A terminal adaptive multi-ssid uplink air interface resource control method and device and medium

Abstract

This invention discloses a terminal-adaptive multi-SSID uplink air interface resource control method, device, and medium, relating to the field of wireless network networking technology. The method isolates and configures virtualized general virtual interfaces and dedicated virtual interfaces with the same name and password through multiple VIF parameters; it adaptively classifies access terminals when uplink data flows meeting specified conditions are captured through service flow monitoring, and finally performs dual-track traffic splitting: for high-compatibility terminals, a standard bootstrap frame is directly sent to them to switch networks; for low-compatibility terminals, the terminal is forcibly de-associated and added to the blacklist of the general virtual interface, while the response channel of the designated dedicated virtual interface is activated, forcibly driving the physical network card of the low-compatibility terminal to converge to the corresponding dedicated virtual interface. This invention achieves MAC layer priority scheduling of terminal uplink traffic by constructing multiple virtual interfaces within a single frequency band of the gateway and binding specific service flows, without the user's awareness.

Description

Technical Field

[0001] This invention relates to the field of wireless network networking technology, and in particular to a terminal adaptive multi-SSID uplink air interface resource control method, device and medium. Background Technology

[0002] In current home and enterprise Wi-Fi wireless network scenarios, latency-sensitive services such as live streaming and cloud computing place extremely high demands on network uplink latency.

[0003] Traditional Wi-Fi Multimedia (WMM) service solutions primarily rely on local queue management on the access point (AP) side to prioritize downlink (AP → STA) traffic. However, due to the distributed channel contention inherent in Wi-Fi networks, the AP cannot directly control the transmission logic on the terminal side (STA → AP). When transmitting high-bit-rate uplink traffic, the terminal still uses a common set of channel contention parameters, leading to severe collisions of sensitive uplink traffic over the air interface. Traditional WMM solutions are thus in a passive state, able to manage downlink but unable to manage uplink.

[0004] Furthermore, the EDCA parameters at the protocol layer are configured based on the terminal as a whole. Even if the AP attempts to modify the terminal's EDCA parameters through the protocol, it will inevitably lead to all uplink traffic of that terminal (including system upgrades and background component synchronization) being elevated in priority simultaneously, making it impossible to achieve fine-grained control at the service flow level and causing internal waste of valuable uplink air interface resources. At the same time, there are a large number of old terminals in the current network with extremely low response rates to the standard 802.11v protocol, causing conventional guidance methods to fail.

[0005] Therefore, the existing technology has at least the following disadvantages: 1. Terminal uplink traffic control is ineffective: The traditional WMM mechanism cannot intervene in the air interface channel contention parameters of the terminal (STA), resulting in severe collisions of uplink video streams for services such as live streaming at the air interface, and unpredictable latency.

[0006] 2. Overly coarse-grained business flow control: EDCA parameters are bound to devices rather than services. Critical uplink flows and non-critical background flows cannot be separated on a single device, leading to increased overhead resource consumption after system-wide privilege escalation. 3. There are blind spots in the guidance of old terminals: The standard 802.11v guidance frame does not have the ability to force scheduling of old terminals, and cannot ensure that uplink high latency sensitive terminals are 100% diverted to the dedicated channel. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide a terminal adaptive multi-SSID uplink air interface resource control method, device and medium. By constructing multiple virtual interfaces within a single frequency band of the gateway and binding specific service flows, the AP can achieve MAC layer priority scheduling of terminal uplink traffic. At the same time, without changing the user's perception, it can achieve standard guidance for highly compatible terminals and directional convergence for less compatible old terminals, thus ensuring the uplink air interface quality of services.

[0008] In a first aspect, the present invention provides a terminal-adaptive multi-SSID uplink air interface resource control method for embedded gateway devices, the method comprising: Multi-SSID uplink parameter isolation configuration process: Virtualize and broadcast at least two virtual interfaces within a single frequency band of the gateway, including at least one general virtual interface and at least one dedicated virtual interface, and write different STA EDCA parameters in the Beacon frames and ProbeResponse frames they send out; the virtual interfaces present to the outside world as Wi-Fi networks with completely identical names and encryption passwords; Business flow monitoring process: Real-time extraction of traffic characteristics of access terminals; when an uplink data flow that meets the specified conditions is captured, a classification prediction instruction is sent to the online learning engine module. Online learning classification process: When a classification prediction instruction is received, the wireless characteristics of the access terminal are obtained, and the compatibility probability value of the access terminal is output through the classification model. Physical layer uplink joint scheduling process: Based on the compatibility probability value output by the engine, a dual-track traffic split is performed: For high-compatibility terminals, a standard 802.11v guide frame is directly sent to them to switch networks; for low-compatibility terminals, the terminal is forcibly deassociated and added to the blacklist of the general virtual interface, while the response channel of the designated dedicated virtual interface is activated, forcibly driving the low-compatibility terminal to converge to the corresponding dedicated virtual interface.

[0009] Furthermore, the method also includes: when it is detected that the uplink data stream under specified conditions has been continuously stopped for a specified time, automatically unblocking the terminal in the blacklist of the general virtual interface, guiding the terminal back to the general virtual interface through the deassociation frame, and releasing the dedicated virtual interface resources.

[0010] Furthermore, during the service flow monitoring process, the compatibility probability value of the access terminal is obtained through the following steps: extracting the terminal's instantaneous RSSI, response latency, and historical guidance success rate as feature vectors to construct a feature matrix; performing a dot product operation between the feature matrix and the hidden layer weight matrix, processing it through an activation function, and then performing a multiplication and addition operation with the output layer weight matrix, finally performing normalization mapping, and inferring and outputting the predicted compatibility probability value.

[0011] Furthermore, the online learning classification process also executes a backpropagation algorithm on the endpoint based on the result of each redirection to dynamically adjust the weight matrix of the classification model.

[0012] Furthermore, during the physical layer uplink joint scheduling process, after the terminal accesses the dedicated virtual interface, the joint physical layer issues a Trigger Frame to perform fine-grained scheduling and isolation of UL-OFDMA resource units in the frequency domain.

[0013] Furthermore, leveraging Wi-Fi 7's multi-link operation capabilities, a designated dedicated virtual interface can be anchored to a 6GHz or high-frequency 5GHz link.

[0014] Secondly, the present invention provides a terminal-adaptive multi-SSID uplink air interface resource control device for embedded gateway devices, the device comprising: The multi-SSID uplink parameter isolation configuration module is used to virtualize and broadcast at least two virtual interfaces within a single frequency band of the gateway, including at least one general virtual interface and at least one dedicated virtual interface. These virtual interfaces present themselves to the outside world as Wi-Fi networks with the same name and encryption password, but each has its own independent hardware BSSID. Business flow monitoring module: Its input end is connected to the data link layer of the general virtual interface in the multi-SSID uplink parameter isolation configuration module. It is used to extract the traffic characteristics of the access terminal in real time through deep packet parsing. When it captures uplink data flow that meets the specified conditions, it sends a classification prediction instruction to the online learning engine module. The online learning engine module is used to receive classification prediction instructions from the service flow monitoring module in real time, then obtain the wireless characteristics of the access terminal, construct a continuous floating-point feature vector, and output the compatibility probability value of the access terminal through the classification model. Physical layer uplink joint scheduling driver module: Its control input is connected to the decision output of the online learning engine module, and its physical control directly acts on the underlying driver of the gateway wireless chip. It is used to perform dual-track traffic splitting based on the compatibility probability value output by the engine: For high-compatibility terminals, it directly sends a standard 802.11v guide frame to them to switch networks; for low-compatibility terminals, it forcibly deassociates the terminal and adds it to the blacklist of the general virtual interface, while activating the response channel of the dedicated virtual interface, forcibly driving the low-compatibility terminal to converge to the dedicated virtual interface.

[0015] Furthermore, the online learning engine module is embedded in the gateway's hardware AI processor, which uses the AI ​​processor's floating-point multiply-accumulate instruction set to perform hardware-level parallel inference and output compatibility probability values.

[0016] Furthermore, the service flow monitoring module is also used to automatically unblock the terminal in the blacklist of the general virtual interface when it is detected that the uplink data flow under specified conditions has been stopped for a specified time, and guide the terminal back to the general virtual interface through the deassociation frame to release the dedicated virtual interface resources.

[0017] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0018] The technical solutions provided in the embodiments of the present invention have at least the following technical effects: 1. Overcoming industry pain points and filling the gap in uplink control: By setting up dedicated virtual interfaces with independent hardware BSSIDs, different traffic priority channels are constructed, overcoming the inherent deficiency of traditional WMM solutions that "can only control downlink QoS and cannot manage terminal uplink priority." Through the combination of "multi-SSID routing isolation + STA EDCA parameter rewriting," the terminal is forced to change its uplink backoff behavior, achieving for the first time fine-grained and absolute priority scheduling of terminal uplink traffic on the AP side.

[0019] 2. Avoid resource consumption caused by whole-machine privilege escalation: Bind the service flow to a dedicated SSID, configure independent and aggressive STA EDCA parameters, and combine physical layer UL-OFDMA trigger frame scheduling to ensure that only uplink data flows that meet the specified conditions (such as "uplink live stream") can enjoy physical layer RU resources, while other background traffic brought by the device is precisely degraded at the physical layer, avoiding air interface chaos caused by traditional whole-machine privilege escalation of EDCA.

[0020] 3. Full compatibility with legacy terminals, eliminating blind spots in the guidance process: The gateway collects terminal response latency, historical success rate, and RSSI metrics in real time and introduces on-side AI for online incremental training to dynamically predict the terminal's compatibility probability with 802.11v wireless guidance. Based on the prediction results, it adaptively switches between a dual-track strategy of "seamless redirection of 11v standard frames" and "directly sending Deauth deassociation frames + strong kicking using a general SSID MAC blacklist" to ensure that devices of all generations can be successfully redirected to the corresponding dedicated channel.

[0021] 4. Seamlessly resolves redirection loops for older terminals: Employing a dynamic multi-VIF architecture with "same name, same encryption" and gateway-side targeted interception technology, it forces low-compatibility older terminals to be redirected without requiring users to manually enter passwords or changing user perception. This eliminates the risk of direct disconnection or dead loops caused by unsaved passwords in traditional solutions.

[0022] 5. Edge-side hardware and software collaborative online closed-loop update: Fully utilize the floating-point acceleration feature of the gateway's built-in AI processor, enabling lightweight neural network classification models to directly perform backpropagation (SGD) on the embedded edge based on guided instant positive and negative feedback (Label), achieving true edge-side online incremental learning and improving the prediction accuracy for unknown / non-standard terminals.

[0023] 6. Joint control and protection across MAC and physical layers: It breaks through the limitations of traditional methods that only modify EDCA parameters or perform physical layer scheduling. It achieves privilege escalation for specific BSSIDs at the MAC layer and performs fine-grained RU isolation in combination with Trigger frames at the physical layer, opening up an absolutely clean dedicated air interface channel for critical services such as live streaming.

[0024] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0026] Figure 1 This is a schematic diagram of the embedded gateway system of the present invention; Figure 2 This is a flowchart illustrating the overall process of the method in Embodiment 1 of the present invention. Figure 3 This is a data flow diagram of online floating-point training and inference on the AI ​​processor side in Embodiment 1 of the present invention; Figure 4 This is a flowchart of the redirection signaling process for old terminals under the same SSID state in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the device in Embodiment 2 of the present invention. Detailed Implementation

[0027] This invention provides a terminal-adaptive multi-SSID uplink air interface resource control method, device, and medium. By constructing multiple virtual interfaces within a single frequency band of the gateway and binding specific service flows, it enables the AP to perform MAC layer priority scheduling of terminal uplink traffic. At the same time, without changing the user's perception, it achieves standard guidance for highly compatible terminals and directional convergence for less compatible older terminals, thus ensuring the uplink air interface quality of services.

[0028] The overall concept of the technical solutions in the embodiments of the present invention is as follows: This invention aims to address the limitations of traditional WMM schemes that cannot control the uplink priority of terminals. It explores how to leverage a multi-SSID architecture to forcibly intervene in and rewrite the uplink air interface contention parameters of the terminal (STA), and combine end-side AI adaptive guidance with physical layer hard scheduling to achieve fine-grained uplink priority scheduling of internal service flows within the terminal.

[0029] Innovation Point 1 (Architecture Level): Broadcast multiple virtual interfaces (VIFs) with the same SSID name and security encryption credentials but different base station identification codes (BSSID) within a single frequency band. By rewriting the STA EDCA parameters of the Beacon frame and Probe Response frame of a specific BSSID and binding a specific service flow, the AP can achieve MAC layer priority scheduling of terminal uplink traffic.

[0030] Innovation Point 2 (Algorithm Level): Utilizing the hardware floating-point instruction set (FP16 / FP32) of the gateway's built-in AI processor, the input is composed of features such as instantaneous RSSI, response latency, and historical boot success rate. The model is incrementally trained and updated through online backpropagation on the terminal side, and the output is the terminal protocol compatibility probability.

[0031] Innovation Point 3 (Mechanism Level): A dual-track directional roaming convergence control mechanism for low-compatibility and outdated terminals, consisting of "unlinking and force kicking + blocking other VIF hardware MAC blacklists with the same name + exclusive uplink VIF probe response".

[0032] Innovation Point 4 (Physical Scheduling Level): In a network environment with the same name and density, after the terminal accesses the dedicated VIF, the MAC layer EDCA parameters are forcibly activated, and the physical layer issues Trigger Frames to specify frequency domain resource units (RUs) for cross-layer joint scheduling of specific latency-sensitive live streams.

[0033] Deployed in an embedded gateway device equipped with a micro AI processor (supporting FP16 / FP32 floating-point instruction sets), see system block diagram. Figure 1 .

[0034] Example 1 This embodiment provides a terminal-adaptive multi-SSID uplink air interface resource control method, such as... Figure 2 As shown, this is for an embedded gateway device. The gateway supports configuration of multiple virtual interfaces (VIF / SSID) in a single frequency band. Preferably, it also supports uplink OFDMA (UL-OFDMA) scheduling, Wi-Fi 6 / Wi-Fi 7 protocols, and multi-link operation (MLO). The method includes: S1. The multi-SSID uplink parameter isolation configuration process virtualizes and broadcasts at least two virtual interfaces (VIF / SSID) within a single frequency band of the gateway, including at least one general virtual interface and at least one dedicated virtual interface, and writes different STA EDCA (terminal-side enhanced distributed channel access) parameters into the Beacon frame and Probe Response frame sent by them; the virtual interface presents itself to the outside world as a Wi-Fi network with the same name and encryption password.

[0035] Typically, a first virtual interface is virtualized as a general-purpose virtual interface, and one or more second virtual interfaces are virtual dedicated interfaces (each virtual interface can connect multiple terminals simultaneously). When there is more than one virtual dedicated interface, the STA EDCA parameters of different interfaces are different, thus defining different priorities (4 VIFs can define 4 different priorities).

[0036] For example, when a dedicated virtual interface (such as VIF1) is used as a dedicated uplink channel, the Beacon frames of this dedicated virtual interface are hard-coded with modified STA EDCA parameters with extremely small contention windows, reducing the CWmin (minimum contention window), CWmax, and AIFSN (arbitration frame interval) of the terminal's dedicated uplink AC_VI (video) and AC_VO (voice) queues to the minimum values ​​of the protocol baseline.

[0037] When a dedicated virtual interface (such as VIF4) is used as a background download channel, the STA EDCA parameter in the Beacon frame of the dedicated virtual interface is increased, significantly increasing its CWmin and AIFSN, making it an "extremely modest" channel at the MAC layer.

[0038] S2. Business Flow Monitoring Process: Deep Packet Analysis (DPI) is performed on the data link layer of the general virtual interface to extract the traffic characteristics of the access terminal in real time. When an uplink data flow that meets specified conditions (such as bursty or high-latency sensitive) is captured, a classification prediction instruction is immediately sent to the online learning engine module. For example, when a terminal initiates an uplink live broadcast or interaction, the monitoring module determines that the terminal is currently in an uplink latency sensitive state based on the continuous large packet uplink throughput and strong interaction cycle characteristics.

[0039] S3. Online learning classification process: Upon receiving a classification prediction instruction, the wireless features of the access terminal are acquired, a continuous floating-point feature vector is constructed, and the compatibility probability value P of the access terminal is output through the classification model. comp This process is seamlessly embedded in the gateway's hardware AI processor, utilizing the AI ​​processor's FP16 / FP32 floating-point multiply-accumulate instruction set to perform hardware-level parallel inference and output a normalized compatibility probability value P. comp .

[0040] In one specific embodiment, the following steps may be included: S31. Feature Matrix Construction: When the uplink traffic monitoring triggers the AI ​​processor's classification and prediction instruction, the processor extracts three continuous wireless features of the terminal from the gateway's physical memory: RSSI (Real-time Received Signal Strength Indicator, single-precision floating-point number FP32, linearly normalized and mapped to the [0.1] interval), ΔT (Average historical interaction response delay, in ms, converted to a continuous floating-point value [0,1] using a bucketing function), P... succ (The success rate of the guidance within the historical sliding window is directly read from the floating-point rate value of [0,1].) A one-dimensional feature matrix is ​​constructed directly in the physical registers of the AI ​​processor from the above three normalized features. =[RSSI,ΔT,P succ ].

[0041] S32, Hardware-level Forward Inference: The AI ​​processor calls its internally integrated FP16 / FP32 floating-point multiply-accumulate (MAC) instruction set to perform a parallel dot product operation between the feature matrix and the hidden layer weight matrix Whidden, read from the parameter storage area. The calculation result is processed by the ReLU activation function through a hardware-level latch, and then compared with the output layer weight matrix W. output After performing multiplication and addition operations, the scalar P is inferred and output through normalization mapping using the Sigmoid activation function. comp If it falls within the |0,1| range, it represents the predicted compatibility probability value of the terminal with the 802.11v bootloader protocol.

[0042] S33, Backpropagation and Gradient Training: The training update engine immediately captures the positive sample label (Y=1) or negative sample label (Y=0) of the final feedback after redirection, and calculates the loss value of the current inference using binary cross-entropy within the AI ​​processor.

[0043] Loss function: Cross-entropy Loss = -Y×Log(P_comp) - (1-Y)×Log(1-P_comp) Then, the stochastic gradient descent (SGD) algorithm is used to perform backpropagation on the physical side to directly calculate the gradient partial derivatives of the weight matrix under the current feature, and immediately update the weight matrix in the parameter storage area by overwriting.

[0044] The gateway's built-in micro AI processor can be used. Please refer to the data flow diagram for the AI ​​processor's edge floating-point online training and inference. Figure 3 .

[0045] S4. Physical Layer Uplink Joint Scheduling Driven Process: Based on the compatibility probability value output by the engine, a dual-track traffic split is executed: For high-compatibility terminals, a standard 802.11v boot frame is directly sent to guide the terminal to seamlessly switch to the designated dedicated virtual interface; for low-compatibility terminals, the gateway skips the 02.11v boot frame attempt and instead sends a deauthentication frame to the terminal to forcibly disconnect it. At the same time, the terminal is added to the blacklist of the general virtual interface, and the response channel of the designated dedicated virtual interface is activated, forcibly driving the physical network card of the low-compatibility terminal to converge to the dedicated BSSID of the corresponding dedicated virtual interface.

[0046] Specifically, the gateway kernel driver has a dynamic blacklist for blocking Probe Requests for non-dedicated BSSIDs, forcibly controlling the hardware RF response status of non-dedicated VIFs. This allows the gateway-side dynamic Probe Request filtering mechanism to be activated, and the MAC addresses of low-compatibility terminals to be written into the hardware-level MAC filtering blacklist of the general virtual interface.

[0047] After the terminal accesses the dedicated uplink virtual interface through the redirection mechanism and activates the high-speed STA EDCA parameters, the gateway further allocates a dedicated resource unit (RU) in the physical frequency domain only for the specific uplink live service flow inside the terminal by issuing a UL-OFDMA trigger frame. Combined with Wi-Fi 7 MLO, the dedicated SSID is hard isolated on the high-frequency link, thereby stripping and suppressing other non-critical uplink traffic inside and outside the terminal at the physical layer.

[0048] To better understand this method, the following specific example illustrates the redirection process of older terminals under SSID synchronization status: Step 1 (Stream Identification Trigger): The terminal (STA) initially associates with the gateway's general interface VIF0 and initiates a high-bitrate uplink video live stream. The gateway's uplink service flow monitoring module acquires this traffic characteristic and adaptively guides the online learning engine to call the AI ​​processor's floating-point computing power acceleration channel for local online inference. It determines that the terminal has an extremely low probability of compatibility with the standard 802.11v protocol and marks it as a "low-compatibility outdated terminal".

[0049] Step 2 (Physical Disconnection and Blacklist Distribution): The gateway VIF0 proactively sends a Deauthentication frame to the terminal, forcibly disconnecting the terminal's current wireless connection. Subsequently, the gateway kernel immediately writes the terminal's MAC address into the hardware-level MAC filtering blacklist within the VIF0 space, causing VIF0 to cease any physical layer handshake responses to the terminal.

[0050] Step 3 (Active Rescan for Networks with the Same Name and Password): After being forcibly kicked out, the Wi-Fi state machine of the terminal's operating system triggers an automatic rescan mechanism. The terminal broadcasts a Probe Request frame over the air to search for previously saved networks with the same SSID.

[0051] Step 4 (Gateway-side Probe Dynamic Response Interception): When the gateway receives the Probe Request frame from the terminal, since VIF0 has already added the terminal to its blacklist, the gateway's underlying chip driver will forcibly discard and intercept VIF0's response to the frame. Simultaneously, the dedicated virtual interface VIF1 (as a dedicated uplink interface) is in an open state, exclusively using VIF1 to reply to the terminal with a Probe Response frame. This response frame carries VIF1's unique STA EDCA IE, which reduces the contention window to the protocol's minimum threshold.

[0052] Step 5 (Directed Convergence and Seamless Access): Since the SSID and encryption password of the two virtual interfaces are completely identical, the terminal's physical layer roaming and reconnection logic will not trigger any "user needs to re-enter password" pop-up prompts. Without the terminal being aware of the connection, its network card driver will be forcibly converged by the gateway's one-way response mechanism, directly sending an Association Request frame to VIF1 and quickly connecting.

[0053] Step 6 (Parameter Activation and Priority Scheduling): VIF1 replies with an Association Response frame, and the terminal successfully associates with the dedicated uplink channel. Once the terminal is associated with VIF1, its Wi-Fi network card must strictly adhere to the ultra-fast EDCA parameters specified in the VIF1 Beacon frame. The backoff time for uplink channel contention initiated by the terminal itself is drastically reduced, obtaining the highest priority uplink "queue-jumping" privilege from the MAC layer mechanism, completely breaking the deadlock that traditional WMM cannot manage uplink priority. To prevent other background flows within the terminal (such as system upgrades) from also jumping the queue, the gateway chip performs precise interception at the physical layer: when it detects that VIF1 has an uplink data transmission requirement, the AP actively sends a Trigger Frame, allocating the optimal and continuous RU (Resource Unit) only for the uplink live stream within this terminal in the frequency domain, realizing centralized timed and frequency-fixed scheduling, and completely eliminating flow collisions within the terminal. It can also utilize the multi-link operation capability of Wi-Fi 7 to hard anchor the dedicated live SSID (VIF 1) to a 6GHz or high-frequency 5GHz link with minimal interference and extremely low latency, achieving absolute zero interference in the physical frequency band.

[0054] Step 7 (Reverse Switchback Release): When the critical uplink service flow of the terminal is detected to stop for a preset time (e.g., 3-5 minutes), the gateway automatically unblocks the hardware filtering blacklist of the general interface VIF0 and guides the terminal back to the general VIF0 through the unassociation frame, thereby releasing the dedicated latency channel resources.

[0055] Please refer to the above signaling interaction process. Figure 4 .

[0056] Based on the same inventive concept, this application also provides an apparatus corresponding to the method in Embodiment 1, as detailed in Embodiment 2.

[0057] Example 2 This embodiment provides a terminal-adaptive multi-SSID uplink air interface resource control device, such as... Figure 5 As shown, for an embedded gateway device, the apparatus includes: The multi-SSID uplink parameter isolation configuration module is used to virtualize and broadcast at least two virtual interfaces within a single frequency band of the gateway, including at least one general virtual interface and at least one dedicated virtual interface, and write different STA EDCA parameters in the Beacon frames and Probe Response frames it sends out; these virtual interfaces present to the outside world as a Wi-Fi network with the same name and encryption password. Business flow monitoring module: Its input end is connected to the data link layer of the general virtual interface in the multi-SSID uplink parameter isolation configuration module. It is used to extract the traffic characteristics of the access terminal in real time through deep packet parsing. When it captures uplink data flow that meets the specified conditions, it sends a classification prediction instruction to the online learning engine module. The online learning engine module is used to receive classification prediction instructions from the service flow monitoring module in real time, then obtain the wireless characteristics of the access terminal, construct a continuous floating-point feature vector, and output the compatibility probability value of the access terminal through the classification model. Physical layer uplink joint scheduling driver module: Its control input is connected to the decision output of the online learning engine module, and its physical control directly acts on the underlying driver of the gateway wireless chip. It is used to perform dual-track traffic splitting based on the compatibility probability value output by the engine: For high-compatibility terminals, it directly sends a standard 802.11v guide frame to them to switch networks; for low-compatibility terminals, it forcibly deassociates the terminal and adds it to the blacklist of the general virtual interface, while activating the response channel of the dedicated virtual interface, forcibly driving the low-compatibility terminal to converge to the dedicated virtual interface.

[0058] Preferably, the online learning engine module is embedded in the hardware AI processor of the gateway, and uses the floating-point multiply-accumulate instruction set of the AI ​​processor to perform hardware-level parallel inference and output compatibility probability values.

[0059] Preferably, the service flow monitoring module is further configured to automatically unblock the terminal in the blacklist of the general virtual interface after detecting that the uplink data flow under specified conditions has been continuously stopped for a specified time, guide the terminal back to the general virtual interface through the deassociation frame, and release the dedicated virtual interface resources.

[0060] By default, user terminals wirelessly associate with the universal interface VIF0 generated by the multi-SSID uplink parameter isolation configuration module. The uplink service flow monitoring module collects the traffic characteristics on this link in real time and outputs its data stream to the adaptive guided online learning engine (physically connected to the AI ​​processor chip of the gateway). The control signals after inference and training by the AI ​​processor are sent to the physical layer uplink joint scheduling driver module (which directly acts on the underlying driver of the gateway's wireless radio frequency chip). The driver layer controls the BSSID hardware switch of different virtual interfaces at the physical layer, ultimately forcing the physical network cards of low-compatibility old terminals to re-converge and associate with the dedicated virtual interface VIF1.

[0061] Since the apparatus described in Embodiment 2 of the present invention is an apparatus used to implement the method of Embodiment 1 of the present invention, those skilled in the art can understand the specific structure and variations of the apparatus based on the method described in Embodiment 1 of the present invention, and therefore will not be described again here. All apparatuses used in the method of Embodiment 1 of the present invention fall within the scope of protection of the present invention.

[0062] Based on the same inventive concept, this application provides a storage medium corresponding to Embodiment 1, as detailed in Embodiment 3.

[0063] Example 3 This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it can implement any of the implementation methods in Embodiment 1.

[0064] Since the computer-readable storage medium described in this embodiment is the same computer-readable storage medium used to implement the method in Embodiment 1 of this application, those skilled in the art can understand the specific implementation methods and various variations of the computer-readable storage medium in this embodiment based on the method described in Embodiment 1 of this application. Therefore, how this computer-readable storage medium implements the method in the embodiments of this application will not be described in detail here. Any computer-readable storage medium used by those skilled in the art to implement the method in the embodiments of this application falls within the scope of protection of this application.

[0065] The technical solutions provided in the embodiments of the present invention have at least the following technical effects: 1. Overcoming industry pain points and filling the gap in uplink control: By setting up dedicated virtual interfaces with independent hardware BSSIDs, different traffic priority channels are constructed, overcoming the inherent deficiency of traditional WMM solutions that "can only control downlink QoS and cannot manage terminal uplink priority." Through the combination of "multi-SSID routing isolation + STA EDCA parameter rewriting," the terminal is forced to change its uplink backoff behavior, achieving for the first time fine-grained and absolute priority scheduling of terminal uplink traffic on the AP side.

[0066] 2. Avoid resource consumption caused by whole-machine privilege escalation: Bind the service flow to a dedicated SSID, configure independent and aggressive STA EDCA parameters, and combine physical layer UL-OFDMA trigger frame scheduling to ensure that only uplink data flows that meet the specified conditions (such as "uplink live stream") can enjoy physical layer RU resources, while other background traffic brought by the device is precisely degraded at the physical layer, avoiding air interface chaos caused by traditional whole-machine privilege escalation of EDCA.

[0067] 3. Full compatibility with legacy terminals, eliminating blind spots in the guidance process: The gateway collects terminal response latency, historical success rate, and RSSI metrics in real time and introduces on-side AI for online incremental training to dynamically predict the terminal's compatibility probability with 802.11v wireless guidance. Based on the prediction results, it adaptively switches between a dual-track strategy of "seamless redirection of 11v standard frames" and "directly sending Deauth deassociation frames + strong kicking using a general SSID MAC blacklist" to ensure that devices of all generations can be successfully redirected to the corresponding dedicated channel.

[0068] 4. Seamlessly resolves redirection loops for older terminals: Employing a dynamic multi-VIF architecture with "same name, same encryption" and gateway-side targeted interception technology, it forces low-compatibility older terminals to be redirected without requiring users to manually enter passwords or changing user perception. This eliminates the risk of direct disconnection or dead loops caused by unsaved passwords in traditional solutions.

[0069] 5. Edge-side hardware and software collaborative online closed-loop update: Fully utilize the floating-point acceleration feature of the gateway's built-in AI processor, enabling lightweight neural network classification models to directly perform backpropagation (SGD) on the embedded edge based on guided instant positive and negative feedback (Label), achieving true edge-side online incremental learning and improving the prediction accuracy for unknown / non-standard terminals.

[0070] 6. Joint control and protection across MAC and physical layers: It breaks through the limitations of traditional methods that only modify EDCA parameters or perform physical layer scheduling. It achieves privilege escalation for specific BSSIDs at the MAC layer and performs fine-grained RU isolation in combination with Trigger frames at the physical layer, opening up an absolutely clean dedicated air interface channel for critical services such as live streaming.

[0071] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0073] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0074] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0075] While specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and not intended to limit the scope of the invention. Equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A terminal-adaptive multi-SSID uplink air interface resource control method, characterized in that, For an embedded gateway device, the method includes: Multi-SSID uplink parameter isolation configuration process: Virtualize and broadcast at least two virtual interfaces within a single frequency band of the gateway, including at least one general virtual interface and at least one dedicated virtual interface, and write different STA EDCA parameters in the Beacon frames and ProbeResponse frames they send out; the virtual interfaces present to the outside world as Wi-Fi networks with completely identical names and encryption passwords; Business flow monitoring process: Real-time extraction of traffic characteristics of access terminals; when an uplink data flow that meets the specified conditions is captured, a classification prediction instruction is sent to the online learning engine module. Online learning classification process: When a classification prediction instruction is received, the wireless characteristics of the access terminal are obtained, and the compatibility probability value of the access terminal is output through the classification model. Physical layer uplink joint scheduling process: Based on the compatibility probability value output by the engine, a dual-track traffic split is performed: For high-compatibility terminals, a standard 802.11v guide frame is directly sent to them to switch networks; for low-compatibility terminals, the terminal is forcibly deassociated and added to the blacklist of the general virtual interface, while the response channel of the designated dedicated virtual interface is activated, forcibly driving the low-compatibility terminal to converge to the corresponding dedicated virtual interface.

2. The method according to claim 1, characterized in that, The method further includes: when it is detected that the uplink data stream under specified conditions has been continuously stopped for a specified time, automatically unblocking the terminal in the blacklist of the general virtual interface, guiding the terminal back to the general virtual interface through the deassociation frame, and releasing the dedicated virtual interface resources.

3. The method according to claim 1, characterized in that: During the service flow monitoring process, the compatibility probability value of the access terminal is obtained through the following steps: extract the terminal's instantaneous RSSI, response latency, and historical boot success rate as feature vectors, and construct a feature matrix; The feature matrix and the hidden layer weight matrix are multiplied together, processed by the activation function, and then multiplied and added together with the output layer weight matrix. Finally, a normalized mapping is performed, and the predicted compatibility probability value is output for inference.

4. The method according to claim 1 or 3, characterized in that: The online learning classification process also executes a backpropagation algorithm on the device side based on the result of each redirection to dynamically adjust the weight matrix of the classification model.

5. The method according to claim 1, characterized in that: During the physical layer uplink joint scheduling process, after the terminal accesses the dedicated virtual interface, the joint physical layer issues a Trigger Frame to perform fine-grained scheduling and isolation of UL-OFDMA resource units in the frequency domain.

6. The method according to claim 1, characterized in that: Leveraging Wi-Fi 7's multi-link operation capabilities, a designated dedicated virtual interface can be anchored to a 6GHz or high-frequency 5GHz link.

7. A terminal-adaptive multi-SSID uplink air interface resource control device, characterized in that, For an embedded gateway device, the apparatus includes: The multi-SSID uplink parameter isolation configuration module is used to virtualize and broadcast at least two virtual interfaces within a single frequency band of the gateway, including at least one general virtual interface and at least one dedicated virtual interface, and write different STA EDCA parameters in the Beacon frames and ProbeResponse frames it sends out; these virtual interfaces present to the outside world as a Wi-Fi network with the same name and encryption password. Business flow monitoring module: Its input end is connected to the data link layer of the general virtual interface in the multi-SSID uplink parameter isolation configuration module. It is used to extract the traffic characteristics of the access terminal in real time through deep packet parsing. When it captures uplink data flow that meets the specified conditions, it sends a classification prediction instruction to the online learning engine module. The online learning engine module is used to receive classification prediction instructions from the service flow monitoring module in real time, then obtain the wireless characteristics of the access terminal, construct a continuous floating-point feature vector, and output the compatibility probability value of the access terminal through the classification model. Physical layer uplink joint scheduling driver module: Its control input is connected to the decision output of the online learning engine module, and its physical control directly acts on the underlying driver of the gateway wireless chip. It is used to perform dual-track traffic splitting based on the compatibility probability value output by the engine: For high-compatibility terminals, it directly sends a standard 802.11v guide frame to them to switch networks; for low-compatibility terminals, it forcibly deassociates the terminal and adds it to the blacklist of the general virtual interface, while activating the response channel of the designated dedicated virtual interface, forcibly driving the low-compatibility terminal to converge to the corresponding dedicated virtual interface.

8. The apparatus according to claim 7, characterized in that: The online learning engine module is embedded in the gateway's hardware AI processor. It uses the AI ​​processor's floating-point multiply-accumulate instruction set to perform hardware-level parallel inference and output compatibility probability values.

9. The apparatus according to claim 7, characterized in that: The service flow monitoring module is also used to automatically unblock the terminal in the blacklist of the general virtual interface when it is detected that the uplink data flow under specified conditions has been stopped for a specified time, and guide the terminal back to the general virtual interface through the deassociation frame to release the dedicated virtual interface resources.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 6.

Images