Dynamic spectrum sensing and anti-interference control method for mining ad hoc network communication system

Abstract

The invention relates to the technical field of communication, and particularly provides a dynamic spectrum sensing and anti-interference control method for a mining ad hoc network communication system, which comprises the following steps of: preprocessing spectrum sensing data to obtain target spectrum sensing data; effective signals and noise are distinguished according to the characteristics of the target spectrum sensing data; judging the frequency band attribute of the effective signal to obtain the information of the frequency band; dividing candidate available frequency bands based on the information of the frequency bands; screening available frequency bands according to various characteristics of the candidate available frequency bands; allocating frequency band resources based on the screened available frequency bands; monitoring the real-time state of the available frequency band, and dynamically accessing the available frequency band; judging whether the available frequency band needs to be switched or not according to the state and environment change of the available frequency band after dynamic access; the reliability and the accuracy of dynamic spectrum sensing of a mining ad hoc network communication system can be improved.

Description

Technical Field

[0001] This invention relates to the field of communication technology, specifically providing a dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system. Background Technology

[0002] Early mining communication systems used wired communication methods, but these were costly to deploy and lacked flexibility. With the development of communication technology, wireless communication systems have become the mainstream trend. Self-organizing network communication systems possess rapid self-organization capabilities, can build communication networks without pre-established infrastructure, and also have automatic recovery functions. Furthermore, with the development of intelligent mining communication, traditional mining communication systems have certain limitations, including low spectrum sensing accuracy, low spectrum resource utilization, unreliable communication, and security issues.

[0003] Traditional wireless communication systems for mining generally rely on fixed spectrum and cannot perform real-time spectrum sensing, making them unsuitable for the complex and dynamic communication environment of mines and unable to transmit communication signals in a timely manner. Traditional anti-interference control, on the other hand, uses static methods based on preset rules, which cannot adapt to the dynamic interference environment of mines. Due to the complex and constantly changing structure of mines, and the presence of strong electromagnetic interference, real-time dynamic spectrum sensing is necessary. Failure to implement timely anti-interference control will affect the reliability of mine-based ad hoc network communication. Summary of the Invention

[0004] This application provides a dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system to solve the problems of low accuracy of dynamic spectrum sensing and unreliable communication.

[0005] This application provides a dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system, the method comprising: Acquire spectrum sensing data for mining applications; The spectrum sensing data is preprocessed to obtain target spectrum sensing data; based on the characteristics of the target spectrum sensing data, effective signals and noise are distinguished. Determine the frequency band attributes of the valid signal to obtain the frequency band information; divide candidate available frequency bands based on the frequency band information; and filter available frequency bands according to various features of the candidate available frequency bands. Based on the selected available frequency bands, allocate frequency band resources; monitor the real-time status of the available frequency bands, and dynamically access the available frequency bands; Based on the status of the available frequency bands and environmental changes after dynamic access, determine whether it is necessary to switch the available frequency bands.

[0006] In some embodiments, distinguishing between valid signals and noise based on the characteristics of the target spectrum sensing data includes: The total energy used in mining is obtained, and the total energy includes signals and noise; Based on the energy characteristics and distribution stability of the target spectrum sensing data, determine whether secondary verification is required; If yes, then distinguish between valid signals and noise based on the results of the secondary verification; if no, then directly distinguish between valid signals and noise.

[0007] In some embodiments, determining the frequency band attribute of the valid signal to obtain information about the frequency band includes: Obtain the power of the received signal in each frequency band of the effective signal, and obtain the signal strength and duration of each frequency band based on the power of the received signal in each frequency band; First, the effective signal is subjected to short Fourier transform using the time-domain smoothing method. Then, the cyclic spectrum of the effective signal is obtained. The modulation type is determined based on the cyclic spectrum. Finally, the modulation type and bandwidth of each frequency band signal are obtained using time-frequency analysis.

[0008] In some embodiments, the process of dividing candidate available frequency bands based on information from the frequency band includes: Obtain the interference power in unlicensed frequency bands and the interference power in licensed frequency bands; When the interference power of the unlicensed frequency band is greater than the first power threshold, the frequency band corresponding to the unlicensed frequency band is removed; otherwise, it is used as a candidate available frequency band. When the interference power of the licensed frequency band is greater than the second power threshold, the frequency band corresponding to the licensed frequency band is removed; otherwise, it is used as a candidate available frequency band.

[0009] In some embodiments, the step of filtering available frequency bands based on multiple features of the candidate available frequency bands includes: Obtain the occupancy status, interference power, spectral entropy, and environmental characteristics of the candidate available frequency bands; The random forest algorithm is used to process the occupancy status, interference power, spectral entropy and environmental characteristics of the candidate available frequency bands. If the output is available, the candidate available frequency bands are selected as available frequency bands.

[0010] In some embodiments, allocating frequency band resources based on the selected available frequency bands; monitoring the real-time status of the available frequency bands; and dynamically accessing the available frequency bands include: Frequency band resources are allocated using a Nash equilibrium algorithm based on the potential energy function and a Staclelberg game theory model, and the selection strategy of the available frequency bands is iteratively optimized using gradient descent. If it is detected that the available frequency band is not occupied by any valid signal, it is determined to be a spectrum hole, and the location and duration of the spectrum hole are recorded, and the spectrum hole is dynamically accessed.

[0011] In some embodiments, determining whether to switch the available frequency band based on the status and environmental changes of the available frequency band after dynamic access includes: After dynamic access, the real-time coordinates of the node location, communication quality indicators, and environmental parameters of the available frequency band are obtained. The communication quality indicators include signal strength, signal-to-noise ratio, bit error rate, and interference intensity. The environmental parameters include gas concentration and temperature. After dynamic access, the real-time coordinates of node locations, communication quality indicators, and environmental parameters of the available frequency bands are used to construct a deep network DQN using the DeepQ-Network deep reinforcement learning method, so as to obtain the Q value for switching frequency bands and the Q value for maintaining the current frequency band. If the absolute value of the difference between the Q value of the switched frequency band and the Q value of the current frequency band is greater than the difference threshold, then the available frequency band needs to be switched.

[0012] The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system provided in this application embodiment features a rapid self-organizing capability and automatic ground recovery function. The mine self-organizing network is suitable for the complex and dynamic communication environment of mines, enabling real-time dynamic spectrum sensing and timely anti-interference processing in case of interference signals. It allows for timely transmission of communication signals. Dynamic spectrum sensing efficiently utilizes spectrum resources, ensuring accuracy in the complex environment of mines, reducing noise and interference from co-channel sources, and ensuring reliable and real-time communication. Anti-interference control ensures reliable communication and guarantees the transmission of personnel positioning and emergency information, ensuring safety in the mining environment. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating the dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system provided in this application embodiment. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0016] The following is combined with Figure 1 The illustrated embodiments describe the technical solution of the present invention: This application provides an embodiment of a dynamic spectrum sensing and anti-interference control method for a mining ad hoc network communication system, referring to... Figure 1 As shown, the dynamic spectrum sensing and anti-interference control method for the mining ad hoc network communication system provided in this embodiment includes the following steps: S110: Acquire spectrum sensing data for mining applications.

[0017] In some embodiments, the implementation of step S110 (acquiring mining spectrum sensing data) may include: It should be noted that the system uses spectrum sensing technology to monitor spectrum utilization in the mine in real time. It can dynamically adjust spectrum resource allocation based on changes in the mine environment (such as equipment additions or removals, signal interference, etc.), thus avoiding resource waste and ensuring communication quality. Dynamic spectrum sensing and resource allocation are performed based on communication nodes in the mine's ad hoc network; interference analysis and control are also conducted to identify spectrum segments with less interference. The spectrum resource allocation results are sent to each communication node using the ad hoc network's communication protocol, and each communication node dynamically adjusts its communication parameters.

[0018] It should be noted that the mine self-organizing network is the support for mine wireless communication. When the communication system is identified as being interfered with, it can automatically request spectrum resources.

[0019] The mining self-organizing mesh network is used for communication, and relevant data is collected by sensors.

[0020] It should be noted that the self-organizing network communication system supports a distributed architecture and automatic networking capabilities. It can be deployed quickly without central control and can automatically repair the impact of changes in network topology. When a node fails, the network can automatically reroute, maintaining communication through other nodes. This ensures stable network operation in complex environments and adapts to the needs of complex mine terrain and emergency scenarios. Since the mine's self-organizing network relies on autonomous node collaboration for communication, node movement causes dynamic changes in the spectrum environment, requiring dynamic spectrum sensing and anti-interference control.

[0021] An FPGA approach is employed to capture transient interference signals from mining equipment. UWB sensors are used to collect key parameters such as signal strength, signal-to-noise ratio, and frequency band occupancy in real time. Gas sensors collect gas data, and micro-vibration sensors collect data. Ultra-wideband UWB technology is used for personnel positioning, enabling timely location of personnel. Sensor nodes are deployed in underground mine roadways to collect spectral data from areas such as tunnel faces and haulage roadways.

[0022] The initial mesh network is constructed based on the geographical location of the BeiDou positioning nodes. Software-defined radio (SDR) is used to collect baseband data in the frequency band and extract received signal strength, power spectral density, etc.; sensors collect environmental parameters.

[0023] It should be noted that, in order to ensure mine safety and avoid safety accidents caused by the occupation of critical spectrum, a dedicated frequency band for emergency use needs to be set up. This band will be integrated into the self-organizing network for mine gas monitoring and personnel location safety monitoring. When a safety warning occurs, the communication priority of the warning information will be increased, and emergency information will be transmitted in a timely manner to ensure the safety of the mine and its personnel.

[0024] The node location is determined by a combination of ultra-wideband (UWB) positioning and time difference of arrival (TDOA) technology; the node's location coordinates are then converted into spatial grid features.

[0025] It should be noted that the spectrum data of a mining ad hoc network mainly includes key information such as spectrum resource occupancy status, spectrum characteristics of interference sources, channel parameters, anti-interference strategy data, network node status, and environmental equipment correlation data. Preprocessing mining spectrum sensing data can suppress high-interference noise, thereby improving spectrum sensing; dynamic spectrum sensing can monitor spectrum resource status in real time and dynamically adjust the application and release of spectrum resources. It can also intelligently predict interference, adjust spectrum resources in advance, and rationally allocate spectrum resources. By monitoring communication status in real time, in case of emergencies such as mid-network interruptions, emergency safety measures can be taken, and dedicated frequency bands can be added to ensure the highest priority safety needs of mining operations.

[0026] S120: Preprocess the mine spectrum sensing data to obtain target spectrum sensing data; distinguish between effective signals and noise based on the characteristics of the target spectrum sensing data.

[0027] In some embodiments, step S120 (preprocessing the mining spectrum sensing data to obtain target spectrum sensing data; distinguishing effective signals from noise based on the characteristics of the target spectrum sensing data) includes: It should be noted that the algorithm processes the node's spectrum information to meet the communication nodes' spectrum requirements; real-time monitoring of the communication quality and spectrum utilization of the communication nodes is necessary. Due to the complex environment of mines, which is susceptible to electromagnetic interference and other disturbances, interference suppression processing is performed before spectrum sensing.

[0028] It should be noted that before spectrum sensing, the received signal needs to be subjected to LMS adaptive filtering. By automatically adjusting the filter coefficients, tracking the time-varying characteristics of the channel, suppressing periodic narrowband interference, and removing high-frequency interference components, the accuracy of spectrum sensing is improved.

[0029] The spectrum sensing information is preprocessed, and the adaptive filtering LMS algorithm (minimum mean square error) is used to suppress noise to obtain the target spectrum sensing data.

[0030] It should be noted that by analyzing communication technology requirements in conjunction with the complex environmental characteristics of mines, the dynamic spectrum sensing effect of the mine's self-organizing network communication system is improved. In response to the real-time changes in the complex mine environment, artificial intelligence algorithms are used to assist spectrum sensing, improving the accuracy of spectrum sensing, thereby enhancing anti-interference control and ensuring the communication of the mine network.

[0031] It should be noted that traditional energy detection algorithms are easily affected by noise in the mining environment. The adaptive threshold energy detection method can determine whether the frequency band is occupied while compressing the time of a single sensing, and can also find a balance between not missing valid signals and not misjudging strong noise.

[0032] Based on the energy characteristics and distribution stability of the target spectrum sensing data, determine whether secondary verification is required; If yes, then distinguish between valid signals and noise based on the results of the secondary verification; if no, then directly distinguish between valid signals and noise.

[0033] The total energy used in mining is obtained, which includes signal and noise; the total energy is processed using an adaptive threshold energy detection method; preliminary node determination results are obtained; if the total energy is greater than the adaptive threshold, the node is preliminarily determined to be occupied by signal. The preliminary judgment results of each node are transmitted to the sink node. The difference between the average distribution and the distribution of each node is calculated using Kullback-Leibler divergence to obtain the average KL divergence. If the average KL divergence is less than the divergence threshold, the valid signal and noise are directly distinguished. If the average KL divergence is greater than the divergence threshold, secondary verification is required to distinguish the valid signal and noise.

[0034] Specifically, an adaptive threshold energy detection method is adopted to dynamically adjust the detection threshold according to the current noise; a sliding window of 35 symbols is used to estimate the noise variance in real time, and the total energy includes both signal and noise.

[0035] The adaptive threshold T is a multiple of the noise variance. , The noise variance is estimated in real time through a sliding window of 35 symbols (which can reflect the current environmental noise level). It is the threshold coefficient (a key adjustable parameter). A value between 2.0 and 5.0 is recommended to determine the correlation strength between the threshold and the noise.

[0036] It should be noted that when the data distribution consistency of nodes is low (large KL divergence), the value can be temporarily increased. (Enhanced anti-interference); When node data distribution consistency is low (small KL divergence), restore the baseline. (Ensure the sensitivity of the detection).

[0037] It should be noted that the sliding window size can be set to 35 symbols. Due to the sudden nature of noise generated in the real-time environment of the mine, such as the vibration noise of the coal mining machine, there may be deviations in the noise variance estimation, resulting in misjudgment of the presence of signals during detection. Cyclic stationary detection is then used for secondary verification to solve the misjudgment problem in complex environments, thereby improving the reliability of detection.

[0038] It should be noted that by using multi-node collaborative spectrum sensing, the information from multiple nodes is processed to determine the spectrum status. First, the sensing of a frequency band by the same node is determined; then, the sensing of a frequency band by different nodes is determined.

[0039] For example, if the current total energy is greater than the adaptive threshold, it is judged as occupied; if multiple nodes (e.g., more than 3 / 5) have a total energy less than the adaptive threshold and no node detects the licensed / unlicensed signal characteristics, it is judged as a spectrum hole.

[0040] It should be noted that in multi-node collaborative spectrum sensing, the sensing results of different nodes on the same frequency band can eliminate the bias that exists in single-node sensing. An adaptive threshold energy detection method is used to determine the sensing of each node. The occupied / idle result is output independently first, and then the output result is transmitted to the sink node through multi-hop routing. Kullback-Leibler divergence is used for evaluation.

[0041] It should be noted that each node outputs a sensing result for the same frequency band, which is a probability distribution of whether the frequency is occupied or idle. KL divergence is used to quantify the differences between these results.

[0042] For example, with 3 or more nodes, calculate the KL divergence of the average distribution and the distribution of each node. If the average KL divergence is less than the divergence threshold of 1, it indicates that the data distribution is highly consistent and the perception results are consistent. In this case, no secondary verification is required, and it can be directly determined as a valid signal for mining. If the average KL divergence is greater than the divergence threshold, it indicates that the data distribution is not consistent and the results are contradictory, such as 2 being occupied and 1 being empty. Secondary verification needs to be triggered.

[0043] A cyclic stationary detection method was used for secondary verification to extract the periodic features of the signal and calculate the correlation value at the cyclic frequency. If the correlation value at the cycle frequency is greater than the frequency threshold, it indicates the existence of a periodic structure and is therefore considered a valid signal; if the correlation value at all cycle frequencies is less than the frequency threshold, it indicates the absence of a periodic structure and is therefore considered noise.

[0044] It should be noted that cyclostationary detection calculates the cyclic autocorrelation function or spectral correlation function of the signal. The effective signal exhibits periodicity and will occur at a specific cyclic frequency. There are non-zero values ​​at any given frequency; while noise is non-periodic, and its cyclic autocorrelation function exists at all cyclic frequencies. All values ​​are close to 0; the frequency threshold is 0.

[0045] S130: Determine the frequency band attributes of the valid signal to obtain frequency band information; divide candidate available frequency bands based on the frequency band information; filter available frequency bands according to various characteristics of the candidate available frequency bands.

[0046] In some embodiments, step S130 (determining the frequency band attributes of the valid signal to obtain frequency band information; dividing candidate available frequency bands based on the frequency band information; filtering available frequency bands according to various features of the candidate available frequency bands) includes: It should be noted that the first step distinguishes between valid signals and noise / interference, the second step determines the frequency band attribute of valid signals, and the third step identifies spectral holes. Both licensed and unlicensed frequency band signals are valid signals; spectral holes are unused intervals within a frequency band that are not occupied by any valid signals.

[0047] The usage status of each frequency band in the mine is monitored in real time. The effective signals and noise used in the mine have already been distinguished. The actual operating frequency band of the effective signals is matched against the industry-defined spectrum rule library to determine the frequency band attributes of the effective signals. When the licensed and unlicensed frequency bands match the corresponding intervals in the rule library, the licensed and unlicensed frequency bands are determined.

[0048] For example, in a mining communication system, licensed frequency bands include the 700MHz-900MHz band specifically for coal mines; unlicensed frequency bands include the 2.4GHz / 5.8GHz industrial bands; and 720-750MHz is the frequency band for video surveillance equipment.

[0049] Obtain the power of the received signal in each frequency band of the valid signal; Specifically, RSSI is calculated based on the power of the received signal in each frequency band. First, the power of the sampled signal is calculated, and then the power of the signal is converted into RSSI in dBm form. The cumulative time when the signal in each frequency band is greater than the threshold is counted to obtain the occupancy time of each frequency band.

[0050] Record the received signal strength (RSSI) and duration of each frequency band; identify the modulation type and bandwidth of the signal; A time-domain smoothing method is employed. First, a short Fourier transform is performed on the effective signal, then the cyclic spectrum of the effective signal is calculated, and the modulation type is determined based on the cyclic spectrum. Time-frequency analysis is then used to obtain the modulation type and bandwidth of the signal in each frequency band.

[0051] It should be noted that monitoring and recording the real-time status of each frequency band helps in the next step of screening available frequency bands.

[0052] It should be noted that due to the complex real-time environment in mines, such as equipment movement and sudden interference, it is necessary to mark the real-time status of each frequency band. The division between unlicensed and licensed frequency bands is determined by relevant management and will not be elaborated upon in detail here.

[0053] Monitor the real-time status of each frequency band and label the real-time status of each frequency band. For spectrum holes: mark the predicted duration of the hole and the stability of the idle state; for unlicensed frequency bands: mark the occupancy intensity, occupancy duration, and the active frequency of the occupant; for licensed frequency bands: mark the modulation type and bandwidth of the signal in each frequency band, and mark the type of licensed equipment and the active period.

[0054] It should be noted that the higher the RSSI value, the greater the intensity of frequency band occupancy.

[0055] It should be noted that due to the dynamic changes in the mine environment, such as equipment movement and interference fluctuations, intelligent sensing technology based on dynamic environmental changes is adopted. For known strong interference in mining scenarios, such as 50Hz power frequency interference and pulse interference from motor equipment, the high-interference frequency bands are first removed, and the remaining frequency bands are retained.

[0056] The preset first power threshold is -80dBm, and the second power threshold is -70dBm; the interference power of unlicensed frequency bands and the interference power of licensed frequency bands are obtained; When the interference power of an unlicensed frequency band exceeds the first power threshold, the corresponding frequency band is removed; otherwise, it is used as a candidate available frequency band and associated with its marked occupancy strength, occupancy duration, and device activity frequency.

[0057] When the interference power of the licensed frequency band is greater than the second power threshold, the corresponding frequency band is removed; otherwise, it becomes a candidate available frequency band and is associated with its marked modulation type, bandwidth, and licensed equipment activity period.

[0058] It's further important to note that available frequency bands must be selected based on specific occupancy status, interference conditions, and environmental characteristics. Several factors influence frequency band availability. Occupancy status (whether it's occupied and the duration of occupation) directly determines whether the band is already being used by other devices. Interference also has a significant impact; even if a band isn't occupied, strong interference or spectrum disorder can lead to unreliable communication. Environmental characteristics also affect frequency band availability; spectrum characteristics vary greatly depending on location (e.g., obstructed areas in mines) and time (e.g., peak operating periods for equipment). Random forests are ensembles of multiple trees, excel at handling binary classification problems, and are adaptable to complex, high-dimensional data scenarios.

[0059] A mine-grade intrinsically safe spectrum analyzer is used to monitor the signal power and modulation scheme of candidate available frequency bands in real time. The RF front-end and ADC of a band sensing device are used to measure the power of unwanted signals within the candidate available frequency bands, distinguishing between useful and interference signals. Spectral entropy is calculated using spectral power distribution; UWB (ultra-wideband) positioning is employed; and time is obtained using the mine's synchronous clock system.

[0060] It should be noted that a feature library of legitimate signals, such as the frequency band and modulation method used in mining, should be pre-entered.

[0061] If there are spectrum holes in the candidate frequency bands, and the predicted duration is ≥10 minutes and the stability meets the standard, then it is a stable and usable frequency band with the first priority.

[0062] For example, the type of interference can be determined by the signal characteristics of the spectrum shape, bandwidth, and modulation method.

[0063] Obtain the occupancy status, interference power, spectral entropy, and environmental characteristics of candidate available frequency bands; The random forest algorithm is used. The inputs are the occupancy status of candidate available frequency bands (whether they are occupied and the duration of occupation), interference power, spectral entropy and environmental characteristics (such as location and time). The output is available or unavailable. If the output is available, it is selected as an available frequency band.

[0064] It should be noted that the 700MHz band (710-720MHz, idle time ≥5s, SNR=22dB, then there is no strong interference) is suitable for matching video surveillance in mines (2MHz bandwidth, QPSK modulation); the 2.4GHz band (2412-2417MHz, idle, subject to slight Wi-Fi interference, SNR=18dB) is suitable for sensor data transmission (narrowband, BPSK modulation).

[0065] S140: Allocate frequency band resources based on the selected available frequency bands; monitor the real-time status of available frequency bands and dynamically access available frequency bands.

[0066] In some embodiments, step S140 (allocating frequency band resources based on the selected available frequency bands; monitoring the real-time status of available frequency bands and dynamically accessing available frequency bands) includes: It should be noted that the spectrum sensing results mainly include six categories: spectrum usage status, interference data, dynamic channel status, anti-interference strategy data, node status, and environmental and equipment correlation data. Noise interference includes, for example, interference from variable frequency motors and power equipment.

[0067] It should be noted that a spectrum-aware database is established for the prediction and dynamic adjustment of spectrum holes. By combining deep learning and game theory models to predict the probability of spectrum holes and interference trends, communication parameters can be adjusted in advance, and dynamic spectrum access strategies can be optimized. For example, the frequency band with the least interference can be selected for data transmission, thereby improving data transmission efficiency.

[0068] It's important to clarify that after identifying available frequency bands, it's still crucial to monitor their status in real-time to maintain good communication quality. Spectrum hole information is used for dynamic spectrum access, allowing for the rational and dynamic utilization of spectrum resources. Based on spectrum sensing results and the real-time status of frequency bands, it's necessary to make real-time decisions about which frequency band an ad hoc network node should access and how it should access it, thereby ensuring communication reliability and security.

[0069] It should be noted that after selecting available frequency bands, it is also necessary to pay attention to the real-time spectrum status of the available frequency bands, and to carry out frequency band resource allocation and anti-interference control processing; based on real-time monitoring of dynamic changes, communication parameters should be adaptively adjusted to ensure smooth communication.

[0070] It should be noted that since multiple sensor nodes in a mine share available frequency bands, excessive competition may lead to conflicts. Each node wants to occupy a low-interference, high-bandwidth frequency band, necessitating a rational and efficient allocation of frequency resources. Distributed game theory models allow each node to allocate available frequency bands based on its local state and surrounding interference, reducing latency and improving spectrum utilization. This is well-suited to the rapid response requirements of large-scale equipment scenarios like mines.

[0071] It should be noted that the Stellarberg game theory model is used, dividing nodes into leader nodes with high perception accuracy and follower nodes with low perception accuracy. The leader node's decisions guide perception; for example, the leader node monitors strong electromagnetic interference at the tunnel face in real time, and the follower nodes adjust their perception strategies based on the leader node's decisions. This avoids redundant and ineffective perception and reduces potential misjudgments that may occur with single-node perception.

[0072] Frequency band resources are allocated using a Nash equilibrium algorithm based on the potential energy function and a Staclelberg game theory model, and gradient descent is used to iteratively optimize the selection strategy of available frequency bands. If the available frequency band is not occupied by any valid signal, it is identified as a spectrum hole, and the location and duration of the spectrum hole are recorded, and the spectrum hole is dynamically accessed.

[0073] It should be noted that the Nash equilibrium algorithm based on the potential energy function is adopted. This algorithm uses gradient descent to iteratively optimize the frequency band selection strategy of each node until each node cannot improve its own revenue by unilaterally changing its strategy.

[0074] It should be noted that the Nash equalization algorithm based on the potential energy function has a fast convergence speed, is suitable for the dynamically changing frequency band environment in mining, and can optimize the selection of frequency bands.

[0075] It should be noted that mining equipment such as excavators, motors, and transformers in the mining environment can generate strong electromagnetic interference. There is also interference caused by environmental characteristics, as well as electromagnetic interference caused by sensor and equipment malfunctions, such as abnormal motor vibration.

[0076] It should be noted that, based on the results of dynamic spectrum sensing, avoidance, suppression, and scheduling methods are used to reduce the impact of interference on the communication system and ensure the quality of communication, especially emergency communication.

[0077] It should be noted that any unoccupied intervals in the detection frequency band that are not occupied by any valid signal are considered as spectral holes, and parameters such as the location and duration of the spectral holes must be recorded.

[0078] The algorithm is based on Long Short-Term Memory (LSTM) network. The input is historical spectrum data such as the frequency band occupancy time, environmental parameters, and device status (node ​​movement speed) in the past 8 minutes. It predicts the location and duration of spectrum holes in the next minute.

[0079] It should be noted that the LSTM algorithm has a prediction accuracy of 89%. For example, if it is predicted that there is a 700MHz frequency band available in the cage shaft area between 16:00 and 16:05, then the spectrum resources can be allocated in advance to make reasonable use of the spectrum resources.

[0080] For example, in mining scenarios, such as cavities that may appear at bends in tunnels, the 700MHz frequency band can be used; opportunistic spectrum access technology can be used to dynamically allocate the 2.4GHz frequency band in the cage shaft to achieve a communication interruption rate of <2%.

[0081] S150: Determine whether it is necessary to switch available frequency bands based on the status of available frequency bands and environmental changes after dynamic access.

[0082] In some embodiments, step S150 (determining whether to switch available frequency bands based on the status of available frequency bands and environmental changes after dynamic access) includes: It should be noted that frequency bands are adjusted in real time based on dynamic spectrum sensing results: when a certain frequency band is interfered with, it is switched to an idle frequency band to achieve reasonable use of spectrum resources.

[0083] It should be noted that this involves real-time monitoring of frequency band usage and dynamic spectrum adjustment. The core function is to monitor the spectrum usage status in real time—whether it is idle or occupied—and assess channel quality. The result of dynamic spectrum sensing is real-time feedback on the interference intensity and primary user status of the frequency band.

[0084] It should be noted that the reinforcement learning method, which monitors the spectrum status and changes in the mine environment in real time and adaptively adjusts communication parameters, is suitable for the dynamic and complex environment of mining.

[0085] It should be noted that, based on the real-time changes in node location, the real-time coordinates of the node location, the communication quality indicators of the current frequency band (signal strength, signal-to-noise ratio, bit error rate, interference intensity), and environmental parameters (gas concentration, temperature), a deep network DQN is constructed using the Deep-Q-Network (DQN) deep reinforcement learning method to obtain the Q-value for maintaining the current frequency band and the Q-value for switching the frequency band. The output result is whether to maintain the current frequency band or switch the frequency band.

[0086] After dynamic access, obtain the real-time coordinates of node locations in available frequency bands, communication quality indicators, and environmental parameters. Communication quality indicators include signal strength, signal-to-noise ratio, bit error rate, and interference intensity. Environmental parameters include gas concentration and temperature. After dynamic access, the real-time coordinates of node locations, communication quality indicators, and environmental parameters of available frequency bands are used to construct a deep network DQN using the DeepQ-Network deep reinforcement learning method, so as to obtain the Q value for switching frequency bands and the Q value for maintaining the current frequency band. If the absolute value of the difference between the Q value of the switched frequency band and the Q value of the current frequency band is greater than the difference threshold, then it is necessary to switch to an available frequency band.

[0087] It should be noted that after the spectrum resources are accessed, it is necessary to continuously monitor the changes in interference in real time. If a new strong interference signal appears in the original frequency band, the "spectrum switching" is triggered, and the interference identification and access process is re-executed. If an abnormal situation occurs in the mine environment, the frequency band is switched to the emergency communication band.

[0088] It should be noted that time-frequency analysis is performed on the acquired spectrum sensing data to extract features such as channel occupancy rate and interference power spectral density; the location and spectrum features are fused into a spatiotemporal feature vector, such as [grid coordinates, channel occupancy rate, interference type]; the spatiotemporal feature vector is input into a CNN algorithm, and the output is an interference probability map for the next 10 seconds. For example, for a transport lane node, if the probability of strong interference occurring within 3 seconds after a certain grid motor starts is predicted to be 85%, then the system will switch to a preset backup channel in advance.

[0089] It should be noted that multi-band collaborative analysis may occur in mines. The processing priority order is emergency early warning, equipment control, video monitoring, and sensor data, with emergency early warning having the highest priority. Emergency signals use dedicated channels to ensure safety.

[0090] It should be noted that if the current 900MHz band is subject to motor interference (SNR=8dB), it is recommended to switch to the 730-740MHz free band within the 700MHz band. If broadband interference is detected, it is recommended to switch from OFDM mode to Frequency Hopping Spread Spectrum (FHSS) mode.

[0091] It should be noted that safety must be the top priority, ensuring spectrum resources for critical operations such as gas monitoring and personnel location. Safety is the highest priority. Dedicated frequency bands can be allocated to ensure uninterrupted communication and guarantee mine safety.

[0092] Set up a dedicated emergency frequency band, with emergency communication using FM signals; and pre-set 2-3 backup channels; if the main channel interference exceeds the limit, quickly switch to the backup channel; For example, when gas levels exceed limits, the communication system quickly switches to the emergency channel, prioritizing the transmission of early warning information and enabling rapid emergency response. Interference-free emergency communication must be ensured. Optimizing the path and reducing transmission latency can meet the requirements of emergency communication.

[0093] It should be noted that the mining environment is linked to safety monitoring. When abnormal spectrum signals are detected, an early warning message is triggered and an emergency plan is activated. When used in emergency rescue scenarios, the dynamic spectrum sensing capability of the self-organizing network is used to quickly establish a temporary communication link to ensure smooth communication of rescue information.

[0094] It should be noted that when the SNR is set to <10dB and the interference probability is >30%, such as around 11MHz, only low-rate emergency signal transmission is allowed. Also, when rescue equipment (such as life detectors and emergency base stations) temporarily accesses the system, increasing spectrum resources, emergency communications have the highest priority. For example, when miners enter high-interference areas, low-power, high-interference-resistance narrowband channels are prioritized to ensure the normal transmission of vital signs data.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dynamic spectrum sensing and anti-interference control method for a mining ad hoc network communication system, characterized in that, include: Acquire spectrum sensing data for mining applications; The spectrum sensing data is preprocessed to obtain the target spectrum sensing data; Based on the characteristics of the target spectrum sensing data, distinguish between valid signals and noise; Determine the frequency band attributes of the valid signal to obtain the frequency band information; divide candidate available frequency bands based on the frequency band information; and filter available frequency bands according to various features of the candidate available frequency bands. Based on the selected available frequency bands, allocate frequency band resources; monitor the real-time status of the available frequency bands, and dynamically access the available frequency bands; Based on the status of the available frequency bands and environmental changes after dynamic access, determine whether it is necessary to switch the available frequency bands.

2. The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system according to claim 1, characterized in that, The preprocessing of the spectrum sensing data to obtain target spectrum sensing data includes: The spectrum sensing data is processed using the adaptive filtering LMS algorithm to suppress noise, thereby obtaining the target spectrum sensing data.

3. The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system according to claim 1, characterized in that, The step of distinguishing between valid signals and noise based on the characteristics of the target spectrum sensing data includes: The total energy used in mining is obtained, and the total energy includes signals and noise; Based on the energy characteristics and distribution stability of the target spectrum sensing data, determine whether secondary verification is required; If yes, then distinguish between valid signals and noise based on the results of the secondary verification; if no, then directly distinguish between valid signals and noise.

4. The dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system according to claim 3, characterized in that, If not, then directly distinguish between valid signals and noise, including: The total energy is processed using an adaptive threshold energy detection method to obtain a preliminary node determination result; if the total energy is greater than the adaptive threshold, the node is preliminarily determined to be occupied by the signal. The preliminary judgment results of each node are transmitted to the sink node, and the difference between the average distribution and the distribution of each node is calculated using Kullback-Leibler divergence to obtain the average KL divergence. If the average KL divergence is less than the divergence threshold, then valid signals and noise can be directly distinguished; if the average KL divergence is greater than the divergence threshold, then secondary verification is required to distinguish valid signals and noise.

5. The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system according to claim 3, characterized in that, If so, then based on the secondary verification results, valid signals and noise are distinguished, including: The periodicity features of the signal are extracted using a cyclic stationary detection method to obtain the correlation value at the cyclic frequency; If the correlation value at the specified cycle frequency is greater than the frequency threshold, it is determined to be a valid signal; if the correlation value at all specified cycle frequencies is less than the frequency threshold, it is determined to be noise.

6. The dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system according to claim 1, characterized in that, The step of determining the frequency band attribute of the valid signal to obtain the frequency band information includes: Obtain the power of the received signal in each frequency band of the effective signal, and obtain the signal strength and duration of each frequency band based on the power of the received signal in each frequency band; First, the effective signal is subjected to short Fourier transform using the time-domain smoothing method. Then, the cyclic spectrum of the effective signal is obtained. The modulation type is determined based on the cyclic spectrum. Finally, the modulation type and bandwidth of each frequency band signal are obtained using time-frequency analysis.

7. The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system according to claim 1, characterized in that, The process of dividing candidate available frequency bands based on information from the frequency band includes: Obtain the interference power in unlicensed frequency bands and the interference power in licensed frequency bands; When the interference power of the unlicensed frequency band is greater than the first power threshold, the frequency band corresponding to the unlicensed frequency band is removed; otherwise, it is used as a candidate available frequency band. When the interference power of the licensed frequency band is greater than the second power threshold, the frequency band corresponding to the licensed frequency band is removed; otherwise, it is used as a candidate available frequency band.

8. The dynamic spectrum sensing and anti-interference control method for a mining self-organizing network communication system according to claim 1, characterized in that, The process of filtering available frequency bands based on multiple features of the candidate available frequency bands includes: Obtain the occupancy status, interference power, spectral entropy, and environmental characteristics of the candidate available frequency bands; The random forest algorithm is used to process the occupancy status, interference power, spectral entropy and environmental characteristics of the candidate available frequency bands. If the output is available, the candidate available frequency bands are selected as available frequency bands.

9. The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system according to claim 1, characterized in that, Based on the selected available frequency bands, frequency band resources are allocated. Monitoring the real-time status of the available frequency bands and dynamically accessing the available frequency bands includes: Frequency band resources are allocated using a Nash equilibrium algorithm based on the potential energy function using the Strackelberg game theory model, and the selection strategy of the available frequency bands is iteratively optimized using gradient descent. If it is detected that the available frequency band is not occupied by any valid signal, it is determined to be a spectrum hole, and the location and duration of the spectrum hole are recorded, and the spectrum hole is dynamically accessed.

10. The dynamic spectrum sensing and anti-interference control method for a mine self-organizing network communication system according to claim 1, characterized in that, The step of determining whether to switch the available frequency band based on the status and environmental changes of the available frequency band after dynamic access includes: After dynamic access, the real-time coordinates of the node location, communication quality indicators, and environmental parameters of the available frequency band are obtained. The communication quality indicators include signal strength, signal-to-noise ratio, bit error rate, and interference intensity. The environmental parameters include gas concentration and temperature. After dynamic access, the real-time coordinates of node locations, communication quality indicators, and environmental parameters of the available frequency bands are used to construct a deep network DQN using the DeepQ-Network deep reinforcement learning method, so as to obtain the Q value for switching frequency bands and the Q value for maintaining the current frequency band. If the absolute value of the difference between the Q value of the switched frequency band and the Q value of the current frequency band is greater than the difference threshold, then the available frequency band needs to be switched.

Images