Signal interference management method and device, management equipment, program product and medium

Abstract

The embodiment of the invention provides a signal interference management method and device, management equipment, a program product and a medium, and is applied to the technical field of network communication. The method comprises the following steps: determining a data acquisition strategy corresponding to a target scene according to the type of the target scene; signal data of wireless network signals in the target scene are collected by adopting the data collection strategy, and the signal data comprise at least one of signal strength information, signal-to-noise ratio information and spectrum occupation information; inputting the signal data to a target network model to obtain a prediction result output by the target network model, the target network model being a Markov model trained based on historical signal data in the target scene; and adjusting network parameters of network equipment in the target scene according to the prediction result. By adopting the method, the problem that an existing interference management scheme is difficult to adapt to a rapidly changing network environment and is easy to misjudge can be solved.

Description

Technical Field

[0001] This application relates to the field of network communication technology, and in particular to a signal interference management method, apparatus, management equipment, program product and medium. Background Technology

[0002] Current wireless network signal interference management (detection and mitigation) mainly relies on two technical approaches:

[0003] (1) Threshold detection system: By setting thresholds for signal strength or signal-to-noise ratio (SNR), the system determines whether interference exists. For example, the system periodically monitors the strength and SNR of the Wi-Fi signal and compares them with preset thresholds. If the signal strength or SNR is lower than the corresponding threshold, interference is considered to exist.

[0004] (2) Rule-based mitigation system: Once interference is detected, network parameters are adjusted according to preset rules to mitigate the interference. Generally, interference is dealt with through manual intervention, such as by users or network administrators adjusting network settings (including but not limited to manually changing channels, adjusting transmission power, or rearranging the location of network equipment).

[0005] However, the aforementioned interference management solutions are difficult to adapt to rapidly changing network environments, and the manpower and time costs required for manual adjustment of network settings are high, resulting in low efficiency and a high risk of misjudgment. Summary of the Invention

[0006] The purpose of this application is to provide a signal interference management method, device, management equipment, program product and medium to solve the problem that existing interference management solutions are prone to misjudgment due to their inability to adapt to rapidly changing network environments.

[0007] One embodiment of this application provides a signal interference management method, the method comprising:

[0008] Based on the type of the target scenario, determine the data acquisition strategy corresponding to the target scenario;

[0009] Using the data acquisition strategy described above, signal data of wireless network signals in the target scenario are acquired, and the signal data includes at least one of signal strength information, signal-to-noise ratio information, and spectrum occupancy information;

[0010] The signal data is input into the target network model to obtain the prediction result output by the target network model, wherein the target network model is a Markov model trained based on historical signal data in the target scene;

[0011] Based on the prediction results, adjust the network parameters of the network devices in the target scenario.

[0012] Optionally, determining the data collection strategy corresponding to the target scenario based on the type of the target scenario includes at least one of the following:

[0013] When the target scene is a public place, the data collection strategy corresponding to the target scene is determined as follows: the collection frequency of the signal data in the target scene is determined based on the pedestrian flow in the target scene.

[0014] When the target scenario is a home environment, the data acquisition strategy corresponding to the target scenario is determined as follows: the signal data is acquired at a first preset frequency;

[0015] When the target scenario is a public office space, the data acquisition strategy for the target scenario is determined as follows: the signal data is acquired at a second preset frequency, where the second preset frequency is greater than the first preset frequency.

[0016] When the target scenario is a personal office space, the data collection strategy for the target scenario is determined as follows: based on the historical network usage information of users in the target scenario, the collection frequency of the signal data in the target scenario is determined, and the historical network usage information reflects the activity level of users in the target scenario using the wireless network at different times.

[0017] Optionally, the prediction results include: signal strength prediction values ​​and / or interference prediction values ​​for each frequency band used to transmit the wireless network signal, wherein the interference prediction values ​​characterize the degree of interference the frequency band will experience in the future, and the network parameters include: the frequency band used by the network device and / or the transmit power of the network device;

[0018] Adjusting the network parameters of the network devices in the target scenario based on the prediction results includes at least one of the following:

[0019] If the first interference prediction value of the frequency band used by the network device exceeds the threshold, and there is a frequency band with an interference prediction value less than the first interference prediction value, the frequency band used by the network device will be switched to the first frequency band.

[0020] If the interference prediction value of the frequency band used by the network device indicates increased interference, the frequency band used by the network device will be switched to the first frequency band.

[0021] If the predicted interference value for the frequency band used by the network device indicates increased interference, the transmit power of the network device shall be reduced.

[0022] The first frequency band is the one with the smallest predicted interference value among all the frequency bands.

[0023] Optionally, the method further includes:

[0024] By abstracting the wireless network environment in the target scenario into a Markov Decision Process (MDP), an initial model is constructed. The state of the MDP represents the network state of the network environment, the action of the MDP represents the adjustment of the network parameters, and the reward function of the MDP guides the model to select actions to suppress interference and / or optimize network performance through positive incentives.

[0025] Based on the historical signal data, the initial model is trained to obtain the target network model.

[0026] Optionally, the method further includes:

[0027] During the training of the initial model, a dual deep Q-Network (DQN) is used to evaluate the value of each action and obtain the value evaluation results.

[0028] Based on the value assessment results, actions are selected during training.

[0029] Optionally, the method further includes:

[0030] The initial model is trained based on the Prioritized Experience Replay (PER) mechanism.

[0031] One embodiment of this application also provides a signal interference management device, the device comprising:

[0032] The first processing module is used to determine the data acquisition strategy corresponding to the target scenario based on the type of the target scenario.

[0033] The data acquisition module is used to acquire signal data of wireless network signals in the target scenario using the data acquisition strategy, wherein the signal data includes at least one of signal strength information, signal-to-noise ratio information, and spectrum occupancy information;

[0034] The model prediction module is used to input the signal data into the target network model and obtain the prediction result output by the target network model, wherein the target network model is a Markov model trained based on historical signal data in the target scene;

[0035] The parameter adjustment module is used to adjust the network parameters of the network devices in the target scenario based on the prediction results.

[0036] Optionally, the first processing module includes:

[0037] The first processing unit is configured to determine, when the target scene is a public place, the data collection strategy corresponding to the target scene is: to determine the collection frequency of the signal data in the target scene based on the pedestrian flow in the target scene;

[0038] The second processing unit is configured to determine, when the target scene is a home environment, the data acquisition strategy corresponding to the target scene is: to acquire the signal data at a first preset frequency;

[0039] The third processing unit is configured to determine, when the target scene is a public office space, the data acquisition strategy corresponding to the target scene is: to acquire the signal data at a second preset frequency, wherein the second preset frequency is greater than the first preset frequency;

[0040] The fourth processing unit is configured to determine the data collection strategy corresponding to the target scenario when the target scenario is a personal office space: based on the historical network usage information of the user in the target scenario, determine the collection frequency of the signal data in the target scenario, wherein the historical network usage information reflects the activity level of the user in the target scenario using the wireless network at different times.

[0041] Optionally, the prediction results include: signal strength prediction values ​​and / or interference prediction values ​​for each frequency band used to transmit the wireless network signal, wherein the interference prediction values ​​characterize the degree of interference the frequency band will experience in the future, and the network parameters include: the frequency band used by the network device and / or the transmit power of the network device;

[0042] The parameter adjustment module includes:

[0043] The first switching unit is configured to switch the frequency band used by the network device to the first frequency band when the first interference prediction value of the frequency band used by the network device exceeds a threshold and there is a frequency band with an interference prediction value less than the first interference prediction value.

[0044] The second switching unit is used to switch the frequency band used by the network device to the first frequency band when the interference prediction value of the frequency band used by the network device indicates increased interference.

[0045] The first adjustment unit is used to reduce the transmit power of the network device when the interference prediction value of the frequency band used by the network device indicates increased interference.

[0046] The first frequency band is the one with the smallest predicted interference value among all the frequency bands.

[0047] Optionally, the device further includes:

[0048] The model building module is used to build an initial model by abstracting the wireless network environment in the target scenario into a Markov Decision Process (MDP). The state of the MDP represents the network state of the network environment, the action of the MDP represents the adjustment of the network parameters, and the reward function of the MDP guides the model to select actions to suppress interference and / or optimize network performance through positive incentives.

[0049] The model training module is used to train the initial model based on the historical signal data to obtain the target network model.

[0050] Optionally, the method further includes:

[0051] The first training module is used to perform value evaluation on each action and obtain value evaluation results during the training of the initial model by employing a dual deep Q-network (DQN).

[0052] Based on the value assessment results, actions are selected during training.

[0053] Optionally, the method further includes:

[0054] The second training module is used to train the initial model based on the priority experience replay (PER) mechanism.

[0055] One embodiment of this application also provides a management device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the program is executed by the processor, it implements the signal interference management method as described in any of the preceding claims.

[0056] One embodiment of this application also provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the signal interference management method as described in any of the preceding claims.

[0057] One embodiment of this application also provides a readable storage medium, wherein a program is stored on the readable storage medium, and when the program is executed by a processor, it implements the steps of the signal interference management method as described in any of the preceding claims.

[0058] At least one of the above-mentioned technical solutions of this application has the following beneficial effects:

[0059] In the signal interference management method of this application embodiment, differentiated data acquisition strategies can be formulated for different target scenario types. Based on this, differentiated signal data acquisition can be achieved. Thus, the target network model trained based on the acquired historical signal data is more suitable for the target scenario, making the prediction results for the target scenario more accurate. Furthermore, based on the prediction results, more precise network parameter adjustments can be made for network devices in the target scenario, avoiding misjudgments. Moreover, since this scheme obtains prediction results based on real-time acquired signal data, it can adapt to rapidly changing network environments. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating the signal interference management method according to one embodiment of this application;

[0061] Figure 2 This is a flowchart illustrating a signal interference management method according to another embodiment of this application;

[0062] Figure 3 This is a flowchart illustrating a signal interference management method according to another embodiment of this application;

[0063] Figure 4 This is a schematic diagram of the structure of a signal interference management device according to one embodiment of this application. Detailed Implementation

[0064] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and are not used to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, and the number of objects is not limited; for example, the first object can be one or more.

[0065] In various embodiments of the present invention, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0066] In the embodiments of this application, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0067] It should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0068] In the embodiments of this application, the term "multiple" refers to two or more, and other quantifiers are similar.

[0069] In addition, the terms "system" and "network" are often used interchangeably in this article.

[0070] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0071] Existing signal interference management solutions have the following shortcomings:

[0072] (1) Since interference sources in the network environment may change with time, location and other factors, statically configured threshold settings cannot flexibly cope with rapidly changing network environments. Therefore, when the network environment changes, existing solutions may not be able to detect new interference sources in a timely manner, nor can they make corresponding adjustments quickly.

[0073] (2) Manual intervention is not only labor-intensive and time-consuming, but human judgment is also prone to bias and inefficient in complex network environments. Although rule-based methods can automate the process to some extent, the rules are fixed and cannot cope with all situations, especially the ever-changing network environment.

[0074] (3) Lacking the ability to learn from historical data and optimize itself, the system cannot improve its detection and mitigation strategies based on past experience, and it is difficult to gradually improve the efficiency and accuracy of detecting and mitigating interference over time.

[0075] This application addresses the problem that existing interference management solutions are prone to misjudgment due to their inability to adapt to rapidly changing network environments, and provides a signal interference management method, apparatus, management equipment, program product, and medium.

[0076] like Figure 1 As shown, one embodiment of the signal interference management method of this application includes:

[0077] Step 101: Determine the data acquisition strategy corresponding to the target scenario based on the type of the target scenario.

[0078] Here, the type of target scenario can be automatically identified by the system (referring to the system implementing the signal interference management method described in this application) or selected by user input. The system can formulate data acquisition strategies for different scenarios. For example, the living room and bedroom may experience signal attenuation due to wall obstruction, while the study usually requires higher network stability, and the kitchen may experience interference due to the large number of appliances. Therefore, it is necessary to determine the data acquisition strategy (such as configuring different signal acquisition frequencies) according to the scenario.

[0079] Step 102: Using the data acquisition strategy, acquire signal data of the wireless network signal in the target scenario. The signal data includes at least one of signal strength information, signal-to-noise ratio information, and spectrum occupancy information.

[0080] Specifically, signal data can be collected through a signal acquisition device, which can monitor Wi-Fi signal strength, signal-to-noise ratio (SNR), and spectrum occupancy in real time.

[0081] like Figure 2 As shown, the collected signal data is transmitted via a Wi-Fi network (such as a home LAN) to a dedicated server: a central processing unit (such as a smart home control center). This central processing unit can be a dedicated computing device or part of an existing smart home control platform, equipped with data processing software for subsequent data preprocessing and model training.

[0082] It should be noted that after acquiring the raw signal data, this raw signal data can be preprocessed to obtain signal data for input into the target network model in the next step. For example, the preprocessing process may include: first, using a bandpass filter to remove high-frequency noise from the signal; then, using a smoothing filter to reduce random noise from the signal; and finally, using standardization to convert all signal data into a standard normal distribution with a mean of 0 and a variance of 1.

[0083] Step 103: Input the signal data into the target network model to obtain the prediction result output by the target network model, wherein the target network model is a Markov model trained based on historical signal data in the target scene.

[0084] Step 104: Adjust the network parameters of the network devices in the target scenario based on the prediction results.

[0085] Through steps 103 to 104, the system can analyze signal data in real time and predict interference, automatically adjusting network parameters to optimize performance.

[0086] In the above steps, a miniature signal acquisition device can be deployed to monitor signal data such as Wi-Fi signal strength, signal-to-noise ratio, and spectrum occupancy in the target scene in real time. It can automatically identify abnormal changes in signal strength and automatically select the best mitigation strategy to adjust network parameters (such as adjusting transmission power, switching channels, etc.) based on real-time signal data by using a trained target network model. The whole process does not require manual intervention from users or network administrators, thus improving the intelligence level of the system.

[0087] In this embodiment, differentiated data acquisition strategies can be formulated for different target scenario types. Based on this, differentiated signal data acquisition can be achieved. Thus, the target network model trained based on the acquired historical signal data is more suitable for the target scenario, making the prediction results for the target scenario more accurate. Furthermore, based on the prediction results, more precise network parameter adjustments can be made for network devices in the target scenario, avoiding misjudgments. Moreover, since this scheme obtains prediction results based on real-time acquired signal data, it can adapt to rapidly changing network environments.

[0088] In some embodiments, determining the data collection strategy corresponding to the target scene based on the type of the target scene includes at least one of the following:

[0089] (1) When the type of the target scene is a public place, the data collection strategy corresponding to the target scene is determined as follows: the collection frequency of the signal data in the target scene is determined according to the flow of people in the target scene.

[0090] It should be noted that in public places, the signal collector can dynamically adjust the sampling frequency based on changes in pedestrian traffic. For example, the number of samplings can be increased during peak hours to better reflect network conditions.

[0091] Specifically, the signal collector can dynamically adjust its sampling frequency based on changes in pedestrian traffic in several ways, including: using devices such as smart cameras, thermal imagers, or Bluetooth beacons to monitor pedestrian traffic. For example, it can count the number of people entering and leaving in real time and record peak periods of pedestrian traffic; it can also link pedestrian traffic data with data from Wi-Fi signal collectors. For instance, when a thermal imager shows that the pedestrian traffic in a certain area exceeds 100 people, the system will automatically adjust the signal sampling frequency for that area. Furthermore, it can set pedestrian traffic thresholds, such as increasing the sampling frequency of the signal collector from once every 5 minutes to once every 2 minutes when there are more than 5 people per square meter.

[0092] Table 1

[0093]

[0094]

[0095] Table 1 above reflects the relationship between pedestrian traffic and data collection frequency at a public place during different time periods. The higher the pedestrian traffic, the higher the data collection frequency. Thus, the system dynamically adjusts the data collection frequency based on real-time pedestrian traffic data to ensure that signal strength data can be acquired more frequently during peak traffic periods, accurately reflecting network conditions.

[0096] (2) When the type of the target scene is a home environment, the data acquisition strategy corresponding to the target scene is determined to be: to acquire the signal data at a first preset frequency.

[0097] It should be noted that the data acquisition strategy in this scenario can also include the deployment of data acquisition points. For example, miniature signal acquisition devices can be deployed in different living areas (such as the living room, bedroom, study, and kitchen). Signal acquisition can be set periodically, for example, once per minute, or triggered when abnormal signal strength fluctuations are detected. This allows for intelligent adjustment of the deployment and acquisition frequency of signal acquisition devices within the home environment, based on the layout of the living space and the presence of obstacles, to ensure uniformity and quality of signal coverage.

[0098] (3) When the type of the target scene is a public office space, the data acquisition strategy corresponding to the target scene is determined as follows: the signal data is acquired at a second preset frequency, where the second preset frequency is greater than the first preset frequency.

[0099] Here, public office spaces can be open office areas, meeting rooms, etc. For example, during meetings, a large number of devices may simultaneously access the network, requiring enhanced signal monitoring and increased signal acquisition frequency. Therefore, in this scenario, the signal acquisition device can frequently collect data to monitor signal fluctuations.

[0100] It should be noted that in public office spaces, the data collection strategy can also be dynamically adjusted based on changes in the number of people on weekdays and non-weekdays (such as adjusting the second preset frequency) to adapt to changes in network usage patterns.

[0101] (4) When the type of the target scenario is a personal office space, the data collection strategy corresponding to the target scenario is determined as follows: based on the historical network usage information of the user in the target scenario, the collection frequency of the signal data in the target scenario is determined, and the historical network usage information reflects the activity of the user in the target scenario in using the wireless network at different times.

[0102] Here, in personal workspaces such as private offices, the data collection frequency can be adjusted based on the individual usage habits of the users in those offices. For example, if a user uses the network more frequently between 10:00 and 11:30 AM, the signal data collection frequency for that time period can be adjusted.

[0103] In this embodiment, different signal acquisition and processing strategies are formulated for different scenarios to adapt to the specific needs of various environments. This scenario-aware method improves the efficiency of signal acquisition and network performance, ensuring that optimal signal quality and interference mitigation can be achieved in different scenarios.

[0104] In some embodiments, the prediction results include: predicted signal strength values ​​and / or predicted interference values ​​for each frequency band used to transmit the wireless network signal, the predicted interference values ​​characterizing the degree of interference the frequency band will experience in the future, and the network parameters include: the frequency band used by the network device and / or the transmit power of the network device.

[0105] It should be noted that, in the embodiments of this application, a trained deep reinforcement learning model (i.e., a target network model) can be used to analyze real-time signal data, predict potential future interference, and formulate countermeasures in advance. Specifically, based on the model's prediction results, the system automatically adjusts network parameters to mitigate interference.

[0106] like Figure 3 As shown, the system can collect the RSSI (Signal Strength Indicator) value received by each access point on a second-by-second basis, monitor SNR (Signal NR), and determine which channels are being used and their signal strength by scanning the surrounding Wi-Fi network. The deep reinforcement learning model (i.e., the target network model) receives RSSI and SNR data from the current and past periods, as well as current channel usage, as input. It then uses this data to predict the interference situation of each channel within a future period (e.g., the next 5 minutes), obtaining the prediction results. These prediction results include the interference probability and expected signal strength for each channel. Based on the model's prediction results, the system selects the optimal adjustment strategy.

[0107] In a specific example, adjusting the network parameters of the network devices in the target scenario based on the prediction result includes at least one of the following:

[0108] (1) If the first interference prediction value of the frequency band used by the network device exceeds the threshold and there is a frequency band with an interference prediction value less than the first interference prediction value, the frequency band used by the network device shall be switched to the first frequency band; wherein, the first frequency band is the frequency band with the smallest interference prediction value among all the frequency bands.

[0109] It's important to note that if there's only one Wi-Fi signal source in the target scenario, the system primarily focuses on the stability of that signal source, such as changes in signal strength and SNR. Therefore, if the signal strength suddenly drops or the SNR decreases, the system will take immediate action. When multiple signal sources exist, in addition to monitoring the primary signal source, the system also needs to monitor interference from other surrounding signal sources. For example, the system can scan for nearby Wi-Fi networks, identify interference sources, and adjust the network parameters of the primary signal source (i.e., the network device) accordingly.

[0110] For example, if the model predicts severe interference on the current channel and there is at least one channel with less interference, the system will automatically switch to that channel. During the switchover process, the system will ensure a smooth transition for all connected devices, avoiding disconnections.

[0111] In one specific embodiment, the system periodically scans surrounding Wi-Fi networks to identify all available channels (i.e., the aforementioned frequency bands) and their signal strengths. When the signal-to-noise ratio (SNR) of the primary signal source is detected to be below a predetermined threshold (e.g., 15 dB), the system initiates an interference detection procedure. Specifically, by comparing the signal strengths of various signal sources, the system identifies the source of the strongest interference. For example, if the signal strength of channel 6 is above -60 dBm, while the signal strength of channel 11 is below -70 dBm, then the interference on channel 6 is determined to be stronger. Based on the location and intensity of the interference source, the system automatically adjusts the parameters of the primary signal source, such as switching to a channel with less interference or adjusting the transmission power.

[0112] Table 2

[0113] Signal Source Channel Signal strength (dBm) Interference intensity (dBm) Adjusting the action Channel 6 -60 -55 Switch channels Channel 11 -62 -65 No action

[0114] Table 2 above illustrates the relationship between channel signal strength, interference intensity, and adjustment actions in an optional example. Assume the primary signal source is currently operating on channel 6, but channel 6 is experiencing severe interference from a nearby Wi-Fi network. The system detects that the SNR of channel 6 has dropped to 10dB, and automatically switches to channel 11 (which experiences relatively less interference) to achieve the initial adjustment to mitigate the interference.

[0115] (2) If the interference prediction value of the frequency band used by the network device indicates that the interference is aggravated, the frequency band used by the network device shall be switched to the first frequency band.

[0116] (3) If the interference prediction value of the frequency band used by the network device indicates that the interference is aggravated, reduce the transmission power of the network device.

[0117] For example, if the target network model predicts that interference on the current channel will worsen, but interference on other channels is less severe, the system may choose to slightly reduce the transmit power of the Wi-Fi router to reduce interference to other devices while maintaining sufficient signal coverage.

[0118] In this embodiment, the system can effectively reduce interference by intelligently adjusting network parameters (such as transmission power, channel switching, etc.) based on real-time signal data and prediction results, ensuring that the network can maintain stable connection and high-quality service under any circumstances, and improving the overall performance of the network.

[0119] In addition, additional signal points can be added when the signal strength is weak. For example, when the system detects that the signal strength at a certain location is lower than a preset threshold (e.g., below -70dBm), the following measures can be taken:

[0120] (1) Automatic addition of collection points: In areas with low signal strength, the system can automatically activate backup collectors or add new collection points by predicting the optimal location through algorithms. For example, if the signal strength in the bedroom is weak, the system can deploy additional collectors in the bedroom to enhance the monitoring of the signal strength in that area.

[0121] (2) Manually adding collection points: In some cases, the system will prompt the user or network administrator to manually add collection points in areas with insufficient signal strength. For example, in a corner with particularly low signal strength, a collector can be added manually.

[0122] (3) Adjust the location of existing acquisition points: The system analyzes the data of existing acquisition points and suggests adjusting their locations to improve signal strength. For example, move the acquisition device closer to the signal source or avoid physical obstacles to improve signal strength.

[0123] In the above embodiments, different signal acquisition and processing strategies are adopted for different scenarios, enabling the system to adapt to signal acquisition and processing in various environmental scenarios. The system can intelligently adjust the deployment and acquisition frequency of the signal acquisition device according to the characteristics of the scenario to ensure the uniformity and quality of signal coverage.

[0124] In some embodiments, when the system detects a sudden drop in signal strength exceeding a preset threshold (e.g., -70 dBm) or a decrease in signal-to-noise ratio (SNR) to a specific level (e.g., below 15 dB), it is considered to have potential interference. At this point, the system immediately activates a deep reinforcement learning model (i.e., the target network model) to make predictions. For example, the system collects the RSSI signal strength indication value received by each access point on a second-by-second basis, monitors SNR, and determines which channels are being used and their signal strengths by scanning the surrounding Wi-Fi network. The target network model receives RSSI and SNR data from the current and past periods, as well as current channel usage, as input, and uses this data to predict interference for each channel in the future period (e.g., the next 5 minutes). Subsequently, based on the policy selection results, the system automatically adjusts network parameters, such as the Wi-Fi router configuration, including transmit power and channel settings. Furthermore, the system can dynamically adjust antenna orientation or add signal amplification devices to further optimize signal coverage.

[0125] In some embodiments, the method further includes: abstracting the wireless network environment in the target scenario into a Markov Decision Process (MDP), constructing an initial model, wherein the state of the MDP represents the network state of the network environment, the action of the MDP represents the adjustment of the network parameters, and the reward function of the MDP guides the model to select actions to suppress interference and / or optimize network performance through positive incentives; and training the initial model based on the historical signal data to obtain the target network model.

[0126] Here, the wireless network environment is modeled as a Markov decision process (MDP), where the state represents the current network state (such as signal strength distribution, device operating frequency, etc.), the action represents the adjustment of network parameters (such as power adjustment, channel switching, etc.), and a reward function is defined to encourage the model to select actions that can reduce interference and improve network performance.

[0127] After building the initial model, the model is trained using historical signal data. By simulating network conditions under different environments, the model learns how to make optimal decisions in various situations, and finally obtains the trained target network model.

[0128] It's important to note that after obtaining the target network model, it can be continuously trained to improve its performance. For example, the system can continuously collect new signal data and feed it back to the target network model for retraining, thereby constantly improving the model's accuracy and adaptability. In this way, even when the network environment changes significantly, the system can quickly adapt to new situations through self-learning. The system can not only adjust according to the current signal conditions but also predict potential future interference and take preventative measures. This proactive approach allows the system to maintain good performance in various environments.

[0129] It should be noted that, in the embodiments of this application, the computational requirements can be reduced by simplifying the model (such as reducing the number of layers and the number of nodes in each layer).

[0130] Specifically, the process of reducing the number of layers includes:

[0131] First, the contribution of each layer is evaluated by training the model and testing the impact of each layer on the final output. For example, sensitivity analysis can be used to identify layers that have little impact on the final result. Next, redundant layers can be removed. For example, if the original model has three fully connected layers (FC layers) namely FC1, FC2, and FC3, and the evaluation shows that the middle layer FC2 has little contribution to the final output, it can be removed. In this way, the simplified model will only have two layers, FC1 and FC3. Finally, after removing the layers, the model needs to be fine-tuned by retraining to ensure that the remaining layers can make up for the effect of the removed layers.

[0132] The process of reducing the number of nodes in each layer includes:

[0133] First, by analyzing the absolute value of the weights of each node in each layer, the importance of the nodes is analyzed, that is, which nodes have a smaller impact on the output. Then, nodes with smaller absolute weight values ​​and their connections are removed. For example, if the original layer has 256 nodes, pruning can reduce it to 128 nodes. Here, L1 regularization can be used to make some weights become zero, thus achieving pruning. After pruning, retraining or fine-tuning techniques are needed to fine-tune and optimize the remaining network to ensure that model performance is not affected.

[0134] In one specific embodiment, the original model structure includes an input layer, intermediate layers (FC1, FC2, and FC3), and an output layer. The input layer receives data such as signal strength and signal-to-noise ratio from a signal acquisition unit; FC1, FC2, and FC3 each contain 256 nodes; the output layer outputs the value of network parameter adjustments. The simplified model structure reduces the number of layers (removing the intermediate layer FC2) and nodes (128 nodes per layer) compared to the original model structure. Retraining the simplified model ensures its performance is comparable to the original model. Thus, by simplifying the model, its complexity is reduced while retaining its core functionality, making the model more lightweight.

[0135] In one specific embodiment, four miniature signal collectors are deployed in a home environment, located in the living room, bedroom, study, and kitchen, respectively. These collectors send signal strength, signal-to-noise ratio, and other data to the central processing unit every 5 minutes. The wireless network environment in this home is then modeled as a Markov decision process, where state St, action At, and reward R are... t+1 This constitutes a complete MDP framework. The specific modeling process is as follows:

[0136] (1) State definition: State St includes data such as signal strength, signal-to-noise ratio, and spectrum occupancy measured by all signal acquisition devices. For example, state St can be represented as the set of signal strengths of four signal acquisition devices, i.e., St = (s1, s2, s3, s4), where si represents the signal strength of the i-th acquisition device.

[0137] (2) Action definition: Action At includes adjusting the transmission power of the Wi-Fi router, switching channels, etc. For example, action At can be represented as a combination of adjusting the transmission power and switching channels, i.e. At = (a1, a2), where a1 represents adjusting the transmission power and a2 represents switching channels.

[0138] (3) Definition of reward function: Reward function R t+1 Encourage efforts to reduce interference and improve network performance. Positive rewards are given for increased signal strength or improved signal-to-noise ratio; conversely, negative rewards are given for decreased signal strength or improved signal-to-noise ratio. The reward function can be expressed as:

[0139]

[0140] In some embodiments, the method further includes: during the training of the initial model, using a dual deep Q-network (DQN) to evaluate the value of each action and obtain the value evaluation result; and selecting actions during training based on the value evaluation result.

[0141] Here, during model training, a deep reinforcement learning model, Dual DQN, can be used to evaluate the value of each action in the current state and select the optimal action. For example, in the initial state, the system selects action a1 to increase the transmission power. After the signal strength is updated, the system obtains a positive reward R according to the reward function. t+1 = +1, and store this experience in the experience replay pool. Wherein, the state transition probability P(S) t+1 |St,At) describes how taking action At in the current state St leads to the next state S. t+1 The probability is calculated. A Deep Q-Network (DQN) is used for training, with the Q-value updated using a dual DQN update rule. Priority experience replay techniques are employed to improve the model's learning efficiency and accuracy.

[0142] Here, assuming the system's initial signal strengths are S1 = 50, S2 = 45, S3 = 60, and S4 = 55, to train a system capable of intelligently detecting and mitigating wireless network signal interference, two neural network evaluation networks can be initialized first. Qeval evaluates the value of each action in the current state, and Qtarget evaluates the value of future states. Next, the system begins real-time monitoring of signal strength, signal-to-noise ratio, and other data, and selects an action At based on the current state St. For example, in an alternative example, the system can choose to increase the transmit power a1 or switch channels a2.

[0143] Suppose the system chooses action a1 to increase the transmission power, and the signal strength is updated as follows: s1′=55, s2′=50, s3′=65, s4′=60. At this point, the signal strength has improved, and the system receives a positive reward R. t+1 =+1, the system will record this experience as: (St=(50,45,60,55),At=(1,0),R=+1,S t+1 = (55, 50, 65, 60) is stored in the experience replay pool.

[0144] In some embodiments, the method further includes training the initial model based on a priority experience replay (PER) mechanism.

[0145] For example, the system samples 100 experiences from the experience replay pool according to priority to update the Q-value. The Q-value is updated using a dual DQN update rule to improve the model's learning efficiency and accuracy. The specific update process can be represented as follows:

[0146] Q′(St,At)=R t+1 +γQtarget(S t+1 ,argamaxQeval(S t+1 ,a))

[0147] Assume R t+1 =+1, γ=0.99, Qtarget(S t+1 ,argmaxaQeval(S t+1 If a)) = 0.8, then the updated Q value is:

[0148] Q'(St,At)=1+0.99×0.8=1.792

[0149] In addition, the system updates the target network every certain number of steps (e.g., 1000 steps) by assigning the weights of the evaluation network Q eval to the target network Q target, in order to further improve the accuracy and adaptability of the model.

[0150] It should be noted that the system continuously collects new signal data and feeds it back to the target network model for retraining, thereby continuously improving the accuracy and adaptability of the target network model. For example, every preset time interval (e.g., weekly), the system will import the latest signal data into the target network model for retraining.

[0151] In the above embodiments, the target network model can dynamically adjust its parameters and structure based on real-time data to adapt to the ever-changing network environment. The dual DQN structure enhances the model's stability and learning efficiency in complex environments. This structure allows the system to more accurately evaluate the value of different actions and select the optimal strategy. Combined with priority experience replay technology, the system can sample based on the importance of experience, further improving learning efficiency and decision-making accuracy, enabling the system to respond more quickly and effectively to wireless network signal interference. This dynamic adjustment capability allows the system to not only perform well during initial deployment but also continuously improve, providing more stable and efficient wireless network services over time.

[0152] It should be noted that, in some optional embodiments of this application, to verify the accuracy of the Markov model, state transition records for each AP (Access Point) can be kept for several consecutive days, including the number of channel handovers and changes in user connection quality. The consistency between the observed state transitions and the Markov model predictions is verified through a chi-square test. Real-time data, such as throughput, latency, and error rate, can also be collected through user feedback and a network performance monitoring system. A performance baseline is set; when the actual performance falls below the baseline, the system automatically adjusts network parameters and reruns the Markov model for prediction. In this way, through precise model testing, the system can more accurately simulate and predict changes in network state, thereby making optimal decisions under various conditions. This improved accuracy enables the system to more reliably identify and mitigate interference in dynamically changing network environments.

[0153] In this embodiment, by verifying model accuracy and setting a performance baseline, the system can automatically adjust network parameters and then rerun the model for prediction. Furthermore, dynamic model adjustment reduces the need for manual intervention, lowers maintenance costs, and improves the system's autonomy and reliability.

[0154] It should also be noted that the signal interference management method described in the above embodiments can not only automatically detect and mitigate signal interference, but also dynamically adapt to network changes in different environments through continuous learning and optimization. Compared with existing solutions, this method has a higher level of intelligence and adaptability, can significantly reduce network failures and improve network speed, and ensure the continuity and reliability of network connections through an automated network parameter adjustment mechanism. It can be applied in multiple fields such as smart homes and enterprise offices, saving maintenance costs and time for enterprise and individual users, and providing effective support for the stability and performance of wireless networks.

[0155] The signal interference management method described in this application employs differentiated data acquisition strategies for different target scenarios, achieving differentiated signal data acquisition. By analyzing real-time signal data through a target network model, the system can continuously provide stable, efficient, and intelligent network interference detection and mitigation services in a constantly changing network environment without manual intervention, ensuring network connectivity continuity and stability, reducing network failures, and improving network speed. Furthermore, through dual DQN and priority experience replay mechanisms, the model can quickly learn and adapt to new interference sources, while optimizing its decision-making process to improve network performance.

[0156] like Figure 4 As shown, one embodiment of this application also provides a signal interference management device, the device comprising:

[0157] The first processing module 410 is used to determine the data acquisition strategy corresponding to the target scene based on the type of the target scene;

[0158] The data acquisition module is used to acquire signal data of wireless network signals in the target scenario using the data acquisition strategy, wherein the signal data includes at least one of signal strength information, signal-to-noise ratio information, and spectrum occupancy information;

[0159] The model prediction module 420 is used to input the signal data into the target network model and obtain the prediction result output by the target network model, wherein the target network model is a Markov model trained based on historical signal data in the target scene;

[0160] The parameter adjustment module 430 is used to adjust the network parameters of the network devices in the target scenario based on the prediction results.

[0161] In this embodiment, differentiated data acquisition strategies can be formulated for different target scenario types. Based on this, differentiated signal data acquisition can be achieved. Thus, the target network model trained based on the acquired historical signal data is more suitable for the target scenario, making the prediction results for the target scenario more accurate. Furthermore, based on the prediction results, more precise network parameter adjustments can be made for network devices in the target scenario, avoiding misjudgments. Moreover, since this scheme obtains prediction results based on real-time acquired signal data, it can adapt to rapidly changing network environments.

[0162] Optionally, the first processing module 410 includes:

[0163] The first processing unit is configured to determine, when the target scene is a public place, the data collection strategy corresponding to the target scene is: to determine the collection frequency of the signal data in the target scene based on the pedestrian flow in the target scene;

[0164] The second processing unit is configured to determine, when the target scene is a home environment, the data acquisition strategy corresponding to the target scene is: to acquire the signal data at a first preset frequency;

[0165] The third processing unit is configured to determine, when the target scene is a public office space, the data acquisition strategy corresponding to the target scene is: to acquire the signal data at a second preset frequency, wherein the second preset frequency is greater than the first preset frequency;

[0166] The fourth processing unit is configured to determine the data collection strategy corresponding to the target scenario when the target scenario is a personal office space: based on the historical network usage information of the user in the target scenario, determine the collection frequency of the signal data in the target scenario, wherein the historical network usage information reflects the activity level of the user in the target scenario using the wireless network at different times.

[0167] Optionally, the prediction results include: signal strength prediction values ​​and / or interference prediction values ​​for each frequency band used to transmit the wireless network signal, wherein the interference prediction values ​​characterize the degree of interference the frequency band will experience in the future, and the network parameters include: the frequency band used by the network device and / or the transmit power of the network device;

[0168] The parameter adjustment module 430 includes:

[0169] The first switching unit is configured to switch the frequency band used by the network device to the first frequency band when the first interference prediction value of the frequency band used by the network device exceeds a threshold and there is a frequency band with an interference prediction value less than the first interference prediction value.

[0170] The second switching unit is used to switch the frequency band used by the network device to the first frequency band when the interference prediction value of the frequency band used by the network device indicates increased interference.

[0171] The first adjustment unit is used to reduce the transmit power of the network device when the interference prediction value of the frequency band used by the network device indicates increased interference.

[0172] The first frequency band is the one with the smallest predicted interference value among all the frequency bands.

[0173] Optionally, the device further includes:

[0174] The model building module is used to build an initial model by abstracting the wireless network environment in the target scenario into a Markov Decision Process (MDP). The state of the MDP represents the network state of the network environment, the action of the MDP represents the adjustment of the network parameters, and the reward function of the MDP guides the model to select actions to suppress interference and / or optimize network performance through positive incentives.

[0175] The model training module is used to train the initial model based on the historical signal data to obtain the target network model.

[0176] Optionally, the method further includes:

[0177] The first training module is used to perform value evaluation on each action and obtain value evaluation results during the training of the initial model by employing a dual deep Q-network (DQN).

[0178] Based on the value assessment results, actions are selected during training.

[0179] Optionally, the method further includes:

[0180] The second training module is used to train the initial model based on the priority experience replay (PER) mechanism.

[0181] The apparatus provided in this application embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, so it will not be described again here.

[0182] It should be noted that the division of units in the embodiments of this application is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units.

[0183] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0184] One embodiment of this application also provides a management device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the program is executed by the processor, it implements the signal interference management method as described in any of the preceding claims.

[0185] The specific implementation of the signal interference management method, which is executed by the program running on the processor of the network device, can be found in the detailed description of the signal interference management method, and will not be repeated here.

[0186] In addition, specific embodiments of this application also provide a readable storage medium storing a program. When executed by a processor, this program implements the various processes of the above-described signal interference management method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here. The readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0187] This application also provides a computer program product, including computer instructions. When executed by a processor, the computer instructions implement the various processes of the above-described signal interference management method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0188] Optionally, embodiments of this application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0189] The computer program product described in this application includes computer instructions that, when executed by a processor, implement the various processes of the method embodiments shown above and achieve the same technical effects. To avoid repetition, these will not be repeated here.

[0190] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0191] It should be noted that many of the functional components described in this specification are referred to as modules / submodules in order to more specifically emphasize the independence of their implementation.

[0192] In this application embodiment, the module / submodule can be implemented in software so that it can be executed by various types of processors. For example, an identified executable code module may include one or more physical or logical blocks of computer instructions, which may be constructed as objects, procedures, or functions. Nevertheless, the executable code of the identified module does not need to be physically located together, but may include different instructions stored in different bits, which, when logically combined, constitute the module and achieve the module's intended purpose.

[0193] In practice, an executable code module can be a single instruction or many instructions, and can even be distributed across multiple different code segments, different programs, and across multiple memory devices. Similarly, operational data can be identified within the module and can be implemented in any suitable form and organized within any suitable type of data structure. This operational data can be collected as a single dataset or distributed across different locations (including different storage devices), and can exist, at least in part, solely as electronic signals within the system or network.

[0194] When a module can be implemented using software, considering the current level of hardware technology, modules that can be implemented in software can be implemented using hardware circuits by those skilled in the art to achieve the corresponding functions, without considering cost. These hardware circuits include conventional very-large-scale integrated circuits (VLSI) or gate arrays, as well as existing semiconductors such as logic chips and transistors, or other discrete components. Modules can also be implemented using programmable hardware devices, such as field-programmable gate arrays, programmable array logic, and programmable logic devices.

[0195] The above describes the preferred embodiments of this application. It should be noted that those skilled in the art can make several improvements and modifications without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A signal interference management method, characterized in that, The method includes: Based on the type of the target scenario, determine the data acquisition strategy corresponding to the target scenario; Using the data acquisition strategy described above, signal data of wireless network signals in the target scenario are acquired, and the signal data includes at least one of signal strength information, signal-to-noise ratio information, and spectrum occupancy information; The signal data is input into the target network model to obtain the prediction result output by the target network model, wherein the target network model is a Markov model trained based on historical signal data in the target scene; Based on the prediction results, adjust the network parameters of the network devices in the target scenario.

2. The signal interference management method according to claim 1, characterized in that, The step of determining the data collection strategy corresponding to the target scenario based on the type of the target scenario includes at least one of the following: When the target scene is a public place, the data collection strategy corresponding to the target scene is determined as follows: the collection frequency of the signal data in the target scene is determined based on the pedestrian flow in the target scene. When the target scenario is a home environment, the data acquisition strategy corresponding to the target scenario is determined as follows: the signal data is acquired at a first preset frequency; When the target scenario is a public office space, the data acquisition strategy for the target scenario is determined as follows: the signal data is acquired at a second preset frequency, where the second preset frequency is greater than the first preset frequency. When the target scenario is a personal office space, the data collection strategy for the target scenario is determined as follows: based on the historical network usage information of users in the target scenario, the collection frequency of the signal data in the target scenario is determined, and the historical network usage information reflects the activity level of users in the target scenario using the wireless network at different times.

3. The signal interference management method according to claim 1, characterized in that, The prediction results include: signal strength prediction values ​​and / or interference prediction values ​​for each frequency band used to transmit the wireless network signal. The interference prediction values ​​characterize the degree of interference that the frequency band will experience in the future. The network parameters include: the frequency band used by the network device and / or the transmit power of the network device. Adjusting the network parameters of the network devices in the target scenario based on the prediction results includes at least one of the following: If the first interference prediction value of the frequency band used by the network device exceeds the threshold, and there is a frequency band with an interference prediction value less than the first interference prediction value, the frequency band used by the network device will be switched to the first frequency band. If the interference prediction value of the frequency band used by the network device indicates increased interference, the frequency band used by the network device will be switched to the first frequency band. If the predicted interference value for the frequency band used by the network device indicates increased interference, the transmit power of the network device shall be reduced. The first frequency band is the one with the smallest predicted interference value among all the frequency bands.

4. The signal interference management method according to claim 1, characterized in that, The method further includes: An initial model is constructed by abstracting the wireless network environment in the target scenario into a Markov Decision Process (MDP). The state of the MDP represents the network state of the network environment, the action of the MDP represents the adjustment of the network parameters, and the reward function of the MDP guides the model to select actions to suppress interference and / or optimize network performance through positive incentives. Based on the historical signal data, the initial model is trained to obtain the target network model.

5. The signal interference management method according to claim 4, characterized in that, The method further includes: During the training of the initial model, a dual deep Q-network (DQN) is used to evaluate the value of each action and obtain the value evaluation results. Based on the value assessment results, actions are selected during training.

6. The signal interference management method according to claim 5, characterized in that, The method further includes: The initial model is trained based on the Priority Experience Replay (PER) mechanism.

7. A signal interference management device, characterized in that, The device includes: The first processing module is used to determine the data acquisition strategy corresponding to the target scenario based on the type of the target scenario. The data acquisition module is used to acquire signal data of wireless network signals in the target scenario using the data acquisition strategy, wherein the signal data includes at least one of signal strength information, signal-to-noise ratio information, and spectrum occupancy information; The model prediction module is used to input the signal data into the target network model and obtain the prediction result output by the target network model, wherein the target network model is a Markov model trained based on historical signal data in the target scene; The parameter adjustment module is used to adjust the network parameters of the network devices in the target scenario based on the prediction results.

8. A management device, characterized in that, It includes a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the signal interference management method as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, Includes computer instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1 to 6.

10. A readable storage medium, characterized in that, The readable storage medium stores a program that, when executed by a processor, implements the steps of the signal interference management method as described in any one of claims 1 to 6.

Images