Environment feature assisted channel multipath intelligent prediction method and device

Abstract

The invention provides a channel multipath intelligent prediction method and device assisted by environmental characteristics. Comprising the steps that channel multipath component geometric information and channel multipath component field intensity information of a to-be-predicted terminal at multiple historical moments, coordinate information corresponding to the to-be-predicted terminal at the multiple historical moments, scene geometric features of a wireless propagation environment where the to-be-predicted terminal is located, and channel multipath component field intensity information of the to-be-predicted terminal at the multiple historical moments are obtained; the coordinate information and the material electromagnetic parameter features of the wireless propagation environment are respectively input into a feature extraction sub-network of the multipath long-period prediction network to obtain three feature maps, and the three feature maps are input into a sequence prediction sub-network of the multipath long-period prediction network to obtain channel multipath component prediction information of the to-be-predicted terminal at a plurality of future moments. According to the embodiment of the invention, the network of the time sequence prediction task oriented to the channel multipath is established, the environmental characteristics are introduced, the influence of the environment on the electromagnetic wave propagation is excavated, the prediction precision is further improved, and the long-period prediction capability is ensured.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to an environmental feature-assisted channel multipath intelligent prediction method and device. Background Art

[0002] Sixth-generation (6G) mobile communication systems are expected to provide intelligent, ultra-reliable, and ubiquitous connectivity, supporting the Internet of Everything (IoE) trend of the future digital society. However, the global coverage and diverse scenarios brought by 6G bring extreme complexity and new challenges to network design and technology optimization. Digital twin technology is considered an innovative tool to overcome these challenges. The propagation behavior of electromagnetic waves and their physical layer operations are mapped in the digital world and defined as the Digital Twin Channel (DTC). DTC is expected to have the ability to predict future channel changes, facilitate the optimization, early warning, and decision-making of various tasks in the network, such as network planning, resource allocation, and link operation, and achieve active adaptation to the complex wireless environment of 6G.

[0003] With the development of artificial intelligence (AI), the use of AI to intelligently predict wireless channels has been widely explored. However, AI-based channel prediction currently focuses on predicting channel parameters at large and small scales, such as path loss and delay spread. There is a lack of methods for predicting the multipath components (MPCs) of the channel at future times, including their geometric and field strength information. Summary of the Invention

[0004] The purpose of the present invention is to provide an environmental feature-assisted channel multipath intelligent prediction method and device to solve the problem that the existing technology based on AI channel prediction still focuses on predicting the large and small scale parameters of the channel, such as path loss, delay spread, etc., and lacks the problem of predicting MPCs at future moments, including the geometric information and field strength information of MPCs.

[0005] To achieve the above objectives, an embodiment of the present invention provides an environmental feature-assisted channel multipath intelligent prediction method, which includes:

[0006] Inputting the channel multipath component geometry information and channel multipath component field strength information of the terminal to be predicted at multiple historical moments into the first feature extraction subnetwork of the multipath long-period prediction network to obtain a first feature map; wherein the first feature map includes the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted;

[0007] Inputting the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long-period prediction network to obtain a second feature map; wherein the second feature map includes the relationship between the geometry of the channel multipath component and the scene geometric features;

[0008] Inputting the coordinate information and the electromagnetic parameter characteristics of the material of the wireless propagation environment into the third feature extraction subnetwork of the multipath long period prediction network to obtain a third feature map; wherein the third feature map includes the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material;

[0009] The first feature map, the second feature map, and the third feature map are input into the sequence prediction subnetwork of the multipath long-period prediction network to obtain channel multipath component prediction information of the terminal to be predicted at multiple future moments; wherein the channel multipath component prediction information includes channel multipath component prediction geometry information and channel multipath component prediction field strength information.

[0010] Optionally, the method, wherein the channel multipath component geometry information and the channel multipath component field strength information at multiple historical moments of the terminal to be predicted are input into a first feature extraction subnetwork of the multipath long-period prediction network to obtain a first feature map, includes:

[0011] Acquire the multipath long period prediction network;

[0012] Acquire the channel multipath component geometry information and the channel multipath component field strength information through a ray tracing simulation model or actual channel measurement;

[0013] The channel multipath component geometry information and the channel multipath component field strength information are input into the first feature extraction subnetwork of the multipath long period prediction network for feature extraction, the relationship between the geometry and field strength of the channel multipath component is mined, and the first feature map is obtained.

[0014] Optionally, in the method, the acquiring the multipath long-period prediction network includes:

[0015] Adopting deep learning architecture, we build a multi-path long-period prediction network to be trained;

[0016] Setting the hyperparameters of the multipath long-period prediction network to be trained and determining the loss function;

[0017] Acquire channel multipath components of a plurality of continuously moving positions and construct a time series data set of the channel multipath components;

[0018] The loss function and the time series data set are used to train the multipath long-period prediction network to be trained, so as to obtain the multipath long-period prediction network.

[0019] Optionally, the method further comprises, before inputting the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long period prediction network to obtain the second feature map:

[0020] Establishing a three-dimensional environmental model of the wireless propagation environment;

[0021] Acquiring, according to the environment model, scatterer information of the wireless propagation environment and motion information of the terminal to be predicted;

[0022] Acquire the scene geometric features and the material electromagnetic parameter features according to the scatterer information;

[0023] The coordinate information is acquired according to the motion information.

[0024] Optionally, in the method, the scatterer information includes one or more of the following:

[0025] Scatterer position;

[0026] Scatter shape;

[0027] Scatter size;

[0028] Electromagnetic parameters of scatterer materials.

[0029] Optionally, the method, wherein the first feature map, the second feature map, and the third feature map are input into the sequence prediction subnetwork of the multipath long period prediction network to obtain channel multipath component prediction information of the terminal to be predicted at multiple future moments, includes:

[0030] Inputting the first feature map, the second feature map, and the third feature map into the sequence prediction subnetwork to obtain splicing features;

[0031] Use neural network models to obtain the evolution law of time series;

[0032] According to the splicing features and the time series evolution law, the channel multipath component prediction information at multiple future moments is obtained.

[0033] Optionally, in the method, the channel multipath component prediction information includes one or more of the following:

[0034] Channel multipath component prediction geometry information;

[0035] Channel multipath component prediction field strength information.

[0036] To achieve the above objectives, an embodiment of the present invention provides an environmental feature-assisted channel multipath intelligent prediction device, which includes:

[0037] A first acquisition module is configured to input the channel multipath component geometry information and channel multipath component field strength information of a terminal to be predicted at multiple historical moments into a first feature extraction subnetwork of a multipath long-period prediction network to obtain a first feature map; wherein the first feature map includes the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted;

[0038] A second acquisition module is configured to input the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long period prediction network to obtain a second feature map; wherein the second feature map includes the relationship between the geometry of the channel multipath component and the scene geometric features;

[0039] A third acquisition module is configured to input the coordinate information and the electromagnetic parameter characteristics of the material of the wireless propagation environment into a third feature extraction subnetwork of the multipath long period prediction network to obtain a third feature map; wherein the third feature map includes the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material;

[0040] The fourth acquisition module is used to input the first feature map, the second feature map and the third feature map into the sequence prediction subnetwork of the multipath long-period prediction network to obtain the channel multipath component prediction information of the terminal to be predicted at multiple future moments; wherein the channel multipath component prediction information includes channel multipath component prediction geometry information and channel multipath component prediction field strength information.

[0041] To achieve the above-mentioned objectives, an embodiment of the present invention provides an electronic device, comprising: a transceiver, a processor, a memory, and a program or instruction stored on the memory and executable on the processor; wherein, when the processor executes the program or instruction, the channel multipath intelligent prediction method assisted by environmental features as described above is implemented.

[0042] To achieve the above-mentioned purpose, an embodiment of the present invention provides a readable storage medium having a program or instruction stored thereon, wherein when the program or instruction is executed by a processor, the steps in the channel multipath intelligent prediction method assisted by environmental features as described above are implemented.

[0043] To achieve the above objectives, an embodiment of the present invention provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, the steps of the above-mentioned environmental feature-assisted channel multipath intelligent prediction method are implemented.

[0044] The beneficial effects of the above technical solution of the present invention are as follows:

[0045] In an embodiment of the present invention, the channel multipath component geometry information, channel multipath component field strength information, coordinate information, scene geometry characteristics of the wireless propagation environment, and material electromagnetic parameter characteristics of the terminal to be predicted at multiple historical moments are input into different feature extraction subnetworks of the multipath long-period prediction network, and multiple feature maps are obtained, including the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted, the scene geometry characteristics, and the scene material electromagnetic parameter characteristics. The multiple feature maps are then input into the sequence prediction subnetwork of the multipath long-period prediction network to obtain channel multipath component prediction information for multiple future moments. The above embodiment builds a network for the time series prediction task of channel multipath. At the same time, in terms of prediction capability, in order to further improve the prediction performance, environmental features, including the geometry and material characteristics of the environment, are introduced to explore the impact of the environment on electromagnetic wave propagation, further improve the prediction accuracy, and ensure the ability of long-period prediction. This improves the problem of the existing technology in AI-based channel prediction, which lacks MPCs for predicting future moments, including geometric information and field strength information of MPCs. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic diagram of an environmental feature-assisted channel multipath intelligent prediction method according to an embodiment of the present invention;

[0047] Figure 2 A schematic diagram of the MPCs geometry and field strength variation rules of the channel multipath intelligent prediction method assisted by environmental features according to an embodiment of the present invention;

[0048] Figure 3 This is a flow chart of the channel multipath intelligent prediction method assisted by environmental features according to an embodiment of the present invention;

[0049] Figure 4 A schematic diagram of a multipath long-period prediction network of the channel multipath intelligent prediction method assisted by environmental characteristics according to an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the evaluation and prediction results of the channel multipath intelligent prediction method assisted by environmental characteristics according to an embodiment of the present invention;

[0051] Figure 6 This is a second schematic diagram of evaluation and prediction results of the channel multipath intelligent prediction method assisted by environmental characteristics according to an embodiment of the present invention;

[0052] Figure 7 This is a schematic diagram of an intelligent channel multipath prediction device assisted by environmental features according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0054] It should be understood that references throughout this specification to "one embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic associated with the embodiment is included in at least one embodiment of the present invention. Therefore, the appearances of "in one embodiment" or "in an embodiment" throughout this specification do not necessarily refer to the same embodiment. Furthermore, these particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0055] In various embodiments of the present invention, it should be understood that the size of the serial numbers of the following processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0056] Additionally, the terms "system" and "network" are often used interchangeably herein.

[0057] In the embodiments provided herein, it should be understood that "B corresponding to A" means that B is associated with A and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B based solely on A; B can also be determined based on A and / or other information.

[0058] For ease of understanding, some contents involved in the embodiments of the present invention are described below:

[0059] like Figure 1 As shown, an environmental feature-assisted channel multipath intelligent prediction method according to an embodiment of the present invention includes:

[0060] S10: Inputting the channel multipath component geometry information and channel multipath component field strength information of the terminal to be predicted at multiple historical moments into a first feature extraction subnetwork of the multipath long-period prediction network to obtain a first feature map; wherein the first feature map includes the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted;

[0061] It should be noted that if Figure 2 As shown, the reflection of electromagnetic waves is taken as an example. The action point of MPCs is defined as the interaction point between the electromagnetic wave and the building, such as the reflection points r1, r2, r in the reflection example. nThe action point of MPCs contains the complete geometric information of MPCs, which can reconstruct the propagation path of electromagnetic waves; the receiving field strength of MPCs is the electric field strength of the path received by the receiving antenna. The action points and receiving field strengths of MPCs of vehicles in different positions are different, but there is an evolution law. Therefore, based on AI models such as Transformer, LSTM, etc., which have powerful sequence law learning capabilities, this law is expected to be effectively learned by neural networks and predict the multipath propagation in the future. Figure 3 As shown, in step S10, historical MPCs (i.e., the channel multipath component geometry information and channel multipath component field strength information at multiple historical moments of the terminal to be predicted) are obtained and input into EAML-net subnetwork 1 (i.e., the first feature extraction subnetwork of the multipath long-period prediction network) to obtain high-dimensional features F1 (i.e., the first feature graph), thereby obtaining the first feature graph that can reflect the relationship between the geometry and field strength of the channel multipath component. Figure 4 As shown, in sub-network 1: multipath feature extraction network, taking the historical k moment as an example, the historical k moment MPC action point coordinate G k (equivalent to the above reflection points r1, r2,., r n 1), and the MPC field strength E at the historical time k k After inputting the multi-layer perceptron, they are input into the multi-head attention mechanism layer to obtain the high-dimensional feature F1 (i.e., the first feature map).

[0062] S20: Input the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long-period prediction network to obtain a second feature map; wherein the second feature map includes the relationship between the geometry of the channel multipath component and the scene geometric features;

[0063] It should be noted that if Figure 2 As shown, at different times, the receiver moves continuously and is located at Rx1, ..., Rx i ,…,Rx n Multiple locations, the receiver is the terminal to be predicted. Figure 4 As shown, in sub-network 2: scene geometric feature extraction network, taking the historical k moment as an example, the scene geometric feature O and the historical k moment Rx coordinate P k (equivalent to Rx1, ..., Rx i ,…,Rx n One of multiple positions) is respectively input into the multi-layer perceptron and then into the multi-layer convolution block to obtain a geometric feature map F2 (that is, the second feature map), thereby obtaining the second feature map that can reflect the relationship between the geometry of the channel multipath component and the scene geometric features.

[0064] S30, inputting the coordinate information and the electromagnetic parameter characteristics of the material of the wireless propagation environment into the third feature extraction subnetwork of the multipath long period prediction network to obtain a third feature map; wherein the third feature map includes the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material;

[0065] It should be noted that if Figure 3 As shown, in step S20 and step S30, the scene geometry (i.e., the scene geometry characteristics of the wireless propagation environment where the terminal to be predicted is located) and the material electromagnetic parameter characteristics (i.e., the material electromagnetic parameter characteristics of the wireless propagation environment) are obtained and respectively input into EAML-net subnetworks 2 and 3 (i.e., the second feature extraction subnetwork and the third feature extraction subnetwork of the multipath long period prediction network) to obtain high-dimensional features F2, F3 (i.e., the second feature map and the third feature map). Figure 4 As shown, in sub-network 3: scene material feature extraction network, taking the historical k moment as an example, the scene material electromagnetic parameter feature I and the historical k moment Rx coordinate P k (equivalent to Rx1, ..., Rx i ,…,Rx n One of multiple positions) is respectively input into the multi-layer perceptron and then into the multi-layer convolution block to obtain the material electromagnetic feature map F3 (that is, the third feature map), thereby obtaining the third feature map that can reflect the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material.

[0066] S40: Inputting the first feature map, the second feature map, and the third feature map into the sequence prediction subnetwork of the multipath long-period prediction network to obtain channel multipath component prediction information for the terminal to be predicted at multiple future moments; wherein the channel multipath component prediction information includes channel multipath component prediction geometry information and channel multipath component prediction field strength information;

[0067] It should be noted that if Figure 3 As shown, in step S40, F1, F2, F3 (i.e., the first feature map, the second feature map, and the third feature map) are input into the EAML-net subnetwork 4 (i.e., the sequence prediction subnetwork of the multipath long-period prediction network) to predict the MPCs (i.e., the channel multipath component prediction information of the terminal to be predicted) at multiple future moments. Figure 4As shown, in subnetwork 4: multipath sequence prediction network, F1, F2, F3 (i.e., the first feature map, the second feature map, the third feature map) are spliced and merged, input into the Transformer time series prediction layer and then into the multi-layer convolution block to obtain the channel multipath component prediction information of the terminal to be predicted at multiple future moments. Taking the future moment t as an example, it is divided into the MPC action point coordinates G at the future moment t t (i.e., the channel multipath component prediction geometry information) and the MPC field strength E at the future time t t (i.e., the channel multipath component predicted field strength information).

[0068] In this embodiment, the channel multipath component geometry information, the channel multipath component field strength information, the coordinate information, the scene geometry characteristics of the wireless propagation environment, and the material electromagnetic parameter characteristics of the terminal to be predicted at multiple historical moments are input into different subnetworks of the multipath long-period prediction network, and multiple feature maps including the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted, the scene geometry characteristics, and the scene material electromagnetic parameter characteristics are obtained respectively. The multiple feature maps are input into the multipath long-period prediction network to obtain the channel multipath component prediction information at multiple future moments. The above embodiment builds a network for the time series prediction task of channel multipath. At the same time, in terms of prediction capability, in order to further improve the prediction performance, environmental features, including the geometry and material characteristics of the environment, are introduced to explore the impact of the environment on electromagnetic wave propagation, further improve the prediction accuracy, and ensure the ability of long-period prediction. This improves the problem of the existing technology in AI-based channel prediction that lacks MPCs for predicting future moments, including geometric information and field strength information of MPCs.

[0069] Optionally, in the method, step S10 includes:

[0070] Acquire the multipath long period prediction network;

[0071] Acquire the channel multipath component geometry information and the channel multipath component field strength information through a ray tracing simulation model;

[0072] The channel multipath component geometry information and the channel multipath component field strength information are input into the first feature extraction subnetwork of the multipath long period prediction network for feature extraction, the relationship between the geometry and field strength of the channel multipath component is mined, and the first feature map is obtained.

[0073] In this embodiment, information C such as the position, shape, size, and material of the scatterer in the wireless propagation environment is obtained. environment ; Antenna configuration C of the transmitter and receiver (ie the terminal to be predicted) antenna, including the number, polarization, gain and transmit power of antennas; the motion parameter C of the terminal to be predicted move , including the velocity vector Acceleration vector i=1,2,…,m, where m is the number of moving objects, and a time series data set of MPCs is constructed. Using ray tracing software such as WirelessInsite and BUPT-RT-SIM, based on C environment 、C antenna 、C move , perform ray tracing simulation, obtain the MPCs geometric information G of the terminal to be predicted in multiple historical moments, taking the historical moment k as an example k (i.e. the geometric information of the channel multipath components) and field strength information E k (i.e. the channel multipath component field strength information). k (i.e. the geometric information of the channel multipath components) and field strength information E k (i.e., the channel multipath component field strength information) is input into the sub-network 1 multipath feature extraction module of LMPE-net (i.e., the first feature extraction sub-network of the multipath long-period prediction network) for automatic feature extraction, during which a 128-head attention mechanism is used to mine G k and E k Finally, the fused MPCs change rule feature F1 (i.e., the first feature map) is input into sub-network 4 (i.e., the sequence prediction sub-network).

[0074] Optionally, in the method, the acquiring the multipath long-period prediction network includes:

[0075] Adopting deep learning architecture, we build a multi-path long-period prediction network to be trained;

[0076] Setting the hyperparameters of the multipath long-period prediction network to be trained and determining the loss function;

[0077] Acquire channel multipath components of a plurality of continuously moving positions and construct a time series data set of the channel multipath components;

[0078] The loss function and the time series data set are used to train the multipath long-period prediction network to be trained, so as to obtain the multipath long-period prediction network.

[0079] In this embodiment, the pytorch deep learning architecture is used to build the LMPE-net network (i.e., the multipath long-term prediction network to be trained), including sub-network 1 multipath feature extraction module, sub-network 2 scene geometry feature extraction module, sub-network 3 scene material feature extraction module and sub-network 4 multipath sequence prediction module. The hyper parameters of LMPE-net are set, and the learning rate is set to 10 -4 , the loss function uses the mean square error of the geometric information G and the field strength information E of the MPCs at the prediction time:

[0080]

[0081] Among them, G m and Represent the true value and predicted value of MPCs geometric information, E m and Represent the true value and predicted value of the MPCs field strength information respectively. The time series data set is composed of multiple (can be more than 10,000) MPCs data under continuously moving positions. A three-dimensional environment model is performed on the scatterers, environment boundaries, etc. in the wireless propagation environment, and the motion parameters of the terminal to be predicted in the scene are set, including the position, shape, size, material, etc. of the scatterers in the wireless propagation environment. environment ; Antenna configuration C of the transmitter and receiver (ie the terminal to be predicted) antenna , including the number, polarization, gain and transmit power of antennas; the motion parameter C of the terminal to be predicted move , including the velocity vector Acceleration vector i=1,2,…,m, where m is the number of moving objects. Using ray tracing software such as WirelessInsite and BUPT-RT-SIM, based on C environment 、C antenna 、C move Perform ray tracing simulation to obtain the geometric information Geo of MPC at different times i and field strength information Ele i , build the time series dataset Set of MPCs MPCs (i.e. the time series data set). The geometric information is the coordinates R of the MPC action point n , n represents the action order; the field strength information is the receiving field strength E of the MPC, that is: The time series data set is used to perform model training on the multipath long-period prediction network to be trained on a graphics card with a memory of more than 6G to obtain the multipath long-period prediction network.

[0082] Optionally, the method, before step S20, further includes:

[0083] Establishing a three-dimensional environmental model of the wireless propagation environment;

[0084] Acquiring, according to the environment model, scatterer information of the wireless propagation environment and motion information of the terminal to be predicted;

[0085] Acquire the scene geometric features and the material electromagnetic parameter features according to the scatterer information;

[0086] The coordinate information is acquired according to the motion information.

[0087] In this embodiment, based on C environment , construct the scene geometry feature O (i.e. the scene geometry feature) and the material electrical parameter feature I (i.e. the material electromagnetic parameter feature). O is defined as tensor[L,W,2], where L and W represent the length and width of the scene, in units of m. O has two channels, which are the boundary feature and height feature of the scene scatterer. I is also defined as tensor[L,W,2], where L and W represent the dielectric constant and conductivity of the scene scatterer material, respectively. Obtain the coordinate information P of the terminal to be predicted at the historical k moment (taking the historical k moment as an example) k , P is defined as tensor[k,3,1], which represents the three-dimensional coordinates of the terminal to be predicted at the historical time k (ie, the coordinate information).

[0088] Optionally, in the method, the scatterer information includes one or more of the following:

[0089] Scatterer position;

[0090] Scatter shape;

[0091] Scatter size;

[0092] Electromagnetic parameters of scatterer materials.

[0093] In this embodiment, information such as the position of the scatterer, the shape of the scatterer, the size of the scatterer, and the electromagnetic parameters of the scatterer material in the wireless propagation environment is obtained and proportional modeling is performed in modeling software such as Sketchup and Blender to obtain the scene geometric characteristics and the material electromagnetic parameter characteristics.

[0094] Optionally, in the method, step S40 includes:

[0095] Inputting the first feature map, the second feature map, and the third feature map into the sequence prediction subnetwork to obtain splicing features;

[0096] Use neural network models to obtain the evolution law of time series;

[0097] According to the splicing features and the time series evolution law, the channel multipath component prediction information at multiple future moments is obtained.

[0098] In this embodiment, the scene geometric feature O (ie, the scene geometric feature) is combined with the historical coordinate P k (i.e. the coordinate information) is input into the scene geometric feature extraction module of the sub-network 2 of LMPE-net (i.e. the second feature extraction sub-network) for automatic feature extraction. During the process, four convolutional layers and one linear layer are used to extract the geometric feature map to obtain the geometric feature map F2 (i.e. the second feature map). The material electromagnetic parameter feature I (i.e. the material electromagnetic parameter feature) is compared with the historical coordinate P k (i.e., the coordinate information) is input into the scene material feature extraction module of sub-network 3 of LMPE-net (i.e., the third feature extraction sub-network) for automatic feature extraction. During this process, four convolutional layers and one linear layer are used to extract the material electromagnetic feature map to obtain the material electromagnetic feature map F3 (i.e., the third feature map). Based on the MPCs change law feature F1 (i.e., the first feature map), the geometric feature map F2 (i.e., the second feature map) and the material electromagnetic feature map F3 (i.e., the third feature map), the neural network model such as Transformer is used to mine the time change law, obtain the time series evolution law, and predict the MPCs (i.e., the channel multipath component prediction information) at multiple moments in the future (taking moment t as an example). Specifically, it includes: 1) inputting F1, F2, and F3 into the multipath sequence prediction module of sub-network 4 of LMPE-net (i.e., the sequence prediction sub-network), performing feature splicing, and obtaining the splicing feature. 2) Based on the spliced features (i.e., the spliced features), the neural network model of the transformer with a two-layer codec architecture is used to mine the time series evolution law and predict the MPCs (i.e., the channel multipath component prediction information) at multiple moments in the future (taking moment t as an example), including G t (i.e., the channel multipath component prediction geometry information) and E t (i.e., the channel multipath component predicted field strength information), completing the entire process of channel multipath component time series prediction.

[0099] Optionally, in the method, the channel multipath component prediction information includes one or more of the following:

[0100] Channel multipath component prediction geometry information;

[0101] Channel multipath component prediction field strength information.

[0102] In this embodiment, the MPCs (i.e., the channel multipath component prediction information) at the future time t include G t(i.e., the channel multipath component prediction geometry information) and E t (i.e., the channel multipath component predicted field strength information).

[0103] It should be noted that the specific steps of evaluating the channel multipath intelligent prediction method assisted by environmental characteristics according to the embodiment of the present invention are as follows:

[0104] 1) Obtain a real-world street scene, such as the one from the open-source OpenStreetMap website. The scene dimensions are 340 meters long, 280 meters wide, and 40 meters high. Define the transmitter as a stationary base station and the receiver as a mobile object, both moving continuously within the scene at a speed of 36 km / h.

[0105] 2) The UE's coordinate features P were constructed based on the Rx positions at 9,000 moments. Ray tracing simulation was then used to calculate the MPC information at each of those 9,000 moments, which was then used to construct the MPCs dataset. The scene's geometric information O and material features I were then constructed based on the 3D model.

[0106] 3) The dataset was divided into training, test, and validation sets in a ratio of 7:1:2. The training set contained 6,300 samples, the test set contained 900 samples, and the validation set contained 1,800 samples. Next, the network was trained on an appropriate platform with a learning rate of 0.0008 and 200 epochs.

[0107] 4) The network based only on traditional transformer is used as the baseline, that is, only sub-network 4. At the same time, the performance of only the multipath feature extraction module plus the multipath sequence prediction module (i.e., sub-networks 1 and 4) and the multipath feature extraction module plus the multipath sequence prediction module plus the scene geometry feature extraction module (i.e., sub-networks 1, 2, and 4) are also trained to show the impact of each module on the prediction performance. The normalized root mean square error (NMSE) is used as an indicator, such as Figure 5 and Figure 6 shown.

[0108] from Figure 5 and Figure 6 It can be seen that the channel multipath intelligent prediction method assisted by environmental features in the embodiment of the present invention reduces the geometric and field strength errors by an average of 31.4% and 40.5% at 10 moments compared with the baseline, and proves that the introduction of scene information can greatly improve the prediction accuracy.

[0109] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0110] 1. Taking into account the need for accurate prediction of propagation multipath information in DTC, this paper fully utilizes the continuity of electromagnetic wave propagation in space and the changing patterns of the action points and field strength of MPCs at different times. Based on an AI model, an intelligent channel multipath prediction method assisted by environmental features is proposed. The proposed Long-Term Multipath Prediction Network assisted by Environment (LMPE-net) overcomes the shortcoming of traditional algorithms that cannot predict MPCs at future times.

[0111] 2. We fully utilize environmental information, such as the position and size of scatterers, and the influence of the electromagnetic parameters of materials on electromagnetic wave propagation, to design scene geometry feature extraction modules and scene material feature extraction modules, which greatly improve the accuracy of MPCs predictions and ensure prediction performance over long periods (such as historical moment: future moment = 1:1).

[0112] 3. Compared with existing solutions, this paper reveals how the movement of moving objects in a dynamic environment changes MPCs, which in turn affects the characteristics of the wireless channel. A fast, accurate, and physically interpretable channel prediction solution is established.

[0113] like Figure 7 To achieve the above-mentioned purpose, an embodiment of the present invention provides an environment feature-assisted channel multipath intelligent prediction device, which includes:

[0114] A first acquisition module 701 is configured to input the channel multipath component geometry information and channel multipath component field strength information of a terminal to be predicted at multiple historical moments into a first feature extraction subnetwork of a multipath long-period prediction network to obtain a first feature map; wherein the first feature map includes the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted;

[0115] A second acquisition module 702 is configured to input the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long-period prediction network to obtain a second feature map; wherein the second feature map includes the relationship between the geometry of the channel multipath component and the scene geometric features;

[0116] A third acquisition module 703 is configured to input the coordinate information and the electromagnetic parameter characteristics of the material of the wireless propagation environment into a third feature extraction subnetwork of the multipath long period prediction network to obtain a third feature map; wherein the third feature map includes the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material;

[0117] The fourth acquisition module 704 is used to input the first feature map, the second feature map and the third feature map into the sequence prediction subnetwork of the multipath long-period prediction network to obtain the channel multipath component prediction information of the terminal to be predicted at multiple future moments; wherein the channel multipath component prediction information includes channel multipath component prediction geometry information and channel multipath component prediction field strength information.

[0118] Optionally, in the device, the first obtaining module 701 includes:

[0119] A first acquiring unit, configured to acquire the multipath long period prediction network;

[0120] A second acquiring unit is configured to acquire the channel multipath component geometric information and the channel multipath component field strength information through a ray tracing simulation model;

[0121] The third acquisition unit is used to input the channel multipath component geometry information and the channel multipath component field strength information into the first feature extraction subnetwork of the multipath long period prediction network for feature extraction, explore the relationship between the geometry and field strength of the channel multipath component, and obtain the first feature map.

[0122] Optionally, in the device, the first acquiring unit includes:

[0123] The first processing component is used to build a multipath long-period prediction network to be trained using a deep learning architecture;

[0124] A first determining component is used to set the hyperparameters of the multipath long-period prediction network to be trained and determine the loss function;

[0125] a second processing component, configured to obtain channel multipath components of a plurality of continuously moving positions and construct a time series data set of the channel multipath components;

[0126] The first acquisition component is used to train the multipath long-period prediction network to be trained using the loss function and the time series data set to acquire the multipath long-period prediction network.

[0127] Optionally, the device further comprises:

[0128] A first processing module is configured to establish a three-dimensional environmental model of the wireless propagation environment;

[0129] a fifth acquisition module, configured to acquire scatterer information of the wireless propagation environment and motion information of the terminal to be predicted based on the environment model;

[0130] a sixth acquisition module, configured to acquire the scene geometric features and the material electromagnetic parameter features according to the scatterer information;

[0131] A seventh acquisition module is configured to acquire the coordinate information according to the motion information.

[0132] Optionally, in the device, the scatterer information includes one or more of the following:

[0133] Scatterer position;

[0134] Scatter shape;

[0135] Scatter size;

[0136] Electromagnetic parameters of scatterer materials.

[0137] Optionally, in the device, the fourth obtaining module 704 includes:

[0138] a fourth acquisition unit, configured to input the first feature map, the second feature map, and the third feature map into the sequence prediction subnetwork to acquire a splicing feature;

[0139] The fifth acquisition unit is used to obtain the time series evolution law using a neural network model;

[0140] A sixth acquisition unit is configured to acquire the channel multipath component prediction information at multiple future moments according to the splicing features and the time series evolution law.

[0141] Optionally, in the apparatus, the channel multipath component prediction information includes one or more of the following:

[0142] Channel multipath component prediction geometry information;

[0143] Channel multipath component prediction field strength information.

[0144] It should be noted here that the above-mentioned device provided by the embodiment of the present invention can implement all the method steps implemented by the above-mentioned method embodiment and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0145] To achieve the above-mentioned objectives, an embodiment of the present invention provides an electronic device, comprising: a transceiver, a processor, a memory, and a program or instruction stored on the memory and executable on the processor; wherein, when the processor executes the program or instruction, the channel multipath intelligent prediction method assisted by environmental features as described above is implemented.

[0146] To achieve the above-mentioned purpose, an embodiment of the present invention provides a readable storage medium having a program or instruction stored thereon, wherein when the program or instruction is executed by a processor, the steps in the channel multipath intelligent prediction method assisted by environmental features as described above are implemented.

[0147] To achieve the above objectives, an embodiment of the present invention provides a computer program product, which includes computer instructions. When the computer instructions are executed by a processor, the steps of the above-mentioned environmental feature-assisted channel multipath intelligent prediction method are implemented.

[0148] It should be further noted that the terminals described in this specification include but are not limited to smartphones, tablet computers, etc., and many functional components described are referred to as modules in order to more particularly emphasize the independence of their implementation methods.

[0149] In embodiments of the present invention, modules can be implemented in software so that they can be executed by various types of processors. For example, an identified executable code module can include one or more physical or logical blocks of computer instructions, for example, which can be constructed as objects, procedures, or functions. Nevertheless, the executable code of the identified module does not need to be physically located together, but can include different instructions stored in different locations, which, when logically combined together, constitute the module and achieve the specified purpose of the module.

[0150] In fact, executable code module can be a single instruction or many instructions, and can even be distributed on a plurality of different code segments, distributed in the middle of different programs, and distributed across a plurality of memory devices.Similarly, operating data can be identified in the module, and can be implemented and organized in the data structure of any appropriate type according to any appropriate form.Described operating data can be collected as a single data set, or can be distributed in different locations (including on different storage devices), and can only be present on a system or network as an electronic signal at least in part.

[0151] When a module can be implemented using software, given the current state of hardware technology, those skilled in the art can build corresponding hardware circuits to implement the corresponding functions of the module, regardless of cost. The hardware circuits may include conventional very large scale integration (VLSI) circuits or gate arrays, as well as existing semiconductors such as logic chips and transistors, or other discrete components. Modules may also be implemented using programmable hardware devices, such as field programmable gate arrays, programmable array logic, or programmable logic devices.

[0152] The above exemplary embodiments are described with reference to the accompanying drawings. Many different forms and embodiments are possible without departing from the spirit and teachings of the present invention. Therefore, the present invention should not be construed as limited to the exemplary embodiments set forth herein. Rather, these exemplary embodiments are provided so that this disclosure will be complete and perfect and will convey the scope of the invention to those skilled in the art. In the drawings, component sizes and relative sizes may be exaggerated for clarity. The terminology used herein is for purposes of describing specific exemplary embodiments only and is not intended to be limiting. As used herein, the singular forms "a," "an," and "the" are intended to include plural forms, unless the context clearly indicates otherwise. It will be further understood that the terms "comprising" and / or "including," when used in this specification, indicate the presence of stated features, integers, steps, operations, components, and / or elements, but do not preclude the presence or addition of one or more other features, integers, steps, operations, components, elements, and / or groups thereof. Unless otherwise indicated, when stated, a range of values includes the upper and lower limits of that range and any subranges therebetween.

[0153] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A channel multipath intelligent prediction method assisted by environmental characteristics, characterized in that: include: Inputting the channel multipath component geometry information and channel multipath component field strength information of the terminal to be predicted at multiple historical moments into the first feature extraction subnetwork of the multipath long-period prediction network to obtain a first feature map; wherein the first feature map includes the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted; Inputting the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long-period prediction network to obtain a second feature map; wherein the second feature map includes the relationship between the geometry of the channel multipath component and the scene geometric features; Inputting the coordinate information and the electromagnetic parameter characteristics of the material of the wireless propagation environment into the third feature extraction subnetwork of the multipath long period prediction network to obtain a third feature map; wherein the third feature map includes the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material; The first feature map, the second feature map, and the third feature map are input into the sequence prediction subnetwork of the multipath long-period prediction network to obtain channel multipath component prediction information of the terminal to be predicted at multiple future moments; wherein the channel multipath component prediction information includes channel multipath component prediction geometry information and channel multipath component prediction field strength information.

2. The method according to claim 1, characterized in that Inputting the channel multipath component geometry information and channel multipath component field strength information of the terminal to be predicted at multiple historical moments into the first feature extraction subnetwork of the multipath long-period prediction network to obtain a first feature map, including: Acquire the multipath long period prediction network; Acquire the channel multipath component geometry information and the channel multipath component field strength information through a ray tracing simulation model; The channel multipath component geometry information and the channel multipath component field strength information are input into the first feature extraction subnetwork of the multipath long period prediction network for feature extraction, the relationship between the geometry and field strength of the channel multipath component is mined, and the first feature map is obtained.

3. The method according to claim 2, characterized in that The acquiring of the multipath long period prediction network comprises: Adopting deep learning architecture, we build a multi-path long-period prediction network to be trained; Setting the hyperparameters of the multipath long-period prediction network to be trained and determining the loss function; Acquire channel multipath components of a plurality of continuously moving positions and construct a time series data set of the channel multipath components; The loss function and the time series data set are used to train the multipath long-period prediction network to be trained, so as to obtain the multipath long-period prediction network.

4. The method according to claim 1, wherein Before inputting the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long period prediction network to obtain the second feature map, the method further includes: Establishing a three-dimensional environmental model of the wireless propagation environment; Acquiring, according to the environment model, scatterer information of the wireless propagation environment and motion information of the terminal to be predicted; Acquire the scene geometric features and the material electromagnetic parameter features according to the scatterer information; The coordinate information is acquired according to the motion information.

5. The method according to claim 4, characterized in that The scatterer information includes one or more of the following: Scatterer position; Scatter shape; Scatter size; Electromagnetic parameters of scatterer materials.

6. The method according to claim 1, characterized in that Inputting the first feature map, the second feature map, and the third feature map into the sequence prediction subnetwork of the multipath long period prediction network to obtain channel multipath component prediction information of the terminal to be predicted at multiple future moments, including: Inputting the first feature map, the second feature map, and the third feature map into the sequence prediction subnetwork to obtain splicing features; Use neural network models to obtain the evolution law of time series; According to the splicing features and the time series evolution law, the channel multipath component prediction information at multiple future moments is obtained.

7. The method according to claim 6, characterized in that The channel multipath component prediction information includes one or more of the following: Channel multipath component prediction geometry information; Channel multipath component prediction field strength information.

8. An environmental feature-assisted channel multipath intelligent prediction device, characterized in that: include: A first acquisition module is configured to input the channel multipath component geometry information and channel multipath component field strength information of a terminal to be predicted at multiple historical moments into a first feature extraction subnetwork of a multipath long-period prediction network to obtain a first feature map; wherein the first feature map includes the relationship between the geometry and field strength of the channel multipath component of the terminal to be predicted; A second acquisition module is configured to input the coordinate information corresponding to the terminal to be predicted at the multiple historical moments and the scene geometric features of the wireless propagation environment in which the terminal to be predicted is located into the second feature extraction subnetwork of the multipath long period prediction network to obtain a second feature map; wherein the second feature map includes the relationship between the geometry of the channel multipath component and the scene geometric features; A third acquisition module is configured to input the coordinate information and the electromagnetic parameter characteristics of the material of the wireless propagation environment into a third feature extraction subnetwork of the multipath long period prediction network to obtain a third feature map; wherein the third feature map includes the relationship between the field strength of the channel multipath component and the electromagnetic parameter characteristics of the scene material; The fourth acquisition module is used to input the first feature map, the second feature map and the third feature map into the sequence prediction subnetwork of the multipath long-period prediction network to obtain the channel multipath component prediction information of the terminal to be predicted at multiple future moments; wherein the channel multipath component prediction information includes channel multipath component prediction geometry information and channel multipath component prediction field strength information.

9. An electronic device comprising: A transceiver, a processor, a memory, and a program or instruction stored in the memory and executable on the processor; characterized in that when the processor executes the program or instruction, the channel multipath intelligent prediction method assisted by environmental features as described in any one of claims 1 to 7 is implemented.

10. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or instruction is executed by a processor, the steps of the channel multipath intelligent prediction method assisted by environmental features as described in any one of claims 1 to 7 are implemented.

11. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the channel multipath intelligent prediction method assisted by environmental features as described in any one of claims 1 to 7.

Images