High-speed rail scene public and private network collaborative optimization method, device, equipment, medium and product

Abstract

The invention provides a high-speed rail scene public and private network collaborative optimization method, device, equipment, medium and product, and the method comprises the steps: determining the network state characteristics of a public and private network based on obtained data which at least comprises the measurement report data of a first terminal accessed to a public network cell and a second terminal accessed to a private network cell, the public and private network state characteristics reflect the network states of the public network and the private network; the network state characteristics of the public network and the private network are input into a trained power adjustment model, power adjustment actions output by the power adjustment model are obtained, the power adjustment model is a model obtained through training based on a reinforcement learning algorithm, and the power adjustment actions comprise power adjustment schemes of all beams of the public network cell; and adjusting the beam power of the public network cell based on the power adjustment action. According to the scheme provided by the invention, independent frequency does not need to be allocated to the private network, beam level adjustment is carried out on the power of the public network cell in real time based on the public network and the private actual state, and the effect of reducing the interference of the public network to the private network is achieved.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to methods, devices, equipment, media and products for collaborative optimization of public and private networks in high-speed rail scenarios. Background Art

[0002] High-speed rail scenarios are key scenarios for wireless network optimization. High-speed rail users are highly concentrated and trains run at high speeds. To ensure the high-speed rail user experience, it is usually necessary to adopt a private network networking method for base stations covering high-speed rail along the line. To ensure the coverage and capacity requirements along the high-speed rail line, it is often necessary to configure a higher priority for the high-speed rail private network. However, there are also public network users in residential areas, ordinary roads and other scenarios around the high-speed rail, so public network base stations will also be built. When the high-speed rail passes through an area with dense public networks, the protection bands of the public network and the high-speed rail private network have fewer available frequencies, and the public and private networks require more carriers at the same time. At this time, the public and private networks will cause serious interference due to frequency overlap, which will have varying degrees of impact on both the public and private networks. Taking a certain operator as an example, both the 5G public network and the high-speed rail private network use the 2.6GHz frequency band. Although the SSB frequency is staggered, the service channels all use a 100M bandwidth of 2515-2615MHz. When the public network load reaches a certain threshold, the public network terminal will cause uplink co-frequency interference to the private network cell, seriously affecting the wireless network performance.

[0003] In the existing technology, the most common method for coordinating interference between high-speed rail public and private networks is to allocate dedicated frequencies or add protection bands for high-speed rail networks. However, wireless frequency resources are precious, and wireless frequency carriers are limited in areas with a large number of users. It is difficult to allocate independent frequencies for private networks. Therefore, in the existing technology, the phenomenon of high-speed rail private networks being interfered with by public networks is still relatively common. Summary of the invention

[0004] The present invention provides a method, device, equipment, medium and product for collaborative optimization of public and private networks in high-speed rail scenarios, so as to solve the defect in the prior art that the high-speed rail private network is interfered by the public network, and reduce the interference of the public network on the high-speed rail private network.

[0005] The present invention provides a method for collaborative optimization of public and private networks in a high-speed rail scenario, comprising: Acquire measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, and determine a network status characteristic of the public and private networks based on the acquired data, wherein the acquired data at least includes the measurement report data of the first terminal and the second terminal, and the network status characteristic of the public and private networks reflects the network status of the public network and the private network; Inputting the public-private network state feature into a trained power adjustment model, and obtaining a power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; The beam power of the public network cell is adjusted based on the power adjustment action.

[0006] According to a method for collaborative optimization of public and private networks in a high-speed rail scenario provided by the present invention, determining the network status characteristics of the public and private networks based on acquired data includes: Determine beam performance data of the public network cell based on the measurement report data of the first terminal, where the beam performance data reflects network performance of each beam of the public network cell; Determine beam interference data based on the measurement report data of the second terminal, where the beam interference data reflects interference of the second terminal by each beam of the public network cell; The network status characteristics of the public-private network are determined based on the extracted data, and the extracted data at least includes the beam performance data and the beam interference data.

[0007] According to a high-speed rail scenario public-private network collaborative optimization method provided by the present invention, before determining the public-private network status characteristics based on the extracted data, the method includes: Obtain base station performance data of the public network base station corresponding to the public network cell and the private network base station corresponding to the private network cell; and combine the beam performance data, the beam interference data and the base station performance data into the extracted data.

[0008] According to a high-speed rail scenario public-private network collaborative optimization method provided by the present invention, the beam interference data is determined based on the measurement report data of the second terminal, including: Determine whether the second terminal is subject to beam interference of the public network cell based on the measurement report data of the second terminal; Obtain proportion values ​​of terminals in the second terminals that are respectively interfered by each beam of the public network cell, and use each of the proportion values ​​as the beam interference data.

[0009] According to a high-speed rail scenario public-private network collaborative optimization method provided by the present invention, determining whether the second terminal is subject to beam interference from the public network cell based on measurement report data of the second terminal includes: When the second terminal accesses the private network cell, and the difference between the RSRP value of the private network cell and the RSRP value of the target beam of the public network cell in the measurement report data of the second terminal is within a preset range, it is determined that the second terminal is interfered by the target beam of the public network cell.

[0010] According to a method for collaborative optimization of public and private networks in a high-speed rail scenario provided by the present invention, when the power adjustment model is trained based on a reinforcement learning algorithm, the reward function is determined based on the channel quality of the public network and the private network after executing the sample power adjustment action output by the power adjustment model.

[0011] The present invention also provides a high-speed rail scenario public-private network collaborative optimization device, comprising: A network status feature acquisition module, configured to acquire measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, and determine a network status feature of a public-private network based on the acquired data, wherein the acquired data at least includes the measurement report data of the first terminal and the second terminal, and the network status feature of the public-private network reflects the network status of the public network and the private network; A model reasoning module, used for inputting the network status characteristics of the public-private network into a trained power adjustment model, and obtaining a power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; A power adjustment module is used to adjust the beam power of the public network cell based on the power adjustment action.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for collaborative optimization of public and private networks in high-speed rail scenarios as described in any one of the above methods is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements the public-private network collaborative optimization method for high-speed rail scenarios as described in any of the above-mentioned methods.

[0014] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for collaborative optimization of public and private networks in high-speed rail scenarios.

[0015] The method, device, equipment, medium and product for coordinated optimization of public and private networks in high-speed rail scenarios provided by the present invention determine the public and private network status characteristics reflecting the network status of the public network and the private network by acquiring data, and the acquired data at least includes measurement report data of the first terminal accessing the public network cell and the second terminal accessing the private network cell. The public and private network status characteristics are input into a power adjustment model trained by a reinforcement learning algorithm, and the power adjustment model outputs a power adjustment scheme at the beam level of the public network cell. The method provided by the present invention does not need to allocate an independent frequency to the private network, but adjusts the beam power of the public network cell in real time based on the actual status of the existing public and private networks, thereby achieving the effect of reducing the interference of the public network to the private network. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 It is one of the flow charts of the public-private network collaborative optimization method for high-speed rail scenarios provided by the present invention.

[0018] Figure 2 This is one of the scenario diagrams of the high-speed rail scenario public-private network collaborative optimization method provided by the present invention.

[0019] Figure 3 This is the second flow chart of the public-private network collaborative optimization method for high-speed rail scenarios provided by the present invention.

[0020] Figure 4 It is a structural schematic diagram of the power adjustment model provided by the present invention.

[0021] Figure 5 This is the second scenario diagram of the high-speed rail scenario public-private network collaborative optimization method provided by the present invention.

[0022] Figure 6 It is a structural diagram of the public-private network collaborative optimization device for high-speed rail scenarios provided by the present invention.

[0023] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] Combine the following Figure 1-Figure 5 The public-private network collaborative optimization method for high-speed rail scenarios provided by the present invention is described. Figure 1 As shown, the high-speed rail scenario public-private network collaborative optimization method includes the following steps: S110, obtaining measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, and determining network status characteristics of the public and private networks based on the obtained data, wherein the obtained data at least includes measurement report data of the first terminal and the second terminal, and the network status characteristics of the public and private networks reflect network statuses of the public network and the private network; S120, inputting the network status characteristics of the public and private networks into the trained power adjustment model, and obtaining the power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; S130. Adjust the beam power of the public network cell based on the power adjustment action.

[0026] like Figure 2 As shown, along the railway, there are both high-speed rail private network cells and public network cells. The high-speed rail private network cells provide network access for high-speed rail fast-moving users, and the public network cells provide network access for mobile users in ordinary scenarios (residential areas, ordinary roads). The public network may interfere with the private network. The method provided by the present invention obtains the measurement report data of the first terminal accessing the public network cell and the second terminal accessing the private network cell, determines the public-private network state characteristics reflecting the network states of the public network and the private network, and inputs them into the power adjustment model obtained by training the reinforcement learning algorithm. The power adjustment model outputs the power adjustment scheme of the public network cell beam level. The method provided by the present invention does not need to allocate independent frequencies for the private network, but adjusts the beam power of the public network cell in real time based on the actual status of the public network and the private network, so as to achieve the effect of reducing the interference of the public network to the private network.

[0027] like Figure 3 As shown, the method provided by the present invention generates a public-private network status feature for reflecting the network status of the public network and the private network by acquiring measurement report (MR) data reported by a first terminal accessing a public network cell and a second terminal accessing a private network cell. Specifically, determining the public-private network status feature based on the acquired data includes: Determine beam performance data of the public network cell based on the measurement report data of the first terminal, where the beam performance data reflects the network performance of each beam of the public network cell; Determine beam interference data based on the measurement report data of the second terminal, where the beam interference data reflects interference of the second terminal by each beam of the public network cell; The network status characteristics of the public and private networks are determined based on the extracted data, where the extracted data at least includes beam performance data and beam interference data.

[0028] The method provided by the present invention describes the beam-level network status of the public network cell from two aspects: beam performance data and beam interference data. The beam performance data reflects the network performance of each beam of the public network cell, while the beam interference data reflects the interference of each beam of the public network cell to the dedicated cell. In this way, the beam-level network status of the public network cell can be accurately described, and the accuracy of the beam-level power adjustment action output subsequently can be improved. When the terminal accesses a wireless cell, it will periodically report a measurement report (MR). The measurement report can obtain the SSB (Synchronization Signal Block) beam number accessed by the current terminal, the RSRP (Reference Signal Receiving Power) value of the accessed cell, and the RSRP value of each beam of the neighboring cell. The beam performance data and beam interference data can be determined through the measurement report data of the first terminal and the second terminal.

[0029] Specifically, determining the beam performance data of the public network cell based on the measurement report data of the first terminal includes: The number of first terminals accessing each beam of the public network cell and the average RSRP value of the first terminals accessing are obtained as beam performance data.

[0030] Assuming that the number of SSB beams configured in the existing network is n (generally, n is 8, and the value of n is related to the network configuration and is not limited), then 2n-dimensional beam performance data can be obtained.

[0031] By obtaining the number of first terminals accessing each beam of the public network cell and the average RSRP value of the first terminals accessing, beam performance data that can more accurately reflect the performance of each beam of the public network cell can be obtained.

[0032] It is understandable that in addition to the RSRP value as an indicator reflecting the beam performance, other indicators that can reflect the beam performance can also be used in the beam performance data.

[0033] Determining beam interference data based on measurement report data of the second terminal includes: Determining whether the second terminal is subject to beam interference of the public network cell based on the measurement report data of the second terminal; The proportion values ​​of the terminals in the second terminals that are respectively interfered by the respective beams of the public network cell are obtained, and the respective proportion values ​​are used as beam interference data.

[0034] In the method provided by the present invention, the interference situation of the second terminal caused by each beam of the public network cell is determined by measuring the report data, and the interference situation of the public network cell to the private network cell is specified to the beam level, thereby improving the accuracy of the subsequent output beam-level frequency adjustment plan for the public network cell.

[0035] More specifically, determining whether the second terminal receives beam interference from a public network cell based on the measurement report data of the second terminal includes: When the second terminal accesses the private network cell, if the difference between the RSRP value of the private network cell and the RSRP value of the target beam of the public network cell in the measurement report data of the second terminal is within a preset range, it is determined that the second terminal is interfered by the target beam of the public network cell.

[0036] In the method provided by the present invention, whether the private network cell network is interfered by a certain beam of the adjacent public network cell is determined by comparing the RSRP of the accessed private network cell with the RSRP of each beam of the adjacent public network cell. Specifically, in the measurement report, the terminal reports the measured RSRP of the currently accessed cell and the RSRP of each beam of the adjacent cell. When the RSRP of the private network cell measured by the second terminal accessing the private network cell is close to the RSRP of a certain beam of the public network cell measured, it is considered that the second terminal is interfered by the beam of the public network cell.

[0037] In one possible implementation, the preset range can be set to 6dB, that is, when the difference between the RSRP of the private network cell measured by the second terminal and the RSRP of the target beam of the public network cell measured is within 6dB, it is determined that the second terminal is interfered by the target beam of the public network cell.

[0038] For each beam of the public network cell, the number of second terminals affected by the beam interference can be obtained. By taking the number of second terminals affected by the beam interference as the numerator and the number of all second terminals as the denominator, a proportional value can be obtained. There are n beams in total, so n proportional values ​​can be obtained. The n proportional values ​​constitute n-dimensional beam interference data.

[0039] From the above description, it can be seen that based on the measurement report data, data of 3n dimensions can be obtained. In a possible implementation, the beam performance data and beam interference data can be directly combined into the public-private network status characteristics. Research has found that the performance indicators on the base station side can also describe the current network status. Therefore, further, in one embodiment of the method provided by the present invention, before determining the public-private network status characteristics based on the extracted data, it includes: Obtain base station performance data of the public network base station corresponding to the public network cell and the private network base station corresponding to the private network cell; The beam performance data, beam interference data and base station performance data are combined to extract data.

[0040] The base station performance data includes the performance indicators of the base station. The base station performance data can also reflect the current network status. By adding the base station performance data to the public and private network status characteristics, the public and private network status characteristics that more accurately reflect the current network status can be obtained, thereby improving the accuracy of the frequency adjustment action generated based on the public and private network status characteristics.

[0041] In a possible implementation, the base station performance data may include the number of access users, PDCP (Packet Data Convergence Protocol) layer rate, PDCP layer packet loss number, PDCP layer delay, and PDCP layer delay jitter. There are base station performance data corresponding to public network cell base stations and private network cell base stations, so there are 5x2 dimensional features in total.

[0042] Combine beam performance data, beam interference data, and base station performance data to form a (3n+5x2)-dimensional feature vector that describes the current public and private network status. As a public and private network status feature.

[0043] After obtaining the public and private network status features, they are input into the trained power adjustment model to obtain the power adjustment action output by the power adjustment model. Based on the power adjustment action, the beam power of the public network cell is adjusted to reduce the interference of the private network to the public network. In order to facilitate model processing, the data in the public and private network status features can be normalized before being input into the power adjustment model.

[0044] In the method provided by the present invention, the power adjustment model is a model trained based on a reinforcement learning algorithm. The reinforcement learning algorithm is a learning algorithm based on an Actor-Critic network, which realizes the coordinated integration of reinforcement learning and high-speed rail public-private network interference, uses the Actor-Critic network to learn to adjust power, and fully and comprehensively describes the public-private network status from two levels: base station statistics and user reports. Sample labeling is not required, which improves learning efficiency.

[0045] Specifically, the power adjustment model is the Actor network in the Actor-Critic network, that is, the policy network, which outputs power adjustment actions according to the current network status of the public and private networks. During the reinforcement learning process, the power adjustment model is trained together with the Critic network in the Actor-Critic network. The Critic network outputs the score of the corresponding adjustment action according to the current network status of the public and private networks.

[0046] In the method provided by the present invention, the power adjustment model outputs a power adjustment plan for each beam of the public network cell. Through reinforcement learning, the power adjustment model can learn the impact of each beam of the public network on the high-speed rail private network, adaptively adjust the power of beams that cause different degrees of interference, realize user migration, and do not adjust the beams that do not cause interference, thereby maximizing the public network performance.

[0047] In order to improve the output efficiency of the power adjustment model, the method provided by the present invention predefines the action space of the power adjustment action output by the power adjustment model, that is, the power adjustment action output by the power adjustment model will only be the action existing in the action space. Specifically, the power adjustment action space is defined , describes the power adjustment action of each public network cell beam. In a possible implementation, That is, each beam can independently increase 1dB, remain unchanged, or decrease 1dB. The action space 3n represents the action of reducing the power of the nth beam by 1dB, 3n+1 represents the action of keeping the power of the nth beam unchanged, 3n+2 represents the action of increasing the power of the nth beam by 1dB, n is the beam number, and the upper limit of the power adjustment of each beam in the action space is controlled within ±6dB. By defining a certain range of action space, the power adjustment model can output power adjustment actions within a predetermined range, reducing the complexity of adjusting the beam power of the private network cell and making the solution more feasible.

[0048] In the method provided by the present invention, when the power adjustment model is trained based on the reinforcement learning algorithm, the reward function is determined based on the channel quality of the public network and the private network after executing the sample power adjustment action output by the power adjustment model. Specifically, the reward function is defined as , Indicates that at the time of the i-th request, the current network status is Take power adjustment action Execute power adjustment action Then, the channel quality indicator (CQI) of the public and private networks is obtained. The CQI can reflect the interference coordination of the current public and private networks. The higher the CQI of the public and private networks, the smaller the interference between the two networks. Therefore, the corresponding reward is calculated based on the CQI. The reward function can be set as: ; in It is the preset benchmark CQI. When the CQI of any public or private network is lower than the benchmark CQI, the reward is negative. When the CQI of both the public and private networks are higher than the benchmark CQI, the reward is positive. The higher the CQI value, the higher the reward. The goal of the system is to minimize the interference between the public and private networks while ensuring a high CQI.

[0049] In the method provided by the present invention, the public-private network status characteristics input into the power adjustment model include data reflecting the public network status and the private network status, and in the process of reinforcement learning, the reward value is determined based on the CQI of the public network and the private network. Therefore, the method provided by the present invention takes into account the network performance of the public network and the private network at the same time, and tries not to reduce the performance of the public network while reducing interference. Compared with traditional methods, better public-private network collaboration can be achieved.

[0050] Considering that both Actor and Critic networks are based on public and private network state feature vectors The hidden feature extraction of the middle layer is the same for both networks, so this solution uses the method of sharing parameter weights for the two networks, and only differs in the final fully connected layer. The Actor (i.e., the power adjustment model) is used as the policy network to output the power adjustment action, and the Critic is used as the value network to output the score. The advantage of this is that it directly reduces the parameters of the network and reduces the risk of overfitting. In a possible implementation, the network structure is as follows Figure 4 As shown in the figure, the first three layers are shared fully connected layers, which transform the public and private network state feature vector Mapped into higher-dimensional hidden features, the Actor strategy network passes the hidden features through a fully connected layer and then outputs the probability of each beam power adjustment through the softmax layer. The Critic value network passes the hidden features through a fully connected layer and then outputs a 1-dimensional value, which is the evaluation value of the current state s.

[0051] The following is a detailed description of the gradient update process in reinforcement learning. Based on the above Actor-Critic network composed of the power adjustment model in the present invention, the PPO2 algorithm is used to update the policy gradient, and the optimization objective function is used. As follows, as the Actor loss (i.e. the loss of the power adjustment model):

[0052] Where E represents expectation, Is the probability ratio, indicating the current network for the current state The output action probability is the same as the network's prediction of the current state before the last update. Output action probability ratio; is a small positive number, such as 0.2; clip() limits the range of probability ratios to and To prevent updates from being too large or too small; is the advantage function, let is the discount factor, which is 0.9, indicating that the reward of future actions is discounted when calculating the action reward. , Represents the evaluation value of the current state estimated by the Critic network.

[0053] Critic loss uses TD loss, TD loss Denoted as: ; In order to obtain immediate rewards after taking actions, reinforcement learning gradually improves its understanding of the dynamics of the environment and makes better decisions based on it by continuously calculating and reducing the TD error.

[0054] Finally, the Actor loss and Critic loss are proportional to each other. Add them together to get the total loss of the network. Because the Actor network needs to maximize the target, it does gradient ascent. So we can get the final loss , use the total loss Loss for gradient update.

[0055] The reinforcement learning module after training is deployed, that is, the Actor network in the Actor-Critic network after training is deployed as a trained power adjustment model. The public network base station and the private network base station can interact with the terminal deployed with the power adjustment model respectively to obtain the measurement report data of the first terminal accessing the public network cell and the second terminal accessing the private network cell, so as to execute the high-speed rail scenario public-private network collaborative optimization method provided by the present invention, and after obtaining the power adjustment action output by the power adjustment model, send it to the public network base station to realize the adjustment of the beam power of the public network cell, such as Figure 5 As shown, after receiving the power adjustment action, the public network base station performs power adjustment or power offset adjustment on each SSB beam of the public network cell in turn.

[0056] The following is a description of the high-speed rail scenario public-private network collaborative optimization device provided by the present invention. The high-speed rail scenario public-private network collaborative optimization device described below and the high-speed rail scenario public-private network collaborative optimization method described above can be referred to each other. Figure 6 As shown, the high-speed rail scenario public-private network collaborative optimization device provided by the present invention includes: The network status feature acquisition module 610 is used to acquire measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, and determine the network status features of the public and private networks based on the acquired data, wherein the acquired data at least includes the measurement report data of the first terminal and the second terminal, and the network status features of the public and private networks reflect the network status of the public network and the private network; The model inference module 620 is used to input the network status characteristics of the public and private networks into the trained power adjustment model to obtain the power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on the reinforcement learning algorithm, and the power adjustment action includes the power adjustment scheme of each beam of the public network cell; The power adjustment module 630 is used to adjust the beam power of the public network cell based on the power adjustment action.

[0057] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute the public-private network collaborative optimization method in the high-speed rail scenario, the method comprising: obtaining measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, determining the network status characteristics of the public-private network based on the obtained data, the obtained data at least including the measurement report data of the first terminal and the second terminal, the network status characteristics of the public-private network reflecting the network status of the public network and the private network; inputting the network status characteristics of the public-private network into the trained power adjustment model, obtaining the power adjustment action output by the power adjustment model, wherein the power adjustment model is a model obtained by training based on the reinforcement learning algorithm, and the power adjustment action includes the power adjustment scheme of each beam of the public network cell; adjusting the beam power of the public network cell based on the power adjustment action.

[0058] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0059] On the other hand, the present invention also provides a computer program product, which includes a computer program, and the computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the high-speed rail scenario public-private network collaborative optimization method provided by the above methods, the method including: obtaining measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, determining public-private network status characteristics based on the obtained data, the obtained data at least including measurement report data of the first terminal and the second terminal, and the public-private network status characteristics reflect the network status of the public network and the private network; inputting the public-private network status characteristics into a trained power adjustment model, and obtaining a power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; adjusting the beam power of the public network cell based on the power adjustment action.

[0060] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the public-private network collaborative optimization method for high-speed rail scenarios provided by the above-mentioned methods, the method comprising: obtaining measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, determining network status characteristics of the public and private networks based on the obtained data, the obtained data at least including the measurement report data of the first terminal and the second terminal, the network status characteristics of the public and private networks reflecting the network status of the public network and the private network; inputting the network status characteristics of the public and private networks into a trained power adjustment model, obtaining a power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; and adjusting the beam power of the public network cell based on the power adjustment action.

[0061] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0062] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for collaborative optimization of public and private networks in high-speed rail scenarios, characterized in that: include: Acquire measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, and determine a network status characteristic of the public and private networks based on the acquired data, wherein the acquired data at least includes the measurement report data of the first terminal and the second terminal, and the network status characteristic of the public and private networks reflects the network status of the public network and the private network; Inputting the public-private network state feature into a trained power adjustment model, and obtaining a power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; The beam power of the public network cell is adjusted based on the power adjustment action.

2. The method for collaborative optimization of public and private networks in high-speed rail scenarios according to claim 1 is characterized in that: The determining of the network status characteristics of the public and private networks based on the acquired data includes: Determine beam performance data of the public network cell based on the measurement report data of the first terminal, where the beam performance data reflects network performance of each beam of the public network cell; Determine beam interference data based on the measurement report data of the second terminal, where the beam interference data reflects interference of the second terminal by each beam of the public network cell; The network status characteristics of the public-private network are determined based on the extracted data, and the extracted data at least includes the beam performance data and the beam interference data.

3. The method for collaborative optimization of public and private networks in high-speed rail scenarios according to claim 2 is characterized in that: Before determining the public-private network status characteristics based on the extracted data, the method includes: Acquire base station performance data of a public network base station corresponding to the public network cell and a private network base station corresponding to the private network cell; The beam performance data, the beam interference data and the base station performance data are combined into the extracted data.

4. The method for collaborative optimization of public and private networks in high-speed rail scenarios according to claim 2 is characterized in that: The determining the beam interference data based on the measurement report data of the second terminal includes: Determine whether the second terminal is subject to beam interference of the public network cell based on the measurement report data of the second terminal; Obtain proportion values ​​of terminals in the second terminals that are respectively interfered by each beam of the public network cell, and use each of the proportion values ​​as the beam interference data.

5. The method for collaborative optimization of public and private networks in high-speed rail scenarios according to claim 2 is characterized in that: The determining, based on the measurement report data of the second terminal, whether the second terminal is subject to beam interference of the public network cell includes: When the second terminal accesses the private network cell, and the difference between the RSRP value of the private network cell and the RSRP value of the target beam of the public network cell in the measurement report data of the second terminal is within a preset range, it is determined that the second terminal is interfered by the target beam of the public network cell.

6. The method for collaborative optimization of public and private networks in high-speed rail scenarios according to claim 1 is characterized in that: When the power adjustment model is trained based on a reinforcement learning algorithm, the reward function is determined based on the channel quality of the public network and the private network after executing the sample power adjustment action output by the power adjustment model.

7. A high-speed rail scenario public-private network collaborative optimization device, characterized in that: include: A network status feature acquisition module, configured to acquire measurement report data of a first terminal accessing a public network cell and a second terminal accessing a private network cell, and determine a network status feature of a public-private network based on the acquired data, wherein the acquired data at least includes the measurement report data of the first terminal and the second terminal, and the network status feature of the public-private network reflects the network status of the public network and the private network; A model reasoning module, used for inputting the network status characteristics of the public-private network into a trained power adjustment model, and obtaining a power adjustment action output by the power adjustment model, wherein the power adjustment model is a model trained based on a reinforcement learning algorithm, and the power adjustment action includes a power adjustment scheme for each beam of the public network cell; A power adjustment module is used to adjust the beam power of the public network cell based on the power adjustment action.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, it implements the method for collaborative optimization of public and private networks in high-speed rail scenarios as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the method for collaborative optimization of public and private networks in high-speed rail scenarios as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by the processor, the method for collaborative optimization of public and private networks in high-speed rail scenarios as described in any one of claims 1 to 6 is implemented.

Images