-organizing network device, comprising: a network architecture determination module configured to determine an initial networking network architecture corresponding to the initial voice data packet according to voice bandwidth information in the initial voice data packet output by the radio device; a voice data transmission module configured to control transmission of the initial voice data packet from the radio device to a relay layer through the initial networking network architecture to obtain a relay voice data packet; a voice data fusion module configured to fuse the relay voice data packet and a relay voice addition packet obtained by the relay layer to obtain a relay voice fusion packet; a network architecture adjustment module configured to adjust the initial networking network architecture according to the relay voice fusion packet to obtain an adjusted networking network architecture; the voice data transmission module is further configured to return to execute the step of controlling transmission of the initial voice data packet from the radio device to the relay layer through the initial networking network architecture to obtain the relay voice data packet, with the relay voice fusion packet as the initial voice data packet and the adjusted networking network architecture as the initial networking network architecture; a voice data reception module configured to obtain an output voice data packet until a voice output instruction for the relay voice data packet is received by the relay layer.

[0006] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements any step of the radio intercom voice transmission MESH ad hoc network method when executing the computer program.

[0007] The radio intercom voice transmission MESH ad hoc network method, device and computer device described above adaptively select an initial networking network architecture according to voice bandwidth information carried by an initial voice data packet, so that the encoding code rate, fragmentation granularity, forward error correction and redundancy strategy are instantaneously matched with the link capacity; subsequently control the data packet to reach a relay layer through the selected architecture, time align the original relay voice data packet with a relay voice addition packet on the relay side and fuse the contents, which significantly improves the anti-packet loss and anti-burst interference capabilities; then adjust the topology and routing according to the fusion result on demand, which dynamically suppresses the jitter, time delay, packet loss rate and retransmission overhead, and maintains high availability and high robustness in a complex electromagnetic / heterogeneous network environment; and the loop execution takes whether the relay layer receives a voice output instruction as a convergence condition, which gradually approaches the optimal transmission strategy and resource allocation in iterations, and the obtained output voice data packet shows obvious improvement in voice clarity, continuity and intelligibility, while the transmission resource consumption is effectively optimized, the energy consumption and bandwidth occupation are reduced, and the instantaneous pressure on the core link is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the embodiments or the related art description will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor based on these drawings.

[0009] Figure 1 An application environment diagram of a radio intercom voice transmission MESH self-organizing network method in an embodiment; Figure 2 A flowchart of a radio intercom voice transmission MESH self-organizing network method in an embodiment; Figure 3 A structural block diagram of a radio intercom voice transmission MESH self-organizing network device in an embodiment; Figure 4 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0010] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0011] The radio intercom voice transmission MESH self-organizing network method provided by the embodiments of the present application can be applied in an application environment as shown in Figure 1 . In the application environment, a terminal 102 communicates with a server 104 through a network. A data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0012] In an exemplary embodiment, as shown in Figure 2 , a radio intercom voice transmission MESH self-organizing network method is provided. The method is described by taking the server in Figure 1 as an example, and includes the following steps 202 to 212. In the method, the following steps are included.

[0013] In step 202, an initial networking network architecture corresponding to an initial voice data packet is determined according to voice bandwidth information in the initial voice data packet output by a radio device.

[0014] In step 204, the initial voice data packet is controlled to be sent from the radio device to a relay layer through the initial networking network architecture, and a relay voice data packet is obtained.

[0015] Step 206, fuse the relay voice data packet and the relay voice addition packet obtained by the relay layer to obtain a relay voice fusion packet.

[0016] Step 208, adjust the initial networking network architecture according to the relay voice fusion packet to obtain an adjusted networking network architecture.

[0017] Step 210, return to execute the step of controlling the initial voice data packet to be sent from the radio device to the relay layer through the initial networking network architecture to obtain the relay voice data packet, taking the relay voice fusion packet as the initial voice data packet and taking the adjusted networking network architecture as the initial networking network architecture.

[0018] Step 212, until the relay layer receives a voice output instruction for the relay voice data packet to obtain an output voice data packet.

[0019] The initial voice data packet is a first round of data packet generated by the radio device at the beginning of the conversation, carrying a voice encoding frame and its timestamp / sequence number / encoding parameter.

[0020] The voice bandwidth information refers to a set of bandwidth-related parameters extracted from the initial voice data packet or its generation process, including target code rate, frame length, allowable packet loss / jitter and delay target, etc.

[0021] The initial networking network architecture is a first round of forwarding topology determined according to the voice bandwidth information and the current link measurement, including relay layer level, path assignment, channel / slot and priority configuration.

[0022] The relay layer is a set of relay nodes located at a specific hop number relative to the source node and undertaking forwarding tasks, which can be organized in layers such as the first layer and the second layer.

[0023] The relay voice data packet is an encapsulation formed after the initial voice data packet is relayed and forwarded by the first layer (or the current layer), with statistical metadata such as link quality and queuing delay generated hop by hop.

[0024] The relay voice addition packet is a supplementary information packet generated or aggregated by the relay layer to enhance transmission reliability and intelligibility, such as redundant fragments, check / reconstruction information or parallel backup frames.

[0025] The relay voice fusion packet is a synthesized data packet obtained by aligning and weighting the relay voice data packet and the relay voice addition packet according to the timestamp, which is used for subsequent forwarding and topology optimization.

[0026] The adjusted networking network architecture is a new round of forwarding configuration obtained by reconstructing the existing topology with minimal disturbance according to the bandwidth / delay / packet loss indicators fed back by the relay voice fusion packet.

[0027] wherein the voice output instruction is an indication signal for starting voice de-encapsulation / playout when the receiving end meets the continuity and the delay threshold or receives the control trigger.

[0028] wherein the output voice data packet is the terminal voice data obtained by de-encapsulation and necessary error concealment processing by the receiving end after the voice output instruction is generated, which can be directly played or sent.

[0029] Specifically, the voice bandwidth information (target code rate, frame length, packet loss tolerance, delay target, etc.) is parsed from the initial voice data packet, and the minimum available bandwidth and the maximum allowable delay of the current voice stream are calculated in combination with the neighbor table and link measurement (SNR / RSSI, ETX, round-trip delay RTT, congestion) maintained by the device. Once the minimum available bandwidth and the maximum allowable delay are obtained, a fast topology selection is performed, that is, when there are multiple high-quality direct links in the neighbor set, a star / single-hop direct link is selected, otherwise, under the premise of meeting the bandwidth and delay constraints, a number of relay nodes are selected according to the rule of "minimum hop count first, and minimum path loss first if there is a conflict", and the relay level and working channel / slot are determined, and an initial networking network architecture (including: relay layer list, forwarding table, frequency band / channel and time slot allocation, priority and queue parameters) is generated.

[0030] The source device timestamps and serializes the voice frames according to the forwarding table, dequeues them according to the priority queue, and sends them on the designated channel / slot. Each relay node uses the "de-duplication + loop suppression" (packet table based on <source ID, sequence number>) and "hop-by-hop aging" strategies for storage and forwarding, while adding the link quality and queuing waiting time measured at this hop and updating the hop count. The MAC layer uses TDMA / CSMA hybrid access and enables small block FEC or retransmission upper limit according to the configuration. The voice frame that reaches the first layer relay and completes a legal forwarding encapsulation forms a relay voice data packet (including the original load + hop-by-hop statistical metadata).

[0031] The relay layer node puts the received relay voice data packet and the relay voice added packet obtained by the layer (such as redundant blocks of the same frame, bypass backup frames, enhanced frequency band segments, or check / reconstruction information) into a jitter buffer and aligns them by timestamp. The aligned frames are subjected to weighted fusion in the transform domain (such as MDCT / FFT amplitude spectrum). The weights are given by the link confidence, packet completeness, and energy masking estimation. The fused frames are re-encapsulated and re-quantized / compressed according to the layer strategy, while the statistical indicators generated during the fusion process (bandwidth occupancy, end-to-end delay estimation, packet loss compensation rate, etc.) are summarized and written. The output is a relay voice fusion packet.

[0032] Under the premise that the first layer relay layer of the initial networking network architecture has been stable, the bandwidth occupation, delay increment, packet loss / retransmission and link quality indicators carried in the relay voice integrated packet are analyzed, and the mismatch degree of the original second layer is determined accordingly. Subsequently, the strategy of "original second layer overall offline, candidate topology unit fast reselection" is adopted, the original second layer corresponding link and assignment are atomically revoked from the forwarding table and time slot / channel plan (placed in the observation reserved state and do not participate in forwarding), and the candidate topology unit is scored and optimized according to the bandwidth compliance, SNR / ETX, node load and interference escape degree, and a new second layer relay layer and its forwarding path and time slot / channel configuration are generated. The boundary of the new second layer relay layer and the first layer relay layer is checked for reachability and loop suppression, and the "minimum disturbance" principle is used to keep the first layer unchanged and replace the next layer. Finally, the new configuration is written into the forwarding table and time slot / channel plan by versioning submission, and the hysteresis and rollback conditions are set to suppress oscillation and ensure fast recovery in abnormal conditions, so as to obtain the adjusted networking network architecture.

[0033] The adjusted networking network architecture obtained in the last step is regarded as a new initial networking network architecture, and the latest relay voice integrated packet is regarded as a new initial voice data packet. Immediately return to the "send to relay layer" process under the protection of versioned configuration. That is, the old second layer still in the observation reserved state is preferentially forwarded along the new forwarding table and the new time slot / channel plan priority (the first layer remains unchanged, only the next hop of the new second layer is used), and the time-limited emptying and session migration (based on the continuity judgment and TTL aging of <source ID, sequence number>) are performed on the old second layer. Maintain jitter buffer and timestamp alignment at the receiving end to avoid voice interruption. At the same time, enable the hysteresis threshold and the minimum holding time to suppress frequent reconstruction, and set the convergence criterion (such as stop iteration when the improvement of key indicators in the last N evaluation windows is less than ε), and if necessary, restore to the last version according to the rollback strategy.

[0034] When the target receiving node (or any designated relay layer / boundary node) detects that it has continuously met the "frame integrity K threshold, jitter ≤ J threshold, end-to-end delay ≤ T threshold" under the current version of the adjusted networking topology, receives explicit control signaling, or has triggered the preset upper limit value of the relay layer of the adjusted networking topology, it generates a voice output instruction. The node performs one-time checking and unpacking (including de-duplication / FEC reconstruction and sequence number gap filling) on the latest frame sequence in the buffer sorted by timestamp, and performs clock drift fine tuning and playback point alignment, then completes decoding and packet loss concealment / comfort noise processing, and outputs continuous output voice data packets (or direct audio streams). At the same time, the final session statistics (total hop count, average / 95th percentile delay, effective recovery rate, packet loss rate, rollback times, etc.) and version number are written into the log and returned to the topology control logic, and if threshold violations occur during output, switch to the last version according to the rollback strategy.

[0035] In the wireless intercom voice transmission MESH self-organizing network method, the initial voice data packet carries voice bandwidth information to adaptively select the initial network architecture, so that the encoding code rate, fragmentation granularity, forward error correction and redundancy strategy are immediately matched with the link capacity; then the control data packet reaches the relay layer through the selected architecture, and the original relay voice data packet and the relay voice addition packet are time-aligned and content-fused on the relay side, which significantly improves the anti-packet loss and anti-burst interference capability; then the topology and routing are adjusted on demand according to the fusion result, to dynamically suppress the jitter, delay, packet loss rate and retransmission overhead, and maintain high availability and high robustness in complex electromagnetic / heterogeneous network environment; the convergence condition is the voice output instruction received by the relay layer, and the optimal transmission strategy and resource allocation are gradually approached in the iteration, and the output voice data packet shows obvious improvement in voice clarity, continuity and intelligibility, while the transmission resource consumption is effectively optimized, the energy consumption and bandwidth occupation are reduced, and the instantaneous pressure on the core link is reduced.

[0036] In an exemplary embodiment, the initial network architecture is adjusted according to the relay voice fusion packet to obtain an adjusted network architecture, including steps 302 to 304. Wherein:

[0037] Step 302, according to the voice bandwidth information of the relay voice fusion packet, the extension architecture of the initial network architecture is selected to obtain a topology expansion network unit.

[0038] Step 304, the initial network architecture and the topology expansion network unit are spliced to obtain the adjusted network architecture.

[0039] Wherein, the extension architecture is an expandable topology scheme composed of candidate nodes / links / small subgraphs outside the initial network architecture, used as a potential supplement or replacement under the bandwidth and delay target.

[0040] Wherein, the topology expansion network unit is a specific deployment unit that can be directly integrated into the initial architecture after optimization according to bandwidth, delay, load and loop suppression rules from the extension architecture set, including corresponding forwarding assignment and channel / slot configuration.

[0041] Specifically, the bandwidth parameters (target code rate / frame length / delay and jitter threshold) and hop-by-hop statistics (link quality, congestion, retransmission rate) reflecting the current business in the relay voice fusion packet are analyzed, and a candidate extension architecture set (which can be a single node, a single link or a small subgraph) is constructed in the current neighbor / link library. Each candidate extension architecture in the candidate extension architecture set is subject to the hard condition of "meeting bandwidth and delay constraints", and the comprehensive cost function (such as w=expected delay increment+ Packet loss risk+ ​Pathloss + Node load The selected topology expansion network unit is combined with the existing topology according to the principle of "minimum disturbance" during the atomic update of the initial networking network architecture, and the existing next layer link is replaced or bypassed if necessary. Then, a new forwarding table, channel / slot plan and priority parameter are generated and issued. Reachability and loop checking (shortest path / minimum loss consistency) are performed on the new edges and existing boundaries before taking effect. Hysteresis and rollback conditions are set at the time of taking effect to suppress oscillation. A short window of health detection (delay / packet loss / jitter) is used to confirm stability after taking effect, thereby forming and enabling the adjusted networking network architecture.

[0042] The selected topology expansion network unit is combined with the existing topology according to the principle of "minimum disturbance" during the atomic update of the initial networking network architecture, and the existing next layer link is replaced or bypassed if necessary. Then, a new forwarding table, channel / slot plan and priority parameter are generated and issued. Reachability and loop checking (shortest path / minimum loss consistency) are performed on the new edges and existing boundaries before taking effect. Hysteresis and rollback conditions are set at the time of taking effect to suppress oscillation. A short window of health detection (delay / packet loss / jitter) is used to confirm stability after taking effect, thereby forming and enabling the adjusted networking network architecture.

[0043] In this embodiment, by selecting the extension architecture according to the bandwidth information of the relay voice fusion packet and splicing it with the existing architecture in the form of a topology expansion network unit, the minimum disturbance elastic expansion and replacement can be realized without interrupting the first layer and the existing stable path. The network dynamically introduces more optimal relays and links under the bandwidth / delay target, thereby reducing end-to-end delay and packet loss, improving effective throughput and voice intelligibility. At the same time, the modular unit splicing facilitates fast convergence and rollback, and can quickly avoid hotspots, balance load and suppress loops when interference or congestion changes, thereby obtaining higher stability and energy efficiency.

[0044] In one exemplary embodiment, the extension architecture of the initial networking network architecture is selected according to the voice bandwidth information of the relay voice fusion packet to obtain a topology expansion network unit, including steps 402 to 408. Among them:

[0045] Step 402, reverse backtracking analysis is performed on the voice bandwidth information and the forwarding statistical data in the relay voice fusion packet to obtain a set of relay layer bottleneck links.

[0046] Step 404, time-frequency occupation prediction analysis is performed on each candidate topology unit of the initial networking network architecture according to the voice bandwidth information and historical interference observation data to obtain each bandwidth compliant candidate unit.

[0047] Step 406, load balancing is solved for each bandwidth compliant candidate unit and the set of relay layer bottleneck links to obtain a tentative expansion network unit.

[0048] Step 408, path reachability adjustment is performed on the tentative expansion network unit according to the minimum hop or minimum path loss rule to obtain a topology expansion network unit.

[0049] Among them, the forwarding statistics are a set of performance metrics collected and reported with the packet in each relay hop transmission process, such as packet loss rate, retransmission times, queuing waiting time, arrival timestamp, SNR / ETX, etc.

[0050] Among them, the reverse backtracking analysis is an analysis method of locating the problem link according to the end-to-end degradation (delay / packet loss, etc.) along the time and path to each hop.

[0051] Among them, the relay layer bottleneck link set is a subset of relay links determined in the backtracking analysis as the most degraded and sustained threshold contribution to end-to-end performance.

[0052] Among them, the historical interference observation data is the interference feature data recorded over time, such as power spectral density, duty cycle, collision rate, and its time series.

[0053] Among them, the candidate topology unit is a basic structural unit that can be used to expand or replace the existing topology, including candidate nodes, candidate links, or small subgraphs.

[0054] Among them, the time-frequency occupation prediction analysis is an analysis based on historical interference and current traffic parameters to predict the availability and bandwidth provided by each frequency band / time slot in the future time window.

[0055] Among them, the bandwidth-compliant candidate unit is a subset of candidate topology units that meet the required bandwidth and confidence conditions under the prediction results and threshold constraints.

[0056] Among them, the load balancing solution is a process of establishing a cost / capacity model between bottleneck demand and candidate resources and solving the optimal or near-optimal allocation to balance the traffic.

[0057] Among them, the proposed expansion network unit is a set of candidate topology units selected after load balancing solution, ready to be incorporated into the existing topology.

[0058] Among them, the minimum hop number is a routing criterion that selects the route with the least number of nodes in the reachable path.

[0059] Among them, the minimum path loss rule is a routing criterion that selects the route with the minimum total propagation loss (or equivalent link cost) in the reachable path.

[0060] Among them, the path reachability adjustment is a process of checking the reachability and constraints of the path after incorporating the proposed unit and making minimum disturbance modifications to ensure connectivity and stability.

[0061] Specifically, extract voice bandwidth information (target code rate / frame length / latency threshold) and hop-by-hop forwarding statistics (packet loss, retransmission, queuing delay, SNR / ETX) from relay voice convergence packets, and timestamp align and abnormality reject the end-to-end latency and packet loss trajectory in the sliding window. Then, score the end-to-end degradation to each hop by maximum increment positioning and score by or comprehensive cost function (such as w= expected latency increment + packet loss risk + path loss + node load) and score, and the link that exceeds the threshold and hits K times in a row is marked as a bottleneck, and a relay layer bottleneck link set is generated.

[0062] At each candidate topology unit (node / link / subgraph) granularity, gather historical interference observation data (power spectral density, duty cycle, same / adjacent channel collision rate) and voice bandwidth information, establish short-term time-frequency availability prediction (such as Kalman / lightweight regression), and obtain future inner available bandwidth upper bound R_max and air interface idle degree , and filter each bandwidth compliant candidate unit that meets the bandwidth and confidence threshold according to the criterion R_req≤R_max·(1− ) and P( ≥η0)≥β.

[0063] Take the relay layer bottleneck link set as one side and each bandwidth compliant candidate unit as the other side, and the edge weight is the comprehensive cost function (such as w= expected latency increment + packet loss risk + path loss + node load) and impose node degree, loop suppression, and concurrent upper limit constraints, and use the minimum cost maximum flow / assignment approximation to solve within the time limit T, output the matching result that can maximize the throughput or minimize the total cost, and the bandwidth compliant candidate unit involved in the matching is the proposed extended network unit.

[0064] Based on the proposed extended network unit, integrate it into the existing initial networking network architecture, and perform reachability and optimality checking on the path connected to the first layer relay layer. The optimality checking prioritizes the "minimum hop count" or, in the event of a conflict, the "minimum path loss" rule to reorganize the next hop and time slot / channel allocation, and eliminates elements that cause loops or violate thresholds (SNR, duty cycle, degree constraint) and performs minimum disturbance fine-tuning. Finally, the element set that meets the reachability and constraints is retained as the topology expansion network unit.

[0065] In this embodiment, by first locating the relay layer bottleneck links through reverse backtracking, then combining historical interference to screen out bandwidth compliant units for time-frequency occupation prediction of candidate topology, and finally achieving near-optimal assignment of resource-demand through load balancing solution, and completing path reachability adjustment under the rules of minimum hop count or minimum path loss, on-demand expansion and replacement can be achieved under the premise of minimum disturbance, significantly reducing end-to-end delay and jitter, suppressing packet loss, improving effective throughput and speech intelligibility, while enhancing robustness to dynamic interference and traffic fluctuations and taking into account energy efficiency and network stability.

[0066] In an exemplary embodiment, reverse backtracking analysis is performed on voice bandwidth information and forwarding statistics in relay voice convergence packets to obtain a set of relay layer bottleneck links, including steps 502 to 508. Among them:

[0067] Step 502, performing a minimum cut analysis on the relay voice convergence packets to obtain a set of candidate bottleneck edges.

[0068] Step 504, performing sliding median filtering on the arrival timestamp sequence in the forwarding statistics to obtain a time delay trajectory sequence.

[0069] Step 506, performing constraint analysis on the voice bandwidth information to obtain a set of bandwidth constraint parameters.

[0070] Step 508, according to the set of bandwidth constraint parameters and the time delay trajectory sequence, performing sparse causal backtracking solution on the set of candidate bottleneck edges to obtain the set of relay layer bottleneck links.

[0071] Among them, the minimum cut analysis is a method of abstracting end-to-end transmission as a flow network and solving the minimum cut under given demand flow to determine the set of key edges that limit throughput.

[0072] Among them, the set of candidate bottleneck edges is a set of link edges that form a dominant restriction on throughput or delay in the current session obtained by minimum cut analysis.

[0073] Among them, the arrival timestamp sequence is the arrival time sequence of the merged data frames at the receiving side according to 〈source ID, sequence number〉.

[0074] Among them, the sliding median filtering is a time series filtering method that replaces the original sequence value with the median in a fixed or adaptive window to suppress impulse noise and outliers.

[0075] Among them, the time delay trajectory sequence is a sequence of end-to-end instantaneous delay changes over time obtained by differencing the arrival times of adjacent frames and smoothing the noise.

[0076] Among them, constraint analysis is the process of converting code rate, frame length, allowable delay / jitter / packet loss, and other service requirements into calculable bandwidth and queue hard / soft constraint parameters.

[0077] wherein the bandwidth constraint parameter set is a parameter set derived from constraint analysis, at least including R_min, T_max, J_max, p_max and necessary redundancy ratio and priority weight.

[0078] wherein the sparse causal backtracking solution is an optimization estimation process of key bottleneck edges by decomposing end-to-end degradation to each hop contribution in sparse regularization under causal order and constraint condition.

[0079] Specifically, the networking network architecture corresponding to the current session is abstracted as a directed flow network, the edge capacity is depicted by the hop-by-hop statistics (such as effective throughput = sending rate x success rate, or the upper bound of available bandwidth converted according to SNR / ETX) carried by the relay voice fusion package, and an end-to-end flow graph is constructed from the source node→the first layer entrance→…→the target receiving node. The minimum cut is solved under the given service demand flow R_req using the minimum cut / maximal flow framework, and the link falling on the minimum cut is regarded as the throughput limiting edge of this round of transmission, which is defined as the candidate bottleneck edge set.

[0080] The arrival timestamp sequence in the forwarding statistics data is monotonically rearranged and de-duplicated based on 〈source ID, sequence number〉, and the time base is aligned with the first hop logical clock, and then the instantaneous end-to-end delay sequence is obtained by differencing adjacent frames. The adaptive window (covering several voice frame periods) performs sliding median filtering to suppress abnormal jitter and spikes caused by burst retransmission, and obtains a smooth and robust delay trajectory sequence.

[0081] The target code rate R, frame length L, maximum allowable delay T_max, jitter J_max and packet tolerance p_max in the voice bandwidth information are analyzed, the lower bound of the bandwidth allocable per hop, the upper limit of the queue waiting and the necessary redundancy / FEC overhead are derived, and the joint constraint parameter set of bandwidth and delay is formed as the bandwidth constraint parameter set P = {R_min, T_max, J_max, p_max, redundancy ratio p, priority weight, etc.}.

[0082] The "delay increment decomposition" model is established with each edge in the candidate bottleneck edge set as a variable, the observed end-to-end delay trajectory sequence is represented as the causal convolution sum of the delay increment of each hop, and the hard constraints (such as the upper limit of single-hop queuing and the lower bound of available bandwidth) given by P and the sparse regularization (encouraging a small number of key edges to bear the main degradation) are added. Through sparse regression / lasso type optimization estimation with constraints, the delay and packet loss contribution of each edge is estimated, the edges with continuous contribution exceeding the threshold and meeting the causal order are selected as the final bottleneck, and the relay layer bottleneck link set is determined.

[0083] In this embodiment, by quickly circumscribing the candidate bottleneck edge with the minimum cut analysis, constructing the anti-noise time delay trajectory with the sliding median filter, quantifying the traffic code rate / time delay and other demand into a unified parameter set with the constraint analysis, and performing sparse causal backtracking solution under the constraint, a small number of key bottleneck links can be accurately positioned while ensuring implementability and causal consistency, so as to realize the minimum disturbance optimization of the topology, significantly reduce the end-to-end delay and packet loss, improve the effective throughput and speech intelligibility, and make the network more robust to interference and load fluctuations, faster convergence, more stable operation and more energy saving.

[0084] In an exemplary embodiment, load balancing is solved for each bandwidth-compliant candidate unit and the set of relay layer bottleneck links to obtain the proposed extended network unit, including steps 602 to 608. Among them:

[0085] Step 602, multi-objective cost embedding is performed on the time slot bearing vector of each bandwidth-compliant candidate unit to obtain weighted capacity information.

[0086] Step 604, according to the set of relay layer bottleneck links, the weighted capacity information is subjected to disturbance robustness processing to obtain robust weighted capacity information.

[0087] Step 606, the robust weighted capacity information is subjected to minimum cost maximum flow solving to obtain assignment scheme data.

[0088] Step 608, the assignment scheme data is subjected to discretization stability shaping to obtain the proposed extended network unit.

[0089] Among them, the time slot bearing vector is a time slot arranged numerical vector representing the effective bandwidth or occupation capacity that a certain candidate unit can provide at each time slot.

[0090] Among them, multi-objective cost embedding is to map multi-objective costs such as time delay increment, packet loss risk, power consumption, and hop number change to the time slot bearing vector for comprehensive measurement of the actual effectiveness of unit bearing.

[0091] Among them, the weighted capacity information is the time slot level resource representation data obtained after multi-objective cost embedding, which carries capacity and cost weight.

[0092] Among them, disturbance robustness processing is a conservative shrinkage of weighted capacity under the consideration of uncertain disturbances such as interference and congestion, so that the solution is still feasible under disturbance.

[0093] Among them, the robust weighted capacity information is the capacity-cost representation data after completing the disturbance robustness processing, which can guarantee availability in the worst case.

[0094] Among them, the minimum cost maximum flow solution is an optimization process to find the flow distribution that meets the maximum throughput and the minimum total cost under the constraints of capacity and cost.

[0095] Among them, the assignment scheme data is the structured result of the carrying capacity, activation flag and boundary interface information of each candidate unit and time slot obtained by the minimum cost maximum flow solution.

[0096] Among them, the discretization stability shaping is a process of converting the continuous carrying solution into a discrete time slot block of "activation / deactivation" and adding hysteresis and protection interval to suppress frequent switching and improve operation stability.

[0097] Specifically, for each bandwidth-compliant candidate unit, its time slot carrying vector x (element is the effective bandwidth / duty cycle that the unit can provide in each time slot) is constructed, and a comprehensive cost function (such as w= is introduced. Expected delay increment + Packet loss risk + Path loss + Node load) is introduced, and the effective capacity of unit flow is obtained by element-by-element mapping of the time slot carrying vector x according to the cost weight. At the same time, the upper limit of node degree, the upper limit of duty cycle and the loop suppression identifier are added to form the weighted capacity information.

[0098] According to the interference / congestion uncertainty corresponding to the set of relay layer bottleneck links, an uncertainty set U (such as capacity reduction coefficient , conflict probability rise ) is established for each bandwidth-compliant candidate unit, and the weighted capacity information is conservatively contracted , and a penalty factor is added to the elements co-located or same-frequency with the bottleneck link, to obtain the robust weighted capacity information that can still be feasible under disturbance.

[0099] Taking the bottleneck demand side (to be shared flow segment) as the source side node and the time slot of the bandwidth-compliant candidate unit as the sink side node, the edge capacity takes the robust weighted capacity , and the edge cost takes the corresponding cost c and adds a hard constraint (node degree, concurrent upper limit, duty cycle / regulation limit, loop prohibited subgraph), to construct a flow network of time slot granularity. The minimum cost maximum flow (or equivalent assignment) is used to obtain the near-optimal flow allocation within the real-time time limit T_max, and the carrying capacity, activation flag and boundary interface information of each bandwidth-compliant candidate unit in each time slot are output to form the assignment scheme data.

[0100] According to the assignment scheme data, the continuous bearing capacity is first quantized into "enable / disable" according to the hysteresis threshold and the minimum holding time, and is divided into discrete time slot blocks and inserted into a guard interval to suppress frequent switching. Then, the constraint projection resolution is performed on the resource conflicts across cells and time slots according to the priority, and at the same time, the only next hop alignment with the first layer boundary and the short loop suppression are completed. If the quantization results in a capacity gap, the alternative cell is enabled from low to high cost until the bandwidth and duty cycle constraints are met. Finally, in a versioned atomic commit manner, the enable cell set and its corresponding time slot and channel configuration and forwarding assignment are generated as the proposed extended network cell.

[0101] In this embodiment, by first embedding the time slot bearing vector into a multi-objective cost to unify the capacity and cost, and then combining the bottleneck link uncertainty to make a disturbance robustness processing to ensure that it is still feasible under the worst case, and then solving the near-optimal resource assignment under the dual constraints of capacity and cost by using the minimum cost maximum flow, and finally converting the continuous solution into a deployable scheme with hysteresis and guard interval through discrete stability shaping, the end-to-end throughput is improved, the delay and packet loss are reduced, the frequent switching is suppressed, and the robustness to interference and load fluctuation is enhanced under the premise of minimum disturbance.

[0102] In one exemplary embodiment, the relay voice data packet and the relay layer obtained relay voice addition packet are fused to obtain a relay voice fusion packet, including steps 702 to 706. Among them:

[0103] Step 702, joint weight analysis is performed on the energy masking spectrum and link confidence data in the relay voice data packet and the relay voice addition packet to obtain fusion weight data.

[0104] Step 704, jitter buffer alignment is performed on the relay voice data packet and the relay voice addition packet to obtain a sequence of voice packet alignment frames.

[0105] Step 706, according to the fusion weight data, the voice packet alignment frame sequence is transformed into a sparse fusion in the transform domain to obtain a relay voice fusion packet.

[0106] Among them, the energy masking spectrum is a spectral representation of the time-frequency energy distribution based on MDCT / FFT and the like and combined with the auditory masking threshold, which is used to indicate which frequency points can be masked by stronger energy components.

[0107] Among them, the link confidence data is a link reliability measure normalized by SNR / ETX, packet loss rate, retransmission count, queuing delay and the like, which is used to reflect the reliability of the voice frame.

[0108] Among them, the joint weight analysis is an analysis process of fusing the energy masking spectrum and the link confidence data according to a preset rule to calculate the time-frequency unit weight.

[0109] wherein the fusion weight data is a time-frequency weight matrix of joint weight analysis output, used to guide the proportion of different source speech frames in each time-frequency unit.

[0110] wherein the jitter buffer alignment is an alignment process of rearranging, deduplicating, filling in missing frames and waiting for out-of-order frames within a buffer area using timestamps and sequence numbers to obtain stable and continuous output.

[0111] wherein the speech packet alignment frame sequence is a frame sequence after jitter buffer alignment, with consistent time base, consistent length and missing gap processed.

[0112] wherein the transform domain sparse fusion is a fusion method of weighting and synthesizing multi-source spectral coefficients in MDCT / FFT and other transform domains according to fusion weights, and combining sparse constraints to suppress noise and false peaks.

[0113] Specifically, the energy masking spectrum (such as MDCT / FFT amplitude spectrum and auditory masking threshold) and the corresponding link confidence data (obtained by normalizing SNR / ETX, packet loss rate, retransmission count and queuing delay) of the relay speech data packet and the relay speech addition packet are extracted, then the above two types of features are aligned in time-frequency grid and 0-1 normalized, and then the weighted logic is used to calculate the fusion weight of each time-frequency unit and apply total energy and smoothing constraints (such as first-order TV regularization in time and frequency directions) to suppress mutations, and finally output the fusion weight data matrix varying with time and frequency.

[0114] The frame sequence of the relay speech data packet and the frame sequence of the relay speech addition packet are written into the jitter buffer area according to the timestamp and sequence number, and the monotonic rearrangement and deduplication are used to ensure sequence consistency, the out-of-order frames are played back within the maximum waiting threshold, and the missing frames are linearly or based on the adjacent spectrum minimum mean square interpolation. Then, the cross-fade processing is performed on the packet boundary to eliminate boundary artifacts, and the speech packet alignment frame sequence with consistent length, consistent time base and missing gap filled is obtained.

[0115] The transform domain decomposition (such as MDCT / FFT) is performed on each frame of the speech packet alignment frame sequence, and the sparse weighted synthesis is performed in each time-frequency unit with the fusion weight data matrix as a mask: After the small amplitude iterative projection combined with sparse constraints and frequency band consistency penalty term to remove false peaks and noise leakage, the amplitude and phase reconstruction and inverse transform are performed to obtain the time domain frame, and the re-quantization / packing is completed according to the encoding strategy and the statistical metadata (bandwidth occupancy, frame rate, end-to-end estimated delay, etc.) of the fusion process is written to form the topology-optimized relay speech fusion packet.

[0116] In this embodiment, by jointly weighting the energy mask spectrum and the link confidence, the system can adaptively give more weight to more reliable and more perceptually valuable components in the time-frequency domain; then, through jitter buffer alignment, it can repair out-of-order and missing packets and unify the time base, significantly reducing the impact of jitter on reconstruction; finally, in the transform domain, it implements sparse fusion to suppress noise and false peaks and reduce boundary artifacts, thereby achieving higher end-to-end intelligibility, lower distortion and packet loss sensitivity, and stable improvement of subjective listening and effective throughput under the premise of low additional latency and minimal topology disturbance.

[0117] In one exemplary embodiment, the method further comprises steps 802 to 806. Among them:

[0118] Step 802, according to the link measurement data and the regulation duty cycle data corresponding to each frequency band of the relay voice fusion packet, cross-band relay potential analysis is performed on each candidate topology unit to obtain a cross-band relay hub set.

[0119] Step 804, time slot alignment is performed on the cross-band relay hub set and the time slot frame structure of each frequency band to obtain a cross-band forwarding time sequence table.

[0120] Step 806, according to the cross-band forwarding time sequence table and the bandwidth time-varying information of the relay voice fusion packet, the initial networking topology is reconstructed in multiple frequency bands to obtain an adjusted networking network architecture.

[0121] Among them, the link measurement data is a quantitative observation of the real-time or historical performance of the wireless link, such as SNR / ETX, delay, packet loss rate, collision rate, and throughput upper bound.

[0122] Among them, the regulation duty cycle data is the maximum power generation / transmission time ratio allowed by each frequency band under local regulatory rules and related limit parameters.

[0123] Among them, the cross-band relay potential analysis is an analysis based on link measurement and duty cycle constraints to evaluate the benefits and costs of relay forwarding between different frequency bands by candidate units to determine their availability.

[0124] Among them, the cross-band relay hub set is a set of nodes or subgraphs selected by cross-band relay potential analysis that are suitable for multi-band forwarding.

[0125] Among them, the time slot frame structure is the frame and time slot arrangement method defined by each frequency band at the physical / MAC layer, as well as its length, guard interval, and transmission / reception rules.

[0126] Among them, time slot alignment is a process of mapping the time slot frame structures of multiple frequency bands to a unified time axis and considering switching delay and guard interval to eliminate collisions.

[0127] The cross-band relay timing table is a specific relay time plan arranged for the selected hub in each frequency band, containing node, frequency band, start and end time slot, and guard interval information.

[0128] The multi-band coupling reconstruction is a process of integrating multi-band capability into the existing topology to form a new configuration under the premise of maintaining the stability of the upper layer, according to the timing table to jointly adjust the routing, time slot and channel.

[0129] Specifically, the link measurement data (such as SNR / ETX, one-way / round-trip delay, packet loss rate, same / adjacent channel conflict rate) and regulatory duty cycle data corresponding to the relay voice fusion package are normalized to a unified dimension by frequency band, and the cross-band relay potential score of each candidate topology unit is calculated: available throughput (subject to duty cycle and interference constraints) - frequency band switching overhead (frequency switching delay, energy consumption, uplink and downlink half-duplex conflict) - stability risk (historical variance, congestion fluctuation), and the node degree and loop suppression constraints are applied, and the cross-band relay hub set is selected from the units with the highest score and no conflict with each other.

[0130] The frame structure and time slot planning of each node in the cross-band relay hub set in different frequency bands are mapped to a unified time axis, considering the local frequency band switching delay, transmit-receive switching guard interval, half-duplex / full-duplex capability and duty cycle upper limit, and a three-step method of alignment-compression-backfill is used to generate a frequency band switching sequence and a transmit-receive time slot block for each hub without causing cross-band collision and violating the duty cycle, and output a cross-band relay timing table containing a four-tuple set of <node, frequency band, start and end time slot, guard interval>.

[0131] The cross-band relay timing table and the bandwidth time-varying information of the relay voice fusion package are solved together, and the initial networking topology is reconstructed according to the principle of "the first relay layer remains unchanged, and the second relay layer is replaced with minimum disturbance", and the cross-band relay hub is preferentially allocated a forwarding table and channel / power parameters within the time slot block allowed by the cross-band relay timing table during the coupling reconstruction process, replacing the inefficient links of the original second relay layer and performing boundary reachability and short loop suppression checking, and if a capacity gap occurs, the backup unit is backfilled in ascending order of cost, and finally the new routing assignment and time slot / channel configuration are effective in a transactional manner, forming an adjusted networking architecture.

[0132] In this embodiment, by first evaluating the cross-band relay potential of the candidate topology unit based on the link measurement and regulatory duty cycle of each frequency band, then accurately aligning the selected hub with the time slot frame structure of each frequency band and forming a cross-band relay timing table, and finally implementing multi-band coupling reconstruction on the initial topology combined with bandwidth time-varying information, the on-demand scheduling and minimum disturbance replacement of spectrum resources can be realized under the compliance premise, significantly reducing the end-to-end delay and packet loss, improving the effective throughput and voice intelligibility, while enhancing the robustness to interference and load fluctuations, suppressing switching oscillation and improving overall energy efficiency and network stability.

[0133] In an exemplary embodiment, according to the link measurement data and the regulation duty cycle data corresponding to each frequency band of the relay voice convergence packet, cross-band relay potential analysis is performed on each candidate topology unit to obtain a cross-band relay hub set, including steps 902 to 908. Among them:

[0134] Step 902, the link measurement data and the regulation duty cycle data corresponding to each frequency band are subjected to chance-constrained confidence projection to obtain a frequency band availability probability cone.

[0135] Step 904, according to the bandwidth time-varying information of the relay voice convergence packet and the frequency band availability probability cone, conditional risk analysis is performed on each candidate topology unit to obtain a cross-band relay potential score set.

[0136] Step 906, Kalman smoothing processing is performed on the time series of the cross-band relay potential score set to obtain a robust score sequence.

[0137] Step 908, the robust score sequence is subjected to submodular function greedy maximization processing under the constraint of independence to obtain a cross-band relay hub set.

[0138] Among them, the chance-constrained confidence projection is to project the random constraints such as "availability not lower than threshold, duty cycle not exceeding limit" into a determined feasible region under a given probability level, so as to be directly used in subsequent optimization.

[0139] Among them, the frequency band availability probability cone is a usable area described in the time-frequency-power-duty cycle dimension with a semi-cone set, representing the probability feasible region of a certain frequency band satisfying the opportunity constraint under uncertain disturbance.

[0140] Among them, the bandwidth time-varying information is a set of bandwidth requirements and coding parameters (such as R(t), frame length, jitter / delay threshold) that describe the change of the business over time.

[0141] Among them, the conditional risk analysis is to evaluate the tail index (such as VaR / CVaR) of the potential loss distribution under a given scenario or condition to quantify the risk exposure of the scheme.

[0142] Among them, the cross-band relay potential score set is a time-varying score result obtained by combining the reachable throughput, switching overhead and stability risk according to the weight, which is used to measure the cross-band relay value of each candidate unit.

[0143] Among them, the Kalman smoothing processing is based on the state space model to perform forward filtering and backward smoothing on the noisy observation to obtain a more stable time series estimation.

[0144] Among them, the robust score sequence is a noise-resistant, time-series-consistent candidate unit score time sequence obtained after Kalman smoothing.

[0145] Where, independent set constraint is a set of mutual exclusion and degree limit imposed on optional units (such as frequency non-parallel, node degree upper limit, loop suppression), which ensures the feasibility of the selection result.

[0146] Where, sub-module function greedy maximum processing is an approximate optimization method that selects elements in order of maximum gain using the marginal diminishing property, and obtains a near-optimal selection set under the above constraints.

[0147] Specifically, the link measurement data (SNR / ETX, packet loss rate, collision rate, latency, etc.) of each frequency band and the regulatory duty cycle data are normalized to a unified dimension, the opportunity constraint level (such as availability probability ≥ β, duty cycle not exceeding threshold θ) is set, the confidence projection is performed on the available area in the time-frequency-power-duty joint space, and the uncertain boundary is described by a semi-cone set, forming a frequency band availability probability cone.

[0148] The bandwidth time-varying information (R(t), frame length, jitter threshold) of the relay voice fusion packet and the frequency band availability probability cone are combined, the conditional risk value (such as CVaR or quantile excess loss) in a given time window Δt is calculated for each candidate topology unit, the reachable throughput, switching delay overhead and stability risk are combined into a revenue-risk score according to the weight, and a cross-frequency relay potential score set is obtained as a time-varying measure for optimization.

[0149] The cross-frequency relay potential score set is regarded as a time-varying state observation containing measurement noise, a first-order or second-order state space model is established, forward Kalman filtering and backward RTS smoothing are performed for each candidate topology unit to suppress score spikes caused by short-term fluctuations and sudden interference, and a robust score sequence with time sequence consistency is output.

[0150] The robust score sequence is defined as a weight to define a coverage / connectivity utility function with marginal diminishing property, sequential greedy selection is performed in the feasible region of the "independent set" constraint (degree constraint, same frequency conflict, loop suppression), the unit with the maximum marginal gain is preferentially selected and projected into the feasible region in real time until there is no feasible gain, and a cross-frequency relay hub set that satisfies the constraints and has an approximately optimal total utility is obtained.

[0151] In this embodiment, the link measurement of multiple frequency bands and the regulatory duty cycle uncertainty are converted into a directly calculable "band availability probability cone" by opportunity constraint confidence projection, and the conditional risk analysis of the candidate topology unit is combined with the bandwidth time-varying information of the relay voice fusion package to form a comparable cross-band relay potential score, and Kalman smoothing is performed on the time sequence to suppress noise and short-term fluctuations, and finally the cross-band relay hub set is screened out by using the submodular function greediness maximization under the independent intensive constraint, which can realize accurate optimization and minimum disturbance replacement of multi-band resources on the whole under the premise of compliance and robustness, significantly improve the end-to-end throughput and reduce the time delay and packet loss, and enhance the robustness and operation stability to interference and load changes.

[0152] It should be understood that although each step in the flowchart involved in each embodiment as described above is shown in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, there is no strict order limitation for the execution of these steps, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0153] Based on the same inventive concept, the embodiments of the present application also provide a radio intercom voice transmission MESH ad hoc network device for implementing the above-mentioned radio intercom voice transmission MESH ad hoc network method. As shown in Figure 3 The device provides a solution to the problem, and the implementation scheme is similar to the implementation scheme described in the above method, so the specific limitations in one or more radio intercom voice transmission MESH ad hoc network device embodiments provided below can refer to the limitations of the radio intercom voice transmission MESH ad hoc network method in the above, and will not be repeated here.

[0154] Each module in the above radio intercom voice transmission MESH ad hoc network device can be realized by software, hardware, and combinations thereof, in whole or in part. Each module described above can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to each module.

[0155] In an exemplary embodiment, a computer device, which can be a server, is provided, and an internal structure diagram of the computer device can be as shown in FIG. 1. Figure 4 The computer device includes a processor, a memory, an input / output interface, and a communication interface. The processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The database of the computer device is configured to store server data. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with terminals outside through a network connection. The computer program is executed by the processor to implement a wireless intercom voice transmission MESH self-organizing network method.

[0156] Those skilled in the art can understand that Figure 4 The structure shown in FIG. 1 is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0157] In an embodiment, a computer device is also provided, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0158] In an embodiment, a computer readable storage medium is provided, which stores a computer program. The computer program is executed by a processor to implement the steps in the above method embodiments.

[0159] In an embodiment, a computer program product or a computer program is provided, which includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the steps in the above method embodiments.

[0160] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0161] It can be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, it can include the processes of the above-mentioned embodiments of each method. Any reference to memory, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided by the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided by the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0162] The technical features of the above embodiments can be combined in any way. To make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present disclosure.

[0163] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. A method for radio intercom voice transmission MESH ad hoc network, characterized in that, The method comprises: determining an initial networking network architecture corresponding to sending an initial voice data packet according to voice bandwidth information in the initial voice data packet output by a radio device; controlling the initial voice data packet to be sent from the radio device to a relay layer through the initial networking network architecture to obtain a relay voice data packet; fusing the relay voice data packet and a relay voice addition packet obtained by the relay layer to obtain a relay voice fusion packet; adjusting the initial networking network architecture according to the relay voice fusion packet to obtain an adjusted networking network architecture; returning to execute the step of controlling the initial voice data packet to be sent from the radio device to the relay layer through the initial networking network architecture to obtain the relay voice data packet, with the relay voice fusion packet as the initial voice data packet and the adjusted networking network architecture as the initial networking network architecture; until a voice output instruction for the relay voice data packet is received by the relay layer to obtain an output voice data packet.

2. The method of claim 1, wherein, The adjusting the initial networking network architecture according to the relay voice fusion packet to obtain an adjusted networking network architecture comprises: selecting an extended architecture of the initial networking network architecture according to voice bandwidth information of the relay voice fusion packet to obtain a topology expansion network unit; splicing the initial networking network architecture and the topology expansion network unit to obtain the adjusted networking network architecture.

3. The method of claim 2, wherein, The selecting an extended architecture of the initial networking network architecture according to voice bandwidth information of the relay voice fusion packet to obtain a topology expansion network unit comprises: performing reverse backtracking analysis on the voice bandwidth information and forwarding statistical data in the relay voice fusion packet to obtain a relay layer bottleneck link set; performing time-frequency occupation prediction analysis on each candidate topology unit for the initial networking network architecture according to the voice bandwidth information and historical interference observation data to obtain each bandwidth-compliant candidate unit; performing load balancing solving on each bandwidth-compliant candidate unit and the relay layer bottleneck link set to obtain a tentative expansion network unit; performing path reachability adjustment on the tentative expansion network unit according to a minimum hop number or minimum path loss rule to obtain the topology expansion network unit.

4. The method of claim 3, wherein, The performing reverse backtracking analysis on the voice bandwidth information and forwarding statistical data in the relay voice fusion packet to obtain a relay layer bottleneck link set comprises: performing minimum cut analysis on the relay voice fusion packet to obtain a candidate bottleneck edge set; performing sliding median filtering on an arrival timestamp sequence in the forwarding statistical data to obtain a time delay trajectory sequence; performing constraint analysis on the voice bandwidth information to obtain a bandwidth constraint parameter set; performing sparse causal backtracking solving on the candidate bottleneck edge set according to the bandwidth constraint parameter set and the time delay trajectory sequence to obtain the relay layer bottleneck link set.

5. The method of claim 3, wherein, The performing load balancing solving on each bandwidth-compliant candidate unit and the relay layer bottleneck link set to obtain a tentative expansion network unit comprises: Multi-objective cost embedding is performed on the time slot bearing vector of each bandwidth compliant candidate unit to obtain weighted capacity information; According to the set of relay layer bottleneck links, the weighted capacity information is perturbation robustness processed to obtain robust weighted capacity information; The robust weighted capacity information is subjected to minimum cost maximum flow solving to obtain assignment scheme data; The assignment scheme data is subjected to discretization stability shaping to obtain the proposed extended network unit.

6. The method of claim 1, wherein, The method further comprises: Joint weight analysis is performed on the energy mask spectrum and link confidence data in the relay voice data packet and the relay voice addition packet to obtain fusion weight data; Jitter buffer alignment is performed on the relay voice data packet and the relay voice addition packet to obtain a sequence of voice packet alignment frames; According to the fusion weight data, transform domain sparse fusion is performed on the sequence of voice packet alignment frames to obtain the relay voice fusion packet.

7. The method of claim 1, wherein, In the case of multiple different frequency bands of the relay voice fusion packet, the method further comprises: According to the link measurement data and regulation duty cycle data corresponding to each frequency band of the relay voice fusion packet, cross-frequency relay potential analysis is performed on each candidate topology unit to obtain a set of cross-frequency relay hubs; Slot alignment is performed on the set of cross-frequency relay hubs and the time slot frame structure of each frequency band to obtain a cross-frequency forwarding time sequence table; According to the cross-frequency forwarding time sequence table and the bandwidth time-varying information of the relay voice fusion packet, multi-frequency coupling reconstruction is performed on the initial networking topology to obtain the adjusted networking network architecture.

8. The method of claim 7, wherein, The method further comprises: Opportunity constraint confidence projection is performed on the link measurement data and regulation duty cycle data corresponding to each frequency band to obtain a frequency band availability probability cone; According to the bandwidth time-varying information of the relay voice fusion packet and the frequency band availability probability cone, conditional risk analysis is performed on each candidate topology unit to obtain a set of cross-frequency relay potential scores; Kalman smoothing is performed on the time series of the set of cross-frequency relay potential scores to obtain a robust score sequence; Submodular function greedy maximization is performed on the robust score sequence under independent set constraints to obtain the set of cross-frequency relay hubs.

9. A wireless radio voice transmission MESH ad hoc network apparatus, characterized by, The device comprises: A network architecture determination module configured to determine an initial networking network architecture corresponding to an initial voice data packet output by a radio device according to voice bandwidth information in the initial voice data packet; A voice data transmission module configured to control the initial voice data packet to be transmitted from the radio device to a relay layer through the initial networking network architecture to obtain a relay voice data packet; A voice data fusion module configured to fuse the relay voice data packet and a relay voice addition packet obtained by the relay layer to obtain a relay voice fusion packet; a network architecture adjusting module, configured to adjust the initial networking network architecture according to the relay voice convergence packet, to obtain an adjusted networking network architecture; a voice data transmission module, further configured to return to execute the step of controlling the initial voice data packet to be sent from the radio device to a relay layer through the initial networking network architecture to obtain a relay voice data packet, with the relay voice convergence packet as the initial voice data packet and the adjusted networking network architecture as the initial networking network architecture; a voice data receiving module, configured to obtain an output voice data packet until a voice output instruction for the relay voice data packet is received by the relay layer.

10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor implements the steps of the method in any one of claims 1 to 8 when executing the computer program.

Images