Adaptive channel access method

Abstract

The invention discloses a self-adaptive channel access method, which specifically comprises the following steps of: 1, an interception stage: monitoring channel energy and a service label of an unlicensed frequency band, identifying a current service type and distributing a service priority weight; 2, a decision-making stage: calculating a preemption factor through a dynamic energy detection threshold value, and judging whether to trigger channel preemption based on the preemption factor; step 3, an access stage: after successful preemption, avoiding conflicts with other protocol signals by pre-distributing time slot codes, and sending data; and 4, a conflict prediction and backoff adjustment stage: predicting a conflict probability based on an LSTM model, performing adaptive backoff window adjustment, and processing an abnormal condition according to an abnormal processing mechanism. According to the method and the device, the high-priority service can preempt the channel more quickly, meanwhile, the ineffective competition of the low-priority service when the channel quality is poor is avoided, the spectrum utilization rate is improved, and meanwhile, the conflict rate and the time delay are reduced.

Description

Technical Field

[0001] The field of the present invention belongs to the field of wireless communication technology, and specifically relates to an adaptive channel access method. Background Art

[0002] In the unlicensed frequency band (such as the 5.8 GHz ISM band), the LBT (Listen Before Talk) mechanism is widely used to avoid conflicts caused by multiple devices transmitting signals simultaneously. The LBT mechanism determines whether to access the channel by listening to whether the channel is idle. If the channel is idle, the device is allowed to send a signal; otherwise, it enters the backoff state. The traditional LBT mechanism uses a fixed energy detection threshold (EDT) and a backoff window. The workflow is as follows: The device listens to the channel energy before transmission. If the energy is lower than the fixed threshold energy detection threshold (EDT), it is determined to be idle and data is sent; otherwise, it enters the backoff state.

[0003] With the explosive growth of differentiated service requirements such as future Ultra Reliable Low Latency Communication (URLLC), Enhance Mobile Broadband (eMBB), and Massive Machine Type Communication (mMTC), the traditional LBT mechanism faces the following defects:

[0004] 1. Mismatch between static threshold and dynamic services: The fixed EDT cannot distinguish service priorities. For example, the URLLC service needs to quickly seize the channel, but a high EDT may cause its waiting time to be too long; while the mMTC service is prone to redundant competition due to a low EDT, resulting in channel congestion.

[0005] 2. Lack of environmental awareness in the backoff mechanism: The existing exponential backoff algorithm (such as the CW adjustment of 802.11ax) only relies on the historical number of collisions and does not combine real-time channel load prediction, resulting in an overly large (resource waste) or overly small (increased conflict) backoff window under high load.

[0006] 3. Increased cross-protocol interference: When heterogeneous protocols such as Wi-Fi and LTE-U / LAA coexist, the fixed CCA (Clear Channel Assessment) mechanism cannot identify protocol characteristics, resulting in cross-system signal collisions (such as Wi-Fi preambles being misjudged as noise). Summary of the Invention

[0007] The object of the present invention is to provide an adaptive channel access method. The present invention enables high-priority services to preempt the channel more quickly, while avoiding ineffective competition of low-priority services when the channel quality is poor, improving the spectrum utilization rate while reducing the collision rate and delay.

[0008] The technical solution of the present invention: An adaptive channel access method specifically includes the following steps:

[0009] Step 1, listening stage: Monitor the channel energy and service tags of the unlicensed band, monitor the channel state information according to the monitored channel energy, identify the current service type and assign service priority weights;

[0010] Step 2, decision-making stage: Based on the channel state information and service priority weights, calculate the preemption factor through the dynamic energy detection threshold, and determine whether to trigger channel preemption based on the preemption factor;

[0011] Step 3, access stage: After successful preemption, avoid conflicts with other protocol signals through pre-allocated time slot codes and send data;

[0012] Step 4, conflict prediction and backoff adjustment stage: Predict the conflict probability based on the LSTM model, perform adaptive backoff window adjustment, and handle abnormal situations according to the exception handling mechanism.

[0013] In the aforementioned adaptive channel access method, in step 1, the service tags include physical layer identifiers of URLLC, eMBB, and mMTC; in the assignment of service priority weights, the weight of URLLC service is 1, the weight of eMBB service is 0.7, and the weight of mMTC service is 0.3.

[0014] In the aforementioned adaptive channel access method, in step 2, the dynamic energy detection threshold EDT current The calculation formula is as follows:

[0015] EDT current = EDT base -α·P priority +β·CSI variation ;

[0016] Among them, EDT base is the basic energy detection threshold, α is the service priority weight factor, P priority is the service priority weight, β is the channel state change rate factor, and CSI variation is the variance of the channel state information change rate calculated based on the sliding window.

[0017] In the aforementioned adaptive channel access method, the variance CSI variation of the channel state information change rate has the following calculation formula:

[0018] CSI variation = Var(CSI window );

[0019] CSI window = [CSI t-N+1 , CSI t-N+2 ,..., CSI t ;

[0020] Among them, CSI window is a sliding window of size N, containing the channel state information values of the most recent N moments; Var() represents the function for calculating variance.

[0021] In the aforementioned adaptive channel access method, the calculation formula for the preemption factor Q is as follows:

[0022]

[0023] Among them, P priority is the service priority weight, and EDT current is the current dynamic energy detection threshold.

[0024] In the aforementioned adaptive channel access method, in step 4, the LSTM conflict prediction model is implemented through the following steps:

[0025] Step 4.1: Collect the historical number of conflicts, channel occupancy rate, and interference pulse characteristics as input features;

[0026] Step 4.2: Use a sliding window to construct time series data, perform normalization and outlier filtering;

[0027] Step 4.3: Deploy a lightweight LSTM model for real-time inference, predict the conflict probability, and determine the risk level.

[0028] In the aforementioned adaptive channel access method, the historical number of conflicts N collision is the number of channel conflicts within the most recent N cycles, the interference pulse characteristic I pulse is the proportion of the detected Wi-Fi preamble characteristics, and the channel occupancy rate represents the calculated proportion of the channel busy state. The calculation formula is:

[0029]

[0030] In the formula, T busy is the channel busy time, and T total is the total channel time.

[0031] In the aforementioned adaptive channel access method, the calculation formula for the conflict probability P collision predicted by the LSTM model is:

[0032] P collision = LSTM(X history , X current );

[0033] Among them, X history represents the historical window data, and X current represents the current moment feature;

[0034] For the historical window data X history , the calculation formula is as follows:

[0035] X history = [X t-4 , X t-3 ,..., X t ;

[0036] Among them, X t represents the feature at time t;

[0037] For the calculation formula of the current moment feature, it is as follows:

[0038]

[0039] Among them, N collision represents the number of historical conflicts, CBR represents the channel occupancy rate, and I pulse represents the interference pulse;

[0040] If the conflict probability P collision is greater than 0.5, the risk level is determined as a high conflict risk;

[0041] If the conflict probability P collision is less than or equal to 0.5, the risk level is determined as a low conflict risk.

[0042] In the aforementioned adaptive channel access method, the adaptive backoff window adjustment includes contention window calculation, service priority coupling, and γ factor dynamic adjustment.

[0043] In the aforementioned adaptive channel access method, for the contention window CW adaptive , the calculation formula is as follows:

[0044] CW adaptive = CW min × [1 + floor(γ × P collision )];

[0045] Among them, CW min represents the minimum contention window size, γ represents the dynamic conflict threshold adjustment factor, and floor() represents the floor function;

[0046] In the foregoing adaptive channel access method, the process of coupling service priorities is that URLLC services preempt the guard time slot with a zero backoff window, and mMTC services use a contention window CW adaptive superimposed binary exponential backoff.

[0047] In the foregoing adaptive channel access method, the formula for calculating the dynamic conflict threshold adjustment due to γ is as follows:

[0048]

[0049] In the foregoing adaptive channel access method, the exception handling mechanism includes conflict oscillation suppression and model failure fallback;

[0050] The conflict oscillation suppression means that when P is greater than 0.9 for 3 consecutive times collision forcefully reset CW adaptive = CW min × 5;

[0051] The model failure fallback means switching to the baseline backoff algorithm when the LSTM model inference times out and / or is abnormal.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The dynamic energy detection threshold mechanism of the present invention makes the access timing of services with different priorities to the channel more reasonable, reducing conflicts caused by unreasonable competition. Through the coupling mechanism of the dynamic energy detection threshold and service priority, high-priority services can quickly seize the channel when the channel conditions permit, avoiding waste of waiting caused by fixed thresholds; low-priority services increase the energy detection threshold when the channel deteriorates, reducing ineffective competition for busy channels, enabling spectrum resources to be more reasonably allocated to different services, thereby improving the overall spectrum utilization rate. For high-priority services such as URLLC, the dynamic energy detection threshold mechanism of the present invention reduces its energy detection threshold, enabling it to preempt the channel first, greatly reducing the transmission delay of high-priority services. Through the LSTM conflict prediction and linear backoff window adjustment mechanism, by accurately predicting the conflict probability and adaptively adjusting the backoff window, the device can compete for the channel more scientifically according to the actual situation, effectively reducing the possibility of conflicts. At the same time, the LSTM conflict prediction mechanism of the present invention allows high-priority services to skip random competition and directly enter the transmission stage when the conflict probability is high, further ensuring low-latency requirements. In the scenario of coexistence between 3GPP LBT and Wi-Fi, the cross-layer collaborative LTT protocol reconstruction plays a key role. The present invention can automatically identify Wi-Fi preambles and dynamically adjust the LBT listening period to avoid conflicts with Wi-Fi signals. In addition, by adaptively adjusting the conflict probability threshold γ value, the present invention optimizes the access strategies of different protocol devices to ensure harmonious coexistence between different protocols and improve the communication stability in heterogeneous network environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a flowchart of the method of the present invention;

[0055] Figure 2 is a module diagram of the method of the present invention;

[0056] Figure 3 is a schematic diagram of the multi-service coexistence application scenario of the present invention;

[0057] Figure 4 is a flowchart of the decision-making stage of the present invention;

[0058] Figure 5 is a flowchart of the dynamic energy detection threshold update of the present invention;

[0059] Figure 6 is a flowchart of the LSTM-based conflict prediction and backoff window length adjustment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The present invention will be further described below with reference to the drawings and embodiments, but it shall not be used as a basis for limiting the present invention.

[0061] Embodiment 1: An adaptive channel access method, as Figure 1 shown, by reconstructing LBT into a three-layer architecture of LTT (Listen-Decide-Access), inserting a "decision layer" between the traditional "listen-transmit", realizing the joint optimization of service priority, channel state, and cross-protocol characteristics. By adding a "decision" decision layer after the "listen" stage of the traditional LBT, combining dynamic service priority and channel state prediction, adopting a flexible energy detection threshold (EDT) adjustment mechanism, and optimizing the channel access strategy by intelligently predicting the collision probability and adaptively backoff window. The modules of the present invention are as Figure 2 shown, including a dynamic energy detection threshold calculation module, a channel energy detection module, and a channel preemption module; the dynamic energy detection threshold calculation module of the adaptive channel access method includes a cross-layer listening module, a channel state monitoring module, and a service priority judgment module; the cross-layer listening module of the adaptive channel access method is used to obtain the current channel energy level, service label, and Wi-Fi preamble characteristics; the channel state monitoring module of the adaptive channel access method is used to monitor the state information of the channel according to the obtained current channel energy level and evaluate the channel quality; the service priority judgment module of the adaptive channel access method judges the priority according to the service label and assigns a priority weight; the dynamic energy detection threshold calculation module of the adaptive channel access method dynamically calculates the energy detection threshold according to the results of the cross-layer listening module, the channel state monitoring module, and the service priority judgment module; the channel energy detection module of the adaptive channel access method detects the energy level in the channel based on the dynamic energy detection threshold and judges the channel state; the channel preemption module of the adaptive channel access method makes a preemption decision according to the result of the channel energy detection module. The specific steps of the present invention are as follows:

[0062] Step 1, listening stage: Monitor the channel energy and service label of the unlicensed band, monitor the channel state information according to the monitored channel energy, identify the current service type, and assign a service priority weight;

[0063] In this step, the service label includes the physical layer identifiers of URLLC, eMBB, and mMTC; as Figure 3 shown, multiple different services (such as URLLC, eMBB, mMTC, etc.) share the unlicensed band for data transmission on the unlicensed band. Due to different service requirements, each service uses different channel access parameters on the unlicensed band.

[0064] In this embodiment, in the assignment of the service priority weight, the weight of the URLLC service is 1, the weight of the eMBB service is 0.7, and the weight of the mMTC service is 0.3.

[0065] During the listening stage, channel information and service information are collected. The physical layer and the MAC layer work collaboratively and in parallel across layers, and the collected information is passed to the decision-making stage to provide a basis for the decisions in the subsequent decision-making stage. The physical layer detects the channel energy and service tags (such as URLLC identifiers), and the MAC layer synchronously receives the Wi-Fi preamble features (such as the OFDM symbol structure).

[0066] Specifically, the physical layer (PHY) listening includes channel energy detection and service tag identification. Channel energy detection is specifically that the physical layer continuously monitors the energy level in the channel, usually measured using metrics such as Received Signal Strength Indicator (RSSI) or Received Signal Strength (RSS). The higher the RSSI / RSS value, the stronger the signal present in the channel, indicating that there may be other devices transmitting data. Service tag identification is specifically that the physical layer attempts to identify whether there is a signal of a specific service in the current channel, which is achieved by parsing specific fields in the received signal. For example, if the device supports URLLC services, the physical layer will attempt to detect whether there is an identifier of the URLLC service in the signal (e.g., a specific physical layer header or control information).

[0067] Specifically, the MAC layer is responsible for receiving and parsing the preamble of the Wi-Fi signal. The preamble is the starting part of the Wi-Fi signal and contains important control information, such as the modulation method and coding rate of the signal. By parsing the preamble, the MAC layer can determine whether there is a Wi-Fi signal in the current channel and obtain the characteristic information of the Wi-Fi signal (e.g., the OFDM symbol structure).

[0068] Step 2: Decision-making stage: Based on the channel state information and service priority weights, calculate the preemption factor through the dynamic energy detection threshold, and determine whether to trigger channel preemption based on the preemption factor, as Figure 4 shown;

[0069] In this step, the decision-making stage decides whether to preempt the channel according to the information collected in the listening stage, comprehensively considering service priorities and channel states.

[0070] Figure 5 The following is the flowchart for updating the current dynamic energy detection threshold, and the specific process is as follows:

[0071] Step 1: Initialize and configure the basic parameters: Set the initial values according to the parameters in the dynamic energy detection threshold calculation formula. EDT baseThe basic energy detection threshold can be configured according to the actual application scenario. This value is usually set according to regulatory restrictions, hardware performance, and empirical values. In this embodiment, it is set to -90 dBm; α is the service priority weight factor, which is used to adjust the impact of priority on EDT. In this embodiment, it is set to 10. β is the channel state change rate factor, which is used to adjust the impact of channel state on EDT. In this embodiment, it is set to 5. N is the size of the configured sliding window, which is used to calculate the variance of the CSI change rate. In this embodiment, it is set to 20, and then the CSI is initialized window is empty.

[0072] Step 2: Periodic or event-triggered update. Periodic update means that the update process is executed every certain period (for example, 10 ms). Event-triggered update means that when the following events are detected, the update process is triggered: a new service arrives and its priority is different from the current service; the channel state changes significantly (for example, the CSI changes exceed a certain threshold).

[0073] Step 3: Obtain the current service priority weight (P priority ), then identify the current service type, and identify the service type to be sent currently (for example, URLLC, eMBB, mMTC) according to the packet header information or through other signaling mechanisms; according to the service type, obtain the corresponding priority weight P priority .

[0074] Step 4: Obtain the current channel state information (CSI), and measure the current channel state information (CSI) through the physical layer. The CSI can be the channel quality indicator (CQI), signal-to-interference-plus-noise ratio (SINR), or received signal strength indicator (RSSI), etc. Add the current CSI value (CSIt) to the sliding window CSI window .

[0075] Step 5: Determine whether the size of CSI window has reached the preset window size N. If the window is full, remove the oldest CSI value in CSI window , and add the current CSI value (CSIt) to the end of CSI window , keeping the window size as N. If the window is not full, directly add the current CSI value (CSIt) to the end of CSI window . CSI window is expressed as follows:

[0076] CSI window = [CSI t-N+1 , CSI t-N+2 ,..., CSI t .

[0077] Step 6: Determine whether the window size meets the condition for calculating variance: Ensure that there are at least two CSI values in CSIwindow to calculate the variance. The specific formula is as follows:

[0078] CSI variation = Var(CSI window );

[0079] where CSI window is a sliding window of size N, containing the channel state information values of the most recent N moments; Var() represents the variance calculation function.

[0080] Step 7: Calculate the dynamic energy detection threshold (EDT current ), and the calculation formula is as follows:

[0081] EDT current = EDT base - α·P priority + β·CSI variation ;

[0082] where EDT base is the basic energy detection threshold, which can be configured according to the actual application scenario; α is the service priority weight factor, used to adjust the impact of priority on EDT; P priority is the service priority weight (i.e., URLLC = 1, eMBB = 0.7, mMTC = 0.3), β is the channel state change rate factor, used to adjust the impact of the channel state on EDT, and CSI variation is the variance of the channel state information change rate calculated based on the sliding window, used for real-time detection of channel degradation.

[0083] Step 8: Update the energy detection threshold EDT current , and apply the calculated EDT current to the physical layer as the energy detection threshold for subsequent channel listening, and end this periodic or event-triggered energy detection threshold update process, waiting for the next update.

[0084] The preemption factor Q is used to measure the priority of preemption of the channel, and the calculation formula is as follows:

[0085]

[0086] where P priority is the service priority weight, indicating the priority of the current service. High-priority services have higher weight values; EDT current is the current dynamic energy detection threshold, indicating the busyness of the current channel. A higher EDT indicates a busier channel.

[0087] Compare the preemption factor Q with the preset threshold Qthreshold Compare. If Q is greater than the preset threshold Q threshold , then trigger the preemption logic, indicating that the current service has a higher priority and the channel is relatively idle, and it can try to preempt the channel.

[0088] Step 3, access phase: After successful preemption, avoid conflicts with other protocol signals by pre-assigning time slot codes and send data;

[0089] In this step, the channel is preempted and data is sent during the access phase, while avoiding conflicts with Wi-Fi signals. Use pre-assigned time slot codes (e.g., OFDM symbol masks) to preempt the channel and avoid conflicts with Wi-Fi signals; after preempting the channel, the device can start sending data. The time slot code is a special signal sequence that can be used to identify different devices or services. By sending a specific time slot code, the device can inform other devices that it is using the channel, thus avoiding conflicts.

[0090] Step 4, conflict prediction and backoff adjustment phase: Predict the conflict probability based on the LSTM model, perform adaptive backoff window adjustment, and handle abnormal situations according to the exception handling mechanism.

[0091] In this step, as Figure 6 shown, the specific processes of predicting the conflict probability and adaptive backoff window adjustment are as follows:

[0092] Step 1: Perform parameter initialization;

[0093] ① LSTM input dimension: Define the input feature dimension (the number of historical conflicts N collision , channel occupancy rate CBR, and interference pulse feature I pulse );

[0094] ② Contention window parameter: CW min is the minimum contention window size (e.g., 16 time slots); γ is the dynamic conflict threshold adjustment factor (initial value: γ = 2);

[0095] ③ Data collection period: Set the statistical period (e.g., collect once every 100 ms);

[0096] ④ LSTM model configuration: Number of hidden layer neurons: 32 (can be adjusted as needed); Time step (timesteps): 5 (data of the previous 5 cycles, can be adjusted as needed); Activation function: tanh / ReLU; Output layer: Sigmoid (conflict probability 0 - 1).

[0097] Step 2: Perform data collection and preprocessing;

[0098] For the number of historical conflicts N collision , channel occupancy rate CBR, and interference pulse feature Ipulse Perform data collection; historical conflict count N collision Indicates counting the number of channel conflicts in the most recent N cycles; interference pulse feature I pulse Is the ratio of the detected Wi-Fi preamble feature (such as the OFDM symbol ratio); the channel occupancy rate CBR represents calculating the ratio of the channel busy state, and the calculation formula is:

[0099]

[0100] Where, T busy Is the channel busy time, T total Is the total channel time.

[0101] Data preprocessing includes sliding window construction, normalization, and outlier filtering; sliding window construction means maintaining a time series window to store historical 5-cycle data; normalization is to perform Min-Max normalization on N collision , CBR, and I pulse Perform Min-Max normalization; outlier filtering means removing noise data caused by instantaneous interference (such as CBR > 95% is regarded as an anomaly).

[0102] Step 3: Train and deploy the LSTM model;

[0103] ① Offline training stage

[0104] Dataset generation: Collect historical conflict event data (input X, label Y = actual conflict result 0 / 1)

[0105] Model training: Loss = Binary Cross-Entropy, Optimizer: Adam;

[0106] Model verification: Evaluate the prediction accuracy through the ROC curve;

[0107] ② Online deployment

[0108] Real-time inference engine: Deploy a lightweight LSTM model (such as TensorFlow Lite) Dynamic parameter update:

[0109] Fixed duration, such as updating the model weights in full every 24 hours;

[0110] Fixed duration, such as incremental update every 1 hour (online learning of the latest conflict patterns);

[0111] Step 4: Perform real-time conflict probability prediction;

[0112] Input feature construction: Extract the features at the current moment: Concatenate the historical window data: X history =[X t-4 ,Xt-3 ,...,X t ;

[0113] Among them, X t represents the feature at time t.

[0114] Through LSTM model inference, calculate the conflict probability:

[0115] P collision = LSTM(X history , X current );

[0116] Then perform probability threshold determination:

[0117] If P collision > 0.5, it is determined as a high conflict risk;

[0118] If P collision ≤ 0.5, it is determined as a low conflict risk.

[0119] Step 5: Adaptive backoff window adjustment;

[0120] Calculate the contention window and dynamically adjust it according to the predicted conflict probability:

[0121] CW adaptive = CW min × [1 + floor(γ × P collision )];

[0122] Among them, CW min represents the minimum contention window size, γ represents the dynamic conflict threshold adjustment factor, and floor() represents the floor function.

[0123] Perform service priority coupling. For high-priority services (URLLC): If P collision > 0.5, skip the random backoff, directly preempt the protection time slot, and at the same time adopt an asymmetric contention strategy (such as a fixed backoff window CW = 0). For low-priority services (mMTC): Force the application of CW adaptive , and superimpose exponential backoff.

[0124] The γ factor is dynamically adjusted and adaptively updated according to the network load:

[0125]

[0126] γ ∈ [1, 5];

[0127] Step 6: Handle abnormal situations according to the abnormal handling mechanism;

[0128] Conflict oscillation suppression: If P is predicted continuously 3 times collision > 0.9, forcefully reset CWadaptive = CW min × 5;

[0129] Model failure fallback: When the LSTM inference times out or is abnormal, switch to the baseline backoff algorithm (such as 802.11ax EDCA).

[0130] The dynamic energy detection threshold mechanism of the present invention makes the access timing of different priority services to the channel more reasonable, reducing conflicts caused by unreasonable competition. Through the coupling mechanism of the dynamic energy detection threshold and service priority, high-priority services can quickly seize the channel when the channel conditions permit, avoiding waste of waiting caused by fixed thresholds; low-priority services increase the energy detection threshold when the channel deteriorates, reducing ineffective competition for busy channels, so that spectrum resources can be more reasonably allocated to different services, thereby improving the overall spectrum utilization rate. For high-priority services such as URLLC, the dynamic energy detection threshold mechanism of the present invention reduces its energy detection threshold, enabling it to preempt the channel first, greatly reducing the transmission delay of high-priority services. Through the LSTM conflict prediction and linear backoff window adjustment mechanism, by accurately predicting the conflict probability and adaptively adjusting the backoff window, the device can compete for the channel more scientifically according to the actual situation, effectively reducing the possibility of conflicts. At the same time, the LSTM conflict prediction mechanism of the present invention allows high-priority services to skip random competition and directly enter the transmission stage when the conflict probability is high, further ensuring low-latency requirements. In the scenario of coexistence of 3GPP LBT and Wi-Fi, the cross-layer collaborative LTT protocol reconstruction plays a key role. The present invention can automatically identify Wi-Fi preambles and dynamically adjust the LBT listening period to avoid conflicts with Wi-Fi signals. In addition, by adaptively adjusting the conflict probability threshold γ value, the present invention optimizes the access strategies of different protocol devices to ensure harmonious coexistence between different protocols and improve the communication stability in heterogeneous network environments.

[0131] In summary, the present invention enables high-priority services to preempt the channel more quickly, while avoiding ineffective competition of low-priority services when the channel quality is poor, improving the spectrum utilization rate while reducing the conflict rate and latency.

Claims

1. An adaptive channel access method, characterized in that, Specifically, it includes the following steps: Step 1, Listening stage: Monitor the channel energy and service tags in the unlicensed frequency band, monitor the channel state information based on the monitored channel energy, identify the current service type and assign service priority weights; Step 2, Decision-making stage: Based on the channel state information and service priority weights, calculate the preemption factor through the dynamic energy detection threshold, and determine whether to trigger channel preemption based on the preemption factor; Step 3, Access stage: After successful preemption, avoid conflicts with other protocol signals by pre-assigning time slot codes and send data; Step 4, Conflict prediction and backoff adjustment stage: Predict the conflict probability based on the LSTM model, perform adaptive backoff window adjustment, and handle abnormal situations according to the exception handling mechanism.

2. The adaptive channel access method according to claim 1, wherein: In Step 1, the service tags include the physical layer identifiers of URLLC, eMBB, and mMTC; in the assignment of service priority weights, the weight of the URLLC service is 1, the weight of the eMBB service is 0.7, and the weight of the mMTC service is 0.

3.

3. The adaptive channel access method according to claim 1, wherein: In step 2, the dynamic energy detection threshold EDT current is calculated as follows: EDT current = EDT base - α·P priority + β·CSI variation ; Among them, EDT base is the basic energy detection threshold, α is the service priority weight factor, P priority is the service priority weight, β is the channel state change rate factor, CSI variation is the variance of the change rate of the channel state information calculated based on the sliding window.

4. The adaptive channel access method according to claim 3, wherein: The variance CSI of the change rate of the channel state information variation is calculated as follows: CSI variation = Var(CSI window ); CSI window = [CSI t-N+1 , CSI t-N+2 ,..., CSI t ; where CSI window is a sliding window of size N that contains the channel state information values for the most recent N time instants; Var() represents the variance calculation function.

5. The adaptive channel access method according to claim 1, wherein: The calculation formula of the preemption factor Q is as follows: Among them, P priority is the business priority weight, and EDT current is the current dynamic energy detection threshold.

6. The adaptive channel access method according to claim 1, characterized in that: Step 4, The LSTM conflict prediction model is implemented through the following steps: Step 4.1, Collect the historical number of conflicts, channel occupancy rate, and interference pulse characteristics as input features; Step 4.2, Use a sliding window to construct time series data, perform normalization and outlier filtering; Step 4.3, Deploy a lightweight LSTM model for real-time inference, predict the conflict probability and determine the risk level.

7. The adaptive channel access method according to claim 6, wherein: The number of historical collisions N collision is the number of channel collisions in the most recent N cycles, and the interference pulse feature I pulse is the proportion of the detected Wi-Fi preamble features. The channel occupancy rate represents the proportion of the busy state of the calculated channel. The calculation formula is as follows: Where, T busy is the channel busy time, and T total is the total channel time.

8. The adaptive channel access method according to claim 7, wherein: The LSTM model predicts the conflict probability P collision and its calculation formula is as follows: P collision = LSTM(X history , X current ); Among them, X history represents historical window data, and X current represents the feature at the current moment; The historical window data X history has the following calculation formula: X history = [X t-4 , X t-3 ,..., X t ; Among them, X t represents the feature at time t; The calculation formula of the current moment feature is as follows: Among them, N collision represents the number of historical conflicts, CBR represents the channel occupancy rate, and I pulse represents an interference pulse; Conflict probability P collision Greater than 0.5, the risk level is determined to be a high conflict risk; Conflict probability P collision Less than or equal to 0.5, the risk level is determined to be a low conflict risk.

9. The adaptive channel access method according to claim 1, wherein: The adaptive backoff window adjustment includes contention window calculation, service priority coupling, and dynamic adjustment of the γ factor.

10. The adaptive channel access method according to claim 9, wherein: The contention window CW adaptive has the following calculation formula: CW adaptive = CW min × [1 + floor(γ × P collision )]; Among them, CW min represents the minimum contention window size, γ represents the dynamic collision threshold adjustment factor, and floor() represents the floor function.

11. The adaptive channel access method according to claim 9, characterized in that: The process of coupling service priorities is that the URLLC service preempts the protected time slot with a zero backoff window, and the mMTC service uses the contention window CW adaptive superimposed with binary exponential backoff.

12. The adaptive channel access method according to claim 9, wherein: The calculation formula of the dynamic conflict threshold adjustment due to γ is as follows:

13. The adaptive channel access method according to claim 1, wherein: The exception handling mechanism includes conflict oscillation suppression and model failure fallback; The conflict oscillation suppression means that when P is greater than 0.9 for three consecutive times collision forcefully reset CW adaptive = CW min × 5; The model failure fallback refers to switching to the baseline backoff algorithm when the LSTM model inference times out and / or is abnormal.

Images