Millimeter-wave communication methods, systems, and computer program products

By introducing low-frequency link-assisted cross-band cooperative beam management into millimeter-wave communication and combining environmental perception data to predict candidate beam range, the problems of high beam training overhead and slow link recovery are solved, achieving efficient beam training and fast link recovery.

CN123122601APending Publication Date: 2026-11-17ZTE CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510628471.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-15
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing millimeter-wave communication suffers from high beam training overhead and slow link recovery, especially in dynamic environments where frequent repetitive training leads to huge signaling and time overhead.

Method used

By establishing a low-frequency link between the access point and the terminal, and using the low-frequency link to assist in the management of the millimeter-wave link, combined with environmental perception data such as inertial measurement units, cameras, and location information, the candidate beam range of the millimeter-wave link is predicted, and the control information of the millimeter-wave link is offloaded through the low-frequency link to achieve cross-frequency band collaborative beam management.

Benefits of technology

It significantly reduces the overhead of millimeter-wave beam training, improves beam training efficiency, and enables fast link adaptation and seamless recovery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122601A_ABST
    Figure CN123122601A_ABST
Patent Text Reader

Abstract

This invention provides a millimeter-wave communication method, system, and computer program product. A first device receives first information from a second device via a low-frequency link. Based on the first information, the first device determines a candidate beam range for the millimeter-wave link communicating with the second device. Based on the candidate beam range, the first device initiates beam training in the millimeter-wave frequency band to determine the millimeter-wave link beam used for millimeter-wave communication. This solves the problems of high beam training overhead and slow link recovery in related technologies, achieving improved beam training efficiency and reduced millimeter-wave link control overhead.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of communications, and more specifically, to a millimeter-wave communication method, system, and computer program product. Background Technology

[0002] The 802.11 series of wireless LAN standards has introduced millimeter-wave (MW) technology, overcoming the high path loss of MW waves with high-gain narrow beams to achieve data rates of multiple Gbps. However, due to the high directivity of MW signals, precise beam alignment is required at both the transmitter and receiver. The existing IEEE 802.11ad / ay standard employs a two-stage beam training process: first, a wide beam is used for sector sweeping (supporting up to 128 discrete directions) to find the optimal coarse beam; then, fine-grained beam optimization is performed based on this coarse beam. The training overhead of this process increases linearly with the number of participating devices and antenna sectors. Frequent repetition of training is required in dynamic environments such as movement or rotation, leading to significant signaling and time overhead. For example, even a slight directional shift can cause a significant drop in link budget or even link failure, greatly increasing the complexity of maintaining MW links.

[0003] Related technologies suffer from problems such as high beam training overhead and slow link recovery. Summary of the Invention

[0004] This invention provides a millimeter-wave communication method, system, and computer program product to at least solve the problems of high beam training overhead and slow link recovery in related technologies.

[0005] According to an embodiment of the present invention, a millimeter-wave communication method is provided, comprising: a first device receiving first information from a second device via a low-frequency link; the first device determining a candidate beam range for a millimeter-wave link communicating with the second device based on the first information; and the first device initiating beam training in a millimeter-wave frequency band based on the candidate beam range to determine a millimeter-wave link beam for millimeter-wave communication.

[0006] According to another embodiment of the present invention, a millimeter-wave communication method is provided, comprising: a second device sending first information to a first device via a low-frequency link; the second device performing beam training in a millimeter-wave frequency band based on a candidate beam range indicated by the first device to determine a millimeter-wave link beam for millimeter-wave communication, wherein the candidate beam range is determined by the first device based on the first information.

[0007] According to another embodiment of the present invention, a millimeter-wave communication system is also provided, comprising: a terminal and an access point (AP), wherein a communication connection between the AP and the terminal is established between a low-frequency link and a millimeter-wave link, the AP receives first information from the terminal through the low-frequency link, and determines a candidate beam range of the millimeter-wave link for communication with the terminal based on the first information; the AP and the terminal initiate beam training in the millimeter-wave frequency band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication.

[0008] According to yet another embodiment of the present invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is configured to perform the steps in any of the above method embodiments when it is run.

[0009] According to yet another embodiment of the present invention, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0010] According to yet another embodiment of the present invention, a computer program product is also provided, comprising a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0011] The present invention provides a millimeter-wave communication method through the above embodiments. A first device receives first information from a second device via a low-frequency link. Based on the first information, the first device determines a candidate beam range for the millimeter-wave link communicating with the second device. Based on the candidate beam range, the first device initiates beam training in the millimeter-wave frequency band to determine the millimeter-wave link beam used for millimeter-wave communication. This method solves the problems of high beam training overhead and slow link recovery in related technologies, achieving improved beam training efficiency and reduced millimeter-wave link control overhead. Attached Figure Description

[0012] Figure 1 This is a hardware structure block diagram of a mobile terminal for a millimeter-wave communication method according to an embodiment of the present invention.

[0013] Figure 2 This is a flowchart of a millimeter-wave communication method according to an embodiment of the present invention;

[0014] Figure 3 This is another flowchart of the millimeter-wave communication method according to an embodiment of the present invention;

[0015] Figure 4 This is a hardware architecture diagram of the millimeter-wave communication method implemented according to an embodiment of the present invention;

[0016] Figure 5 This is a flowchart of beam training according to an embodiment of the present invention;

[0017] Figure 6 This is a flowchart illustrating the transmission of control information via an anchored link according to an embodiment of the present invention;

[0018] Figure 7 This is a flowchart of beam training based on historical CSI assistance according to an embodiment of the present invention;

[0019] Figure 8 This is a flowchart of an embodiment of the present invention for optimizing millimeter-wave beam training based on IMU attitude data;

[0020] Figure 9 This is a flowchart of vision-assisted millimeter-wave beam selection according to an embodiment of the present invention;

[0021] Figure 10 This is a flowchart of dynamic beam training based on machine learning, according to an embodiment of the present invention.

[0022] Figure 11 This is a flowchart of multi-terminal collaborative beam training according to an embodiment of the present invention;

[0023] Figure 12 This is a flowchart of millimeter-wave link quality monitoring and recovery according to an embodiment of the present invention;

[0024] Figure 13 This is a flowchart illustrating the selection of a backup beam using an intelligent selection algorithm, according to an embodiment of the present invention. Detailed Implementation

[0025] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0027] With the introduction of Multi-Link Operation (MLO) and sub-7GHz anchored links in next-generation Wi-Fi standards (such as 802.11be and the developing 802.11bq), researchers have proposed schemes to utilize low-frequency bands to assist millimeter-wave link management. One approach involves establishing an "anchored link" between the access point (AP) and the station (STA), offloading some control information from the millimeter-wave link to the sub-7GHz link to alleviate the overhead of millimeter-wave beacons and management frames. Utilizing the sub-7GHz Wi-Fi link to predict the optimal millimeter-wave beam and rate settings in real-time enables rapid link adaptation. Passively inferring the communication direction using sub-7GHz signals significantly reduces the overhead of millimeter-wave beam training without requiring additional protocol overhead. Furthermore, machine learning methods combining historical Channel State Information (CSI) and various sensor information (such as Inertial Measurement Unit (IMU), camera, and location information) can be used to predict the optimal beam direction for millimeter-wave links.

[0028] The methods and embodiments provided in this application can be executed on a mobile terminal, a computer terminal, or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal for a millimeter-wave communication method according to an embodiment of the present invention. Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0029] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the millimeter-wave communication method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0030] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0031] This embodiment provides a millimeter-wave communication method. Figure 2 This is a flowchart of a millimeter-wave communication method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps:

[0032] In step S202, the first device receives first information from the second device via a low-frequency link.

[0033] In one exemplary embodiment, the first device and the second device are multi-link devices, the first device includes an access point (AP) device, and the second device includes a terminal device.

[0034] In one exemplary embodiment, the first information includes at least one of the following: channel information acquired by the second device via a low-frequency link; broadcast measurement information acquired by the second device via a low-frequency link; historical channel information of the second device; attitude data information of the inertial measurement unit (IMU) of the second device; visual information of the second device; and position information of the first device and / or the second device.

[0035] In this embodiment of the invention, the first device determines a candidate beam, a candidate beam range, or a candidate beam set based on first information from the second device. The first information includes, but is not limited to, at least one of the following: channel information, broadcast measurement information, and environmental awareness information. The environmental awareness information includes, but is not limited to, at least one of the following: IMU attitude data information, visual information, and location information.

[0036] In one exemplary embodiment, the low-frequency link includes a sub-7GHz link.

[0037] In one exemplary embodiment, before the first device receives first information from the second device via a low-frequency link, the method further includes: the first device and the second device establishing a communication connection on the low-frequency link.

[0038] In this embodiment of the invention, both the first device and the second device have a sub-7GHz link and a millimeter-wave link.

[0039] In step S204, the first device determines the candidate beam range of the millimeter-wave link communicating with the second device based on the first information.

[0040] In one exemplary embodiment, in response to the first information including IMU attitude data information, the first device receives a first predicted candidate beam range from a second device, wherein the first predicted candidate beam range is determined by the second device based on the IMU attitude data information.

[0041] In this embodiment of the invention, the IMU can sense the orientation and motion state of the terminal (i.e., the second device mentioned above) and predict the current possible communication path by combining it with the prior antenna orientation. Adjusting the millimeter-wave beam training direction according to the IMU attitude can limit the training search range to a smaller angular region, thereby reducing the number of training sessions and latency of the millimeter-wave link.

[0042] In one exemplary embodiment, in response to the first information including visual information, the method further includes: the first device receiving a second predicted candidate beam range from the second device, wherein the second predicted candidate beam range is determined by the second device based on the visual information.

[0043] In this embodiment of the invention, based on vision-assisted millimeter-wave beam selection, the cameras of the first and / or second devices can capture images of the surrounding scene and identify the location of the communication peer or environmental obstacles through visual algorithms. Based on the image processing results, the relative position and line-of-sight direction between the devices can be directly inferred, thereby predicting the most likely millimeter-wave beam direction. This method does not rely entirely on channel detection, significantly reducing the number of beam training iterations and improving stability.

[0044] Step S206: The first device initiates beam training in the millimeter-wave frequency band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication.

[0045] In one exemplary embodiment, a first device initiates beam training in the millimeter-wave band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication. This includes: the first device initiating sector-level sweep (SLS) beam training in the millimeter-wave band and receiving short sector sweep (SSW) feedback from a second device; based on the SSW feedback, the first device and the second device perform interactive beam training via a beam refinement protocol (BRP) to determine the millimeter-wave link beam for millimeter-wave communication.

[0046] In this embodiment of the invention, beam training on a millimeter-wave link includes coarse alignment and fine optimization stages. Through a hierarchical training logic of coarse scanning and fine optimization, both efficiency and accuracy can be balanced.

[0047] In one exemplary embodiment, after the first device initiates beam training in the millimeter-wave band based on the candidate beam range and determines the millimeter-wave link beam for millimeter-wave communication, the method further includes: updating historical channel information based on the results of the beam training.

[0048] In this embodiment of the invention, the CSI database is updated based on the training results for future use.

[0049] In one exemplary embodiment, the method further includes: the first device sending control information of the millimeter-wave link to the second device via the anchor link, and receiving millimeter-wave link capability information of the second device, wherein the anchor link is a sub-7GHz link configured by the first device.

[0050] In this embodiment of the invention, control information is transmitted via an anchor link. By offloading control information from the millimeter-wave link to the sub-7GHz anchor link, broadcast and beacon overhead on the millimeter-wave channel can be reduced. The AP (i.e., the first device mentioned above) configures and notifies each terminal (i.e., the second device mentioned above) to establish an anchor connection on a specific sub-7GHz link. All control information such as beam training requests, feedback, and link status reports for the millimeter-wave link is exchanged through this anchor link. Only necessary beam alignment and data transmission are performed on the millimeter-wave link, thereby significantly reducing the transmission burden of beacons and management frames. This is applicable to multi-link device systems where the AP and multiple terminals can communicate simultaneously on multiple sub-7GHz and millimeter-wave links. This embodiment demonstrates how to utilize a sub-7GHz link as an "anchor link" to assist in managing the control and beam training process of the millimeter-wave link.

[0051] In one exemplary embodiment, in response to a first device providing millimeter-wave connectivity to a plurality of second devices, the method further includes: the first device determining, based on the first information of each second device, the candidate beam range of the millimeter-wave link communicating with each second device; the first device coordinating and triggering beam training of the plurality of second devices on a low-frequency link, and determining the millimeter-wave link beam for millimeter-wave communication for each second device sequentially or in parallel.

[0052] In this embodiment of the invention, the first device establishes a training schedule based on the CSI and status of each second device, triggering beam training of multiple second devices sequentially or in parallel to prevent mutual interference and improve training efficiency. The first device can dynamically arrange the training order and parameters of different second devices based on information collected on the sub-7GHz link (such as the location of each terminal or historical beams).

[0053] This invention provides a millimeter-wave communication method. A first device receives first information from a second device via a low-frequency link. Based on the first information, the first device determines a candidate beam range for the millimeter-wave link communicating with the second device. Based on the candidate beam range, the first device initiates beam training in the millimeter-wave frequency band to determine the millimeter-wave link beam used for millimeter-wave communication. This method solves the problems of high beam training overhead and slow link recovery in related technologies, achieving improved beam training efficiency and reduced millimeter-wave link control overhead.

[0054] This embodiment provides a millimeter-wave communication method. Figure 3 This is another flowchart of the millimeter-wave communication method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps:

[0055] In step S302, the second device sends the first information to the first device via a low-frequency link.

[0056] In one exemplary embodiment, the first information includes at least one of the following: channel information acquired by the second device via a low-frequency link; broadcast measurement information acquired by the second device via a low-frequency link; historical channel information of the second device; attitude data information of the inertial measurement unit (IMU) of the second device; visual information of the second device; and location information of the first device and / or the second device.

[0057] In this embodiment of the invention, the first device determines a candidate beam, a candidate beam range, or a candidate beam set based on first information from the second device. The first information includes, but is not limited to, at least one of the following: channel information, broadcast measurement information, and environmental awareness information. The environmental awareness information includes, but is not limited to, at least one of the following: IMU attitude data information, visual information, and location information.

[0058] In one exemplary embodiment, the low-frequency link includes a sub-7GHz link.

[0059] Step S304: The second device performs beam training in the millimeter-wave frequency band based on the candidate beam range indicated by the first device to determine the millimeter-wave link beam for millimeter-wave communication, wherein the candidate beam range is determined by the first device based on the first information.

[0060] In one exemplary embodiment, the second device performs beam training in the millimeter-wave band based on the candidate beam range indicated by the first device to determine the millimeter-wave link beam for millimeter-wave communication, including: the second device receiving an instruction for sector-level scan (SLS) beam training from the first device and sending short sector scan (SSW) feedback to the first device; the second device and the first device performing interactive beam training through the beam refinement protocol (BRP) to determine the millimeter-wave link beam for millimeter-wave communication.

[0061] In one exemplary embodiment, the method further includes: the second device selecting a millimeter-wave reference beam from historical millimeter-wave beams based on historical channel information and current first information, and sending an initial beam configuration to the first device based on the millimeter-wave reference beam, so that the first device determines the candidate beam range based on the initial beam configuration.

[0062] In this embodiment of the invention, the terminal (i.e., the second device mentioned above) and the AP (i.e., the first device mentioned above) maintain their respective CSI databases, recording and updating the optimal beam index for the corresponding direction at the end of each communication. When the terminal establishes a link again, it first queries the historical database: if the best beam record matching the scenario is found, it directly uses that beam for communication on the millimeter-wave link; otherwise, it performs conventional sector frequency sweep training to select the best beam.

[0063] In this embodiment of the invention, by maintaining a CSI history database and recording the optimal beam configuration during previous communication processes, the terminal can directly call upon historical information as initial training values ​​when re-establishing a connection, significantly reducing the overhead of initial beam scanning. This approach is particularly suitable for scene revisiting and stable environmental conditions, and can significantly improve the speed of link establishment.

[0064] In an exemplary embodiment, in response to the first information including IMU attitude data information, the method further includes: a second device determining attitude direction information relative to the first device based on the IMU attitude data information; the second device determining a first predicted candidate beam range based on the attitude direction information, and sending the first predicted candidate beam range to the first device.

[0065] In one exemplary embodiment, in response to the first information including visual information, the method further includes: the second device determining the position information of the first device based on the visual information, and determining the three-dimensional vector direction of the first device relative to the second device based on the position information of the first device; the second device determining the second predicted candidate beam range based on the three-dimensional vector direction, and sending the second predicted candidate beam range to the first device.

[0066] In one exemplary embodiment, the method further includes: the second device performing machine learning training based on real-time first information to determine whether to retrain the beam; if retraining is required, the second device and the first device performing beam training based on the newly determined candidate beam range; otherwise, maintaining the current millimeter-wave link.

[0067] In this embodiment of the invention, by introducing machine learning decision logic, the adaptive adjustment of the beam training strategy is realized in a complex mobile environment, avoiding blind and repeated training and reducing the control overhead of the millimeter-wave link.

[0068] The machine learning-based dynamic beam training in this embodiment of the invention inputs real-time metrics such as CSI history and signal-to-noise ratio, along with environmental awareness data, into the training decision model. The terminal can then predict when the link needs retraining and what training method (e.g., coarse or fine training) should be used. This avoids frequent and ineffective retraining, enabling millimeter-wave communication to maintain high performance while reducing unnecessary control overhead.

[0069] In one exemplary embodiment, the method further includes: a second device monitoring the quality index of the millimeter-wave link in real time; if the quality index is less than a preset index threshold, the second device notifies the first device of the millimeter-wave link anomaly via a low-frequency link, and determines a new millimeter-wave beam for millimeter-wave communication based on the first information; otherwise, the current millimeter-wave link is maintained.

[0070] In this embodiment of the invention, the terminal continuously monitors the signal quality of the millimeter-wave link. When a link quality degradation exceeding a threshold is detected, the terminal immediately sends an indication to the access point (AP) and switches to a sub-7GHz link to maintain the connection. Then, based on existing CSI or sensor data, a backup beam direction is quickly selected, and beam training is performed again after the obstacle disappears to restore millimeter-wave communication. This process minimizes the impact of interruptions on services.

[0071] In this embodiment of the invention, the terminal continuously monitors the quality of the millimeter-wave link. Once a sharp drop in signal or link failure is detected, the terminal immediately switches to the sub-7GHz link to maintain basic communication. Simultaneously, historical CSI and environmental awareness information are used to quickly select an alternative beam direction and attempt to rebuild the millimeter-wave link, achieving rapid link recovery and seamless switching.

[0072] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0073] This embodiment also provides a millimeter-wave communication system, characterized in that it includes: a terminal and an access point (AP), a communication connection between the AP and the terminal is established between a low-frequency link and a millimeter-wave link, the AP receives first information from the terminal through the low-frequency link, and determines the candidate beam range of the millimeter-wave link for communication with the terminal based on the first information; the AP and the terminal initiate beam training in the millimeter-wave frequency band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication.

[0074] This embodiment also provides a millimeter-wave communication device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0075] In this embodiment of the invention, the millimeter-wave communication device includes: a multi-link AP and a multi-link terminal. The multi-link AP has a first sub-7GHz transceiver module and a first millimeter-wave transceiver module, and the multi-link terminal has a second sub-7GHz transceiver module, a second millimeter-wave transceiver module, and an environment perception module. The environment perception module is used to acquire terminal attitude, visual, or location information. The device also includes: a beam selection module and a beam training control module, used to predict candidate beams of the millimeter-wave link based on CSI transmitted from the sub-7GHz link and environment perception data, and to control beam training on the millimeter-wave link.

[0076] In this embodiment of the invention, the millimeter-wave communication device further includes a historical CSI storage unit, which is used to save historical channel state information and corresponding millimeter-wave optimal beam index between the multi-link terminal and the multi-link AP, so that the terminal can call it when reconnecting.

[0077] In this embodiment of the invention, the environmental perception module of the millimeter-wave communication device includes one or more of an IMU sensor, a camera, or a positioning module, for providing input to the machine learning unit; the millimeter-wave communication device also includes a machine learning unit, which outputs a beam training strategy based on CSI history, sensor information, and environmental data.

[0078] In this embodiment of the invention, the millimeter-wave communication device further includes a sub-7GHz anchor link module. The multi-link AP uses the anchor link to carry the control frames of the millimeter-wave link and only sends simplified beacons on the millimeter-wave link. The multi-link terminal must first connect to the anchor link before it can participate in millimeter-wave communication.

[0079] In this embodiment of the invention, the beam training control module of the millimeter-wave communication device can dynamically adjust the frequency, search range, or training mode of beam training according to the instructions of the machine learning unit, thereby achieving adaptive beam management.

[0080] In this embodiment of the invention, the millimeter-wave communication device may also include different modules, and the naming and functional division of the modules may be selected in different ways according to the actual situation, without specific limitations.

[0081] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0082] This invention also provides a network device, which includes a receiver, a transmitter, and a processor. The network device is used to perform the steps of the above-described millimeter-wave communication method embodiments through at least one of the receiver, transmitter, and processor.

[0083] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to perform the steps in any of the above method embodiments when executed.

[0084] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0085] Embodiments of the present invention also provide an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0086] In one exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0087] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0088] In one exemplary embodiment, the computer program product described above includes a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the methods described in various embodiments of this application.

[0089] Specific examples in this embodiment can be found in the examples described in the above embodiments and exemplary implementations, and will not be repeated here.

[0090] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0091] To enable those skilled in the art to better understand the technical solutions of the present invention, the following description is provided in conjunction with different embodiments.

[0092] Example 1

[0093] In this embodiment of the invention, in order to solve the problems of high beam training overhead and slow link recovery, a millimeter-wave communication method is provided. Based on a cross-band collaborative beam management mechanism, under the premise that both the AP and the terminal have sub-7GHz links and millimeter-wave links, the sub-7GHz links are used to assist millimeter-wave beam training, and the beam is intelligently selected in combination with environmental perception information.

[0094] In this embodiment of the invention, both the AP and the terminal have low-frequency links and millimeter-wave links, wherein the sub-7GHz link is a type of low-frequency link.

[0095] Multi-link cooperative beam training mechanism: The terminal (non-AP STA) and the AP perform initial communication and channel sounding on a sub-7GHz link. The terminal collects Channel State Information (CSI) or measurement reports through this link. Based on the coarse information obtained from the sub-7GHz link, i.e., the first information in the above embodiment, the AP and the terminal jointly estimate the candidate direction range of the millimeter-wave beam. Then, beam training (including coarse frequency sweeping and fine adjustment) is performed only on the millimeter-wave link within this candidate range, thereby significantly reducing the search space, decreasing the number of probe and control frames on the millimeter-wave link, and improving training efficiency. For example, this mechanism avoids aimless scanning of the entire space, thus accelerating the establishment of the millimeter-wave link.

[0096] Intelligent beam selection mechanism: The terminal combines its historical CSI and environmental perception data, such as inertial measurement unit sensor data, camera visual information, or positional information, to perceive the environment and uses a machine learning model to predict the available beams for the millimeter-wave link. This machine learning model takes CSI history, signal-to-noise ratio, and terminal attitude data as input, outputs the best or candidate beam set, and dynamically adjusts the training strategy accordingly (e.g., choosing coarse or fine training, adjusting the training frequency, or the search angle). Through adaptive strategies, redundant training iterations and latency can be further reduced, enabling rapid response to environmental changes and maintaining link quality.

[0097] Anchored Link Design: In a multi-link system, the AP can be configured with a sub-7GHz link as the anchor link for the millimeter-wave link, used to carry control information for the millimeter-wave link (such as training triggers, feedback, status reports, etc.). This significantly reduces the beacon and management overhead of the millimeter-wave channel. The AP instructs each terminal on the direction of beam training through the anchor link, requiring only simplified beacon broadcasting on the millimeter-wave link to maintain link synchronization and improve system resource utilization efficiency.

[0098] Figure 4 This is a hardware architecture diagram of the millimeter-wave communication method implemented according to an embodiment of the present invention, such as... Figure 4 As shown, the system implementing millimeter-wave communication includes an access point (AP) and a terminal, both of which are multi-link devices. The AP includes a beam selection module and a millimeter-wave beam training module. The beam selection module is responsible for dynamically selecting the optimal beam direction to improve signal quality. The millimeter-wave beam training module is used for beam alignment and channel optimization in high-frequency bands (e.g., 60 GHz). The terminal has a symmetrical structure to the AP and also includes a beam selection module and a millimeter-wave beam training module, enabling bidirectional collaboration.

[0099] In this embodiment of the invention, the system implementing the millimeter-wave communication method supports a dual-link cooperation mechanism, such as... Figure 4As shown, it simultaneously supports sub-7GHz links (low-frequency band) and millimeter-wave links (high-frequency band, such as 60GHz). The sub-7GHz link is used to provide wide coverage and stable control signaling transmission (such as channel state feedback and beam configuration instructions). The channel state feedback can be carried in the first information in the above embodiment. The beam configuration instructions can carry the first and second predicted candidate beam ranges, etc., as described in the above embodiment. In this embodiment, the sub-7GHz link has the advantage of strong penetration and is suitable for basic connections in non-line-of-sight (NLoS) scenarios.

[0100] For millimeter-wave links, in this embodiment of the invention, they are used to handle high-speed data transmission, relying on beamforming technology to combat path loss. They offer advantages such as ultra-large bandwidth, supporting extremely low latency and ultra-high throughput (eMBB scenarios in 5G NR).

[0101] In this embodiment of the invention, the beam selection module dynamically adjusts the millimeter-wave beam direction by exchanging Channel State Information (CSI) via the sub-7GHz link. For example, the terminal uses the first information feedback signal strength to allow the AP to select the optimal beam to establish the millimeter-wave link.

[0102] In this embodiment of the invention, the millimeter-wave beam training module performs beam sweeping to align the transceiver beams during initial access or channel changes.

[0103] Through this invention, the sub-7GHz link transmits beam training results, reducing millimeter-wave signaling overhead. Sub-7GHz ensures reliability, while millimeter-wave enhances capacity, achieving a dual advantage in coverage and speed. Modular design allows for dynamic adaptation to mobile scenarios (such as rapid beam switching when a user rotates their terminal).

[0104] The system and method for implementing millimeter-wave communication provided in this invention achieve efficient beam alignment and rapid reconnection in millimeter-wave communication by constructing a two-layer beam management mechanism of cross-band collaborative training and intelligent selection. Using the technical solution of this invention, the time required for millimeter-wave beam training and control overhead can be significantly reduced, while maintaining or even improving link throughput and robustness. It is suitable for indoor high-speed Wi-Fi: millimeter-wave transmission of 4K video, sub-7GHz maintaining control connection. Factory Automated Guided Vehicle (AGV) communication: low-frequency band controlling AGV movement, high-frequency band real-time transmission of sensor data. Augmented Reality / Virtual Reality (AR / VR) devices: millimeter-wave low latency ensures interactive experience, sub-7GHz prevents connection loss due to obstruction, and other scenarios.

[0105] Example 2

[0106] In this embodiment, the basic process of sub-7GHz cooperative beam training is introduced. Figure 5 This is a flowchart of the beam training process according to an embodiment of the present invention, as follows: Figure 5 As shown, it includes the following steps:

[0107] Step S501: The terminal sends a measurement report CSI_Report.

[0108] In this embodiment, the terminal sends a measurement report to the access point (AP) via a sub-7GHz control channel (low-frequency band, wide coverage, and strong penetration). This measurement report is a CSI measurement report. The CSI measurement report provides the AP with preliminary channel information for subsequent millimeter-wave beam prediction.

[0109] In this embodiment, the CSI measurement report includes the current channel status (such as signal strength, phase, and multipath information).

[0110] In this embodiment, the CSI measurement report may include beamforming feedback, such as a recommended beam orientation or precoding matrix. The recommended beam orientation, etc., may correspond to the first and second predicted candidate beam ranges, etc., in the embodiments described above.

[0111] Step S502: AP predicts millimeter-wave beam.

[0112] In this embodiment, the AP predicts the optimal beam direction (Angle of Arrival / Angle of Departure, AoA / AoD) in the millimeter-wave band based on the CSI measurement report fed back by the terminal, using spatial correlation algorithms (such as compressed sensing and deep learning models). That is, the candidate beam range in the above embodiment can be used to narrow down the range of subsequent beam training through prediction, reducing scanning overhead. This avoids the high latency of omnidirectional beam scanning and improves beam alignment efficiency.

[0113] Step S503: AP initiates beam training for sector-level scanning SLS.

[0114] In this embodiment, the AP initiates beam training based on the millimeter-wave channel by sending training signals, such as null data packet (NDP) frames, in different directions (sectors) through a Transmit Sector Sweep (TXSS). The terminal listens to and measures the signal-to-noise ratio (SNR) or received signal strength (RSSI) of each sector.

[0115] Step S503 can preliminarily determine the optimal direction of the transmitted beam (Tx Beam).

[0116] Step S504: The terminal provides feedback on the optimal beam ID.

[0117] In this embodiment, the terminal sends a Short Sector Scan Feedback (SSW-Feedback) frame to the AP via a millimeter-wave channel or a sub-7GHz control channel to feed back the optimal beam ID. The optimal beam ID includes the optimal direction and the sector ID of the received beam. The SSW-Feedback frame may also contain channel quality metrics such as signal-to-noise ratio (SNR) and throughput estimation.

[0118] Step S504 locks in the coarse-grained beam pair, providing a basis for subsequent fine-grained adjustment (Beam Refinement Protocol, BRP).

[0119] Step S505: Fine adjustment of BRP.

[0120] In this embodiment, the AP and the terminal enter the BRP phase and perform interactive training (such as sending / receiving short training sequences) through the millimeter wave channel to further optimize the beam angle (AoA / AoD) and beamwidth.

[0121] In this embodiment, the angle and width of the beam can be optimized using an adaptive algorithm (such as gradient descent) to improve angular resolution.

[0122] Step S505 eliminates beam deviation (such as offset caused by user micro-movement) and improves directional transmission accuracy.

[0123] Step S506: Millimeter wave data transmission.

[0124] In this embodiment, the AP and terminal use optimized beams for high-speed data transmission (such as 8K video and large file transfers) to facilitate cyclic monitoring and adjustment of the millimeter-wave link. During transmission, the process periodically returns to step S501 (updating the CSI via the sub-7GHz control channel) to dynamically adapt to channel changes (such as obstruction or user movement), effectively addressing the volatility of millimeter waves and maintaining a high-throughput, low-latency millimeter-wave link.

[0125] In this embodiment, the terminal first initiates a channel measurement or broadcast process on the sub-7GHz link to collect CSI; based on this information, the AP and the terminal jointly estimate the optimal direction of the millimeter wave, and then perform beam training and fine selection only within the predicted direction range on the millimeter wave link, thereby reducing the number of frequency sweeps across the entire space.

[0126] The beam training process provided in this embodiment acquires coarse channel information and estimates the millimeter-wave transmission direction through a sub-7GHz link, significantly reducing the beam search range of the millimeter-wave link, decreasing the number of millimeter-wave channel scans and training overhead, thereby accelerating the millimeter-wave link establishment speed. It is suitable for typical indoor Wi-Fi environments where both the AP and the terminal support sub-7GHz and millimeter-wave multi-link communication, such as scenarios where terminals like laptops or smartphones establish connections with multi-link APs.

[0127] In this embodiment, by dividing the beam training process into a layered beam training process, consisting of a coarse-grained (SLS) stage and a fine-grained (BRP) stage, efficiency and accuracy can be balanced.

[0128] Example 3

[0129] This embodiment describes the millimeter-wave communication method of the present invention, which transmits control information of the millimeter-wave link via an anchored link. In this embodiment, a low-frequency anchored link (sub-7GHz) ensures the reliability of control signaling, while a high-frequency millimeter-wave link enables high-speed data transmission.

[0130] Figure 6 This is a flowchart illustrating the transmission of control information via an anchored link according to an embodiment of the present invention, such as... Figure 6 As shown, it includes the following steps:

[0131] Step S601: Configure the sub-7GHz anchor link.

[0132] In this embodiment, the AP configures a low-rate, high-coverage anchor link in the sub-7GHz band (such as 2.4GHz / 5 / 6GHz). For example, it selects one of multiple sub-7GHz links as the anchor link and instructs each terminal to connect to the link in the beacon frame or management frame.

[0133] The configuration of the anchor link includes channel allocation, power control, and broadcast parameters for basic signaling frames (such as Beacon frames). By configuring the anchor link, a stable transmission channel can be provided for the control signaling of the subsequent millimeter-wave link, avoiding signaling interruptions caused by millimeter-wave obstruction or movement.

[0134] Step S602: Connect and maintain the anchor link.

[0135] In this embodiment, the mobile terminal scans the Beacon frames of the sub-7GHz channel, completes association and authentication with the AP, and each terminal establishes and maintains a connection with the anchored link according to the AP's instructions, ensuring that control information is always exchanged through this link before using the millimeter wave link.

[0136] Both the terminal and the AP maintain link activity through periodic keep-alive signals, such as acknowledgments (ACKs), to ensure that the control link is always available and to provide backup for the establishment and maintenance of the millimeter-wave link.

[0137] Step S603: Send millimeter-wave link control information.

[0138] In this embodiment, when millimeter-wave beam training needs to be performed, the AP and the terminal exchange control information such as training trigger and beam index through the anchor link, without needing to broadcast these frames on the millimeter-wave link.

[0139] The AP sends control signaling for the millimeter-wave link to the terminal, including but not limited to: a list of available millimeter-wave channels (such as the 60GHz band), beam training parameters (such as scan interval and training sequence length), and security configurations (such as encryption keys).

[0140] Step S603 can provide the terminal with the initial configuration of the millimeter-wave link, reducing the overhead of subsequent blind scanning.

[0141] Step S604: Return millimeter-wave link control information.

[0142] In this embodiment, the mobile terminal responds via the sub-7GHz link, feeding back millimeter-wave link capability information to the AP, such as: supported millimeter-wave frequency bands and bandwidth, antenna array size (e.g., 4×4 MIMO), and current channel status (e.g., interference level).

[0143] Step S604 helps the AP optimize the resource configuration of the millimeter-wave link (such as beamwidth and modulation scheme).

[0144] Step S605: Simplify beacon broadcasting.

[0145] In this embodiment, the AP performs a simplified beacon broadcast (e.g., containing only basic information) on the millimeter-wave link, and the terminal receives it and reports information such as link quality through the anchor link.

[0146] In this embodiment, a simplified beacon frame, such as a null data packet (NDP), contains only the necessary time synchronization and beam training information. The beacon frame is broadcast in wide beam (Quasi-Omni mode) or predefined sector beam.

[0147] In this embodiment, the terminal quickly locks onto the AP's millimeter-wave signal by listening to the beacon, reducing initial access latency.

[0148] Step S606: Guide millimeter-wave beam training.

[0149] In this embodiment, based on feedback collected on the anchored link, the AP guides the terminal to perform millimeter-wave beam training in a specified direction, and then updates the optimal beam configuration upon completion.

[0150] In this embodiment, the AP guides the terminal to perform millimeter-wave beam training in a specified direction, including the following:

[0151] 1. Coarse-grained beam alignment: The AP and terminal test beams in different directions through sector-level scanning SLS to select the optimal Tx / Rx beam pair.

[0152] 2. Fine-grained optimized BRP: Based on feedback CSI, dynamically adjust the beam angle and width (e.g., narrowing it from 30° to 5°).

[0153] Step S606 can overcome the high path loss of millimeter waves and maximize the gain of directional transmission.

[0154] In this embodiment, during the deployment phase, the AP selects a sub-7GHz link as the anchor link for the millimeter-wave link and notifies the terminal to connect. All subsequent control frames related to the millimeter-wave link (such as beam training triggers, feedback, link reports, etc.) are transmitted through the anchor link. The AP performs simplified beacon broadcasting on the millimeter-wave link (e.g., sending only minimal identification information), and the terminal still provides feedback through the anchor link after receiving the beacon. Through centralized control, the AP can instruct the terminal on the anchor link to perform beam training in which directions, thereby further reducing the overhead on the millimeter-wave link. The sub-7GHz anchor link achieves cross-band control information sharing, alleviating the control overhead of the millimeter-wave channel.

[0155] This invention transmits control information via an anchor link. By offloading control information from the millimeter-wave link to the sub-7GHz anchor link, broadcast and beacon overhead on the millimeter-wave channel can be reduced. The AP configures and notifies each terminal to establish an anchor connection on a specific sub-7GHz link. All control information such as beam training requests, feedback, and link status reports for the millimeter-wave link is exchanged through this anchor link. Only necessary beam alignment and data transmission are performed on the millimeter-wave link, thereby significantly reducing the transmission burden of beacons and management frames. This is applicable to multi-link device systems where the AP and multiple terminals can communicate simultaneously on multiple sub-7GHz and millimeter-wave links. This embodiment demonstrates how to utilize a sub-7GHz link as an "anchor link" to assist in managing the control and beam training process of the millimeter-wave link.

[0156] In this embodiment, the sub-7GHz anchor link ensures uninterrupted control signaling, while the millimeter-wave link focuses on high-speed data transmission, making it more reliable. Pre-exchanging configuration information via the sub-7GHz band reduces the scanning time for millimeter-wave beam training, resulting in greater efficiency. If the millimeter-wave link fails due to obstruction, it can be immediately reinitialized via the anchor link, dynamically adapting to changes in the millimeter-wave link. Step description:

[0157] Example 4

[0158] In this embodiment of the invention, the terminal selects a millimeter-wave reference beam from the historical millimeter-wave beams based on historical channel information and current first information, and sends an initial beam configuration to the AP based on the millimeter-wave reference beam, so that the AP can determine the candidate beam range based on the initial beam configuration.

[0159] In this embodiment, the beam training process assisted by historical CSI is introduced.

[0160] Figure 7 This is a flowchart of beam training based on historical CSI assistance according to an embodiment of the present invention, such as... Figure 7 As shown, it includes the following steps:

[0161] Step S701: The terminal searches for reference beam records in the CSI historical database.

[0162] In this embodiment, the terminal performs routine beam training and communication with the AP. After the communication ends, the current best transmit / receive (Tx / Rx) beam index is stored in the CSI historical database. When the terminal attempts to establish a link with the same AP again, it first queries the CSI historical database to see if there is a record of the best beam under the corresponding environment (such as location, time, channel fingerprint) or location, which will serve as the millimeter-wave reference beam.

[0163] In this embodiment, the retrieval criteria may include: location tags (such as GPS coordinates, indoor beacon IDs), timestamps (such as interference patterns during peak hours on weekdays), and channel fingerprints (such as multipath features, obstacle distribution).

[0164] In this embodiment, the terminal sends the initial beam configuration to the AP based on the millimeter-wave reference beam, reusing the historical best beam configuration to reduce the initial training overhead.

[0165] Step S702: Determine if a reference record has been found?

[0166] The terminal selects a millimeter-wave reference beam from the historical millimeter-wave beams. If so, it proceeds to step S703, directly using the historical beam as the millimeter-wave reference beam. The terminal first attempts to communicate using this millimeter-wave reference beam on the millimeter-wave link and performs a rapid link quality check. If not, the reference beam performance is insufficient or it is ineffective, then it proceeds to step S704 to perform a new round of full-beam scanning.

[0167] In this embodiment, the CSI database is updated after the communication process is completed, regardless of the situation, for future retrieval.

[0168] In this embodiment, if the historical records have a sufficiently high spatial correlation (e.g., similar location) or temporal correlation (e.g., same time period) with the current scene, they are deemed valid.

[0169] Step S703: Use the reference beam as the initial beam for training.

[0170] The terminal uses the beam direction (AoA / AoD), beamwidth, and precoding matrix from historical records as the initial beam configuration to directly initiate millimeter-wave link training. The AP performs fine-tuning (such as ±5 angular offset testing) based on the terminal's initial beam configuration.

[0171] The latency of sector-level scanning SLS is reduced by using a reference beam as the initial beam for training (traditional methods require testing dozens of beam directions).

[0172] Step S704: Perform beam scanning on the millimeter-wave link.

[0173] In the event of insufficient or failed reference beam performance, the AP and terminal perform beam scanning based on the millimeter-wave link, including the following:

[0174] 1. Coarse scan: The AP sends sector-level training sequences, and the terminal measures the signal strength in each direction, such as Received Signal Strength Indication (RSSI) or Signal-to-Noise Ratio (SNR).

[0175] 2. Beam pairing: Both sides exchange sector sweep feedback frames (SSW-Feedback) to lock the optimal Tx / Rx beam pair.

[0176] In this embodiment, performing a new round of full-beam scanning is applicable to: entirely new environments (such as entering a room for the first time), or when historical records are invalid (such as changes in furniture layout).

[0177] Step S705: Perform refined training and confirm the optimal beam.

[0178] In this embodiment, during the BRP phase, the AP sends a short training sequence (Training-Receive / Transmit, TRN-R / T), and the terminal reports the phase deviation to optimize the beam angle and width. During the confirmation phase, both parties confirm the optimal beam through an ACK frame and enter data transmission mode.

[0179] In this embodiment, by refining the training and confirming the optimal beam, the beamwidth can be narrowed to 1° to 3° (the initial sector beam is usually 15° to 30°).

[0180] Step S706: Update the CSI database to record the new millimeter-wave beam.

[0181] In this embodiment, the terminal or AP writes the current optimal beam parameters (such as beam ID, SNR, and location label) into the CSI database. This may trigger online training of machine learning models, for example, using a Long Short-Term Memory (LSTM) network to predict the beam's changing trend as it moves.

[0182] Updating the CSI database and recording new millimeter-wave beams has the following long-term value: forming an environment-beam mapping knowledge base that can be directly accessed in similar scenarios in the future.

[0183] In this embodiment, the terminal and the AP maintain their respective CSI databases, recording and updating the optimal beam index for the corresponding direction at the end of each communication. When the terminal establishes a link again, it first queries the historical database: if the best beam record matching the scenario is found, it directly uses that beam for communication on the millimeter-wave link; otherwise, it performs regular sector frequency sweep training to select the best beam.

[0184] In this embodiment, by maintaining a CSI history database and recording the optimal beam configuration during previous communication processes, the terminal can directly call upon historical information as initial training values ​​when re-establishing a connection, significantly reducing the overhead of initial beam scanning. This approach is particularly suitable for scene revisiting and stable environmental conditions, and can significantly improve the speed of link establishment.

[0185] In this embodiment of the invention, the beam training process based on historical CSI is suitable for communication between the AP and the repeatedly accessed terminal in a relatively fixed environment where the terminal repeatedly enters the same area (such as a home or office scenario). This embodiment demonstrates how to accelerate the beam training process using historical channel state information (CSI).

[0186] In this embodiment, intelligent beam reuse is achieved by reducing initial training time by more than 70% using historical CSI data. During the beam training phase, a hierarchical training logic of coarse scanning and fine optimization is used to balance efficiency and accuracy. By establishing a self-learning database, such as through machine learning, a closed-loop update mechanism is implemented to adapt to dynamic environments (such as changes in pedestrian flow or new obstacles).

[0187] Typical scenarios applicable to beam training processes based on historical CSI include, but are not limited to: Museum AR navigation: When visitors revisit, historical beams are directly invoked, connecting to millimeter-wave 8K video streams in seconds. Subway station intelligent monitoring: Different beam configurations are used during peak hours to avoid obstruction by crowds. Factory robot inspection: The optimal beam for each location is recorded, enabling seamless switching during movement.

[0188] In this embodiment, by utilizing historical CSI information, the terminal can skip large-scale sector frequency sweeps or part of the training phase, thereby quickly establishing a millimeter-wave link.

[0189] Example 5

[0190] In this embodiment of the invention, in response to the first information including IMU attitude data information, the terminal determines the attitude direction information relative to the AP based on the IMU attitude data information; the terminal determines the first predicted candidate beam range based on the attitude direction information and sends the first predicted candidate beam range to the AP.

[0191] Figure 8 This is a flowchart illustrating the optimization of millimeter-wave beam training based on IMU attitude data according to an embodiment of the present invention, as shown below. Figure 8As shown, it includes the following steps:

[0192] Step S801: The terminal detects IMU attitude data.

[0193] In this embodiment, the terminal initiates a sub-7GHz link connection and detects IMU attitude data in real time via an accelerometer / gyroscope. The accelerometer is used to collect and detect three-axis acceleration, i.e., dynamic displacement, while the gyroscope is used to collect and detect angular velocity, i.e., rotation direction. The sampling rate is typically ≥100Hz.

[0194] During the data acquisition process described above, timestamps and sensor error calibration values ​​(such as gyroscope zero bias) are recorded simultaneously.

[0195] In this embodiment, the terminal detects IMU attitude data to provide raw data for subsequent attitude change analysis, such as detecting the rotation angle or swing amplitude of the user's handheld terminal.

[0196] Step S802: Analyze attitude changes to predict possible communication directions.

[0197] In this embodiment, the terminal calculates its attitude direction (azimuth and pitch angles) relative to the AP based on the angle change of the IMU. The calculation method can be Kalman filtering and machine learning methods.

[0198] In this embodiment, the terminal performs attitude calculation, including converting IMU data into the terminal's three-dimensional spatial orientation (e.g., pitch angle 30°, azimuth angle 45°) through quaternion or Euler angle conversion.

[0199] In this embodiment, the terminal performs motion prediction, including: short-term prediction: predicting the terminal's orientation within the next 100ms based on the integral of angular velocity. Long-term trend prediction: learning user behavior patterns (such as frequent leftward turning) using an LSTM model. Output result: predicting the range of the signal arrival angle AoA of the AP relative to the terminal (such as ±15° sector), thus obtaining the first prediction candidate beam range.

[0200] Step S803: Narrow the millimeter-wave beam training range to the predicted sector.

[0201] In this embodiment, the terminal sends a beam training range suggestion, i.e., the first predicted candidate beam range, to the AP through the sub-7GHz control channel. Based on the IMU calculation results, the terminal and the AP limit the search range of the millimeter wave beam training to the corresponding directional sector and its neighborhood.

[0202] In this embodiment, the range of the first predicted candidate beam includes, but is not limited to:

[0203] Predict the sector center angle (e.g., azimuth 120°); sector width (e.g., ±20°, determined by prediction confidence); AP dynamically adjusts the millimeter-wave beam scanning range, testing only the beam direction within the target sector.

[0204] Step S804: Perform focused beam training on the millimeter-wave link.

[0205] In this embodiment, both the terminal and the AP perform beam training (sector sweep and fine adjustment) on the millimeter-wave link for this limited angle region to select the optimal beam.

[0206] In step S804, focusing beam training on the millimeter-wave link includes coarse alignment and fine optimization phases.

[0207] Coarse alignment: The AP transmits the training sequence (TRN-R / T) within the predicted sector, and the terminal measures the SNR of each beam.

[0208] Fine-tuning: If the initial prediction deviation is large, initiate BRP to fine-tune the beam direction. Use the gradient descent algorithm to gradually approximate the optimal beam angle.

[0209] In this embodiment, anti-interference design is also implemented: if a sudden occlusion is detected (such as a user waving), the IMU data is re-verified and the process returns to step S801.

[0210] Step S805: Select and use the best beam for communication.

[0211] After establishing millimeter-wave communication, the terminal continuously monitors IMU data. If the attitude changes significantly, beam prediction and training are re-executed. The AP and the terminal confirm the optimal beam ID (e.g., Tx beam #12 + Rx beam #7) and lock this configuration for data transmission.

[0212] In this embodiment, the AP and the terminal simultaneously initiate a keep-alive mechanism, including the following:

[0213] Periodic IMU monitoring (e.g., detecting attitude changes every 50ms); dynamic beam tracking: if the attitude change exceeds the threshold (e.g., rotating 5°), re-trigger prediction (return to step 2).

[0214] In this embodiment, the terminal calculates a coarse azimuth and pitch angle relative to the AP based on inertial sensor data (such as gyroscope and accelerometer). Based on the calculation results, the system performs beam training and fine optimization only within the predicted directional sector on the millimeter-wave link. If the environment is complex or the prediction error is large, further fine-tuning can be performed by combining historical CSI data.

[0215] In this embodiment, the IMU can sense the terminal's orientation and motion state, and predict the current possible communication path based on the prior antenna orientation. Adjusting the millimeter-wave beam training direction according to the IMU's attitude can limit the training search range to a smaller angular region, thereby reducing the number of training iterations and latency of the millimeter-wave link.

[0216] The process of optimizing millimeter-wave beam training based on IMU attitude data in this embodiment of the invention is applicable to mobile device scenarios with a built-in inertial measurement unit (IMU), such as when a user is moving indoors while holding a mobile phone or wearing a head-mounted device. This embodiment demonstrates how to reduce the millimeter-wave beam training search space using the directional information provided by the IMU.

[0217] The process of optimizing millimeter-wave beam training based on IMU attitude data in this invention has the following technical advantages:

[0218] 1. Low power consumption: The IMU consumes only 0.1mW, which is far lower than the power consumption of millimeter-wave full scan.

[0219] 2. Low latency: Beam training time is reduced from the traditional 20ms to less than 2ms.

[0220] 3. High robustness: Dynamically adapts to scenarios such as human occlusion and terminal rotation.

[0221] Typical application scenarios for the millimeter-wave beam training process based on IMU attitude data in this invention include, but are not limited to: VR games: When the user's head turns, the millimeter-wave beam tracks the orientation of the head-mounted display (HMD) in real time. Vehicle communication: Predicting the base station direction based on the vehicle's turning angle reduces beam switching interruptions. UAV inspection: Combining IMU and GPS data to maintain millimeter-wave link stability.

[0222] This embodiment utilizes IMU information to predict the communication direction, effectively reducing blind scanning of millimeter-wave links and improving the real-time performance and efficiency of beam training.

[0223] Example 6

[0224] In this embodiment of the invention, in response to the first information including visual information, the terminal determines the location information of the AP based on the visual information, and determines the three-dimensional vector direction of the AP relative to the terminal based on the location information of the AP; the terminal determines the second predicted candidate beam range based on the three-dimensional vector direction, and sends the second predicted candidate beam range to the AP.

[0225] Figure 9 This is a flowchart of vision-assisted millimeter-wave beam selection according to an embodiment of the present invention, such as... Figure 9 As shown, it includes the following steps:

[0226] Step S901: The terminal camera captures an image of the environment.

[0227] In this embodiment, the terminal's camera captures images of the current environment and sends them to the visual processing unit for preprocessing.

[0228] In this embodiment, the terminal is equipped with a high frame rate camera (such as a 120fps RGB or infrared camera) and a millimeter-wave antenna array is activated simultaneously.

[0229] In this embodiment, the terminal captures panoramic images of the environment in real time, focusing on, but not limited to, the following: AP location features (such as beamforming antenna panels of a specific shape); obstacle distribution (such as the outlines of people and furniture).

[0230] In this embodiment, the overall segment also needs to collect additional metadata: timestamp, camera focal length / distortion parameters, and IMU attitude data (if fused).

[0231] In this embodiment, visual-radio frequency joint sensing is achieved by matching the image resolution with the millimeter-wave wavelength (~5mm).

[0232] Step S902: Visual algorithm identifies target location.

[0233] In this embodiment, the location of the communication peer device or indicator sign in the image is identified and located using visual algorithms (such as deep learning object detection).

[0234] In this embodiment, the algorithm architecture of the visual algorithm includes:

[0235] Target detection: Locate the AP antenna panel based on the model, and output the pixel coordinates (x,y) and confidence score.

[0236] Depth estimation: Binocular visual parallax method (if dual cameras are available); monocular deep learning to estimate the straight-line distance from the terminal to the AP.

[0237] Coordinate transformation: Mapping the image coordinate system (2D) to the terminal body's 3D spatial coordinate system through perspective transformation.

[0238] Output (Second Predicted Candidate Beam Range): 3D vector direction (azimuth) of AP relative to the terminal. Pitch angle θ, distance d).

[0239] Step S903: Determine the candidate beam direction for the millimeter-wave link.

[0240] Based on the identification results, the system calculates the relative spatial positions (including distance and angle) between devices to determine the most likely millimeter-wave communication path direction.

[0241] In this embodiment, based on the visual output From the perspective of querying the codebook pre-stored in the terminal.

[0242] In this embodiment, if a partial obstruction such as glass is detected, the diffraction beam is activated (e.g., ±10° deflection).

[0243] In this embodiment, candidate beams are selected by combining historical CSI data (e.g., prioritizing high SNR beams at the same location within the past 24 hours).

[0244] In this embodiment, anti-interference design is also implemented to exclude beam directions that are visually identified as dynamic obstacles (such as a person waving).

[0245] Step S904: Perform targeted beam training on the millimeter-wave link.

[0246] A candidate beam list is generated according to a defined directional range, and beam training is performed on the millimeter-wave link for this list to select the beam pair with the best signal.

[0247] In this embodiment, the above training process adopts the following training protocol:

[0248] 1. Coarse Alignment: The AP only sends the training sequence (TRN-R / T) in the direction of the candidate beam, skipping other directions. The terminal measures the RSSI / SNR of each candidate beam and feeds back the IDs of the best 3 beams.

[0249] 2. Fine-tuning: Perform BRP on the top three beams: send phase-adjustable probe signals and fine-tune the beams to 0.5° accuracy.

[0250] The time consumption of this embodiment is compared with that of the traditional solution as follows: Traditional full scan: 64 beams × 2ms = 128ms; This embodiment: 4 beams × 2ms + fine adjustment 5ms = 13ms (time consumption reduced by 90%).

[0251] Step S905: Select and use the optimal beam.

[0252] A millimeter-wave link is established based on the training results, and visual information is continuously referenced in subsequent updates.

[0253] In this embodiment, the selection criteria for the best beam are as follows: the final beam is selected by combining instantaneous quality (SNR≥25dB) and stability (jitter <3dB in the past 5 seconds).

[0254] In this embodiment, a keep-alive mechanism is set, such as continuous visual monitoring: the AP position is detected every 200ms, and if the offset exceeds the beamwidth (e.g., 5°), the process is retried.

[0255] In this embodiment, millimeter-wave-assisted sensing is used, for example, by detecting terminal micro-movements through Doppler frequency shift of received signals.

[0256] Table 1 shows a comparison of the advantages of the vision-assisted millimeter-wave beam selection in this embodiment with traditional methods.

[0257] Table 1. Comparison of advantages between vision-assisted millimeter-wave beam selection and traditional methods.

[0258] Dimension Conventional scheme Vision-aided scheme Training time consumption 100+ms <15ms Mobile adaptability Depend on periodic full scan Real-time vision tracking Anti-occlusion capability Blind handover Obstacle-identification-based beam evading

[0259] Typical application scenarios for vision-assisted millimeter-wave beam selection in this embodiment include, but are not limited to: shopping mall navigation robots: visually identifying the position of millimeter-wave APs on the ceiling to quickly establish links; factory AGVs: predicting the optimal beam direction by identifying reflective markers on the equipment; smart home VR: head-mounted cameras tracking the position of APs in the living room to maintain 8K video streams.

[0260] This embodiment utilizes visual perception to provide prior information for millimeter-wave beam training, which can further reduce the reliance on traditional channel measurements in complex environments and improve transmission reliability and speed.

[0261] This embodiment uses a vision-assisted millimeter-wave beam selection method. The terminal or access point (AP) camera acquires real-time scene images, and the visual computing module executes target detection and localization algorithms to determine the location of the peer device in space. Based on the location inference results, the system generates a set of available beams for the millimeter-wave link and optimizes the training scheme. Then, beam training is performed on the millimeter-wave link using this set, and the optimal beam is selected.

[0262] This embodiment utilizes a vision-assisted millimeter-wave beam selection method. The camera captures images of the surrounding scene and uses visual algorithms to identify the location of the communication peer or environmental obstacles. Based on the image processing results, the relative position and line-of-sight direction between devices can be directly inferred, thereby predicting the most likely millimeter-wave beam direction. This method does not rely entirely on channel detection, significantly reducing the number of beam training iterations and improving stability.

[0263] This embodiment of vision-assisted millimeter-wave beam selection is applicable to environments where access points (APs) or terminals are equipped with cameras, such as indoor routers equipped with panoramic cameras or AR devices. This embodiment demonstrates the use of visual information to assist in millimeter-wave link beam management.

[0264] Example 7

[0265] In this embodiment of the invention, the terminal performs machine learning training based on real-time first information to determine whether to retrain the beam. If retraining is required, the terminal and the AP perform beam training based on the newly determined candidate beam range. Otherwise, the current millimeter-wave link is maintained.

[0266] Figure 10 This is a flowchart of dynamic beam training based on machine learning, as described in an embodiment of the present invention. Figure 10 As shown, it includes the following steps:

[0267] Step S1001: The terminal collects CSI data in real time.

[0268] In this embodiment, the terminal collects link quality indicators (such as RSSI / CSI) and environmental awareness data in real time during the communication process.

[0269] The specific content to be collected is as follows:

[0270] Channel State Information (CSI) includes multipath component amplitude / phase, delay spread, Doppler shift, etc.

[0271] Environmental perception data: terminal location (GPS or Ultra-Wideband (UWB)), IMU motion status, and historical performance indicators of millimeter-wave links.

[0272] In practical implementation, compressed sensing technology can be used to reduce CSI feedback overhead (e.g., compressing from 256 dimensions to 32 dimensions). Time synchronization accuracy needs to reach the μs level.

[0273] Step S1002: Input the machine learning model to predict training requirements.

[0274] The collected data is fed into a machine learning prediction model, which outputs whether the beam needs to be retrained based on previous experience.

[0275] In this embodiment, retraining is triggered when the prediction confidence is >90% and the current SNR is <15dB, and the prediction threshold is dynamically adjusted (e.g., the tolerance is lowered by 5% in mobile scenarios).

[0276] Step S1003: Retraining determination.

[0277] In this embodiment, if the model determines that retraining is required, the system selects the appropriate training scheme (such as fast fine training or full-range scanning) based on the model's suggestion and executes it on the millimeter-wave link.

[0278] In this embodiment, the retraining determination can refer to the determination factors shown in Table 2.

[0279] Table 2 Examples of Retraining Determination Factors

[0280]

[0281]

[0282] Step S1004: Millimeter wave beam retraining.

[0283] In this embodiment, if the model determines that no training is needed, it continues to use the current beam for communication while continuing to monitor the link status.

[0284] In this embodiment, the above training process is based on the following hierarchical training protocol:

[0285] 1. Coarse search phase (2ms): Based on the potential beam direction output in step 2, perform a restricted sector scan (e.g., ±30° range); use a binary search algorithm to quickly locate the top 3 candidate beams.

[0286] 2. Fine-tuning phase (1ms): Perform BRP on candidate beams; use orthogonal training sequences to eliminate multi-user interference.

[0287] Step S1005: Maintain the current beam.

[0288] In this embodiment, after each training or detection process, the system updates the model's input data and the CSI database to continuously optimize the decision-making.

[0289] In this embodiment, the following keep-alive mechanism is designed:

[0290] Continuously monitor the CSI change rate and set dual threshold triggers: soft threshold (SNR drops by 3dB): start background scanning; hard threshold (SNR < 10dB): force interrupt the current transmission.

[0291] In this embodiment, dynamic power compensation is designed: the transmit power is adjusted in reverse according to the link quality.

[0292] Step S1006: Database update and strategy optimization.

[0293] In this embodiment, a knowledge base is constructed. Incremental learning is used to build the knowledge base, employing the Federated Learning framework. The terminal only uploads model gradients, protecting privacy; the model update cycle is adaptively adjusted (ranging from 1 minute to 24 hours).

[0294] In this embodiment, typical application scenarios of machine learning-based dynamic beam training include, but are not limited to, smart factory AGVs: predictive beam switching is triggered the instant the robotic arm is occluded, and the training trigger point is optimized by learning from historical trajectories; VR competitive games: maintaining sub-millisecond beam tracking when the head rotates rapidly, and preloading beam configuration according to the game scene.

[0295] This embodiment introduces machine learning decision logic to achieve adaptive adjustment of beam training strategy in complex mobile environments, avoiding blind and repeated training and reducing the control overhead of millimeter-wave links.

[0296] In this embodiment, the machine learning-based dynamic beamforming training involves the terminal continuously monitoring channel quality and mobility status, and inputting the collected data into a pre-trained machine learning model. The model outputs whether beamforming training is needed and the training scheme parameters, such as the training interval or angle adjustment range. If needed, the terminal initiates the corresponding beamforming training process; otherwise, it maintains the current beam and continues communication. After training is complete, the model and CSI records are updated based on new data.

[0297] In this embodiment, the machine learning-based dynamic beam training inputs real-time metrics such as CSI history and signal-to-noise ratio, along with environmental awareness data, into the training decision model. The terminal can then predict when the link needs retraining and what training method (e.g., coarse or fine training) should be used. This avoids frequent and ineffective retraining, allowing millimeter-wave communication to maintain high performance while reducing unnecessary control overhead.

[0298] The machine learning-based dynamic beam training in this embodiment is suitable for scenarios where both the terminal and the access point (AP) have certain computing capabilities and can run lightweight machine learning models. This embodiment demonstrates how the terminal dynamically decides whether to retrain the beam and adjust the training strategy based on real-time link quality and environmental changes.

[0299] Example 8

[0300] In this embodiment of the invention, in response to an AP providing millimeter-wave connectivity to multiple terminals, the AP determines the candidate beam range of the millimeter-wave link communicating with each terminal based on the first information of each terminal; the AP coordinates and triggers beam training of multiple terminals on the low-frequency link, and determines the millimeter-wave link beam for millimeter-wave communication for each terminal sequentially or in parallel.

[0301] Figure 11 This is a flowchart of multi-terminal collaborative beam training according to an embodiment of the present invention, such as... Figure 11 As shown, it includes the following steps:

[0302] Step S1101: The AP collects the CSI and location information of all terminals.

[0303] In this embodiment, the AP collects the CSI or location information of each STA (terminal) through the sub-7GHz link. In actual implementation, it can also be combined with other sensing information.

[0304] In this embodiment, the data collection content includes, but is not limited to:

[0305] CSI: Millimeter-wave channel matrix (amplitude / phase) reported by each terminal; PowerDelay Profile (PDP) of multipath components; Doppler shift;

[0306] Location information: Absolute coordinates (UWB / GPS positioning, accuracy ±10cm); azimuth / elevation angle relative to the AP (based on RSSI fingerprint or AoA estimation).

[0307] In this embodiment, data is collected periodically through the sub-7GHz control channel, and compression feedback technology (such as Type II CSI codebook) is used to reduce overhead.

[0308] Step S1102: Generate a training scheduling plan for each terminal.

[0309] In this embodiment, the AP determines the training strategy based on the situation of each terminal, for example, prioritizing the training of terminals with better channel quality or faster movement.

[0310] In this embodiment, the core logic of the training strategy scheduling algorithm is as follows:

[0311]

[0312] In this embodiment, the AP determines the training strategy based on the situation of each terminal. It maximizes parallel training, for example, allowing spatially isolated terminals to train synchronously, thus ensuring time efficiency. Fairness is maintained by ensuring minimal training opportunities for edge terminals. Low energy consumption is maintained by suppressing terminal power consumption caused by frequent training.

[0313] Step S1103: Trigger beam training for each terminal in sequence.

[0314] In this embodiment, the AP triggers the beam training process of each terminal sequentially via millimeter-wave link or broadcast command according to a predetermined plan. The execution process is as follows:

[0315] 1. AP broadcast trigger frame: includes target terminal ID, training time slot, and candidate beam range.

[0316] 2. Terminal Response Training: Sector-level scanning (SLS) and beam optimization (BRP) are completed within a specified time slot. Orthogonal training sequences are used for different terminals to avoid interference.

[0317] 3. Result feedback: The terminal returns the optimal beam ID and SNR through the control channel.

[0318] In this embodiment, the following conflict handling is performed: for example, when a terminal that has not completed training is detected, the time slot is automatically delayed to the next cycle in order to dynamically adjust the time slot.

[0319] Step S1104: Summarize the results and update the optimal beam for each terminal.

[0320] In this embodiment, each terminal performs beam scanning according to the AP instructions, and after completion, feeds back the measurement results to the AP via a sub-7GHz link or a millimeter-wave link.

[0321] In this embodiment, the AP updates the beam configuration database of each terminal based on the feedback data and adjusts the subsequent training plan as needed. The database update content is shown in Table 3.

[0322] Table 3. Example of database update content

[0323] Field Explanation Update policy Best beam ID Tx / Rx beam number Instant coverage Beam stability Historical SNR variance Sliding average Effective period According to mobile speed calculation Adaptive survival time

[0324] In this embodiment, the AP updates the terminal beam configuration in batches via multicast management frames. Predictive beam tracking is enabled for high-speed mobile terminals (>3m / s).

[0325] Typical application scenarios for multi-terminal collaborative beam training in this embodiment include, but are not limited to:

[0326] Sports stadium live streaming: APs uniformly schedule millimeter-wave beams from thousands of mobile phones to avoid beam conflicts in the audience area. Factory digital twin: Optimal beams are pre-configured based on the real-time location of AGVs to reduce training interruptions. Intelligent transportation hubs: Dynamic priority scheduling (emergency vehicle terminals are trained first).

[0327] The multi-terminal collaborative beam training in this embodiment improves the overall training efficiency while minimizing co-channel interference and resource conflicts.

[0328] In this embodiment of multi-terminal collaborative beam training, the AP first collects environmental and channel information for each terminal via a sub-7GHz link. Based on this information, the AP generates a beam training schedule (e.g., time-series or grouped parallel). According to the schedule, the AP sequentially notifies the terminals to perform beam training via a millimeter-wave link. Each terminal executes sector scanning according to the instructions and, after completing training, feeds back the results via a sub-7GHz link or a millimeter-wave link. Finally, the AP summarizes the results and updates the optimal beam for each terminal.

[0329] In this embodiment of multi-terminal collaborative beam training, the AP establishes a training schedule based on the CSI and status of each terminal, triggering beam training of multiple terminals sequentially or in parallel to prevent mutual interference and improve training efficiency. The AP can dynamically arrange the training order and parameters of different terminals based on information collected on the sub-7GHz link (such as the location of each terminal or historical beams).

[0330] This embodiment of multi-terminal collaborative beam training is applicable to scenarios where an AP provides millimeter-wave connectivity to multiple terminals, and each terminal can exchange information via a sub-7GHz link. This embodiment illustrates a method for an AP to coordinate multiple terminals to perform beam training.

[0331] Example 9

[0332] In this embodiment of the invention, the terminal monitors the quality indicators of the millimeter-wave link in real time; if the quality indicators are less than a preset threshold, the terminal notifies the AP of the millimeter-wave link anomaly through the low-frequency link, and determines a new millimeter-wave beam for millimeter-wave communication based on the first information; otherwise, the current millimeter-wave link is maintained.

[0333] Figure 12 This is a flowchart of millimeter-wave link quality monitoring and recovery according to an embodiment of the present invention, such as... Figure 12 As shown, it includes the following steps:

[0334] Step S1201: Monitor the quality of the millimeter-wave link.

[0335] In this embodiment, the terminal continuously monitors the quality indicators of the link during millimeter-wave communication, such as received signal strength or frame error rate.

[0336] In this embodiment, the monitored quality indicators include, but are not limited to:

[0337] Real-time signal quality: Signal-to-noise ratio (SNR), Received signal strength (RSSI), Bit error rate (BER);

[0338] Physical layer parameters: Doppler frequency shift (detection terminal movement), delay spread (multipath interference);

[0339] Business requirements: The data type currently being transmitted (e.g., 8K video requires an SNR ≥ 25dB).

[0340] In this embodiment, the terminal can perform a rapid measurement every 10ms (the 802.11ay standard supports sub-millisecond CSI feedback). In this embodiment, the terminal can use a sliding window algorithm to eliminate transient fluctuation interference.

[0341] Step S1202: Link quality is below the threshold?

[0342] If the link quality is lower than a preset threshold, the terminal immediately switches to the sub-7GHz link and notifies the AP of the millimeter-wave link anomaly through this link.

[0343] In this embodiment, the dynamic threshold settings are shown in Table 4.

[0344] Table 4. Example of Dynamic Threshold Settings

[0345] Scene type Decision threshold Trigger action Static scene SNR < 15 dB Start optimization Mobile scene SNR < 20 dB Immediate switch Emergency service BER>1e-4 Forced switch

[0346] If the link quality is below the threshold, then step S1203: switch to the sub-7GHz link to maintain communication.

[0347] The terminal uses historical CSI database and environmental data to select an alternative beam direction, such as the last confirmed path direction, in preparation for link reconstruction.

[0348] In this embodiment, the handover protocol for maintaining communication on the sub-7GHz link includes the following:

[0349] 1. Fast association: Reassociation is completed within 50ms using pre-stored sub-7GHz channel parameters (such as Basic Service Set Identifier BSSID);

[0350] 2. Service migration: Maintain uninterrupted TCP / IP sessions; adaptive speed reduction (e.g., from 10Gbps on millimeter wave to 1Gbps on sub-7GHz).

[0351] In this embodiment, the advantages of the above-mentioned handover protocol include: handover latency <100ms (meeting ITU-T G.114 voice service requirements).

[0352] If the link quality is not lower than the threshold, then step S1204: continue millimeter-wave link communication.

[0353] After the obstacle disappears or the path is restored, the terminal initiates beam training on the millimeter-wave link in the selected beam direction to attempt to restore millimeter-wave communication.

[0354] In this embodiment, preventative optimizations are performed: the beamwidth is fine-tuned (e.g., widened from 5° to 8° to counteract slight movement); and redundant transmission is performed: dual-link concurrency is enabled for critical data packets (e.g., Automatic Repeat Request (ARQ)).

[0355] Step S1205: Select a backup beam using historical CSI / environmental awareness data.

[0356] In this embodiment, after step S1203 is executed to switch to the sub-7GHz link to maintain communication when the link quality is below the threshold, a backup beam is selected through a smart selection algorithm.

[0357] In this embodiment, selecting a backup beam using an intelligent selection algorithm includes: retrieving the top 3 beam configurations at the same location within the last hour based on spatial correlation matching; and obstacle prediction.

[0358] Figure 13 This is a flowchart illustrating the selection of a backup beam using an intelligent selection algorithm, as described in an embodiment of the present invention.Figure 13 As shown, based on historical CSI / environmental perception data and spatial correlation matching, the top 3 beam configurations at the same location within the last hour are retrieved, and in the event of a matching failure, obstacle prediction-assisted detection is initiated via LiDAR.

[0359] The result of selecting backup beams, for example, output: generate 3 candidate beam directions (primary + 2 backup).

[0360] Step S1206: Perform beam training on the millimeter-wave link to restore communication.

[0361] In this embodiment, after reconstruction is complete, the terminal can return to the millimeter-wave link for data transmission, or switch links according to network requirements. In this embodiment, beam training to restore communication on the millimeter-wave link can adopt a layered recovery strategy as shown in Table 5.

[0362] Table 5. Examples of Layered Recovery Strategies

[0363] Stage Time consumption Accuracy Applicable scenario Fast scan 2ms ±5° Emergency recovery BRP optimization 5ms ±0.5° High QoS service

[0364] In this embodiment, compared to the traditional method which requires 20ms for complete training, the method in this embodiment can recover in as little as 2ms.

[0365] In this embodiment of the millimeter-wave link quality monitoring and recovery process, the terminal continuously monitors the signal quality of the millimeter-wave link. When a link quality degradation exceeding a threshold is detected, the terminal immediately sends an indication to the access point (AP) and switches to a sub-7GHz link to maintain the connection. Then, based on existing CSI or sensor data, a backup beam direction is quickly selected, and beam training is performed again after the obstacle disappears to restore millimeter-wave communication. This process minimizes the impact of interruptions on services.

[0366] The millimeter-wave link quality monitoring and recovery process in this embodiment continuously monitors the millimeter-wave link quality through the terminal. Once a sharp drop in signal or link failure is detected, the terminal immediately switches to the sub-7GHz link to maintain basic communication. Simultaneously, it quickly selects an alternative beam direction and attempts to rebuild the millimeter-wave link using historical CSI and environmental awareness information, achieving rapid link recovery and seamless switching.

[0367] This embodiment of the millimeter-wave link quality monitoring and recovery process is applicable when a millimeter-wave link experiences a momentary interruption due to movement, obstruction, or other reasons, such as a user moving and obstructing the antenna or environmental changes causing the disappearance of an unobstructed path. This embodiment demonstrates how to utilize a sub-7GHz link to assist in rapid handover and recovery of millimeter-wave communication.

[0368] The millimeter-wave link quality monitoring and recovery process in this embodiment has the following advantages:

[0369] 1. Seamless switching: Millimeter wave and sub-7GHz links work together to ensure "zero-aware" interruption;

[0370] 2. Intelligent prediction: Historical data reuse reduces beam recovery time by 90%;

[0371] 3. Dynamic Adaptation: Automatically adjust switching strategies based on business needs.

[0372] The millimeter-wave link quality monitoring and recovery process in this embodiment has typical application scenarios including but not limited to: surgical robot control: immediately switching to sub-7GHz to ensure command transmission when millimeter-wave transmission is interrupted; intelligent warehousing and logistics: AGV quickly restores millimeter-wave connection through historical beam data; VR live streaming vehicle: automatically selects anti-obstruction beam direction when the vehicle moves; through this embodiment, the system can quickly locate and repair millimeter-wave link faults while maintaining basic connectivity, achieving an uninterrupted communication experience.

[0373] In summary, the millimeter-wave communication method provided in this invention, based on a cross-band cooperative beam management mechanism, establishes a cooperative control mechanism between the sub-7GHz link and the millimeter-wave link. Channel information or environmental awareness information is acquired on the sub-7GHz link to assist in beam training and selection for the millimeter-wave link. The terminal combines historical CSI data with environmental information obtained from built-in sensors (such as IMU, camera, and positioning information) and utilizes a machine learning model to dynamically predict and determine the available beam range of the millimeter-wave link, thereby optimizing the training strategy. This mechanism significantly improves beam training efficiency and reduces millimeter-wave link control overhead. Related embodiments illustrate the system architecture and beam training process through example flowcharts, effectively improving millimeter-wave communication performance.

[0374] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A millimeter-wave communication method, characterized in that, include: The first device receives first information from the second device via a low-frequency link; The first device determines the candidate beam range of the millimeter-wave link communicating with the second device based on the first information; The first device initiates beam training in the millimeter-wave band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication.

2. The method according to claim 1, characterized in that, The first device and the second device are multi-link devices. The first device includes an access point (AP) device; the second device includes a terminal device.

3. The method according to claim 1, characterized in that, The first information includes at least one of the following: The second device obtains channel information through the low-frequency link; The second device acquires broadcast measurement information via the low-frequency link; The historical channel information of the second device; The attitude data information of the inertial measurement unit (IMU) of the second device; The visual information of the second device; Location information of the first device and / or the second device.

4. The method according to claim 1, characterized in that, The low-frequency links include sub-7GHz links.

5. The method according to claim 1, characterized in that, Before the first device receives the first information from the second device via a low-frequency link, the method further includes: The first device and the second device establish a communication connection on the low-frequency link.

6. The method according to claim 1, characterized in that, The first device initiates beam training in the millimeter-wave frequency band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication, including: The first device initiates sector-level scanning (SLS) beam training in the millimeter-wave band and receives short sector scanning (SSW) feedback from the second device. Based on the SSW feedback, the first device and the second device perform interactive beam training via the Beam Refinement Protocol (BRP) to determine the millimeter-wave link beam for millimeter-wave communication.

7. The method according to claim 6, characterized in that, After the first device initiates beam training in the millimeter-wave frequency band based on the candidate beam range to determine the millimeter-wave link beam for millimeter-wave communication, the method further includes: Based on the results of the beam training, update the historical channel information.

8. The method according to claim 1, characterized in that, Also includes: The first device sends control information for the millimeter-wave link to the second device via the anchor link and receives capability information of the millimeter-wave link from the second device, wherein the anchor link is a sub-7GHz link configured by the first device.

9. The method according to claim 1, characterized in that, In response to the first information including IMU attitude data information, the method further includes: The first device receives a first predicted candidate beam range from the second device, wherein the first predicted candidate beam range is determined by the second device based on the IMU attitude data information.

10. The method according to claim 1, characterized in that, In response to the first information including visual information, the method further includes: The first device receives a second predicted candidate beam range from the second device, wherein the second predicted candidate beam range is determined by the second device based on the visual information.

11. The method according to claim 1, characterized in that, In response to a first device providing millimeter-wave connectivity to a plurality of second devices, the method further includes: The first device determines the candidate beam range of the millimeter-wave link communicating with each of the second devices based on the first information of each of the second devices. The first device coordinates and triggers beam training of multiple second devices on the low-frequency link, and determines the millimeter-wave link beam for each second device for millimeter-wave communication in sequence or in parallel.

12. A millimeter-wave communication method, characterized in that, include: The second device sends the first information to the first device via a low-frequency link; The second device performs beam training in the millimeter-wave band based on the candidate beam range indicated by the first device to determine the millimeter-wave link beam for millimeter-wave communication, wherein the candidate beam range is determined by the first device based on the first information.

13. The method according to claim 12, characterized in that, The first information includes at least one of the following: The second device obtains channel information through the low-frequency link; The second device acquires broadcast measurement information via the low-frequency link; The historical channel information of the second device; The attitude data information of the inertial measurement unit (IMU) of the second device; The visual information of the second device; Location information of the first device and / or the second device.

14. The method according to claim 12, characterized in that, The low-frequency links include sub-7GHz links.

15. The method according to claim 12, characterized in that, The second device performs beam training in the millimeter-wave band based on the candidate beam range indicated by the first device to determine the millimeter-wave link beam for millimeter-wave communication, including: The second device receives an instruction for sector-level scanning SLS beam training from the first device and sends short sector scanning SSW feedback to the first device; The second device and the first device perform interactive beam training through the Beam Refinement Protocol (BRP) to determine the millimeter-wave link beam for millimeter-wave communication.

16. The method according to claim 12, characterized in that, Also includes: The second device selects a millimeter-wave reference beam from the historical millimeter-wave beams based on the historical channel information and the current first information, and sends an initial beam configuration to the first device based on the millimeter-wave reference beam, so that the first device determines the candidate beam range based on the initial beam configuration.

17. The method according to claim 12, characterized in that, In response to the first information including IMU attitude data information, the method further includes: The second device determines its attitude orientation relative to the first device based on IMU attitude data. The second device determines the range of the first predicted candidate beam based on the attitude direction information and sends the range of the first predicted candidate beam to the first device.

18. The method according to claim 12, characterized in that, In response to the first information including visual information, the method further includes: The second device determines the position information of the first device based on the visual information, and determines the three-dimensional vector direction of the first device relative to the second device based on the position information of the first device; The second device determines the second predicted candidate beam range based on the three-dimensional vector direction and sends the second predicted candidate beam range to the first device.

19. The method according to claim 12, characterized in that, Also includes: The second device performs machine learning training based on the first information in real time and determines whether to retrain the beam. In cases where beam training is required, the second device and the first device perform beam training based on the redefined candidate beam range; Otherwise, maintain the current millimeter-wave link.

20. The method according to claim 12, characterized in that, Also includes: The second device monitors the quality indicators of the millimeter-wave link in real time; If the quality index is less than a preset threshold, the second device notifies the first device of the millimeter-wave link anomaly via the low-frequency link, and determines a new millimeter-wave beam for millimeter-wave communication based on the first information; otherwise, the current millimeter-wave link is maintained.

21. A millimeter-wave communication system, characterized in that, include: The terminal and the access point (AP) establish communication connections between the AP and the terminal via low-frequency and millimeter-wave links. The AP receives first information from the terminal through the low-frequency link and determines the candidate beam range of the millimeter-wave link communicating with the terminal based on the first information. The AP and the terminal initiate beam training in the millimeter-wave band based on the candidate beam range to determine the millimeter-wave link beam used for millimeter-wave communication.

22. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the method described in any one of claims 1 to 11, or implements the steps of the method described in any one of claims 12 to 20.

23. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method described in any one of claims 1 to 11, or the steps of the method described in any one of claims 12 to 20.

24. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method described in any one of claims 1 to 11, or the steps of the method described in any one of claims 12 to 20.