Model training method and device, model application method and device, communication equipment, communication system and storage medium

Abstract

The invention provides a model training and application method and device, equipment and a storage medium, and the method comprises the steps: determining a training sample set which comprises at least one group of sample data, and the sample data comprises first information and / or second information; wherein the first information is channel state information determined by a sensing signal receiving end based on a received sensing signal, and the first information is used for determining a sensing result of a sensing target; the second information is actual state information of the sensing target; training a first model based on the training sample set; wherein the first model can determine at least one type of sensing result of the sensing target based on the first information. According to the method and the device, the technical problems of limited application scene, relatively high complexity, relatively poor sensing accuracy under the condition of multiple sensing targets and the like when the sensing result is calculated by a sensing algorithm are solved, the sensing accuracy is ensured, the sensing complexity is reduced, and the sensing application scene is widened.

Description

Model training, application method and device, communication equipment, communication system, storage medium Technical Field

[0001] The present disclosure relates to the field of communication technology, and in particular to model training, application methods and devices, communication equipment, communication systems, and storage media. Background Art

[0002] With the continuous development of artificial intelligence (AI) technology, its application areas are becoming more and more extensive. Among them, the 3rd Generation Partnership Project (3GPP) has also introduced AI technology.

[0003] Summary of the Invention

[0004] The present disclosure proposes model training, application methods and devices, communication equipment, communication systems, and storage media.

[0005] According to a first aspect of an embodiment of the present disclosure, a model training method is proposed, which is performed by a first device. The method includes:

[0006] Determine a training sample set, where the training sample set includes at least one set of sample data, and the sample data includes first information and / or second information; wherein the first information is channel state information determined by a perception signal receiving end based on a received perception signal, and the first information is used to determine a perception result of a perception target; and the second information is actual state information of the perception target;

[0007] A first model is trained based on the training sample set; wherein the first model can determine at least one type of perception result of the perception target based on the first information.

[0008] According to a second aspect of an embodiment of the present disclosure, a model application method is provided, which is executed by a second device. The method includes:

[0009] determining first information, where the first information is channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine a perception result of the perception target;

[0010] determining a first model, where the first model can determine at least one type of perception result of the perception target based on the first information;

[0011] A perception result of the perception target is determined based on the first information and the first model.

[0012] According to a third aspect of an embodiment of the present disclosure, a first device is provided, including:

[0013] a processing module, configured to determine a training sample set, the training sample set including at least one set of sample data, the sample data including first information and / or second information; wherein the first information is channel state information determined by the perception signal receiving end based on the received perception signal, the first information being used to determine a perception result of a perception target; and the second information is actual state information of the perception target;

[0014] The processing module is further used to train a first model based on the training sample set; wherein the first model can determine at least one type of perception result of the perception target based on the first information.

[0015] According to a fourth aspect of the embodiments of the present disclosure, a second device is provided, characterized by comprising:

[0016] a processing module, configured to determine first information, where the first information is channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine a perception result of the perception target;

[0017] The processing module is further configured to determine a first model, wherein the first model can determine at least one type of perception result of the perception target based on the first information;

[0018] The processing module is further used to determine a perception result of the perception target based on the first information and the first model.

[0019] According to a fifth aspect of an embodiment of the present disclosure, a communication device is provided, including:

[0020] one or more processors;

[0021] The processor is used to call instructions to enable the communication device to execute any one of the methods described in the first aspect to the second aspect.

[0022] According to the sixth aspect of an embodiment of the present disclosure, a communication system is proposed, characterized in that it includes a first device and a second device, wherein the first device is configured to implement the model training method described in the first aspect, and the second device is configured to implement the model application method described in the second aspect.

[0023] According to the seventh aspect of the embodiment of the present disclosure, a storage medium is proposed, which stores instructions, and is characterized in that when the instructions are executed on a communication device, the communication device executes any one of the methods described in the first to second aspects. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above and / or additional aspects and advantages of the present disclosure will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0025] FIG1A is a schematic diagram of the architecture of some communication systems provided by embodiments of the present disclosure;

[0026] FIG1B is a schematic diagram illustrating the architecture of a sensing LoS path, a sensing NLoS path, a clutter LOS path, and a clutter NLOS path in an ISAC simulation scenario according to an embodiment of the present disclosure;

[0027] FIG1C is a schematic diagram of an architecture for determining a perception result based on a perception algorithm according to an embodiment of the present disclosure;

[0028] FIG2A is a flow chart of a model training method provided in yet another embodiment of the present disclosure;

[0029] FIG2B is a schematic diagram showing a first model for outputting all perception results according to an embodiment of the present disclosure;

[0030] FIG2C is a schematic diagram showing different first models for outputting a type of perception result according to an embodiment of the present disclosure;

[0031] FIG2D is a schematic diagram of a flow chart of a model application method provided in yet another embodiment of the present disclosure;

[0032] FIG3 is a flow chart of a model training method provided in yet another embodiment of the present disclosure;

[0033] FIG4 is a flow chart of a model application method provided in yet another embodiment of the present disclosure;

[0034] FIG5 is a schematic structural diagram of a first device provided by an embodiment of the present disclosure;

[0035] FIG6 is a schematic structural diagram of a second device provided by an embodiment of the present disclosure;

[0036] FIG7A is a schematic structural diagram of a communication device provided by an embodiment of the present disclosure;

[0037] FIG7B is a schematic structural diagram of a chip provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0038] The embodiments of the present disclosure provide model training, application methods and devices, communication equipment, communication systems, and storage media.

[0039] In a first aspect, an embodiment of the present disclosure provides a model training method, which is performed by a first device. The method includes:

[0040] Determine a training sample set, where the training sample set includes at least one set of sample data, and the sample data includes first information and / or second information; wherein the first information is channel state information determined by a perception signal receiving end based on a received perception signal, and the first information is used to determine a perception result of a perception target; and the second information is actual state information of the perception target;

[0041] A first model is trained based on the training sample set; wherein the first model can determine at least one type of perception result of the perception target based on the first information.

[0042] In the above embodiment, a model training method is provided to achieve successful training of the first model. The first model can determine at least one type of perception result of the perception target based on the input first information. It can be seen that the first model trained by the training method of the present disclosure is used to perceive the perception target, that is, the present disclosure provides a training method for a "perception model" so that the model can be successfully trained. Then, perception can be achieved based on the model in the future without having to calculate the perception result using the perception algorithm as in the related art. This effectively solves the technical problems such as the limited application scenarios when the perception algorithm calculates the perception result, the high complexity, and the poor perception accuracy in the case of multiple perception targets, thereby ensuring the perception accuracy, reducing the perception complexity, and broadening the perception application scenarios.

[0043] In conjunction with some embodiments of the first aspect, in some embodiments, the perception result of the perception target includes at least one of the following categories:

[0044] a distance of the perceived target determined based on the first information;

[0045] a speed of the perceived target determined based on the first information;

[0046] The angle of the perceived target is determined based on the first information.

[0047] In combination with some embodiments of the first aspect, in some embodiments, the first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of perception results.

[0048] In the above embodiment, the specific categories of perception results are defined, and the correlation between the first information and different types of perception results is defined. This allows the trained first model to successfully output at least one type of perception result based on the first information, thereby realizing model-based perception, ensuring perception accuracy, reducing perception complexity, and broadening perception application scenarios.

[0049] In conjunction with some embodiments of the first aspect, in some embodiments, the second information includes at least one of the following:

[0050] the actual distance of the perceived target;

[0051] the actual speed of the perceived target;

[0052] The actual angle of the perceived target.

[0053] In the above embodiment, the specific content of the second information is limited so that the first model can be successfully trained based on the second information, ensuring that the trained first model can accurately perceive the perception target and ensure the perception accuracy.

[0054] In conjunction with some embodiments of the first aspect, in some embodiments, training the first model based on the training sample set includes:

[0055] Determining at least one type of perception result that needs to be determined by the first model;

[0056] Determining third information based on the first information, the third information being: sub-information in the first information used to determine at least one dimension of the at least one type of perception result;

[0057] Inputting the third information into the first model to obtain an output result of the first model;

[0058] Calculating a loss function based on an output result of the first model and the second information;

[0059] The parameters of the first model are adjusted based on the loss function, and the third information is input into the adjusted first model to obtain the output result of the first model, and then the loss function of the first model is calculated until the first condition is met, and the training is determined to be completed.

[0060] With reference to some embodiments of the first aspect, in some embodiments, determining third information based on the first information includes:

[0061] Setting sub-information in the first information that is not used to determine at least one dimension of the at least one type of perception result to a fixed value, and determining the first information after being set to the fixed value as the third information; and / or

[0062] Sub-information for determining at least one dimension of the at least one type of perception result is extracted from the first information as the third information.

[0063] In conjunction with some embodiments of the first aspect, in some embodiments, the at least one dimension includes one or more of the following dimensions:

[0064] Orthogonal frequency division multiplexing OFDM symbol dimension;

[0065] Subcarrier dimension;

[0066] Receive antenna port dimensions.

[0067] In conjunction with some embodiments of the first aspect, in some embodiments, the first condition includes at least one of the following:

[0068] The training period of the first model reaches a first threshold;

[0069] The first model converged.

[0070] In the above embodiment, a specific method for training the first model is provided so that the training of the first model can be successfully implemented. Perception can then be implemented based on the model without having to calculate the perception results using the perception algorithm as in the related art. This effectively solves technical problems such as limited application scenarios when the perception algorithm calculates the perception results, high complexity, and poor perception accuracy in the case of multiple perception targets, thereby ensuring perception accuracy, reducing perception complexity, and broadening perception application scenarios.

[0071] In conjunction with some embodiments of the first aspect, in some embodiments, the first model may determine a type of perception result of the perception target based on the first information, and different first models are used to determine different types of perception results; or

[0072] The first model can determine all perception results of the perception target based on the first information.

[0073] In combination with some embodiments of the first aspect, in some embodiments, when the first model can determine all perception results of the perception target based on the first information, the first information is the same as the third information.

[0074] In the above embodiments, it is limited that different first models can respectively determine a type of perception result, or a first model can be used to determine multiple types of perception results. This provides multiple implementable methods for the method of "determining the perception result based on the first model", thereby improving the flexibility of the perception method.

[0075] In a second aspect, an embodiment of the present disclosure provides a model application method, which is executed by a second device. The method includes:

[0076] determining first information, where the first information is channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine a perception result of the perception target;

[0077] determining a first model, where the first model can determine at least one type of perception result of the perception target based on the first information;

[0078] A perception result of the perception target is determined based on the first information and the first model.

[0079] In the above embodiment, a method for applying a first model is provided, which can be used to determine the perception results of a perception target. That is, perception can be achieved based on this first model, eliminating the need to calculate the perception results using a perception algorithm as in related technologies. This effectively addresses technical issues such as the limited application scenarios, high complexity, and poor perception accuracy in the case of multiple perception targets when using perception algorithms to calculate perception results. This ensures perception accuracy, reduces perception complexity, and broadens the application scenarios of perception.

[0080] In conjunction with some embodiments of the second aspect, in some embodiments, the perception result of the perception target includes at least one of the following categories:

[0081] a distance of the perceived target determined based on the first information;

[0082] a speed of the perceived target determined based on the first information;

[0083] The angle of the perceived target is determined based on the first information.

[0084] In combination with some embodiments of the second aspect, in some embodiments, the first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of perception results.

[0085] In the above embodiment, the specific categories of perception results are defined, and the correlation between the first information and different types of perception results is defined. This enables the first model to successfully output at least one type of perception result based on the first information, thereby realizing model-based perception, ensuring perception accuracy, reducing perception complexity, and broadening perception application scenarios.

[0086] In conjunction with some embodiments of the second aspect, in some embodiments, determining the perception result of the perception target based on the first information and the first model includes:

[0087] Determining at least one type of perception result that needs to be determined by the first model;

[0088] Determining third information based on the first information, the third information being: sub-information in the first information used to determine at least one dimension of the at least one type of perception result;

[0089] The third information is input into the first model to obtain the at least one type of perception result output by the first model.

[0090] With reference to some embodiments of the second aspect, in some embodiments, determining third information based on the first information includes:

[0091] Setting sub-information in the first information that is not used to determine at least one dimension of the at least one type of perception result to a fixed value, and determining the first information after being set to the fixed value as the third information; and / or

[0092] Sub-information for determining at least one dimension of the at least one type of perception result is extracted from the first information as the third information.

[0093] In conjunction with some embodiments of the second aspect, in some embodiments, the at least one dimension includes one or more of the following dimensions:

[0094] Orthogonal frequency division multiplexing OFDM symbol dimension;

[0095] Subcarrier dimension;

[0096] Receive antenna port dimensions.

[0097] In the above embodiment, a specific method for outputting at least one type of perception result based on the first model is provided, so that the perception of the perception target can be accurately achieved based on the first model, thereby ensuring the perception accuracy.

[0098] In conjunction with some embodiments of the second aspect, in some embodiments, the first model may determine a type of perception result of the perception target based on the first information, and different first models are used to determine different types of perception results; or

[0099] The first model can determine all perception results of the perception target based on the first information.

[0100] In combination with some embodiments of the second aspect, in some embodiments, when the first model can determine all perception results of the perception target based on the first information, the first information is the same as the third information.

[0101] In the above embodiments, it is limited that different first models can respectively determine a type of perception result, or a first model can be used to determine multiple types of perception results. This provides multiple implementable methods for the method of "determining the perception result based on the first model", thereby improving the flexibility of the perception method.

[0102] In a third aspect, an embodiment of the present disclosure provides a first device, including:

[0103] a processing module, configured to determine a training sample set, the training sample set including at least one set of sample data, the sample data including first information and / or second information; wherein the first information is channel state information determined by the perception signal receiving end based on the received perception signal, the first information being used to determine a perception result of a perception target; and the second information is actual state information of the perception target;

[0104] The processing module is further used to train a first model based on the training sample set; wherein the first model can determine at least one type of perception result of the perception target based on the first information.

[0105] In conjunction with some embodiments of the third aspect, in some embodiments, the perception result of the perception target includes at least one of the following categories:

[0106] a distance of the perceived target determined based on the first information;

[0107] a speed of the perceived target determined based on the first information;

[0108] The angle of the perceived target is determined based on the first information.

[0109] In combination with some embodiments of the third aspect, in some embodiments, the first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of perception results.

[0110] In conjunction with some embodiments of the third aspect, in some embodiments, the second information includes at least one of the following:

[0111] the actual distance of the perceived target;

[0112] the actual speed of the perceived target;

[0113] The actual angle of the perceived target.

[0114] In conjunction with some embodiments of the third aspect, in some embodiments, training the first model based on the training sample set includes:

[0115] Determining at least one type of perception result that needs to be determined by the first model;

[0116] Determining third information based on the first information, the third information being: sub-information in the first information used to determine at least one dimension of the at least one type of perception result;

[0117] Inputting the third information into the first model to obtain an output result of the first model;

[0118] Calculating a loss function based on an output result of the first model and the second information;

[0119] The parameters of the first model are adjusted based on the loss function, and the third information is input into the adjusted first model to obtain the output result of the first model, and then the loss function of the first model is calculated until the first condition is met, and the training is determined to be completed.

[0120] In conjunction with some embodiments of the third aspect, in some embodiments, determining third information based on the first information includes:

[0121] Setting sub-information in the first information that is not used to determine at least one dimension of the at least one type of perception result to a fixed value, and determining the first information after being set to the fixed value as the third information; and / or

[0122] Sub-information for determining at least one dimension of the at least one type of perception result is extracted from the first information as the third information.

[0123] In conjunction with some embodiments of the third aspect, in some embodiments, the at least one dimension includes one or more of the following dimensions:

[0124] Orthogonal frequency division multiplexing OFDM symbol dimension;

[0125] Subcarrier dimension;

[0126] Receive antenna port dimensions.

[0127] In conjunction with some embodiments of the third aspect, in some embodiments, the first condition includes at least one of the following:

[0128] The training period of the first model reaches a first threshold;

[0129] The first model converged.

[0130] In conjunction with some embodiments of the third aspect, in some embodiments, the first model may determine a type of perception result of the perception target based on the first information, and different first models are used to determine different types of perception results; or

[0131] The first model can determine all perception results of the perception target based on the first information.

[0132] In combination with some embodiments of the third aspect, in some embodiments, when the first model can determine all perception results of the perception target based on the first information, the first information is the same as the third information.

[0133] In a fourth aspect, an embodiment of the present disclosure provides a second device, including:

[0134] a processing module, configured to determine first information, where the first information is channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine a perception result of the perception target;

[0135] The processing module is further configured to determine a first model, wherein the first model can determine at least one type of perception result of the perception target based on the first information;

[0136] The processing module is further used to determine a perception result of the perception target based on the first information and the first model.

[0137] In conjunction with some embodiments of the fourth aspect, in some embodiments, the perception result of the perception target includes at least one of the following categories:

[0138] a distance of the perceived target determined based on the first information;

[0139] a speed of the perceived target determined based on the first information;

[0140] The angle of the perceived target is determined based on the first information.

[0141] In combination with some embodiments of the fourth aspect, in some embodiments, the first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of perception results.

[0142] In conjunction with some embodiments of the fourth aspect, in some embodiments, determining the perception result of the perception target based on the first information and the first model includes:

[0143] Determining at least one type of perception result that needs to be determined by the first model;

[0144] Determining third information based on the first information, the third information being: sub-information in the first information used to determine at least one dimension of the at least one type of perception result;

[0145] The third information is input into the first model to obtain the at least one type of perception result output by the first model.

[0146] In conjunction with some embodiments of the fourth aspect, in some embodiments, determining the third information based on the first information includes:

[0147] Setting sub-information in the first information that is not used to determine at least one dimension of the at least one type of perception result to a fixed value, and determining the first information after being set to the fixed value as the third information; and / or

[0148] Sub-information for determining at least one dimension of the at least one type of perception result is extracted from the first information as the third information.

[0149] In conjunction with some embodiments of the fourth aspect, in some embodiments, the at least one dimension includes one or more of the following dimensions:

[0150] Orthogonal frequency division multiplexing OFDM symbol dimension;

[0151] Subcarrier dimension;

[0152] Receive antenna port dimensions.

[0153] In conjunction with some embodiments of the fourth aspect, in some embodiments, the first model may determine a type of perception result of the perception target based on the first information, and different first models are used to determine different types of perception results; or

[0154] The first model can determine all perception results of the perception target based on the first information.

[0155] In combination with some embodiments of the fourth aspect, in some embodiments, when the first model can determine all perception results of the perception target based on the first information, the first information is the same as the third information.

[0156] In a fifth aspect, an embodiment of the present disclosure proposes a communication device, which includes: one or more processors; one or more memories for storing instructions; wherein the processor is used to call the instructions so that the communication device executes the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0157] In the sixth aspect, an embodiment of the present disclosure proposes a communication system, which includes: a first device and a second device; wherein the first device is configured to execute the method described in the first aspect and the optional implementation of the first aspect, and the second device is configured to execute the method described in the second aspect and the optional implementation of the second aspect.

[0158] In the seventh aspect, an embodiment of the present disclosure proposes a storage medium, which stores instructions. When the instructions are executed on a communication device, the communication device executes the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0159] In an eighth aspect, an embodiment of the present disclosure proposes a program product. When the program product is executed by a communication device, the communication device executes the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0160] In a ninth aspect, an embodiment of the present disclosure proposes a computer program, which, when executed on a computer, enables the computer to execute the method described in the first aspect, the optional implementation of the first aspect, the second aspect, and the optional implementation of the second aspect.

[0161] It is understandable that the above-mentioned terminals, network devices, communication devices, communication systems, storage media, program products, and computer programs are all used to execute the methods proposed in the embodiments of the present disclosure. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects of the corresponding methods and will not be repeated here.

[0162] The present disclosure provides invention titles. In some embodiments, the terms model training, application method, information processing method, information sending method, and information receiving method are interchangeable; the terms communication device, information processing device, information sending device, and information receiving device are interchangeable; and the terms information processing system, communication system, information sending system, and information receiving system are interchangeable.

[0163] The embodiments of the present disclosure are not exhaustive and are merely illustrative of some embodiments, and are not intended to be a specific limitation on the scope of protection of the present disclosure. In the absence of contradiction, each step in a certain embodiment can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a certain embodiment can also be implemented as an independent embodiment, and the order of the steps in a certain embodiment can be arbitrarily exchanged. In addition, the optional implementation methods in a certain embodiment can be arbitrarily combined; in addition, the embodiments can be arbitrarily combined. For example, some or all steps of different embodiments can be arbitrarily combined, and a certain embodiment can be arbitrarily combined with the optional implementation methods of other embodiments.

[0164] In each embodiment of the present disclosure, unless otherwise specified or provided for by logic, the terms and / or descriptions between the embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form a new embodiment based on their inherent logical relationships.

[0165] The terms used in the embodiments of the present disclosure are only for the purpose of describing specific embodiments and are not intended to limit the present disclosure.

[0166] In the embodiments of the present disclosure, unless otherwise specified, elements expressed in the singular, such as "a", "an", "the", "above", "said", "the", "the", etc., may mean "one and only one", or "one or more", "at least one", etc. For example, when using articles such as "a", "an", "the" in English in translation, the noun following the article may be understood as a singular expression or a plural expression.

[0167] In the embodiments of the present disclosure, “plurality” refers to two or more.

[0168] In some embodiments, the terms "at least one of", "at least one of", "at least one of", "one or more", "a plurality of", "multiple", etc. can be used interchangeably.

[0169] In the embodiments of the present disclosure, descriptions such as “at least one of A, B, C…”, “A and / or B and / or C…”, etc. include the situation where any one of A, B, C… exists alone, and also include any combination of any multiple of A, B, C…, and each situation can exist alone; for example, “at least one of A, B, C” includes the situation where A exists alone, B exists alone, C exists alone, the combination of A and B, the combination of A and C, the combination of B and C, and the combination of A, B, and C; for example, A and / or B includes the situation where A exists alone, B exists alone, and the combination of A and B.

[0170] In some embodiments, descriptions such as "in one case A, in another case B," or "in response to one case A, in response to another case B," may include the following technical solutions depending on the situation: executing A independently of B (in some embodiments, A); executing B independently of A (in some embodiments, B); selectively executing A and B (in some embodiments, selecting between A and B); and executing both A and B (in some embodiments, A and B). The same applies when there are more branches, such as A, B, and C.

[0171] The prefixes such as "first" and "second" in the embodiments of the present disclosure are only used to distinguish different description objects and do not constitute any restriction on the position, order, priority, quantity or content of the description objects. For the statement of the description object, please refer to the description in the context of the claims or embodiments, and the use of prefixes should not constitute unnecessary restrictions. For example, if the description object is a "field", the ordinal number before the "field" in "first field" and "second field" does not limit the position or order between the "fields". "First" and "second" do not limit whether the "fields" they modify are in the same message, nor do they limit the order of the "first field" and "second field". For another example, if the description object is a "level", the ordinal number before the "level" in "first level" and "second level" does not limit the priority between the "levels". For another example, the number of description objects is not limited by the ordinal number and can be one or more. Taking "first device" as an example, the number of "devices" can be one or more. In addition, the objects modified by different prefixes can be the same or different. For example, if the description object is "device", then the "first device" and the "second device" can be the same device or different devices, and their types can be the same or different; for another example, if the description object is "information", then the "first information" and the "second information" can be the same information or different information, and their contents can be the same or different.

[0172] In some embodiments, “including A,” “comprising A,” “used to indicate A,” and “carrying A” can be interpreted as directly carrying A or indirectly indicating A.

[0173] In some embodiments, terms such as "in response to...", "in response to determining...", "in the case of...", "at the time of...", "when...", "if...", "if...", etc. can be used interchangeably.

[0174] In some embodiments, terms such as "greater than", "greater than or equal to", "not less than", "more than", "more than or equal to", "not less than", "higher than", "higher than or equal to", "not less than", and "above" can be replaced with each other, and terms such as "less than", "less than or equal to", "not greater than", "less than", "less than or equal to", "not more than", "lower than", "lower than or equal to", "not higher than", and "below" can be replaced with each other.

[0175] In some embodiments, devices, etc. can be interpreted as physical or virtual, and their names are not limited to the names recorded in the embodiments. Terms such as "device", "equipment", "device", "circuit", "network element", "node", "function", "unit", "section", "system", "network", "chip", "chip system", "entity", and "subject" can be used interchangeably.

[0176] In some embodiments, "network" can be interpreted as devices included in the network (eg, access network equipment, core network equipment, etc.).

[0177] In some embodiments, the terms "access network device (AN device)", "radio access network device (RAN device)", "base station (BS)", "radio base station" "fixed station", "node", "access point", "transmission point (TP)", "reception point (RP)", "transmission / reception point (TRP)", "panel", "antenna panel", "antenna array", "cell", "macro cell", "small cell", "femto cell", "pico cell", "sector", "cell group", "carrier", "component carrier", "bandwidth part (BWP)" and the like may be used interchangeably.

[0178] In some embodiments, the terms "terminal", "terminal device", "user equipment (UE)", "user terminal", "mobile station (MS)", "mobile terminal (MT)", subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, etc. can be used interchangeably.

[0179] In some embodiments, the access network device, the core network device, or the network device can be replaced by a terminal. For example, the various embodiments of the present disclosure can also be applied to a structure in which the communication between the access network device, the core network device, or the network device and the terminal is replaced by communication between multiple terminals (for example, it can also be called device-to-device (D2D), vehicle-to-everything (V2X), etc.). In this case, it can also be set as a structure in which the terminal has all or part of the functions of the access network device. In addition, language such as "uplink" and "downlink" can also be replaced by language corresponding to communication between terminals (for example, "side"). For example, uplink channels, downlink channels, etc. can be replaced by side channels, and uplinks, downlinks, etc. can be replaced by side links.

[0180] In some embodiments, the terminal may be replaced by an access network device, a core network device, or a network device. In this case, the access network device, the core network device, or the network device may have a structure that has all or part of the functions of the terminal.

[0181] In some embodiments, obtaining data, information, etc. may comply with the laws and regulations of the country where the data is obtained.

[0182] In some embodiments, data, information, etc. may be obtained with the user's consent.

[0183] In addition, each element, each row, or each column in the table of the embodiment of the present disclosure can be implemented as an independent embodiment, and the combination of any elements, any rows, and any columns can also be implemented as an independent embodiment.

[0184] The correspondences shown in the tables of the present disclosure can be configured or predefined. The values ​​of the information in each table are merely examples and can be configured to other values, which are not limited by the present disclosure. When configuring the correspondences between information and parameters, it is not necessarily required to configure all the correspondences shown in each table. For example, in the tables of the present disclosure, the correspondences shown in certain rows may not be configured. For another example, appropriate deformation adjustments can be made based on the above tables, such as splitting, merging, etc. The names of the parameters shown in the titles of the above tables may also adopt other names that can be understood by the communication device, and the values ​​or representations of the parameters may also adopt other values ​​or representations that can be understood by the communication device. When implementing the above tables, other data structures may also be used, such as arrays, queues, containers, stacks, linear lists, pointers, linked lists, trees, graphs, structures, classes, heaps, hash tables or hash tables, etc.

[0185] The predefined in the present disclosure may be understood as defined, predefined, stored, pre-stored, pre-negotiated, pre-configured, solidified, or pre-burned.

[0186] Figure 1A is a schematic diagram of the architecture of a communication system according to an embodiment of the present disclosure. As shown in Figure 1A, the communication system 100 may include a perception signal receiving end, a first device, and a second device; wherein the first device can be used for model training, and the second device can be used for model application; and the perception signal receiving end and the first device can be the same or different devices; the perception signal receiving end and the second device can be the same or different devices; the first device and the second device can be the same or different devices. Optionally, the perception signal receiving end, the first device, and the second device can be terminals or network devices. Optionally, the above-mentioned network device can include at least one of an access network device and a core network device.

[0187] In some embodiments, the terminal includes, for example, a mobile phone, a wearable device, an Internet of Things device, a car with communication function, a smart car, a tablet computer, a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal device in industrial control, a wireless terminal device in self-driving, a wireless terminal device in remote medical surgery, a wireless terminal device in a smart grid, a wireless terminal device in transportation safety, a wireless terminal device in a smart city, and at least one of a wireless terminal device in a smart home, but is not limited thereto.

[0188] In some embodiments, the access network device is, for example, a node or device that accesses a terminal to a wireless network. The access network device may include an evolved NodeB (eNB), a next generation evolved NodeB (ng-eNB), a next generation NodeB (gNB), a node B (NB), a home node B (HNB), a home evolved nodeB (HeNB), a wireless backhaul device, a radio network controller (RNC), a base station controller (BSC), a base transceiver station (BTS), a base band unit (BBU), a mobile switching center, a base station in a 6G communication system, an open base station (Open RAN), a cloud base station (Cloud RAN), a base station in other communication systems, and at least one of an access node in a wireless fidelity (WiFi) system, but is not limited thereto.

[0189] In some embodiments, the technical solution of the present disclosure can be applied to the Open RAN architecture. In this case, the interfaces between or within the access network devices involved in the embodiments of the present disclosure can be transformed into internal interfaces of the Open RAN, and the processes and information interactions between these internal interfaces can be implemented through software or programs.

[0190] In some embodiments, the access network device can be composed of a centralized unit (CU) and a distributed unit (DU), where the CU can also be called a control unit. The CU-DU structure can be used to split the protocol layer of the access network device, with the functions of some protocol layers centrally controlled by the CU, and the functions of the remaining part or all of the protocol layers distributed in the DU, which is centrally controlled by the CU, but is not limited to this.

[0191] In some embodiments, the core network device may be a device including one or more network elements, or may be multiple devices or a group of devices, each including all or part of one or more network elements. The network element may be virtual or physical. The core network, for example, includes at least one of an Evolved Packet Core (EPC), a 5G Core Network (5GCN), and a Next Generation Core (NGC). Alternatively, the core network device may also be a location management function network element. Exemplarily, the location management function network element includes a location server (location server), which may be implemented as any one of the following: Location Management Function (LMF), Enhanced Serving Mobile Location Centre (E-SMLC), Secure User Plane Location (SUPL), and Secure User Plane Location Platform (SUPLLP).

[0192] It can be understood that the communication system described in the embodiment of the present disclosure is for the purpose of more clearly illustrating the technical solution of the embodiment of the present disclosure, and does not constitute a limitation on the technical solution proposed in the embodiment of the present disclosure. Ordinary technicians in this field can know that with the evolution of the system architecture and the emergence of new business scenarios, the technical solution proposed in the embodiment of the present disclosure is also applicable to similar technical problems.

[0193] The following embodiments of the present disclosure may be applied to the communication system 100 shown in FIG1A , or a portion thereof, but are not limited thereto. The entities shown in FIG1A are illustrative only. The communication system may include all or part of the entities shown in FIG1A , or may include other entities other than those shown in FIG1A . The number and form of the entities may be arbitrary. The connection relationship between the entities is illustrative only. The entities may be connected or disconnected, and the connection may be in any manner, including direct or indirect, wired or wireless.

[0194] The embodiments of the present disclosure can be applied to Long Term Evolution (LTE), LTE-Advanced (LTE-A), LTE-Beyond (LTE-B), SUPER 3G, IMT-Advanced, 4th generation mobile communication system (4G), 5th generation mobile communication system (5G), 5G new radio (NR), future radio access (FRA), new radio access technology (RAT), new radio (NR), new radio access (NX), future generation radio access (FX), Global System for Mobile Communications (GSM (registered trademark)), CDMA2000, Ultra Mobile Broadband (UMB), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, Ultra-WideBand (UWB), Bluetooth (registered trademark), Public Land Mobile Network (PLMN) networks, Device-to-Device (D2D) systems, Machine-to-Machine (M2M) systems, Internet of Things (IoT) systems, Vehicle-to-Everything (V2X), systems utilizing other model training and application methods, and next-generation systems based on and extending these systems. Furthermore, multiple systems may be combined (for example, LTE or LTE-A combined with 5G) for application.

[0195] Optionally, in a communication system, it is often necessary to sense a sensing target to determine a sensing result of the sensing target. This sensing result may include, for example, the distance, speed, angle, etc. of the sensing target. Currently, the Integrated Sensing and Communication (ISAC) system can implement sensing and detection functions based on communication. It is one of the most important research directions in wireless communications and a key enabling technology in 5G-A (5G-Advanced) and future 6G mobile communications. Therefore, an ISAC system (or an ISAC simulation scenario) is often used to achieve sensing of the sensing target. Optionally, the ISAC simulation scenario typically includes a sensing signal transmitter, a sensing signal receiver, a sensing target, and an obstruction. The sensing signal transmitter can be used to transmit a sensing signal, and the sensing signal transmitted by the sensing signal transmitter can be directly transmitted to the sensing signal receiver, or the sensing signal can reach the sensing signal receiver after being reflected or refracted by the sensing target and / or obstruction; the obstruction may be, for example, another terminal or debris. Among them, considering that the transmission paths of the perception signals may be different, the following types of signals are mainly considered in the ISAC simulation scenario, namely: perception line of sight (LoS) path, perception non-line of sight (NLoS) path, clutter LOS path, and clutter NLOS path. Optionally, Figure 1B is a schematic diagram of the architecture of the perception LoS path, perception NLoS path, clutter LOS path, and clutter NLOS path in the ISAC simulation scenario according to an embodiment of the present disclosure. As shown in Figure 1B, the perception signal transmitter and the perception signal receiver are both base stations, the perception target is the vehicle in Figure 1B, and there is an obstruction between the base station and the vehicle. Among them, the transmission path without obstruction between the base station and the vehicle is the perception LOS path, the transmission path with obstruction between the base station and the vehicle is the perception NLOS path, the transmission path without obstruction between the base station and the obstruction is the clutter LOS path, and the transmission path with obstruction between the base station and the obstruction is the clutter NLOS path. Furthermore, after receiving the sensing signals of different paths, the sensing signal receiving end can measure the channel state information of the different paths based on the sensing signals of the different paths, and determine the sensing result of the sensing target based on the channel state information. The channel state information can be, for example, a channel state information matrix H, which is the superposition of the sensing target channel (such as the sensing LOS path, the sensing NLOS path, etc.), the clutter channel (such as the clutter LOS path, the clutter NLOS path, etc.) and the channel noise. The channel state information matrix H generally includes three dimensions: subcarrier, orthogonal frequency division multiplexing (OFDM) symbol, and receiving antenna port.Specifically, the channel state information matrix H represents the channel state information and is an M×N×P complex matrix, where M is the number of OFDM symbols, N is the number of subcarriers, and P is the number of receive antenna ports. The element in the channel state information matrix H at the position of the mth OFDM symbol, the nth subcarrier, and the pth receive antenna port is a complex number representing the impact on the amplitude and phase of the signal as it propagates from the transmit antenna to the receive antenna: H. m,n,p =C m,n,p =a m,n,p +i×b m,n,p .

[0196] Where C is a complex number, a is the real part of the complex number, b is the imaginary part of the complex number, a and b themselves are real numbers, and i is the imaginary unit.

[0197] Optionally, in some embodiments, the phase shifts in the three dimensions of the channel state information matrix H do not affect each other, and different types of perception results of the perception target can be obtained based on the phase changes in the three dimensions. For example, the distance estimation of the perception target can be completed based on the phase difference caused by the inter-subcarrier delay, the speed estimation of the perception target can be completed based on the phase difference caused by the Doppler effect between OFDM symbols, and the angle estimation of the perception target can be completed based on the phase changes between the received signals of multiple antenna ports. In addition, a three-dimensional sensing algorithm such as a three-dimensional detection algorithm (3D-MUSIC) and a 3D-ESPRIT can be used to obtain the perception results of the three parameters of the distance, speed, and angle of the perception target based on the channel state information matrix H. Optionally, Figure 1C is a schematic diagram of the architecture for determining the perception results based on the above-mentioned sensing algorithm (such as the aforementioned three-dimensional sensing algorithm) according to an embodiment of the present disclosure. As shown in Figure 1C, the above-mentioned sensing algorithm can be used to calculate the channel state information matrix H to obtain the distance perception result, speed perception result, and angle perception result of the perception target.

[0198] However, although the perception algorithm can achieve high accuracy in perceiving target distance, speed, angle, and other information in some ISAC application scenarios, the following major problems still exist:

[0199] (1) The applicability of perception algorithms to specific scenarios is limited

[0200] The performance of some perception algorithms can be significantly degraded by factors such as clutter and noise in the scene. For example, when clutter is introduced, the maximum eigenvalue estimated by the 3D ESPRIT algorithm may correspond to a clutter scatterer, degrading the performance of distance, velocity, and angle estimation. In the case of multiple perceived targets, the direct path of non-perceived targets can have a certain impact on the perception of perceived targets. When the perceived target is far away, the impact of noise is particularly significant.

[0201] (2) The computational complexity of the perception algorithm is high

[0202] In some cases, achieving highly accurate perception results requires a high level of algorithm complexity, hindering their practical application. For example, the 3D MUSIC and 3D ESPRIT algorithms can simultaneously calculate and obtain target distance, speed, and angle perception results, while achieving low perception errors. However, these algorithms require complex operations such as eigenvalue decomposition and spatial spectrum calculation, resulting in a high overall computational complexity.

[0203] (3) Accuracy in multi-target perception needs to be improved

[0204] In practical applications of synaesthesia integration, there may be multiple perception targets, necessitating the design of a multi-target perception algorithm to obtain perception results for multiple different perception targets. Existing multi-target perception algorithms often have errors when estimating the number of perception targets, resulting in large errors in the algorithm's estimation results.

[0205] Therefore, a perception method based on an AI model has been introduced. By inputting the channel state information of the perceived target into the AI ​​model, the AI ​​model directly outputs the perception result. This solves the technical problems of existing perception algorithms, such as limited application scenarios, high complexity, and poor perception accuracy in the case of multiple perceived targets. However, how to implement perception based on this AI model, as well as how to train and apply it, are urgent issues that need to be addressed.

[0206] FIG2A is an interactive diagram of a model training method according to an embodiment of the present disclosure. As shown in FIG2A , the embodiment of the present disclosure relates to a model training method for a communication system 100, the method comprising:

[0207] Step 2101: The first device determines a training sample set.

[0208] Optionally, the first device may be a device for performing model training.

[0209] Optionally, the training sample set may be data used by the first device to train a model. The training sample set may include at least one set of sample data, and the sample data may include the first information and / or the second information.

[0210] Optionally, the above-mentioned first information may be channel state information determined by the perception signal receiving end based on the received perception signal, and the first information may be used to determine the perception result of the perception target. Optionally, the first information may be the channel state information matrix H described before the embodiment of Figure 2A. For a detailed introduction to the "channel state information matrix H", please refer to the description of the above embodiment. In addition, in some embodiments, the first information may include sub-information of at least one dimension, and sub-information of different dimensions may be used to determine different types of perception results. Among them, the perception result of the perception target includes at least one of the following categories: the distance of the perception target, the speed of the perception target, and the angle of the perception target. Furthermore, in some embodiments, when the first information is the channel state information matrix H described previously in the embodiment of FIG. 2A , the first information may include sub-information of three dimensions, and the sub-information of the three dimensions may be, respectively: sub-information corresponding to the “OFDM symbol” dimension (i.e., information related to the “OFDM symbol” dimension in the channel state information matrix H), sub-information corresponding to the “subcarrier” dimension (i.e., information related to the “subcarrier” dimension in the channel state information matrix H), and sub-information corresponding to the “receiving antenna port” dimension (i.e., information related to the “receiving antenna port” dimension in the channel state information matrix H). The sub-information corresponding to the “OFDM symbol” dimension can be used to determine the speed of the perceived target, the sub-information corresponding to the “subcarrier” dimension can be used to determine the distance of the perceived target, and the sub-information corresponding to the “receiving antenna port” dimension can be used to determine the angle of the perceived target. It should be noted that the "distance of the perceived target, the speed of the perceived target, and the angle of the perceived target" in the above content are only some examples of the perception results of the perceived target. The perception results may also include other results, such as: the perception results of the perceived target may also include the position of the perceived target, etc. The present disclosure does not make specific limitations on this, and other types of perception results are also within the scope of protection of the present disclosure.

[0211] Optionally, the above-mentioned second information may be actual state information of the perceived target; in some embodiments, the second information may include at least one of the following: actual distance of the perceived target, actual speed of the perceived target, actual angle of the perceived target. Optionally, the second information may be used as a label in a training sample set, for example. It should be noted that the "actual distance of the perceived target, actual speed of the perceived target, actual angle of the perceived target" in the above content are only some examples of the content included in the second information. The second information may also include other actual state information of the perceived target, such as: the second information may also include the actual position of the perceived target, etc. The present disclosure does not make specific limitations on this, and other actual state information of the perceived target that the second information may include is also within the scope of protection of the present disclosure.

[0212] Optionally, the aforementioned training sample set may be determined by a perception signal receiving end. Furthermore, in some embodiments, the first device may be a different device from the perception signal receiving end, or the first device may be the same device as the perception signal receiving end. When the first device and the perception signal receiving end are different devices, the training sample set may be sent by the perception signal receiving end to the first device. When the first device and the perception signal receiving end are the same device, the training sample set is determined by the first device based on the received perception signal.

[0213] Alternatively, in some other embodiments, the second information in the aforementioned training sample set may be sent by the sensing target to the first device.

[0214] Furthermore, in some embodiments, the first device can be used to train a first model, wherein the first model can be used to implement perception. The first model can be, for example, an AI perception model. Furthermore, the phrase "the first model can be used to implement perception" can be embodied as follows: the first model can determine at least one type of perception result of a perceived target based on the first information. Specifically, in some embodiments, the first model can determine a type of perception result of a perceived target based on the first information, and different first models are used to determine different types of perception results. For example, assume there are three first models, namely, first model #1, first model #2, and first model #3. First model #1 can determine the distance of a perceived target based on the first information, first model #2 can determine the speed of a perceived target based on the first information, and first model #3 can determine the angle of a perceived target based on the first information. Alternatively, in other embodiments, the first model can determine all perception results of a perceived target based on the first information, that is, the first model can determine the distance, speed, and angle of a perceived target based on the first information. Alternatively, in some other embodiments, the first model may also determine multiple categories of perception results of the perception target based on the first information, but not all perception results, and the perception results determined by different first models are different; for example, assuming there are three first models, namely first model #1, first model #2, and first model #3, wherein first model #1 can determine the distance and speed of the perception target based on the first information, first model #2 can determine the distance and angle of the perception target based on the first information, and first model #3 can determine the speed and angle of the perception target based on the first information. Alternatively, in some other embodiments, the number of categories of perception results determined by different first models may be different. For example, some first models may determine one category of perception results of the perception target based on the first information, and other first models may determine multiple categories of perception results or all perception results of the perception target based on the first information. This disclosure does not specifically limit this.

[0215] For example, in some embodiments, FIG2B is a schematic diagram of a first model used to output all perception results according to an embodiment of the present disclosure, and FIG2C is a schematic diagram of different first models used to output a type of perception results according to an embodiment of the present disclosure. As shown in FIG2B , the AI ​​perception model (i.e., the aforementioned first model) can output the distance perception result, speed perception result, and angle perception result of the perception target based on the channel matrix H (i.e., the aforementioned channel state information matrix H). And, as shown in FIG2C , there are three first models, namely, a distance perception model, a speed perception model, and an angle perception model. The distance perception model is used to output the distance perception result of the perception target based on the distance perception information in the channel state information matrix H (i.e., the sub-information corresponding to the aforementioned "subcarrier" dimension), the speed perception model is used to output the speed perception result of the perception target based on the speed perception information in the channel state information matrix H (i.e., the sub-information corresponding to the aforementioned "OFDM symbol" dimension), and the angle perception model is used to output the angle perception result of the perception target based on the angle perception information in the channel state information matrix H (i.e., the sub-information corresponding to the aforementioned "receiving antenna port" dimension).

[0216] Step 2102: The first device determines at least one type of perception result that needs to be determined by the first model.

[0217] Optionally, in some embodiments, the specific categories of the at least one type of perception results that the first model needs to determine may be predefined by the protocol, or may be determined autonomously by the first device.

[0218] Step 2103: The first device determines third information based on the first information.

[0219] Optionally, in some embodiments, the first device may determine the third information from the first information based on at least one type of perception result that needs to be determined by the first model. The third information may be: sub-information of at least one dimension in the first information used to determine at least one type of perception result (i.e., at least one type of perception result that needs to be determined by the first model in step 2102). For example, when the at least one type of perception result that needs to be determined by the first model includes the speed and distance of the perceived target, the third information may be the sub-information corresponding to the "OFDM symbol" dimension and the sub-information corresponding to the "subcarrier" dimension in the first information.

[0220] Optionally, the method for the first device to determine the third information based on the first information may include at least one of the following:

[0221] Method 1: directly extracting sub-information of at least one dimension for determining at least one type of perception result from the first information as the third information.

[0222] Method 2: The sub-information of at least one dimension in the first information that is not used to determine at least one type of perception result is set to a fixed value (such as 0 or 1), and the first information after being set to the fixed value is determined as the third information.

[0223] For example, assuming that at least one type of perception result to be determined by the first model is: distance perception result, the first device can retain the data in the first information (i.e., the channel state information matrix H) that can reflect the distance characteristics of the perception target, i.e., N complex data on the subcarrier dimension, and set the OFDM symbol dimension data M and the receiving antenna port dimension data P in the first information (i.e., the channel state information matrix H) to fixed values, for example, let M=1, P=1, so as to determine the above-mentioned third information. Or

[0224] As another example, assuming that at least one type of perception result to be determined by the first model is speed perception result, the first device can retain the data in the first information (i.e., the channel state information matrix H) that can reflect the perception target speed characteristics, i.e., M complex data on the OFDM symbol dimension, and set the data N on the subcarrier dimension and the data P on the receiving antenna port dimension to fixed values, for example, let N=1, P=1, so as to determine the above-mentioned third information. Or

[0225] As another example, assuming that at least one type of perception result that the first model needs to determine is: angle perception result, the first device can retain the data in the first information (i.e., the channel state information matrix H) that can reflect the angle characteristics of the perception target, that is, P complex data in the receiving antenna port dimension, and set the data M in the OFDM symbol dimension and the data N in the subcarrier dimension to fixed values, for example, let M=1, N=1, so as to determine the above-mentioned third information.

[0226] It should be noted that, in other embodiments, the first device may also adopt other methods to determine the third information based on the first information, and this disclosure does not specifically limit this.

[0227] Optionally, in some embodiments, the above-mentioned methods for determining the third information can also be combined with each other. For example, assuming that the at least one perception result required to be determined by the first model includes a distance perception result and a speed perception result, the first device can use the above-mentioned method one to determine the sub-information used to determine the distance perception result from the first information, and use the above-mentioned method two to determine the sub-information used to determine the speed perception result from the first information, and then combine the determined sub-information into the third information.

[0228] It should also be noted that, in some embodiments, when the first model needs to determine all perception results of the perception target, the first information is the same as the third information.

[0229] Step 2104: The first device inputs the third information into the first model to obtain an output result of the first model.

[0230] Optionally, the output result of the first model may include at least one type of perception result determined by the first model based on the third information.

[0231] For example, assuming that the at least one type of perception result that needs to be determined by the first model determined by the first device in the above step 2102 is the speed, angle or distance of the perceived target, then the output result of the first model here is: the speed, angle or distance of the perceived target determined by the first model based on the third information; or, assuming that the at least one type of perception result that needs to be determined by the first model determined by the first device in the above step 2102 is the speed, angle and distance of the perceived target, then the output result of the first model here is: the speed, angle and distance of the perceived target determined by the first model based on the third information.

[0232] Step 2105: The first device calculates a loss function based on the output result of the first model and the second information.

[0233] Optionally, the first device can calculate the loss function by comparing the error between the output result of the first model and the actual state of the perceived target indicated by the second information. For example, assuming that the output result of the first model includes the speed of the perceived target, the first device can calculate the loss function by comparing the error between the speed of the perceived target output by the first model and the actual speed of the perceived target in the second information.

[0234] Step 2106: The first device adjusts the parameters of the first model based on the loss function, inputs the third information into the adjusted first model to obtain the output result of the first model, and then calculates the loss function of the first model until the first condition is met, and determines that the training is completed.

[0235] Optionally, the first condition may include, for example, at least one of the following:

[0236] The training cycle (or training times) of the first model reaches a first threshold;

[0237] The first model converges. The convergence of the first model can be understood as, for example, when the loss function of the first model decreases to a value smaller than a certain value, the model is considered to have converged.

[0238] Optionally, steps 2102-2106 above provide a method for training the first model by the first device in an embodiment of the present disclosure. It should be noted that, in other embodiments, the first device may also train the first model based on other methods. For example, the first device may train the first model based on a stochastic gradient descent (SGD) method. This disclosure does not impose any specific limitations on this.

[0239] Step 2107: The first device and the second device are different devices, and the first device sends fourth information to the second device.

[0240] Optionally, the second device may be a device for applying the first model, or may be an inference device for the first model. Furthermore, the fourth information may be used to indicate model parameters of the first model trained by the first device. After the second device receives the fourth information, the second device may deploy the first model based on the fourth information, so that the second device can subsequently successfully apply the first model for perception.

[0241] It should be noted that, in some embodiments, the first device and the second device may also be the same device.

[0242] In the above embodiment, a model training method is provided to achieve successful training of the first model. The first model can determine at least one type of perception result of the perception target based on the input first information. It can be seen that the first model trained by the training method of the present disclosure is used to perceive the perception target, that is, the present disclosure provides a training method for a "perception model" so that the model can be successfully trained. Then, perception can be achieved based on the model in the future without having to calculate the perception result using the perception algorithm as in the related art. This effectively solves the technical problems such as the limited application scenarios when the perception algorithm calculates the perception result, the high complexity, and the poor perception accuracy in the case of multiple perception targets, thereby ensuring the perception accuracy, reducing the perception complexity, and broadening the perception application scenarios.

[0243] The model training method involved in the embodiments of the present disclosure may include at least one of steps 2101 to 2107. For example, step 2101 may be implemented as an independent embodiment, step 2102 may be implemented as an independent embodiment, step 2103 may be implemented as an independent embodiment, and step 2101+S2102 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0244] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0245] FIG2D is an interactive diagram of a model application method according to an embodiment of the present disclosure. As shown in FIG2D , the embodiment of the present disclosure relates to a model application method for a communication system 100, the method comprising:

[0246] Step 2201: The second device determines the first information.

[0247] The first information may be channel state information determined by the perception signal receiving end based on the received perception signal, and the first information may be used to determine the perception result of the perception target; for a detailed introduction to the first information, reference may be made to the above embodiment description.

[0248] Furthermore, in some embodiments, the aforementioned first information may be determined by the perception signal receiving end. In some embodiments, the second device may be a different device from the perception signal receiving end, or the second device may be the same device as the perception signal receiving end. When the second device and the perception signal receiving end are different devices, the first information may be sent by the perception signal receiving end to the second device. When the second device and the perception signal receiving end are the same device, the first information is determined by the second device based on the received perception signal.

[0249] Step 2202: The first device and the second device are different devices, and the first device sends fourth information to the second device.

[0250] Optionally, the first device may be a device for training the first model, and the second device may be a device for applying the first model, or may be an inference device for the first model. For a detailed introduction to the first model, reference may be made to the above embodiment.

[0251] Furthermore, the fourth information can be used to indicate the model parameters of the first model trained by the first device. For a detailed introduction to the fourth information, please refer to the description of the above embodiment.

[0252] It should be noted that, in some embodiments, the first device and the second device may also be the same device.

[0253] Step 2203: The second device determines the first model.

[0254] Optionally, when the first device and the second device are different devices, the second device can determine the first model based on the fourth information sent by the first device; when the first device and the second device are the same device, the second device can directly determine the first model after training the first model.

[0255] Step 2204: The second device determines at least one type of perception result that needs to be determined by the first model.

[0256] Step 2205: The second device determines third information based on the first information.

[0257] Step 2206: The second device inputs the third information into the first model to obtain at least one type of perception result output by the first model.

[0258] For a detailed description of steps 2204-2206, please refer to the above embodiment description.

[0259] A method for applying a first model is provided, which can be used to determine perception results of a perception target. Specifically, perception can be achieved based on the first model, eliminating the need to calculate perception results using a perception algorithm as in related technologies. This effectively addresses technical issues such as the limited application scenarios, high complexity, and poor perception accuracy in the case of multiple perception targets when using perception algorithms to calculate perception results. This ensures perception accuracy, reduces perception complexity, and broadens perception application scenarios.

[0260] The model training method involved in the embodiments of the present disclosure may include at least one of steps 2201 to 2206. For example, step 2201 may be implemented as an independent embodiment, step 2202 may be implemented as an independent embodiment, step 2203 may be implemented as an independent embodiment, and step 2201+S2202 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0261] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0262] In addition, it should be noted that the present disclosure does not limit the specific structure of the first model. The specific structure of the first model depends on the design and deployment in the actual application process. The present disclosure mainly provides a method for training and applying the first model.

[0263] Figure 3 is an interactive diagram of a model training and application method according to an embodiment of the present disclosure. As shown in Figure 3, the embodiment of the present disclosure relates to a model training method for a first device, the method comprising:

[0264] Step 3101: Determine a training sample set.

[0265] Step 3102: Train a first model based on the training sample set.

[0266] Optionally, the training sample set includes at least one set of sample data, and the sample data includes first information and / or second information; wherein, the first information is the channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine the perception result of the perception target; the second information is the actual state information of the perception target.

[0267] Optionally, the first model may determine at least one type of perception result of the perception target based on the first information.

[0268] Optionally, the perception result of the perception target includes at least one of the following categories:

[0269] a distance of the perceived target determined based on the first information;

[0270] a speed of the perceived target determined based on the first information;

[0271] The angle of the perceived target is determined based on the first information.

[0272] Optionally, the first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of perception results.

[0273] Optionally, the second information includes at least one of the following:

[0274] the actual distance of the perceived target;

[0275] the actual speed of the perceived target;

[0276] The actual angle of the perceived target.

[0277] Optionally, the training a first model based on the training sample set includes:

[0278] Determining at least one type of perception result that needs to be determined by the first model;

[0279] Determining third information based on the first information, the third information being: sub-information in the first information used to determine at least one dimension of the at least one type of perception result;

[0280] Inputting the third information into the first model to obtain an output result of the first model;

[0281] Calculating a loss function based on an output result of the first model and the second information;

[0282] The parameters of the first model are adjusted based on the loss function, and the third information is input into the adjusted first model to obtain the output result of the first model, and then the loss function of the first model is calculated until the first condition is met, and the training is determined to be completed.

[0283] Optionally, determining third information based on the first information includes:

[0284] Setting sub-information in the first information that is not used to determine at least one dimension of the at least one type of perception result to a fixed value, and determining the first information after being set to the fixed value as the third information; and / or

[0285] Sub-information for determining at least one dimension of the at least one type of perception result is extracted from the first information as the third information.

[0286] Optionally, the at least one dimension includes one or more of the following dimensions:

[0287] Orthogonal frequency division multiplexing OFDM symbol dimension;

[0288] Subcarrier dimension;

[0289] Receive antenna port dimensions.

[0290] Optionally, the first condition includes at least one of the following:

[0291] The training period of the first model reaches a first threshold;

[0292] The first model converged.

[0293] Optionally, the first model may determine a type of perception result of the perception target based on the first information, and different first models are used to determine different types of perception results; or

[0294] The first model can determine all perception results of the perception target based on the first information.

[0295] Optionally, when the first model can determine all perception results of the perception target based on the first information, the first information is the same as the third information.

[0296] For a detailed description of steps 3101 - 3102 , please refer to the above embodiment description.

[0297] The model training and application methods involved in the embodiments of the present disclosure may include at least one of steps 3101 and 3102. For example, step 3101 may be implemented as an independent embodiment, step 3102 may be implemented as an independent embodiment, and step 3101+S3102 may be implemented as an independent embodiment, but the present invention is not limited thereto.

[0298] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0299] Figure 4 is an interactive diagram of a model training and application method according to an embodiment of the present disclosure. As shown in Figure 4, the embodiment of the present disclosure relates to a model application method for a second device, the method comprising:

[0300] Step 4101: Determine the first information.

[0301] Step 4102: Determine the first model.

[0302] Step 4103: Determine a perception result of the perception target based on the first information and the first model.

[0303] Optionally, the first information is channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine a perception result of the perception target.

[0304] Optionally, the first model may determine at least one type of perception result of the perception target based on the first information.

[0305] Optionally, the perception result of the perception target includes at least one of the following categories:

[0306] a distance of the perceived target determined based on the first information;

[0307] a speed of the perceived target determined based on the first information;

[0308] The angle of the perceived target is determined based on the first information.

[0309] Optionally, the first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of perception results.

[0310] Optionally, determining a perception result of the perception target based on the first information and the first model includes:

[0311] Determining at least one type of perception result that needs to be determined by the first model;

[0312] Determining third information based on the first information, the third information being: sub-information in the first information used to determine at least one dimension of the at least one type of perception result;

[0313] The third information is input into the first model to obtain the at least one type of perception result output by the first model.

[0314] Optionally, determining third information based on the first information includes:

[0315] Setting sub-information in the first information that is not used to determine at least one dimension of the at least one type of perception result to a fixed value, and determining the first information after being set to the fixed value as the third information; and / or

[0316] Sub-information for determining at least one dimension of the at least one type of perception result is extracted from the first information as the third information.

[0317] Optionally, the at least one dimension includes one or more of the following dimensions:

[0318] Orthogonal frequency division multiplexing OFDM symbol dimension;

[0319] Subcarrier dimension;

[0320] Receive antenna port dimensions.

[0321] Optionally, the first model may determine a type of perception result of the perception target based on the first information, and different first models are used to determine different types of perception results; or

[0322] The first model can determine all perception results of the perception target based on the first information.

[0323] Optionally, when the first model can determine all perception results of the perception target based on the first information, the first information is the same as the third information.

[0324] For a detailed description of steps 4101-4103, please refer to the above embodiment.

[0325] The model training and application methods involved in the embodiments of the present disclosure may include at least one of steps 4101 to 4103. For example, step 4101 may be implemented as an independent embodiment, and step 4102 may be implemented as an independent embodiment, but are not limited thereto.

[0326] In this embodiment or example, unless there is any contradiction, each step can be independent, arbitrarily combined or exchanged in order, the optional methods or optional examples can be arbitrarily combined, and can be arbitrarily combined with any steps of other embodiments or other examples.

[0327] The following is an exemplary introduction to the above method.

[0328] The AI-based synaesthesia integration implementation method proposed in the present disclosure mainly completes the perception of the distance, speed, and angle information of the target. To train the AI ​​perception model, it is necessary to first collect a training data set, which contains a certain number of data samples. Each data sample is composed of a channel state information matrix H obtained by measuring the perception target and the actual distance d, speed v, and angle θ information of the perception target. As mentioned above, the channel matrix H of the perception target is an M×N×P complex matrix, which contains the distance, speed, and angle information of the perception target. Among them, the data change in the subcarrier dimension corresponds to the distance information of the perception target, the data change in the receiving antenna port dimension corresponds to the angle information of the target, and the data change in the OFDM symbol dimension corresponds to the speed information of the target. The actual distance d, speed v, and angle θ information of the perception target are labels in the model training data set.

[0329] During the model training process, the perception results output by the model, including distance speed angle The error between the label distance d, velocity v, and angle θ of the training set is used as the loss function. Model training is performed through methods such as stochastic gradient descent (SGD) to reduce the loss function, so that the model learns the mapping relationship between the input channel matrix and the output perception result. Model training is completed after a specific number of training cycles is reached or the model accuracy meets certain requirements.

[0330] Since the present disclosure mainly considers obtaining information on the three parameters of the perceived target, namely, distance, speed, and angle, and different dimensions of the perceived target channel state information matrix are related to different types of parameter information, the present disclosure proposes two target perception methods based on AI models.

[0331] (1) Simultaneous perception model of distance, speed, and angle

[0332] The channel state information matrix H is directly used as the input data of the AI ​​perception model to train an AI perception model, which also outputs the distance of the target. speed angle The three types of perception results are shown in Figure 2B above.

[0333] (2) Distance, speed, and angle perception models

[0334] Combined with the perception algorithm, the data of different dimensions in the channel matrix are used to perceive the target distance, speed, and angle. Based on a part of the data in the channel matrix H, the three types of perception results of distance, speed, and angle are estimated respectively. Three AI perception models are trained to output distance and speed respectively. speed angle The three types of perception results are shown in Figure 2C above.

[0335] Data preprocessing refers to data extraction from the channel matrix H, specifically: mainly retaining the information in the channel matrix used for target distance perception, that is, the data in the subcarrier dimension, for training data of the distance perception model; mainly retaining the information in the channel matrix used for target speed perception, that is, the data in the OFDM symbol dimension, for training data of the speed perception model; mainly retaining the information in the channel matrix used for target angle perception, that is, the data in the receiving antenna port dimension, for training data of the angle perception model.

[0336] The distance, speed, and angle simultaneous perception model uses the channel state information matrix as input data for the AI ​​model. The channel state information matrix contains data feature information about the distance, speed, and angle of the perceived target. The AI ​​model can extract features based on the input channel state information matrix data and then output the perception results.

[0337] For the distance, speed, and angle perception models, it is necessary to first extract data based on the channel matrix. That is, to extract specific data corresponding to the perception result type from the original data, which can reflect the characteristics of the perceived target distance, speed, angle, etc. Specifically:

[0338] The original channel matrix is ​​a complex matrix H, which contains complex data in three dimensions: OFDM symbols, subcarriers, and receiving antenna ports.

[0339] (1) Retain data in the channel matrix that can reflect the target distance characteristics, that is, data in the subcarrier dimension, as input data for the distance perception model. The processing method includes but is not limited to: setting M and P to certain values, for example, setting M = 1 and P = 1, and obtaining N complex data in the subcarrier dimension.

[0340] (2) Retaining data in the channel matrix that can reflect the target speed characteristics, that is, data in the OFDM symbol dimension, as input data for the speed perception model. The processing method includes but is not limited to: setting N and P to certain values, for example, setting N = 1 and P = 1, and obtaining M complex data in the OFDM symbol dimension.

[0341] (3) Retaining data in the channel matrix that can reflect the target angle characteristics, that is, data in the receiving antenna port dimension, as input data for the angle perception model. The processing method includes but is not limited to: setting M and N to specific values, for example, setting M = 1 and N = 1, and obtaining P complex data in the receiving antenna port dimension.

[0342] This disclosure only provides one possible way of data preprocessing.

[0343] The AI-based target perception solution proposed in this disclosure includes, but is not limited to, three types of information: distance, speed, and angle of the perceived target. These three types of information are the most commonly used types of perception parameters in synaesthesia integration applications and can meet the needs of most application scenarios. Using the same ideas as this solution, the perception of other types of parameters can also be achieved. Moreover, when perceiving different types of results, different data feature extraction methods can be selected, and different methods can be combined with each other.

[0344] In addition, the specific structure of the AI ​​perception model proposed in this disclosure for perceiving target distance, speed, angle and other information depends on the design and deployment in the actual application process.

[0345] As can be seen from the above, in the application scenario of integrated communication and perception, the strong learning and modeling capabilities of neural network models can be utilized to train an AI perception model based on a certain amount of data, thereby completing the perception of target distance, speed, angle, and other information. To address the shortcomings of existing traditional perception algorithms, such as limited scenario applicability, high computational complexity, and low accuracy in multi-target perception, this disclosure uses perception datasets to train an AI neural network model, achieving optimizations in perception accuracy, computational complexity, and scope of application, thereby promoting the further development of integrated communication and perception technology.

[0346] Furthermore, the present disclosure proposes a method for perceiving target distance, speed, angle and other information based on an AI neural network model for integrated communication and perception application scenarios. In different integrated communication and perception application scenarios, an AI perception model is trained based on a data set consisting of the channel state information matrix of the perceived target and its actual distance, speed, angle and other parameter information, which can achieve target distance, speed and angle perception with high accuracy. The AI-based target perception solution has a large scope of application, which is conducive to solving the shortcomings of traditional perception algorithms such as limited application scenarios, high complexity, and poor perception accuracy in multi-target situations, thereby promoting the development and application of integrated communication and perception technology.

[0347] The embodiments of the present disclosure further provide an apparatus for implementing any of the above methods. For example, an apparatus is provided, comprising units or modules for implementing each step performed by a terminal in any of the above methods. For another example, another apparatus is provided, comprising units or modules for implementing each step performed by a network device (e.g., an access network device, a core network function node, a core network device, etc.) in any of the above methods.

[0348] It should be understood that the division of the various units or modules in the above device is merely a division of logical functions. In actual implementation, they may be fully or partially integrated into a physical entity, or they may be physically separated. In addition, the units or modules in the device may be implemented in the form of a processor calling software: for example, the device includes a processor, the processor is connected to a memory, and the memory stores instructions. The processor calls the instructions stored in the memory to implement any of the above methods or implement the functions of the various units or modules of the above device, wherein the processor is, for example, a general-purpose processor, such as a central processing unit (CPU) or a microprocessor, and the memory is a memory within the device or a memory outside the device. Alternatively, the units or modules in the device can be implemented in the form of hardware circuits, and the functions of some or all of the units or modules can be realized by designing the hardware circuits. The above-mentioned hardware circuits can be understood as one or more processors; for example, in one implementation, the above-mentioned hardware circuit is an application-specific integrated circuit (ASIC), which realizes the functions of some or all of the above units or modules by designing the logical relationship of the components in the circuit; for example, in another implementation, the above-mentioned hardware circuit can be realized by a programmable logic device (PLD). Taking a field programmable gate array (FPGA) as an example, it can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by configuring the configuration file, thereby realizing the functions of some or all of the above units or modules. All units or modules of the above devices can be realized in the form of software called by the processor, or in the form of hardware circuits, or in part by the form of software called by the processor, and the rest by hardware circuits.

[0349] In the embodiments of the present disclosure, the processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), or a digital signal processor (DSP). In another implementation, the processor can implement certain functions through the logical relationship of the hardware circuit. The logical relationship of the above-mentioned hardware circuit is fixed or reconfigurable. For example, the processor is a hardware circuit implemented by an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and implementing the hardware circuit configuration can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units or modules. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), etc.

[0350] FIG5 is a schematic diagram of the structure of the first device proposed in an embodiment of the present disclosure. As shown in FIG5 , it includes:

[0351] a processing module, configured to determine a training sample set, the training sample set including at least one set of sample data, the sample data including first information and / or second information; wherein the first information is channel state information determined by the perception signal receiving end based on the received perception signal, the first information being used to determine a perception result of a perception target; and the second information is actual state information of the perception target;

[0352] The processing module is further used to train a first model based on the training sample set; wherein the first model can determine at least one type of perception result of the perception target based on the first information.

[0353] Optionally, the processing module is used to execute the steps related to "processing" performed by the first device in any of the above methods. The first device further includes a transceiver module, which can be used to execute the steps related to "transmitting and receiving" performed by the first device in any of the above methods. Detailed description is omitted here.

[0354] FIG6 is a schematic diagram of the structure of the second device proposed in an embodiment of the present disclosure. As shown in FIG6 , it includes:

[0355] a processing module, configured to determine first information, where the first information is channel state information determined by the perception signal receiving end based on the received perception signal, and the first information is used to determine a perception result of the perception target;

[0356] The processing module is further configured to determine a first model, wherein the first model can determine at least one type of perception result of the perception target based on the first information;

[0357] The processing module is further used to determine a perception result of the perception target based on the first information and the first model.

[0358] Optionally, the processing module is used to execute the steps related to "processing" performed by the second device in any of the above methods. The second device further includes a transceiver module, which can be used to execute the steps related to "transmitting and receiving" performed by the second device in any of the above methods. Detailed description is omitted here.

[0359] Figure 7A is a schematic diagram of the structure of a communication device 7100 proposed in an embodiment of the present disclosure. Communication device 7100 can be a network device (e.g., an access network device, a core network device, etc.), a terminal (e.g., a user equipment, etc.), a chip, a chip system, or a processor that supports a network device to implement any of the above methods, or a chip, a chip system, or a processor that supports a terminal to implement any of the above methods. Communication device 7100 can be used to implement the methods described in the above method embodiments. For details, please refer to the description of the above method embodiments.

[0360] As shown in Figure 7A, the communication device 7100 includes one or more processors 7101. The processor 7101 can be a general-purpose processor or a dedicated processor, for example, a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, and the central processing unit can be used to control the communication device (such as a base station, a baseband chip, a terminal device, a terminal device chip, a DU or a CU, etc.), execute programs, and process program data. The processor 7101 is used to call instructions to enable the communication device 7100 to perform any of the above methods.

[0361] In some embodiments, the communication device 7100 further includes one or more memories 7102 for storing instructions. Optionally, all or part of the memories 7102 may be located outside the communication device 7100.

[0362] In some embodiments, the communication device 7100 further includes one or more transceivers 7103. When the communication device 7100 includes one or more transceivers 7103, the communication steps such as sending and receiving in the above method are performed by the transceiver 7103, and the other steps are performed by the processor 7101.

[0363] In some embodiments, a transceiver may include a receiver and a transmitter, which may be separate or integrated. Optionally, the terms transceiver, transceiver unit, transceiver, and transceiver circuit may be used interchangeably; the terms transmitter, transmitting unit, transmitter, and transmitting circuit may be used interchangeably; and the terms receiver, receiving unit, receiver, and receiving circuit may be used interchangeably.

[0364] Optionally, the communication device 7100 further includes one or more interface circuits 7104, which are connected to the memory 7102. The interface circuits 7104 may be configured to receive signals from the memory 7102 or other devices, and may be configured to send signals to the memory 7102 or other devices. For example, the interface circuits 7104 may read instructions stored in the memory 7102 and send the instructions to the processor 7101.

[0365] The communication device 7100 described in the above embodiment may be a network device or a terminal, but the scope of the communication device 7100 described in the present disclosure is not limited thereto, and the structure of the communication device 7100 may not be limited by FIG. 7a. The communication device may be an independent device or may be part of a larger device. For example, the communication device may be: 1) an independent integrated circuit IC, or a chip, or a chip system or subsystem; (2) a collection of one or more ICs, optionally, the above IC collection may also include a storage component for storing data or programs; (3) an ASIC, such as a modem; (4) a module that can be embedded in other devices; (5) a receiver, a terminal device, an intelligent terminal device, a cellular phone, a wireless device, a handheld device, a mobile unit, an in-vehicle device, a network device, a cloud device, an artificial intelligence device, etc.; (6) others, etc.

[0366] 7B is a schematic diagram of the structure of a chip 7200 proposed in an embodiment of the present disclosure. If the communication device 7100 can be a chip or a chip system, please refer to the schematic diagram of the structure of the chip 7200 shown in FIG7B , but the present disclosure is not limited thereto.

[0367] The chip 7200 includes one or more processors 7201 , and the processor 7201 is used to call instructions so that the chip 7200 executes any of the above methods.

[0368] In some embodiments, chip 7200 further includes one or more interface circuits 7202, which are connected to memory 7203. Interface circuit 7202 can be used to receive signals from memory 7203 or other devices, and can be used to send signals to memory 7203 or other devices. For example, interface circuit 7202 can read instructions stored in memory 7203 and send the instructions to processor 7201. Optionally, the terms interface circuit, interface, transceiver pin, and transceiver are interchangeable.

[0369] In some embodiments, the chip 7200 further includes one or more memories 7203 for storing instructions. Alternatively, all or part of the memories 7203 may be located outside the chip 7200.

[0370] The present disclosure also proposes a storage medium having instructions stored thereon. When the instructions are executed on the communication device 7100, the communication device 7100 executes any of the above methods. Optionally, the storage medium is an electronic storage medium. Optionally, the storage medium is a computer-readable storage medium, but is not limited thereto and may also be a storage medium readable by other devices. Optionally, the storage medium may be a non-transitory storage medium, but is not limited thereto and may also be a temporary storage medium.

[0371] The present disclosure also provides a program product, which, when executed by the communication device 7100, enables the communication device 7100 to perform any of the above methods. Optionally, the program product is a computer program product.

[0372] The present disclosure also proposes a computer program, which, when executed on a computer, causes the computer to perform any one of the above methods.

[0373] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer programs. When the computer program is loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present disclosure are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer program can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a high-density digital video disc (DVD)), or a semiconductor medium (eg, a solid state disk (SSD)).

[0374] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0375] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0376] The above description is merely a specific embodiment of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this disclosure should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure should be based on the scope of protection of the claims.

Claims

1. A model training method, characterized in that, Performed by a first device, the method includes: Determine a training sample set, the training sample set including at least one set of sample data, the sample data including first information and / or second information; wherein, the first information is channel state information determined by a sensing signal receiving end based on received sensing signals, and the first information is used to determine a sensing result of a sensing target; the second information is actual state information of the sensing target; Train a first model based on the training sample set; wherein, the first model can determine at least one type of sensing result of the sensing target based on the first information.

2. The method according to claim 1, wherein The sensing result of the sensing target includes at least one of the following: The distance of the sensing target determined based on the first information; The speed of the sensing target determined based on the first information; The angle of the sensing target determined based on the first information.

3. The method according to claim 1 or 2, characterized in that, The first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of sensing results.

4. The method according to any one of claims 1 to 3, characterized in that The second information includes at least one of the following: The actual distance of the sensing target; The actual speed of the sensing target; The actual angle of the sensing target.

5. The method according to any one of claims 1-4, characterized in that, The training the first model based on the training sample set includes: Determine at least one type of sensing result that the first model needs to determine; Determine third information based on the first information, the third information being: sub-information of at least one dimension in the first information used to determine the at least one type of sensing result; Input the third information into the first model to obtain an output result of the first model; Calculate a loss function based on the output result of the first model and the second information; Adjust parameters of the first model based on the loss function, input the third information into the adjusted first model to obtain an output result of the first model, and then calculate the loss function of the first model until a first condition is met, at which point it is determined that the training is complete.

6. The method according to claim 5, wherein The determining the third information based on the first information includes: Set sub-information of at least one dimension in the first information that is not used to determine the at least one type of sensing result to a fixed value, and determine the first information after being set to the fixed value as the third information; and / or Extract sub-information of at least one dimension in the first information used to determine the at least one type of sensing result as the third information.

7. The method according to any one of claims 3-5, characterized in that, The at least one dimension includes one or more of the following dimensions: Orthogonal frequency division multiplexing (OFDM) symbol dimension; Subcarrier dimension; Receiving antenna port dimension.

8. The method according to claim 5, wherein The first condition includes at least one of the following: The training cycle of the first model reaches a first threshold; The first model converges.

9. The method according to any one of claims 1-8, characterized in that, The first model can determine one type of sensing result of the sensing target based on the first information, and different first models are used to determine different types of sensing results; or The first model can determine all sensing results of the sensing target based on the first information; Wherein When the first model can determine all sensing results of the sensing target based on the first information, the first information is the same as the third information.

10. A model application method, characterized in that Performed by a second device, the method includes: Determine first information, where the first information is channel state information determined by a sensing signal receiving end based on received sensing signals, and the first information is used to determine a sensing result of a sensing target; The first information is used to determine the sensing result of the sensing target; Determine a first model, where the first model can determine at least one type of sensing result of the sensing target based on the first information; Determine the sensing result of the sensing target based on the first information and the first model.

11. The method according to claim 10, wherein The sensing result of the sensing target includes at least one of the following: The distance of the sensing target determined based on the first information; The speed of the sensing target determined based on the first information; The angle of the sensing target determined based on the first information.

12. The method according to claim 10 or 11, characterized in that, The first information includes sub-information of at least one dimension, and sub-information of different dimensions is used to determine different types of sensing results.

13. The method according to any one of claims 10-12, characterized in that, The determining the sensing result of the sensing target based on the first information and the first model includes: Determine at least one type of sensing result that the first model needs to determine; Determine third information based on the first information, where the third information is: sub-information of at least one dimension in the first information that is used to determine the at least one type of sensing result; Input the third information into the first model to obtain the at least one type of sensing result output by the first model.

14. The method according to claim 13, wherein The determining the third information based on the first information includes: Set sub-information of at least one dimension in the first information that is not used to determine the at least one type of sensing result to a fixed value, and determine the first information after being set to the fixed value as the third information; and / or Extract sub-information of at least one dimension in the first information that is used to determine the at least one type of sensing result as the third information.

15. The method according to any one of claims 12-14, characterized in that The at least one dimension includes one or more of the following dimensions: OFDM symbol dimension; Sub-carrier dimension; Receiving antenna port dimension.

16. The method according to any one of claims 10 to 15, characterized in that, The first model can determine one type of sensing result of the sensing target based on the first information, and different first models are used to determine different types of sensing results; or The first model can determine all sensing results of the sensing target based on the first information; Wherein When the first model can determine all sensing results of the sensing target based on the first information, the first information is the same as the third information.

17. A first device, characterized in that, Includes: ​ ​ 18. A second device, characterized in that, ​ ​ The processing module is further configured to determine a first model, where the first model can determine at least one type of perception result of the perception target based on the first information; The processing module is further configured to determine the perception result of the perception target based on the first information and the first model.

19. A communication device, characterized in that, Comprising: One or more processors; A memory coupled to the processor, where instructions are stored on the memory, and when the instructions are executed by the processor, the Communication device executes the method according to any one of claims 1 to 9 or claims 10 to 16.

20. A communication system, characterized in that, Including a first device and a second device, where the first device is configured to implement the method according to any one of claims 1 to 9, and the second device is configured to implement the method according to any one of claims 10 to 16.

21. A storage medium, the storage medium stores instructions, characterized in that, When the instructions run on the communication device, the communication device is caused to execute the method according to any one of claims 1 to 9 or claims 10 to 16.

Images