Interference coordination method and system for heterogeneous networks based on dynamic spectrum sharing

Abstract

This invention relates to the field of wireless communication anti-interference, and discloses a method and system for collaborative suppression of heterogeneous network interference based on dynamic spectrum sharing. The method constructs a dynamic spectrum sharing mechanism for heterogeneous networks: network nodes are deployed in a master-slave architecture, including one master node and N slave nodes, where N≥0; the number of slave nodes is limited according to different application scenarios; spectrum prediction input data is collected; an LSTM spectrum demand prediction model is constructed: based on data with a training dataset size of ≥80,000 records from step two, the model is trained to predict the spectrum demand of each node within the next 10 minutes, with a prediction accuracy of ≥85%; look-ahead spectrum adjustment: if the overlap rate of a certain frequency band is predicted to be ≥28% within the next 5 minutes, the master node issues a frequency band switching command to the conflicting node 3 minutes in advance, scheduling it to an idle channel, thereby reducing temporary spectrum conflicts by ≥70%. This invention can improve overall network capacity and communication stability according to different application scenarios.

Description

Technical Field

[0001] This invention relates to the field of wireless communication anti-interference technology, specifically to a method and system for collaborative suppression of heterogeneous network interference based on dynamic spectrum sharing. Background Technology

[0002] In today's communications field, with the rapid development of various wireless devices and services, spectrum resources are becoming increasingly scarce. To meet the ever-growing communication demands, heterogeneous networks have emerged, which improve spectrum utilization and system capacity by integrating networks of different types, coverage areas, and transmission characteristics. However, the mutual interference between different network nodes in heterogeneous networks has become a key factor restricting their performance improvement.

[0003] Traditional spectrum allocation methods mostly employ fixed frequency band allocation, which lacks flexibility in complex and ever-changing communication scenarios. For example, in scenarios with multiple transmitters coexisting, such as film shooting or concerts, the spectrum requirements of different devices are highly dynamic and random. At a concert, services such as lighting control signals and audio transmission experience peak and off-peak frequency usage at different times. If a fixed frequency band allocation is used, signal interruptions due to frequency band overlap can easily occur, severely impacting the performance.

[0004] While existing interference suppression technologies can alleviate interference problems to some extent, they still have many drawbacks. Some technologies only address a single type of interference, failing to comprehensively address the complex and diverse interference situations in heterogeneous networks. Others do not adequately consider the efficient use of spectrum resources when suppressing interference, resulting in low spectrum utilization. Furthermore, some interference suppression methods rely on centralized control by a central node, which not only increases signaling overhead but also reduces the network's self-organizing capabilities and flexibility, making it difficult to adapt to scenarios requiring rapid deployment and dynamic adjustments in the field.

[0005] This invention addresses the problems existing in the prior art. Summary of the Invention

[0006] The present invention aims to provide a method and system for collaborative suppression of heterogeneous network interference based on dynamic spectrum sharing, so as to solve the problems of severe network interference and insufficient communication stability in existing complex environments.

[0007] To solve the above problems, the present invention adopts the following technical solution: Option 1: A heterogeneous network interference cooperative suppression method based on dynamic spectrum sharing, including the following steps: Step 1: Construct a dynamic spectrum sharing mechanism for heterogeneous networks: Deploy network nodes in a master-slave architecture, including 1 master node and N slave nodes, where N≥0. The number of slave nodes is limited according to different application scenarios. When N=0, spectrum data is obtained through the master node. Slave nodes upload spectrum usage information to the master node every 150ms, including the occupancy rate of each frequency band and the real-time transmit power of the node. The sampling interval for the occupancy rate of each frequency band is 5 minutes, with an accuracy of ±1.5%. The real-time transmit power of the node ranges from -35dBm to 20dBm, with an accuracy of ±0.8dBm. Step 2, collect spectrum prediction input data: obtain historical spectrum data and real-time network status data for the past 24 hours; among which, the historical spectrum data includes the frequency band occupancy rate curve and transmit power fluctuation value recorded every 5 minutes, and the node location error of the real-time network status data is ≤3m, and the service type priority is: main service 1.0 / auxiliary service 0.6; Step 3: Construct an LSTM spectrum demand prediction model: Train the model based on the training dataset of Step 2 with a size of ≥80,000 data points, and predict the spectrum demand of each node in the next 10 minutes with a prediction accuracy of ≥85%; Step 4, Forward-looking spectrum adjustment: If it is predicted that the overlap rate of a certain frequency band will be ≥28% within the next 5 minutes, the master node will issue a frequency band switching instruction to the conflicting node 3 minutes in advance to schedule it to an idle channel, thereby reducing temporary spectrum conflicts by ≥70%; Idle channel judgment criteria: occupancy rate ≤4%.

[0008] Beneficial effects: 1. Architectural Flexibility and Scenario Adaptability: A dynamic spectrum sharing mechanism is built through a master-slave architecture (1 master node + N slave nodes, N≥0). The number of slave nodes can be flexibly adjusted according to the application scenario, providing a basic framework for subsequent differentiated optimization for different scenarios such as film and television shooting and concerts, avoiding the limitations of traditional fixed node configurations that are difficult to adapt to multiple scenarios. Specifically, when N=0, only the master node (i.e., the host) is used to obtain spectrum data, and the host has both search and reception functions.

[0009] 2. Precise Spectrum Information Interaction: Spectrum usage information is uploaded from nodes every 150ms. Combined with quantitative parameters of frequency band occupancy (sampling interval 5 minutes, accuracy ±1.5%) and transmit power (range -35dBm to 20dBm, accuracy ±0.8dBm), the master node can accurately grasp the spectrum status of the entire network in real time. Compared with the traditional second-level upload interval, the timeliness of spectrum information is improved by 67%, providing data support for precise interference suppression.

[0010] 3. Proactive Interference Avoidance: Based on the LSTM model (training data ≥80,000 records, prediction accuracy ≥85%), the spectrum demand for the next 10 minutes is predicted, and conflicting nodes with an overlap rate ≥28% are scheduled to idle channels (occupancy rate ≤4%) 3 minutes in advance, reducing temporary spectrum conflicts by ≥70%. This solves the signal interruption problem caused by traditional "post-event response" interference suppression and ensures the continuity of real-time services such as lighting control for film and television shooting and audio transmission for concerts.

[0011] 4. Improved resource utilization efficiency: The idle channel judgment standard (occupancy rate ≤ 4%) ensures that spectrum resources are not idle. Combined with the prediction logic of historical data and real-time status, it avoids blind frequency band allocation and lays the foundation for subsequent spectrum utilization improvement (such as an overall improvement of 45%). At the same time, the master-slave architecture reduces data redundancy in the whole network broadcast and reduces signaling overhead by more than 50%.

[0012] This invention constructs a heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing. By interacting with spectrum usage information in real time, it collaboratively adjusts the transmission power and frequency band to achieve efficient spectrum utilization when multiple networks coexist, reduce inter-network interference, and improve overall network capacity and communication stability, thereby meeting the communication needs of complex scenarios such as film and television shooting and concerts.

[0013] Preferably, in step one, the number of nodes is limited differently depending on the application scenario, specifically including: Enclosed space: The number of slave nodes is limited to 8-16, of which core slave nodes, including the main camera and lighting control console, account for ≥50% of the total number of slave nodes. Their transmission power accuracy is improved to ±0.5dBm, and the frequency band occupancy sampling interval is shortened to 3 minutes. The proportion of auxiliary slave nodes is ≤50%, and low-cost communication modules can be used. Their frequency band occupancy sampling interval is allowed to be relaxed to 10 minutes. Open space: The number of slave nodes is limited to 16-32, of which mobile slave nodes, including mobile stage lights and wireless microphones, account for ≥60%, and their positional error requirement is increased to ≤2m; the proportion of fixed slave nodes is ≤40%, and a simplified version of the spectrum interaction protocol can be used to reduce data transmission volume; In both scenarios, the core slave node has a higher priority for anti-interference than the auxiliary slave node and the fixed slave node, and the communication quality of the core node is guaranteed first in the event of spectrum conflict.

[0014] Beneficial effects: By limiting the number and parameters of slave nodes in specific scenarios, the core node's power accuracy of ±0.5dBm ensures signal transmission reliability of ≥99.5% in high-stability film and television shooting scenarios. In large-coverage concert scenarios, the 2m positional error of mobile nodes ensures dynamic tracking accuracy. The low-cost configuration and relaxed parameters of auxiliary nodes, while meeting basic anti-interference requirements (collision reduction ≥60%), reduce the overall system cost by 15-20%, achieving a scenario-based balance of "precise anti-interference - cost optimization".

[0015] Based on the master-slave architecture and business priorities, the system enables scenario-based classification and parameter differentiation configuration of slave nodes. In film and television shooting scenarios, the power accuracy of the core slave node within ±0.5dBm can ensure the reliability of the main camera signal by ≥99.5%, and the sampling interval of the auxiliary slave node can be relaxed to reduce hardware costs. In concert scenarios, the positional error of the mobile slave node within ≤2m ensures the accuracy of dynamic spectrum adjustment, and the protocol of the fixed slave node is simplified to reduce data redundancy. The overall system deployment cost is reduced by 15-20%. While meeting the anti-interference requirements of different scenarios (conflict reduction ≥60%), the system avoids "one-size-fits-all" high-cost configurations or low-performance compromises.

[0016] Preferably, the hierarchical power regulation steps based on node classification are as follows: The first step is to define five levels of transmit power standards: Level 1 Energy Saving Level: -35dBm to -25dBm, Level 2 Basic Level: -24dBm to -15dBm, Level 3 Standard Level: -14dBm to -5dBm, Level 4 Enhanced Level: -4dBm to -10dBm, and Level 5 Highest Level: 11dBm to -20dBm. The second step is to calculate the degree of spectrum conflict: the frequency band overlap rate of adjacent nodes, including master nodes, slave nodes, and different types of slave nodes, is calculated as (overlapping frequency band width / total used frequency band width) × 100%. Conflict levels are then classified as: low conflict <20%, medium conflict 20%-40%, and high conflict >40%. The third step is to dynamically adjust the power: when there is a high conflict, the power of the auxiliary slave node and the fixed slave node is reduced by 1-2 levels, while the core slave node and the mobile slave node are maintained at level 3 or above; when there is a medium conflict, the power of the auxiliary slave node and the fixed slave node is reduced by 1 level; when there is a low conflict, the current level is maintained; after adjustment, the network interference is reduced by ≥40%, and the node energy consumption is reduced by ≥25%.

[0017] Beneficial effects: This invention combines service priority, slave node classification, and a 5-level power standard to achieve precise matching of "conflict level - node type - power adjustment". In high-conflict scenarios, only the power of auxiliary / fixed slave nodes is reduced, which can ensure that the signal-to-noise ratio of the main service carried by the core / mobile slave node (such as the main camera for film and television shooting, and the wireless microphone for concerts) is ≥26dB, avoiding the interruption of the main service signal. At the same time, the 25% reduction in node power consumption is suitable for battery life-sensitive scenarios such as outdoor shooting and open-air concerts, and the 40% reduction in inter-network interference meets the needs of multi-node parallel communication, achieving the dual goals of "anti-interference and power consumption optimization".

[0018] Preferably, the game-theoretic collaborative steps for node classification are as follows: The first step is to construct a game theory model: the node communication quality with a packet loss rate of ≤0.8% is used as the payoff function, and the strategy space includes switching frequency bands and reducing power; among them, there are ≥3 available idle channels for switching frequency bands, and the power reduction ranges from 1 to 3 levels; the nodes include master nodes and various types of slave nodes; The second step is strategy iteration and optimization: update the node strategy every 300ms until all nodes reach Nash equilibrium, with a convergence time of ≤1.5s; The third step is global allocation of master nodes: based on the balance results, the core slave nodes and mobile slave nodes are allocated low-collision channels with an overlap rate of <15% to ensure that the overall network capacity is increased by ≥30% and the continuous uninterrupted time is increased by ≥20%.

[0019] Beneficial effects: By building a game theory model based on node classification, a strategy iteration cycle of 300ms and a convergence time of 1.5s can quickly achieve spectrum allocation balance in scenarios such as simultaneous filming by multiple crews and competition among multiple devices in concerts, avoiding the aggravation of interference caused by "power escalation" between nodes. At the same time, by prioritizing the allocation of low-collision channels to core / mobile slave nodes, a 30% increase in network capacity can support large-scale access of 32 slave nodes in a concert, and a 20% improvement in communication stability ensures zero interruption of real-time services such as lighting control in film and television shooting and audio transmission in concerts, balancing the benefits of individual nodes with overall network performance.

[0020] Preferably, it further includes a self-interference removal step for node classification: The first step, dual-band spectrum sensing: using sensing radio technology, scanning the 2.4GHz ISM band in the range of 2400-2483.5MHz and the 5.8GHz ISM band in the range of 5725-5825MHz, with a scanning interval of ≤40ms, a scanning bandwidth of ≥22MHz, and an interference signal identification sensitivity of ≤-92dBm; The second step is self-interference identification: If the signal strength difference of co-channel interference between the core slave nodes or between the core slave node and the mobile slave node is ≤5dB, the self-interference removal process is triggered. The third step is frequency band reallocation: the master node allocates a new frequency band to the interfering node within 80ms. The new frequency band is ≥25MHz away from the original frequency band. Priority is given to ensuring that the new frequency band occupancy rate of the core slave node and the mobile slave node is ≤3%. After reallocation, the signal isolation between nodes is ≥32dB.

[0021] Beneficial effects: Through dual-band sensing and rapid redistribution, the high sensitivity of -92dBm can identify weak interference signals in the complex electromagnetic environment of concerts (self-interference recognition rate ≥98%), and the 80ms redistribution delay is suitable for dynamic switching scenarios of multi-camera shooting in film and television, avoiding picture interruption caused by self-interference. At the same time, it prioritizes ensuring that the new frequency band occupancy rate of core / mobile slave nodes is ≤3%, and the 32dB signal isolation can ensure that the bit error rate of main services (such as high-definition video and real-time audio) is ≤10^-6, solving the problem of performance degradation of main services caused by "indiscriminate redistribution" in traditional self-interference removal.

[0022] Preferably, the power adjustment step further includes conflict response optimization based on the accuracy of the spectrum information: The first step, rapid collision detection: A sliding window algorithm is used to calculate the frequency band overlap rate of adjacent nodes in real time, with a detection delay of ≤45ms and a detection accuracy maintained at ±1.5%; The second step is to adjust the response in stages: In the case of low conflict, the fixed slave node maintains its current power and only updates the spectrum information upload frequency to 200ms; in the case of medium conflict, the auxiliary slave node power is reduced by 1 level, and the master node verifies the interference reduction effect at 100ms intervals; in the case of high conflict, the fixed slave node power is reduced by 2 levels, and the core slave node power is compensated to level 4 to ensure that the main service signal strength fluctuation is ≤3dB.

[0023] Beneficial effects: Based on graded power control, the conflict response mechanism is optimized. The 45ms fast detection delay can quickly trigger adjustments in high-conflict scenarios between adjacent studios during film and television shooting, avoiding interference propagation. The 100ms effect verification during medium-conflict scenarios ensures accurate power adjustment (interference reduction deviation ≤5%). During high-conflict scenarios, the power compensation of the core slave node can prevent a sudden drop in the main service signal strength (fluctuation ≤3dB). At the same time, the fixed 200ms information upload frequency adjustment of the slave node reduces signaling overhead. Overall, a closed loop of "fast response - accurate control - main service protection" is achieved, and the interference reduction effect is stably maintained at over 40%.

[0024] Preferably, the strategy iteration also includes scenario convergence optimization: The first step is to optimize the iteration parameters based on the scenario: In the closed space scenario, the strategy iteration weight of the core slave node is increased to 1.2, the weight of the auxiliary slave node is 0.8, and the convergence time is compressed to 1.2s; In the open space scenario, the strategy iteration weight of the mobile slave node is increased to 1.1, the weight of the fixed slave node is 0.9, and the search range of the idle channel is expanded to 5. The second step is to verify the balance result: After each iteration converges, the master node verifies that the packet loss rate of the core slave node and the mobile slave node is ≤0.8%. If it does not meet the requirement, the iteration is repeated until the standard is met.

[0025] Beneficial effects: The game-theoretic iteration parameters are optimized for different scenario characteristics. In the closed-space film and television shooting scenario, a convergence time of 1.2 seconds can quickly resolve conflicts between equipment from multiple film crews, and the weight of 1.2 for the core slave node ensures that the main camera signal meets the priority. In the open-space concert scenario, the search range of 5 idle channels improves the frequency band switching flexibility of the mobile slave node, and a weight of 1.1 ensures the stability of the wireless microphone signal. At the same time, the equalization result verification mechanism ensures that the packet loss rate of the core / mobile slave node is consistently ≤0.8%, avoiding the problem of the main service quality not meeting the standard after iterative convergence, and further improving the scenario adaptability and reliability of game-theoretic collaboration. Preferably, frequency band reallocation also includes enhanced isolation of service priorities: The first step is to define the isolation level requirements: the signal isolation between the core slave nodes is ≥35dB, the isolation between the core slave node and the auxiliary slave node is ≥32dB, and the isolation between the mobile slave node and the fixed slave node is ≥30dB. The second step is dynamic frequency band selection: Prioritize selecting channels from the idle channels with an interval of ≥30MHz from the interference frequency band and assign them to the core slave node, channels with an interval of ≥25MHz and assign them to the mobile slave node, and channels with an interval of ≥20MHz and assign them to the auxiliary / fixed slave node.

[0026] Beneficial effects: By employing tiered isolation requirements and dynamic frequency band selection, the 35dB isolation of the core slave nodes ensures no signal crosstalk between main cameras in film and television shooting (bit error rate ≤10^-7). The 25MHz spacing requirement for mobile slave nodes adapts to the mobile scenarios of concert equipment (isolation remains ≥30dB when the location changes). The 20MHz spacing of auxiliary / fixed slave nodes balances the isolation effect and channel utilization. At the same time, relying on the idle channel resources of claim 1, the frequency band reallocation success rate is ≥99%, avoiding delays in self-interference removal caused by insufficient channels, and further enhancing the transmission stability of the main service signals.

[0027] Option 2: A heterogeneous network interference cooperative suppression system based on dynamic spectrum sharing, employing the heterogeneous network interference cooperative suppression method described above, including: Master-Slave Spectrum Interaction Module: Configured with 1 master node and N slave nodes, where N≥0. The number of slave nodes is limited according to scenario classification. When N=0, spectrum data is obtained through the master node. Every 150ms, the slave nodes transmit spectrum usage information to the master node: frequency band occupancy rate ±1.5%, transmit power -35dBm to 20dBm±0.8dBm; module construction delay ≤450ms. Multi-dimensional data acquisition module: Collects historical spectrum data and real-time network status data over the past 24 hours, with a data error of ≤1%; LSTM prediction scheduling module: Based on a model trained with ≥80,000 data points, it outputs a spectrum demand prediction for the next 10 minutes with a prediction delay of ≤90ms; when the predicted frequency band overlap rate is ≥28%, it issues a handover instruction to an idle channel with an occupancy rate of ≤4% 3 minutes in advance. Game-theoretic collaborative allocation module: Built-in Nash equilibrium solution algorithm, iterates node strategy once every 300ms, convergence time ≤1.5s; based on the equilibrium result, it prioritizes the allocation of low-collision channels with an overlap rate of <15% to the core slave node and mobile slave node, improving network capacity by ≥30%.

[0028] Beneficial effects: This system uses four modules and corresponding methodological steps to achieve rapid deployment in closed-space film and television shooting and open-space concerts with a module construction latency of 450ms. The prediction latency of 90ms ensures timely forward scheduling (reducing temporary conflicts by ≥70%). The game-theoretic collaborative allocation module, combined with node classification, can quickly achieve spectrum balance in multiple scenarios. The 30% increase in network capacity supports large-scale node access. At the same time, the data acquisition module's ±1% error ensures accurate subsequent prediction and control. The overall system achieves a systematic implementation of the methodology, taking into account anti-interference performance, deployment efficiency, and scenario adaptability.

[0029] Preferably, it further includes: Hierarchical power control module: Stores 5 levels of transmit power standards, receives the collision level signal from the collision detection module, and adjusts the slave node power within 90ms to reduce inter-network interference by ≥40% and node power consumption by ≥25%; Dual-band self-interference removal module: Configured with 2.4GHz+5.8GHz scanning sub-modules, generating frequency band reallocation instructions within 80ms to ensure that the isolation between core slave nodes is ≥35dB and the isolation between mobile slave nodes is ≥32dB.

[0030] Beneficial effects: This system is supplemented with a power regulation and self-interference cancellation module, featuring a fast response of 45ms conflict detection + 90ms power adjustment, which can quickly reduce interference (deviation ≤5%) in high-conflict scenarios with adjacent studios; the dual-band scanning submodule's high sensitivity of -92dBm ensures comprehensive identification of weak interference (identification rate ≥98%), and the 80ms redistribution delay avoids interruption of video shooting; at the same time, the core slave node's 35dB isolation and 25% energy consumption reduction achieve a system-level balance of "anti-interference, energy consumption, and signal quality", adapting to complex multi-node coexistence scenarios.

[0031] Advantages of this invention: 1. Multi-scenario anti-interference accuracy: Through a combination strategy of "scenario-based node configuration + forward prediction + dynamic adjustment", it not only solves the problem of interference between adjacent studios caused by the small enclosed space and dense equipment in film and television shooting, but also copes with the dynamic spectrum conflict caused by the movement and large number of equipment in the open space of concerts. It reduces the spectrum overlap signal conflict rate by more than 60% in various scenarios, covering the two core requirements of high stability and high flexibility.

[0032] 2. Cost and performance synergistic optimization: Supports differentiated constraints on slave nodes for different scenarios (such as high-precision configuration of core nodes for film and television shooting and low-cost simplification of auxiliary nodes). While ensuring the anti-interference performance of core business (such as the signal reliability of the main camera ≥99.5%), the overall system deployment cost is reduced by 15-20%, avoiding the drawbacks of traditional "one-size-fits-all" high-cost configuration or low-performance compromise.

[0033] 3. High efficiency in both spectrum and energy consumption: The dynamic spectrum sharing mechanism breaks the fixed frequency band allocation mode, increasing spectrum utilization by 45%, which is suitable for the flexible needs of temporary addition and removal of equipment in film and television shooting and the movement of equipment in concerts; the combination of hierarchical power control and LSTM prediction reduces node energy consumption by 25%, which is in line with the trend of green communication and is especially suitable for battery life-sensitive scenarios such as outdoor concerts and outdoor shooting.

[0034] 4. Network self-organization and stability: The hierarchical information interaction and game theory collaboration mechanism under the master-slave architecture (convergence time ≤1.5s) eliminates the need for excessive centralized control by the central node, reduces signaling overhead and improves network adaptability. Even when devices are temporarily connected (≤10 devices / minute) or the scenario changes dynamically, communication stability can still be improved by 20%, meeting the needs of rapid on-site deployment. Attached Figure Description

[0035] Figure 1 This is a flowchart of the heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing according to the present invention.

[0036] Figure 2 This is a logic block diagram of the heterogeneous network interference collaborative suppression system based on dynamic spectrum sharing, as described in this invention. Detailed Implementation

[0037] The following detailed description illustrates the specific implementation method: As attached Figure 1 As shown: A heterogeneous network interference cooperative suppression method based on dynamic spectrum sharing includes the following steps: 1. Infrastructure and Data Acquisition: First, a dynamic spectrum sharing mechanism for heterogeneous networks is constructed. Then, spectrum prediction input data is collected to build an LSTM spectrum demand prediction model. First, a heterogeneous network dynamic spectrum sharing mechanism is constructed, deploying network nodes in a master-slave architecture, including 1 master node and N slave nodes (N≥0). The number of slave nodes is limited according to the application scenario: In enclosed spaces (such as film studios), the number of slave nodes is 8-16, with core slave nodes such as main cameras and lighting control consoles accounting for ≥50%, transmit power accuracy ±0.5dBm, and frequency band occupancy sampling interval of 3 minutes; auxiliary slave nodes account for ≤50%, using low-cost communication modules, and frequency band occupancy sampling interval of 10 minutes. In open spaces (such as concert venues), the number of slave nodes is 16-32, with mobile slave nodes such as mobile stage lights and wireless microphones accounting for ≥60%, position error ≤2m; fixed slave nodes account for ≤40%, and a simplified spectrum interaction protocol is used to reduce data transmission volume. Every 150ms, all slave nodes upload spectrum usage information to the master node, including the frequency band occupancy rate (sampling interval 5 minutes, accuracy ±1.5%) and the node's real-time transmit power (range -35dBm to 20dBm, accuracy ±0.8dBm).

[0038] Next, spectrum prediction input data is collected, including historical spectrum data (including frequency band occupancy curves and transmit power fluctuations recorded every 5 minutes) and real-time network status data (node ​​location error ≤ 3m, service type priority is "main service 1.0 / auxiliary service 0.6"), based on these data of ≥ 80,000 records, an LSTM spectrum demand prediction model is constructed to predict the spectrum demand of each node in the next 10 minutes, ensuring a prediction accuracy of ≥ 85%.

[0039] 2. Forward-looking scheduling and multi-dimensional interference suppression If the overlap rate of a certain frequency band is predicted to be ≥28% within the next 5 minutes, the master node sends a frequency band switching command to the conflicting node 3 minutes in advance, scheduling it to an idle channel with an occupancy rate of ≤4%, thereby reducing temporary spectrum conflicts by ≥70%. Simultaneously, for different scenarios and slave node classifications, multi-dimensional interference suppression strategies are implemented: Tiered power control: Define 5 levels of transmit power standards (Level 1: -35 to 25dBm, Level 2: -24 to 15dBm, Level 3: -14 to 5dBm, Level 4: -4 to 10dBm, Level 5: 11 to 20dBm). Calculate the frequency band overlap rate of adjacent nodes (including master and slave nodes, and different types of slave nodes) (formula: (overlapping frequency band width / total used frequency band width) × 100%). Classify conflict levels into low (<20%), medium (20%-40%), and high (>40%) based on the overlap rate. In high conflict, auxiliary and fixed slave nodes reduce power by 1-2 levels, while core and mobile slave nodes maintain Level 3 or higher. In medium conflict, auxiliary and fixed slave nodes reduce power by 1 level. In low conflict, maintain the current level. After adjustment, inter-network interference is reduced by ≥40%, and node power consumption is reduced by ≥25%.

[0040] Game-theoretic collaboration: Using a packet loss rate ≤ 0.8% as the node communication quality benefit function, the policy space includes "switching frequency bands (with ≥ 3 available idle channels)" and "reducing power (at least 1-3 levels)". Participating nodes include master nodes and various types of slave nodes. Node policies are updated every 300ms and converge to a Nash equilibrium state within 1.5s. Based on the equilibrium result, the master node prioritizes allocating low-collision channels with an overlap rate < 15% to core slave nodes and mobile slave nodes, ensuring an overall network capacity increase of ≥ 30% and a continuous uninterrupted time increase of ≥ 20%.

[0041] Self-interference cancellation: Sensing radio technology is used to scan the 2.4GHz ISM band (2400-2483.5MHz) and the 5.8GHz ISM band (5725-5825MHz), with a scanning interval ≤40ms, bandwidth ≥22MHz, and interference identification sensitivity ≤-92dBm. If co-channel interference (signal strength difference ≤5dB) is detected between core slave nodes or between a core slave node and a mobile slave node, the master node allocates a new frequency band to the interfering node within 80ms (with an interval ≥25MHz from the original frequency band), prioritizing the core slave node and mobile slave node to ensure that the new frequency band occupancy rate is ≤3%, and the signal isolation between nodes after reallocation is ≥32dB.

[0042] 3. Strategy optimization and effect enhancement Based on the above, further optimize the interference suppression effect: Collision response optimization: A sliding window algorithm is used to calculate the frequency band overlap rate of adjacent nodes in real time, with a detection delay of ≤45ms and an accuracy of ±1.5%. In the case of low collisions, the fixed slave node maintains its current power and only uploads the spectrum information to update the frequency to 200ms. In the case of medium collisions, after the auxiliary slave node power is reduced by 1 level, the master node verifies the interference reduction effect at 100ms intervals. In the case of high collisions, the fixed slave node power is reduced by 2 levels, and the core slave node power is compensated to level 4 to ensure that the main service signal strength fluctuation is ≤3dB.

[0043] Game-theoretic convergence optimization: In closed-space scenarios, the strategy iteration weight of the core slave node is increased to 1.2 and that of the auxiliary slave node is 0.8, reducing the convergence time to 1.2 seconds. In open-space scenarios, the strategy iteration weight of the mobile slave node is increased to 1.1 and that of the fixed slave node is 0.9, expanding the search range of idle channels to 5. After each iteration converges, the master node verifies the communication quality (packet loss rate ≤ 0.8%) of the core slave node and the mobile slave node. If the quality is not met, the iteration is repeated until the standard is met.

[0044] Isolation Enhancement: Set graded isolation requirements (≥35dB between core and slave nodes, ≥32dB between core and auxiliary slave nodes, ≥30dB between mobile and fixed slave nodes), and dynamically select idle channels: Prioritize allocating channels with an interval of ≥30MHz from interfering frequency bands to core slave nodes, channels with an interval of ≥25MHz to mobile slave nodes, and channels with an interval of ≥20MHz to auxiliary / fixed slave nodes.

[0045] The heterogeneous network interference cooperative suppression system based on dynamic spectrum sharing of the present invention includes the following: 1. Configuration of core functional modules like Figure 2 As shown, the system fully implements the above method and includes four core modules connected to the central processing unit: Master-Slave Spectrum Interaction Module: Configured with 1 master node and N slave nodes (N≥0). When N=0, spectrum data is acquired solely by the master node, which has both search and receive capabilities. The "master" in this context refers to the function of acquiring data, while the "slave" refers to the function of collecting data. The slave components can exist as code within the actual host hardware or within a physical slave device. When N=0, the slave components reside within the master node, allowing the master node to acquire spectrum data independently without the need for slave nodes.

[0046] The number of slave nodes is limited according to the scenario (8-16 in enclosed spaces, 16-32 in open spaces). Slave nodes transmit spectrum usage information (band occupancy rate ±1.5%, transmit power -35dBm to 20dBm ±0.8dBm) to the master node every 150ms. The module construction delay is ≤450ms, ensuring that the master node has real-time knowledge of the entire network's spectrum status.

[0047] Multi-dimensional data acquisition module: Collects historical spectrum data from the past 24 hours (records frequency band occupancy curve and transmit power fluctuation value every 5 minutes) and real-time network status data (node ​​location error ≤3m, service priority "main service 1.0 / auxiliary service 0.6"), with a data acquisition error ≤1%, providing high-quality input data for subsequent prediction models.

[0048] LSTM Prediction and Scheduling Module: Based on ≥80,000 data points, an LSTM model is trained to output the spectrum demand prediction results for the next 10 minutes (accuracy ≥85%), with a prediction latency ≤90ms. When the overlap rate of a certain frequency band is predicted to be ≥28% within the next 5 minutes, a frequency band switching command is issued to the conflicting node 3 minutes in advance, scheduling it to an idle channel with an occupancy rate ≤4%, reducing temporary spectrum conflicts by ≥70%.

[0049] Game-theoretic collaborative allocation module: It incorporates a Nash equilibrium solution algorithm, updates node policies every 300ms, and converges to an equilibrium state within 1.5s. Based on the equilibrium results, it prioritizes allocating low-collision channels with an overlap rate of <15% to core slave nodes and mobile slave nodes, ensuring a network capacity increase of ≥30% and a continuous uninterrupted time improvement of ≥20%.

[0050] 2. Extended Function Module Configuration To enhance interference suppression, the system can be additionally configured with two expansion modules: The graded power control module stores five levels of transmit power standards and receives the collision level signal output by the collision detection unit (using a sliding window algorithm, with a detection delay of ≤45ms and an accuracy of ±1.5%). Within 90ms, it adjusts the power of slave nodes: in case of high collisions, it reduces the power of auxiliary / fixed slave nodes by 1-2 levels and maintains the power of core / mobile slave nodes at ≥3 levels; in case of medium collisions, it reduces the power of auxiliary / fixed slave nodes by 1 level; and in case of low collisions, it maintains the power. Ultimately, it achieves a reduction of ≥40% in inter-network interference and a reduction of ≥25% in node energy consumption.

[0051] Dual-band self-interference cancellation module: Configured with a 2.4GHz+5.8GHz dual-band scanning submodule (scanning interval ≤40ms, bandwidth ≥22MHz, interference identification sensitivity ≤-92dBm). When co-channel interference (signal strength difference ≤5dB) is detected between core and slave nodes or between core and mobile slave nodes, a frequency band reallocation command is generated within 80ms (the new frequency band is ≥25MHz apart from the original frequency band). This ensures that the isolation between core and slave nodes is ≥35dB and between mobile slave nodes is ≥32dB, prioritizing the new frequency band occupancy rate of core / mobile slave nodes to be ≤3%.

[0052] Specifically, Example 1 This example illustrates a scenario of multiple devices working together in adjacent studios during film and television shooting.

[0053] This example is for a scenario where two film crews are shooting in adjacent studios (enclosed space, each studio area is 200 square meters). (The two sheds are 5 meters apart.) The main camera, lighting control console, temporary listening devices, and other equipment need to be deployed simultaneously. The specific steps are as follows: Step 1: Build a master-slave spectrum sharing architecture Deploy one master node (located in the control room between the two sheds) and 12 slave nodes, of which: There are 6 core slave nodes (accounting for 50%): each connected to 2 main cameras (service priority 1.0), 2 lighting control consoles (service priority 1.0), and 2 director monitors (service priority 1.0), with a transmit power accuracy of ±0.5dBm and a frequency band occupancy sampling interval of 3 minutes. There are 6 auxiliary slave nodes (accounting for 50%): each connected to 4 temporary listening devices (service priority 0.6) and 2 field recorder walkie-talkies (service priority 0.6), using low-cost communication modules, with a frequency band occupancy sampling interval of 10 minutes; Every 150ms, all slave nodes upload spectrum usage information to the master node, including the occupancy rate (accuracy ±1.5%) of the eight channels in the 2.4GHz ISM band (2400-2483.5MHz) and transmit power (range -35dBm to 20dBm, accuracy ±0.8dBm).

[0054] Step 2: Collect spectrum prediction data Collect historical spectrum data from the past 24 hours (recording occupancy curves and transmit power fluctuations for 8 channels every 5 minutes, such as the daily average occupancy of the main camera channel f1 (2412MHz) at 35% and power fluctuations at ±2dB), as well as real-time network status data: Node position error ≤3m (achieved through UWB positioning module); Service type priority: main camera and lighting control console are 1.0, temporary listening device and walkie-talkie are 0.6.

[0055] Step 3: Train the LSTM spectrum prediction model An LSTM model is trained based on ≥80,000 historical and real-time data (including spectrum data from the past 10 similar shooting sessions) to predict the spectrum demand of each node within the next 10 minutes with an accuracy of 86%. For example, it predicts that "the overlap rate of channel f2 (2417MHz) will reach 32% within the next 5 minutes".

[0056] Step 4: Forward-looking spectrum conflict avoidance The master node sends a frequency band switching command to the two auxiliary slave nodes (temporary listeners) occupying channel f2 3 minutes in advance, and schedules them to the idle channel f7 (2442MHz) with an occupancy rate of 3% (≤4%), thereby reducing temporary spectrum conflicts by 72%.

[0057] Step 5: Graded power regulation Define five levels of transmit power standards: Level 1 (-35 to 25dBm), Level 2 (-24 to 15dBm), Level 3 (-14 to 5dBm), Level 4 (-4 to 10dBm), and Level 5 (11 to 20dBm). The overlap rate of channels f1 and f3 between the two core slave nodes (master cameras) was detected to be 42%. Dynamic adjustment: The power of the auxiliary slave node (walkie-talkie) is reduced by 2 levels from level 3 (-8dBm) to level 1 (-30dBm), while the core slave node (main camera) is maintained at level 4 (5dBm) (in accordance with step seven of claim 3). After the adjustment, the interference between the two booths is reduced by 45% (≥40%), and the energy consumption of the auxiliary node is reduced by 27% (≥25%).

[0058] Step 6: Game Theory Collaboration Construct a game theory model: with "packet loss rate ≤ 0.8%" as the payoff function, the strategy space includes "switching to f4 / f6 / f8 (3 idle channels)" and "reducing power by 1-3 levels"; The strategy iterates once every 300ms and reaches Nash equilibrium in 1.4s; The master node prioritizes allocating channel f1 with an overlap rate of 12% to the core slave nodes, resulting in an overall network capacity increase of 31%.

[0059] Step 7: Self-interference cancellation Dual-band sensing: scanning 2.4GHz (2400-2483.5MHz) and 5.8GHz (5725-5825MHz), scanning interval 40ms, bandwidth 22MHz, interference identification sensitivity -93dBm; Interference between two master cameras (core slave nodes) on the same frequency was detected, with a signal strength difference of 4dB. Frequency band reallocation: The master node allocates a new channel f5 (2432MHz) to one of the master cameras within 78ms. The new channel occupancy rate is 2%, and the signal isolation between the two master cameras is 36dB after reallocation.

[0060] This embodiment of the system is deployed in the control room and two studios at the shooting location, and includes the following modules: Master-slave spectrum interaction module Configure one master node (industrial-grade gateway, model: RG450) and 12 slave nodes (the core slave node uses the high-precision wireless module XR240, and the auxiliary slave nodes use the low-cost module XR110). The module setup delay is 420ms. The slave nodes transmit spectrum data (channel occupancy rate ±1.5%, transmit power -35-20dBm±0.8dBm) to the master node every 150ms.

[0061] Multi-dimensional data acquisition module Data was collected using a data acquisition unit (model: DC200): Historical spectrum data: Records the occupancy and power fluctuations of 8 channels every 5 minutes (stored on a local SD card with a capacity of 64GB). Real-time network status: The node position error is ensured to be ≤3m by using a UWB positioning module (accuracy ±0.5m). Service priority is set by software tags (main service 1.0 / auxiliary 0.6). Data acquisition error: 0.8%.

[0062] LSTM Predictive Scheduling Module The system incorporates an LSTM model (85,000 training data points) deployed on the edge computing unit of the master node (2 TOPS computing power), with a prediction latency of 85ms (≤90ms). When the predicted channel overlap rate is ≥28%, a handover command is sent to an idle channel 3 minutes in advance via the LoRa protocol (occupancy rate ≤4%).

[0063] Game Theory Collaborative Allocation Module The built-in Nash equilibrium solution algorithm (implemented based on the Python pulp library) iterates the strategy once every 300ms, with a convergence time of 1.4s; it prioritizes the allocation of channels with an overlap rate of <15% to core slave nodes, thereby improving network capacity by 31%.

[0064] Graded power regulation module It stores a 5-level power standard (fixed in the slave node firmware) and connects to a collision detection unit (using a sliding window algorithm, with a detection delay of 43ms and an accuracy of ±1.5%). Upon receiving a high collision signal, it adjusts the auxiliary slave node power within 90ms (accuracy ±0.8dBm), ultimately reducing interference by 45% and energy consumption by 27%.

[0065] Dual-band self-interference cancellation module It is equipped with dual-band antennas (2.4GHz gain 5dBi, 5.8GHz gain 6dBi), a scan interval of 40ms, and a sensitivity of -93dBm; it generates a redistribution command (new band interval 26MHz) within 80ms, and the isolation between core and slave nodes is 36dB.

[0066] This embodiment has the following advantages: Main business stability assurance: The signal reliability of the core slave node (main camera) reaches 99.6% (packet loss rate 0.4%≤0.8%), with no interruption of the picture, and the response delay of the lighting control console command is <100ms, which meets the requirements of "real-time monitoring and precise lighting control" in film and television shooting; Interference suppression effect: Interference between two booths is reduced by 45%, self-interference elimination rate between core nodes is 98%, temporary spectrum conflict is reduced by 72%, and the problem of "image noise caused by signal crosstalk" of adjacent booth equipment is solved. Cost and energy consumption optimization: The auxiliary slave nodes adopt low-cost modules, reducing the overall system deployment cost by 18% (≥15%) and reducing the energy consumption of the auxiliary nodes by 27%, which is suitable for outdoor shooting with "susceptibility to battery life and controllable budget". Deployment efficiency: The system module build latency is 420ms, and the slave node access time is <300ms, which meets the dynamic requirements of "quick scene building and instant activation" in shooting scenarios.

[0067] Example 2: This example is a large-scale outdoor concert scene.

[0068] This embodiment is designed for a 10,000-person outdoor concert scenario (open space, stage area 500 square meters). (The audience area has a radius of 100m) and requires the deployment of equipment such as wireless microphones, mobile stage lights, and ambient sound transmitters. The specific steps are as follows: Step 1: Build a master-slave spectrum sharing architecture Deploy 1 master node (located on the stage side control console) and 24 slave nodes (N=24, conforming to the limitation of claim 2 "16-32 slave nodes in a concert scene"), wherein: 15 mobile slave nodes (62.5% ≥ 60%): connected to 8 wireless microphones (service priority 1.0) and 7 mobile follow spotlights (service priority 1.0), with a position error ≤ 2m (positioning via GPS + IMU) and a transmission power accuracy of ±0.8dBm; Nine fixed slave nodes (accounting for 37.5% ≤ 40%): connecting six ambient sound transmitters (service priority 0.6) and three fixed audience area speakers (service priority 0.6), using a simplified spectrum interaction protocol (data transmission volume reduced by 40%). All slave nodes upload spectrum usage information (occupancy rate of 12 channels in the 5.8GHz ISM band (5725-5825MHz), accuracy ±1.5%; transmit power -35-20dBm, accuracy ±0.8dBm) to the master node every 150ms.

[0069] Step 2: Data Acquisition and Forward-Looking Scheduling Collect historical data from the past 24 hours (record the occupancy rate of 12 channels every 5 minutes, such as the daily average occupancy rate of the commonly used wireless microphone channel f9 (5750MHz) at 40%), and real-time data: mobile slave node location error 1.8m≤2m, service priority 1.0 / 0.6; The LSTM model was trained based on 90,000 data points, with a prediction accuracy of 87% ≥ 85%, predicting that "the overlap rate of channel f10 (5760MHz) will reach 30% ≥ 28% in the next 5 minutes". The master node switches two fixed slave nodes (audio transmitters) to channel f12 (5800MHz) with an occupancy rate of 2%≤4% 3 minutes in advance, reducing temporary collisions by 75%≥70%.

[0070] Step 3: Graded power regulation The power standard is the same as in Example 1. The mobile slave node (wireless microphone) maintains level 4 (8dBm), and the fixed slave node (speaker) is initially at level 3 (-6dBm). The overlap rate of channel f8 (5740MHz) between the fixed slave node in the audience area and the moving slave node on the stage was detected to be 28% (with a mid-level collision rate of 20%-40%). Adjustment: Fixed slave node power is reduced by 1 level to level 2 (-12dBm), and mobile slave node is maintained at level 4. After adjustment, interference is reduced by 42% ≥ 40%, and energy consumption is reduced by 26% ≥ 25%.

[0071] Step 4: Game Theory Collaboration The payoff function of the game model is "packet loss rate ≤ 0.8%", and the strategy space includes "switching to f7 / f8 / f11 (3 idle channels)" and "reducing power by 1-3 levels". The strategy iterates every 300ms and converges to Nash equilibrium in 1.3s. The master node prioritizes allocating channel f9 with an overlap rate of 13% < 15% to mobile slave nodes, resulting in a 32% increase in network capacity.

[0072] Step 5: Conflict Response Optimization The overlapping rate is detected using a sliding window algorithm with a detection delay of 44ms and an accuracy of ±1.5%. After the conflict adjustment, the interference reduction effect was verified at a master node interval of 100ms (from 28% to 12%, with a deviation of ≤5%). In cases of high conflict (such as a sudden private WiFi use in the audience area causing an overlap rate of 45%), the power of fixed slave nodes is reduced by 2 levels to level 1 (-32dBm), and the power of mobile slave nodes is compensated to level 5 (15dBm), with signal strength fluctuation of 2.8dB.

[0073] Step 6: Iterative Game Theory Optimization Scenario-based iteration parameters: mobile slave node iteration weight 1.1, fixed slave node 0.9, and the idle channel search range is expanded to 5 (f6-f10). The convergence time was reduced to 1.2s; Post-balanced check: If the packet loss rate of the mobile slave node is 0.6% ≤ 0.8%, it meets the standard; if it does not meet the standard (e.g., the packet loss rate of a wireless microphone is 1.2%), iterate once again and reduce it to 0.5%.

[0074] This embodiment of the system is deployed in the concert control console, stage equipment, and audience area equipment, and includes the following modules: Master-slave spectrum interaction module The master node uses an industrial-grade wireless controller (model: WC500). Among the 24 slave nodes: 15 mobile slave nodes use high-precision modules (model: WM300, supporting GPS positioning), and 9 fixed slave nodes use simplified modules (model: WM100, reducing data volume by 40% after protocol simplification). The module construction latency is 430ms≤450ms, and the slave nodes upload spectrum data (5.8GHz channel occupancy ±1.5%, power ±0.8dBm) every 150ms.

[0075] Multi-dimensional data acquisition module A multi-mode data acquisition unit (model: DC300) is used, which integrates GPS (mobile node position error 1.8m≤2m) and a spectrum analyzer (records 12 channel data every 5 minutes); service priority is configured through console software, and the data error is 0.9%≤1%.

[0076] LSTM Predictive Scheduling Module Deployed on the edge server of the control console (4 TOPS computing power), with 90,000 LSTM model training data, prediction latency of 82ms≤90ms; look-ahead scheduling instructions are issued through the 5.8GHz frequency band, with a switching response time of <200ms.

[0077] Game Theory Collaborative Allocation Module Built-in Nash equilibrium algorithm (implemented in C++, iteration period 300ms), mobile slave node iteration weight 1.1, idle channel search range 5; convergence time 1.3s≤1.5s, network capacity improved by 32%.

[0078] Graded power regulation module The power standard is fixed in the slave node firmware, and the collision detection unit uses a real-time spectrum analysis chip (model: SA200, detection delay 44ms≤45ms); power adjustment is achieved through PWM signal, with adjustment delay 88ms≤90ms, interference reduction of 42%, and energy consumption reduction of 26%.

[0079] Dual-band self-interference cancellation module It is equipped with a 2.4GHz + 5.8GHz dual-band scanning antenna (gain 8dBi), a scanning interval of 40ms, and a sensitivity of -94dBm ≤ -92dBm; the redistribution command issuance delay is 75ms ≤ 80ms, and the isolation between mobile slave nodes is 33dB ≥ 32dB.

[0080] Advantages of this embodiment: Large-scale node adaptation: 24 slave nodes work stably together, network capacity is increased by 32%, and 8 wireless microphones + 7 spotlights can work in parallel without signal congestion, meeting the "multi-device-high concurrency" requirements of concerts; Dynamic scene response: The position error of the moving slave node is 1.8m≤2m, the game iteration is 300ms / time, the conflict detection is 44ms / time, and it can adapt to the movement of the follow spot and the changes in the interference in the audience area in real time. The signal strength fluctuation is ≤2.8dB, and the wireless microphone voice is smooth. Interference and energy consumption balance: Interference is reduced by 42% during collisions, energy consumption of fixed slave nodes is reduced by 26%, and the self-interference removal rate of the 5.8GHz band is 99%, solving the problem of "signal interruption caused by complex outdoor electromagnetic environment". Operation and maintenance efficiency: The system deployment takes only 30 minutes, and the nodes are plug-and-play, adapting to the needs of "quick setup and instant performance" for concerts, reducing manual debugging time by 60%.

[0081] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for cooperative suppression of interference in heterogeneous networks based on dynamic spectrum sharing, characterized in that, Includes the following steps: Step 1: Construct a dynamic spectrum sharing mechanism for heterogeneous networks: Deploy network nodes in a master-slave architecture, including 1 master node and N slave nodes, where N > 0. The number of slave nodes is limited according to different application scenarios. Slave nodes upload spectrum usage information to the master node every 150ms. This information includes the occupancy rate of each frequency band and the real-time transmit power of the node. The sampling interval for the occupancy rate of each frequency band is 5 minutes, with an accuracy of ±1.5%. The real-time transmit power of the node ranges from -35dBm to 20dBm, with an accuracy of ±0.8dBm. Step 2, collect spectrum prediction input data: obtain historical spectrum data and real-time network status data from the past 24 hours; among which, the historical spectrum data includes frequency band occupancy curves and transmit power fluctuation values ​​recorded every 5 minutes; Step 3: Construct an LSTM spectrum demand prediction model: Based on the data trained in Step 2, predict the spectrum demand of each node within the next 10 minutes, including the frequency band overlap rate. Step 4, Forward spectrum adjustment: If it is predicted that the overlap rate of a certain frequency band is ≥28% at a certain time point within the next 5 minutes, the master node will issue a frequency band switching instruction to the conflicting node 3 minutes in advance of that time point and schedule it to an idle channel; In step one, the number of nodes is limited differently based on the application scenario, specifically including: Enclosed space: The number of slave nodes is limited to 8-16, of which core slave nodes, including the main camera and lighting control console, account for ≥50% of the total number of slave nodes. Their transmission power accuracy is improved to ±0.5dBm, and the frequency band occupancy sampling interval is shortened to 3 minutes. The proportion of auxiliary slave nodes is ≤50%, and low-cost communication modules can be used. Their frequency band occupancy sampling interval is allowed to be relaxed to 10 minutes. Open space: The number of slave nodes is limited to 16-32, of which mobile slave nodes, including mobile stage lights and wireless microphones, account for ≥60%, and their positional error requirement is increased to ≤2m; The hierarchical power regulation steps based on node classification are as follows: The first step is to define five levels of transmit power standards: Level 1 Energy Saving Level: -35dBm to -25dBm, Level 2 Basic Level: -24dBm to -15dBm, Level 3 Standard Level: -14dBm to -5dBm, Level 4 Enhanced Level: -4dBm to -10dBm, and Level 5 Highest Level: 11dBm to -20dBm. The second step is to calculate the degree of spectrum conflict: the frequency band overlap rate of adjacent nodes, including master nodes, slave nodes, and different types of slave nodes, is calculated as (overlapping frequency band width / total used frequency band width) × 100%. Conflict levels are then classified as: low conflict <20%, medium conflict 20%-40%, and high conflict >40%. The third step is to dynamically adjust the power: when there is a high conflict, the power of the auxiliary slave node and the fixed slave node is reduced by 1-2 levels, while the core slave node and the mobile slave node are maintained at level 3 or above; when there is a medium conflict, the power of the auxiliary slave node and the fixed slave node is reduced by 1 level; when there is a low conflict, the current level is maintained; after adjustment, the network interference is reduced by ≥40%, and the node energy consumption is reduced by ≥25%.

2. The heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing according to claim 1, characterized in that, The game-theoretic collaborative steps for node classification are as follows: The first step is to construct a game theory model: the node communication quality with a packet loss rate of ≤0.8% is used as the payoff function, and the strategy space includes switching frequency bands and reducing power; among them, there are ≥3 available idle channels for switching frequency bands, and the power reduction ranges from 1 to 3 levels; the nodes include master nodes and various types of slave nodes; The second step is strategy iteration and optimization: update the node strategy every 300ms until all nodes reach Nash equilibrium, with a convergence time of ≤1.5s; The third step is global allocation of master nodes: based on the balance results, the core slave nodes and mobile slave nodes are allocated low-collision channels with an overlap rate of <15% to ensure that the overall network capacity is increased by ≥30% and the continuous uninterrupted time is increased by ≥20%.

3. The heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing according to claim 1, characterized in that, It also includes a self-interference removal step for node classification: The first step, dual-band spectrum sensing: using sensing radio technology, scanning the 2.4GHz ISM band in the range of 2400-2483.5MHz and the 5.8GHz ISM band in the range of 5725-5825MHz, with a scanning interval of ≤40ms, a scanning bandwidth of ≥22MHz, and an interference signal identification sensitivity of ≤-92dBm; The second step is self-interference identification: If the signal strength difference of co-channel interference between the core slave nodes or between the core slave node and the mobile slave node is ≤5dB, the self-interference removal process is triggered. The third step is frequency band reallocation: the master node allocates a new frequency band to the interfering node within 80ms. The new frequency band is ≥25MHz away from the original frequency band. Priority is given to ensuring that the new frequency band occupancy rate of the core slave node and the mobile slave node is ≤3%. After reallocation, the signal isolation between nodes is ≥32dB.

4. The heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing according to claim 1, characterized in that, The power adjustment steps also include collision response optimization based on the accuracy of the spectrum information: The first step, rapid collision detection: A sliding window algorithm is used to calculate the frequency band overlap rate of adjacent nodes in real time, with a detection delay of ≤45ms and a detection accuracy maintained at ±1.5%; The second step is to adjust the response in stages: In the case of low conflict, the fixed slave node maintains its current power and only updates the spectrum information upload frequency to 200ms; in the case of medium conflict, the auxiliary slave node power is reduced by 1 level, and the master node verifies the interference reduction effect at 100ms intervals; in the case of high conflict, the fixed slave node power is reduced by 2 levels, and the core slave node power is compensated to level 4 to ensure that the main service signal strength fluctuation is ≤3dB.

5. The heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing according to claim 2, characterized in that, Strategy iteration also includes scenario convergence optimization: The first step is to optimize the iteration parameters based on the scenario: In a closed space scenario, the strategy iteration weight of the core slave node is increased to 1.2, the weight of the auxiliary slave node is 0.8, and the convergence time is compressed to 1.2s; In open space scenarios, the policy iteration weight of mobile slave nodes is increased to 1.1, the weight of fixed slave nodes is 0.9, and the search range of idle channels is expanded to 5. The second step is to verify the balance result: After each iteration converges, the master node verifies that the packet loss rate of the core slave node and the mobile slave node is ≤0.8%. If it does not meet the requirement, the iteration is repeated until the standard is met.

6. The heterogeneous network interference collaborative suppression method based on dynamic spectrum sharing according to claim 3, characterized in that, Frequency band reallocation also includes enhanced isolation of service priorities: The first step is to define the isolation level requirements: the signal isolation between the core slave nodes is ≥35dB, the isolation between the core slave node and the auxiliary slave node is ≥32dB, and the isolation between the mobile slave node and the fixed slave node is ≥30dB. The second step is dynamic frequency band selection: Prioritize selecting channels from the idle channels with an interval of ≥30MHz from the interference frequency band and assign them to the core slave node, channels with an interval of ≥25MHz and assign them to the mobile slave node, and channels with an interval of ≥20MHz and assign them to the auxiliary / fixed slave node.

7. A heterogeneous network interference cooperative suppression system based on dynamic spectrum sharing, characterized in that, The heterogeneous network interference cooperative suppression method as described in claim 1 includes: Master-Slave Spectrum Interaction Module: Configured with 1 master node and N slave nodes, where N≥0. The number of slave nodes is limited according to scenario classification. When N=0, spectrum data is obtained through the master node. Every 150ms, the slave nodes transmit spectrum usage information to the master node: frequency band occupancy rate ±1.5%, transmit power -35dBm to 20dBm±0.8dBm; module construction delay ≤450ms. Multi-dimensional data acquisition module: Collects historical spectrum data and real-time network status data over the past 24 hours, with a data error of ≤1%; LSTM prediction scheduling module: Based on a model trained with ≥80,000 data points, it outputs a spectrum demand prediction for the next 10 minutes with a prediction delay of ≤90ms; when the predicted frequency band overlap rate is ≥28%, it issues a handover instruction to an idle channel with an occupancy rate of ≤4% 3 minutes in advance. Game-theoretic collaborative allocation module: Built-in Nash equilibrium solution algorithm, iterates node strategy once every 300ms, convergence time ≤1.5s; based on the equilibrium result, it prioritizes the allocation of low-collision channels with an overlap rate of <15% to the core slave node and mobile slave node, improving network capacity by ≥30%.

8. The heterogeneous network interference cooperative suppression system based on dynamic spectrum sharing according to claim 7, characterized in that, Also includes: Hierarchical power control module: Stores 5 levels of transmit power standards, receives the collision level signal from the collision detection module, and adjusts the slave node power within 90ms to reduce inter-network interference by ≥40% and node power consumption by ≥25%; Dual-band self-interference removal module: Configured with 2.4GHz+5.8GHz scanning sub-modules, generating frequency band reallocation instructions within 80ms to ensure that the isolation between core slave nodes is ≥35dB and the isolation between mobile slave nodes is ≥32dB.

Images