Method and device for determining input information of AI model, equipment, medium and product

Abstract

The invention discloses an AI model input information determination method and device, equipment, a medium and a product. The method comprises the following steps: receiving first information; determining measurement starting time according to the first information; reporting second information to second equipment according to the measurement starting time and the first information; wherein the second information is used for generating AI model input information. According to the scheme, the second information used for generating the AI model input information is determined according to the measurement starting time and the first information, so that the second information sent by different first devices meets a certain rule, the AI model input information generated by the second devices based on the second information has consistency, and the performance of the AI model is guaranteed.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a method, apparatus, device, medium and product for determining input information of an AI model. Background Technology

[0002] In research on the integration of Artificial Intelligence (AI) / Machine Learning (ML) with air interface technology, AI technology can be used to solve some complex problems in communication. This is achieved by deploying AI models with corresponding functions in terminal or network-side devices. However, currently, the input information provided by different devices for AI models is inconsistent, resulting in poor performance of the AI ​​models. Summary of the Invention

[0003] This application provides a method, apparatus, device, medium, and product for determining AI model input information, which solves the problem that the lack of consistency in current AI model input information leads to poor AI model performance.

[0004] In a first aspect, embodiments of this application provide a method for determining AI model input information, applied to a first device, the method comprising:

[0005] Receive the first message;

[0006] Based on the first information, determine the measurement start time;

[0007] Based on the measurement start time and the first information, second information is reported to the second device; wherein, the second information is used to generate AI model input information.

[0008] The measurement start time includes at least one of the following:

[0009] The time is determined based on the symbol of the reference signal;

[0010] The time of the sampling point closest to the time of the first path detected by the first device;

[0011] The measurement start time configured for the second device;

[0012] The boundary of the starting symbol where the reference signal detected by the first device is located;

[0013] The time of the first sampling point in the detection path information.

[0014] The first information includes at least one of the following:

[0015] First offset information;

[0016] Reference signal configuration information;

[0017] The first indication information is used to indicate how the measurement start time is determined.

[0018] The determination of the measurement start time based on the first information includes at least one of the following:

[0019] The measurement start time is determined based on the symbol of the reference signal as determined by the reference signal configuration information;

[0020] The measurement start time is determined based on the first offset information;

[0021] Based on the target parameters, target information for determining the measurement start time is selected from the first information, and the measurement start time is determined based on the target information; wherein, the target parameters are parameters provided by the second device, pre-configured parameters, or pre-defined parameters;

[0022] The measurement start time is determined using the measurement start time determination method indicated by the first indication information.

[0023] The determination of the measurement start time using the measurement start time determination method indicated by the first indication information includes at least one of the following:

[0024] When the first indication information indicates that the measurement start time is determined based on the symbol start time, the measurement start time is determined according to the symbol offset and / or time slot offset in the reference signal configuration information;

[0025] When the first indication information indicates that the measurement start time is determined based on the detected first path, path detection is performed, and the measurement start time is determined according to the time of the detected first path;

[0026] When the first indication information indicates that the measurement start time is determined based on the offset information configured by the second device, the measurement start time is determined according to the first offset information;

[0027] When the first indication information indicates that the measurement start time is determined based on the symbol of the reference signal detected by the first device, the measurement start time is determined according to the boundary of the starting symbol of the reference signal detected by the first device.

[0028] When the first indication information indicates that the measurement start time is determined based on the detection path information, the time of the first sampling point in the detection path information is determined as the measurement start time.

[0029] The second information includes at least one of the following:

[0030] The AI ​​model is input with relevant information, wherein, when the relevant information input to the AI ​​model includes time information, the time information is offset time information based on a reference time;

[0031] Timestamp information, which is related to a reference signal;

[0032] Quality indication information.

[0033] The reference time includes at least one of the following:

[0034] Reference time for uplink relative arrival time (UL-RTOA);

[0035] Reference transmission and access point TRP reference time;

[0036] The subframe boundary where the local timing reception reference signal is located.

[0037] Secondly, embodiments of this application provide a method for determining AI model input information, applied to a second device, the method comprising:

[0038] Send first information to the first device, the first information being used to determine the measurement start time;

[0039] The system receives second information sent by the first device. The second information is generated based on the measurement start time and the first information. The second information is used to generate AI model input information.

[0040] The measurement start time includes at least one of the following:

[0041] The time is determined based on the symbol of the reference signal;

[0042] The time of the sampling point closest to the time of the first path detected by the first device;

[0043] The measurement start time configured for the second device;

[0044] The boundary of the starting symbol where the reference signal detected by the first device is located;

[0045] The time of the first sampling point in the detection path information.

[0046] The first information includes at least one of the following:

[0047] First offset information;

[0048] Reference signal configuration information;

[0049] The first indication information is used to indicate how the measurement start time is determined.

[0050] The second information includes at least one of the following:

[0051] The AI ​​model is input with relevant information, wherein, when the relevant information input to the AI ​​model includes time information, the time information is offset time information based on a reference time;

[0052] Timestamp information, which is related to a reference signal;

[0053] Quality indication information.

[0054] The reference time includes at least one of the following:

[0055] UL-RTOA reference time;

[0056] Refer to the reference time of TRP;

[0057] The subframe boundary where the local timing reception reference signal is located.

[0058] Thirdly, embodiments of this application provide an apparatus for determining AI model input information, applied to a first device, the apparatus comprising:

[0059] The receiving module is used to receive the first information;

[0060] The determination module is used to determine the measurement start time based on the first information sent by the second device;

[0061] The sending module is used to report second information to the second device based on the measurement start time and the first information, wherein the second information is used to generate AI model input information.

[0062] Fourthly, embodiments of this application provide an apparatus for determining AI model input information, applied to a second device, the apparatus comprising:

[0063] The sending module is used to send first information to the first device, wherein the first information is used to determine the measurement start time;

[0064] The receiving module is used to receive second information sent by the first device. The second information is generated based on the measurement start time and the first information, and the second information is used to generate AI model input information.

[0065] Fifthly, embodiments of this application provide an AI model input information determination device, comprising: a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the AI ​​model input information determination method as described in the first aspect, or implements the AI ​​model input information determination method as described in the second aspect.

[0066] Sixthly, embodiments of this application provide a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for determining AI model input information as described in the first aspect, or implements the method for determining AI model input information as described in the second aspect.

[0067] In a seventh aspect, embodiments of this application provide a computer program product, including computer instructions, which, when executed by a processor, implement the method for determining AI model input information as described in the first aspect, or implement the method for determining AI model input information as described in the second aspect.

[0068] The beneficial effects of the above-mentioned technical solution of this application are:

[0069] In the method for determining AI model input information according to this application embodiment, firstly, first information is received; secondly, a measurement start time is determined based on the first information; and finally, second information is reported to a second device based on the measurement start time and the first information. The second information is used to generate AI model input information. In this way, the second information used to generate AI model input information can be determined based on the measurement start time and the first information, ensuring that the second information sent by different devices meets certain rules. This guarantees the consistency of the AI ​​model input information generated by the second device based on the second information, thus ensuring the performance of the AI ​​model. Attached Figure Description

[0070] Figure 1 One of the flowcharts illustrating the method for determining input information for an AI model according to an embodiment of this application;

[0071] Figure 2 This is a schematic diagram showing the starting position of the measurement window as the starting position of the symbol in an embodiment of this application;

[0072] Figure 3 This is a schematic diagram showing the time rounding of the path detected by the first device, with the starting position of the measurement window in this embodiment of the application.

[0073] Figure 4 This is a schematic diagram showing the measurement start time within the CP range in the embodiments of this application;

[0074] Figure 5 A second schematic flowchart illustrating the method for determining input information for the AI ​​model in this application embodiment;

[0075] Figure 6 One of the structural schematic diagrams of the device for determining AI model input information in embodiments of this application;

[0076] Figure 7A second schematic diagram of the structure of the device for determining input information of the AI ​​model in an embodiment of this application;

[0077] Figure 8 A schematic diagram of the structure of the device for determining input information for the AI ​​model in this application embodiment. Detailed Implementation

[0078] To make the technical problems, technical solutions, and advantages of this application clearer, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments. In the following description, specific details such as particular configurations and components are provided merely to aid in a comprehensive understanding of the embodiments of this application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Furthermore, for clarity and brevity, descriptions of known functions and structures have been omitted.

[0079] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0080] In the various embodiments of this application, it should be understood that the sequence number of each process described below does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0081] In addition, the terms "system" and "network" are often used interchangeably in this article.

[0082] In the embodiments provided in this application, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0083] The following, with reference to the accompanying drawings, provides a detailed description of the methods, apparatus, devices, media, and products for determining AI model input information according to embodiments of this application.

[0084] It's important to note that in AI-integrated air interface positioning, beam management, Channel State Information (CSI) feedback, and CSI prediction, the input to the AI ​​model needs to be standardized to some extent. When the model input is channel estimation-related information, it can be based on sampling granularity or channel path. When the model input is based on sampling, it's necessary to determine the sampling window start time, the window length for selecting sampling points, and the number of samples for resampling. However, there are currently no unified regulations, and there's no clear solution for how to report this information after it's generated.

[0085] Based on the above-mentioned technologies, embodiments of this application provide a method for determining AI model input information. This method is applied to a first device, such as a User Equipment (UE), NR NodeB (gNB), Transmission Reception Point (TRP), Onboard Unit (OBU), Roadside Unit (RSU), etc. Figure 1 As shown, the method includes:

[0086] Step 101: Receive first information. Specifically, this step involves the first device receiving first information sent by the second device. More specifically, the first information is used to assist in generating the second information described below; therefore, the first information can also be referred to as auxiliary information.

[0087] Step 102: Determine the measurement start time based on the first information. Specifically, the measurement start time is the start time of the measurement window of the first device. Therefore, this step can also be described as: determining a first time based on the first information, where the first time is / is used to indicate the start time of the measurement window of the first device. Additionally, the second information mentioned below carries the measurement results of the (reference) signal; that is, the measurement results of the reference signal detected within the measurement window are used to generate the second information. This step determines the measurement start time based on the first information, ensuring that different first devices perform signal measurements within a measurement window whose start time is determined based on the measurement start time sent by the second device, thus improving the consistency of signal measurements.

[0088] Step 103: Based on the measurement start time and the first information, report second information to the second device; wherein, the second information is used to generate AI model input information. Specifically, this step includes: generating second information based on the measurement start time and the first information, and reporting the second information to the second device. Generating the second information based on the measurement start time and the first information specifically involves, for example: determining the start time of the measurement window based on the measurement start time; determining the measurement window based on the length of the measurement window carried in the first information and the start time of the measurement window; then selecting the sampling information to be reported within the measurement window based on the selection rules and / or predefined selection rules of the channel sampling information carried in the first information, and generating the second information based on the selected sampling information to be reported. That is, the time information included in the channel estimation carried in the second information is based on the time information of the sampling points; furthermore, reporting the second information to the second device can specifically be done by sending the second information to the second device in the same way as the second information carried in the first information.

[0089] The above steps specifically generate second information based on the measurement start time and the first information, so that the generated second information meets certain rules, thereby making the consistency of the AI ​​model input information determined based on the second information relatively good.

[0090] It should be noted that the AI ​​model input information mentioned above can specifically be the input information from at least one of the training, inference, and supervision phases. This ensures that the AI ​​model input information at different stages conforms to the same rule, exhibiting consistency and thus improving the performance of the AI ​​model.

[0091] In the method for determining AI model input information in this application embodiment, firstly, a first device receives first information; secondly, the first device determines a measurement start time based on the first information; obtains a signal measurement result; and finally, the first device reports second information to a second device based on the measurement start time and the first information. The second information is used to generate AI model input information. Thus, the first device can start measuring a reference signal from the measurement start time and obtain second information for generating AI model input information based on the signal measurement result and the first information, ensuring that the second information meets certain rules. Furthermore, the first device sends the second information to the second device, ensuring that the AI ​​model input information generated by the second device based on different second information sent by the first device is consistent, thereby guaranteeing the performance of the AI ​​model.

[0092] As a specific implementation, the measurement start time includes at least one of the following:

[0093] (1) The time determined based on the symbol of the reference signal; here, firstly, the symbol of the reference signal is, for example, the start time of the symbol of the reference signal. Secondly, the symbol of the reference signal is, for example, the start time of the starting symbol of the reference signal. Thirdly, the time determined based on the symbol of the reference signal specifically refers to: a time offset configured based on the reference time, wherein the time offset is the time offset between the symbol of the reference signal and the reference time, that is: configuring the measurement start time based on the time interval between the start time of the symbol of the reference signal and the reference time; wherein, this time interval may be different for different devices (such as TRP). Fourthly, the time determined based on the symbol of the reference signal can be implicitly included in the configuration information of the reference signal, such as being included in the time-frequency resources of the reference signal, which the TRP knows and maintains, and which is at the sampling point of the time-domain channel estimation. That is to say, the measurement start time is obtained according to the position of the symbol of the reference signal, and the measurement start time is at / coincides with the sampling point. Fifthly, the time determined based on the symbol of the reference signal specifically refers to: the measurement start time determined based on the symbol of the reference signal and the length of the cyclic prefix (CP) of the symbol. The measurement start time determined in this step is a reference. Figure 2 .

[0094] (2) The time of the sampling point that is close to the time of the first path detected by the first device; Here, the sampling point that is close to the time of the first path detected is, for example: if the time of the first path detected coincides with the sampling point, then the sampling point that is close to the time of the first path detected is the sampling point that coincides with the time of the first path detected; if the time of the first path detected does not coincide with the sampling point, then the sampling point that is close to the time of the first path detected is: the previous sampling point adjacent to the time of the first path detected or the next sampling point adjacent to the time of the first path detected. Specifically, the one closest to the "time of the first path detected" is selected from "the previous sampling point" and "the next sampling point"; where the time of the first path detected is the reference point, and "the previous" and "the next" correspond to "left" and "right" on the time axis, that is: the previous sampling point is the sampling point that is left adjacent to the time of the first path detected on the time axis, and the next sampling point is the sampling point that is right adjacent to the time of the first path detected on the time axis.

[0095] Based on this, "the time of the sampling point closest to the time of the first path detected by the first device" can also be described as: the time obtained by rounding down the time of the first path of the wireless channel detected by the first device. Here, "rounding down" means: if the time of the detected path is not at a sampling point, then the time is rounded down to the nearest (previous) or (next) sampling point; that is, the time of the nearest sampling point is used as the measurement start time. Specifically, a diagram illustrating the rounding down of the time of the path detected by the first device as the measurement start time (the start time of the measurement window) is shown below. Figure 3 As shown.

[0096] In addition, the aforementioned "first path" (or wireless channel first path) refers to the information of the wireless channel through which the reference signal propagates after the receiving end (first device) receives the reference signal. The reference signal may propagate through multiple wireless channels, and the first path is the first path detected in time.

[0097] (3) The measurement start time configured for the second device; here, the second device can configure a fixed offset time or a dynamic offset time based on the reference time. That is, the measurement start time is determined based on the reference time and the fixed or dynamic offset time configured by the second device. Specifically, the second device can configure the measurement start time (the start time of the measurement window) according to scene information, such as the distance between the transmitting and receiving reference signal devices in the scene. In addition, the second device can configure the same measurement start time for different first devices (such as TRP) or configure different start times for different first devices.

[0098] (4) The boundary of the starting symbol where the reference signal detected by the first device is located.

[0099] (5) Detect the time of the first sampling point in the path information.

[0100] Here, it should be noted that, in the specific implementation methods described above, firstly, the reference signals include, for example, a downlink positioning reference signal (DL-PRS), an uplink sounding reference signal (UL-SRS), a sidelink positioning reference signal (SL-PRS), a channel sounding signal (SRS-pos) used for positioning, or a channel state information reference signal (CSI-RS). Secondly, the reference time includes, for example, at least one of the following: a reference time (T0+t) of the uplink relative time of arrival (UL-RTOA). SRS ), and the reference time of TRP (T subframeRexi The third method for determining the start time of the measurement window can be configuration-based, taking the measurement method "time determined based on the symbol where the reference signal is located" and / or "time of the sampling point near the time of the first path detected by the first device" as an example. For example, it can be configured based on the information carried in the first information, or implicitly configured according to other parameters, where "other parameters" are, for example, "target parameters" in the following text.

[0101] Based on the specific implementation method described above, further, the first device starts at the configured measurement start time, within the measurement window (e.g., Figure 2 or Figure 3 The detection is performed within the Nt window (in the image), and the Nt with the strongest power is selected. ’ Each sampling point (e.g.) Figure 2 or Figure 3 (The three sampling points indicated by the dashed arrows in the Nt window corresponding to TRP1) / the sampling point with the highest power and the Nt adjacent to the sampling point with the highest power. ’ -1 sampling points (e.g.) Figure 2 or Figure 3 The information of the three sampling points (indicated by the dashed arrow in the Nt window corresponding to TRP2) is reported. Here, "reporting" means reporting in the second information.

[0102] As an optional implementation, the first information includes at least one of the following:

[0103] First offset information; here, the first offset information is, for example, the time offset configured by the second device based on the reference time, wherein the time offset is a fixed offset time or a dynamic offset time configured by the second device.

[0104] Reference signal configuration information; for example, the reference signal configuration information includes at least one of the following: reference signal resource information, reference signal identification (ID) information, and reference signal periodic information. As mentioned above, the reference signal here includes, for example, DL PRS, UL-SRS, SL PRS, SRS-pos, or CSI-RS.

[0105] The first indication information is used to indicate how the measurement start time is determined.

[0106] It should be noted that, based on one or more of the first information, including first offset information, reference signal configuration information, and first indication information, the first device may determine the measurement start time based on the first offset information, or based on the reference signal configuration information, or based on the first indication information and the first offset information, or based on the first indication information and the reference signal configuration information, etc. The process of determining the measurement start time based on the first information will be explained in detail later.

[0107] Here, we will illustrate the implicit determination of the measurement start time based on the reference signal configuration information with an example:

[0108] This is determined based on the configuration in NR-DL-PRS-Info; the configuration in NR-DL-PRS-Info is as follows:

[0109] The downlink positioning reference signal resource slot offset (dl-PRS-Resource SlotOffset) field specifies the starting slot of the DL-PRS Resource with respect to the corresponding DL-PRS-Resource Set Slot Offset.

[0110] The downlink positioning reference signal resource symbol offset (dl-PRS-Resource SlotOffset) field specifies the starting symbol of the DL-PRS resource within a slot determined by dl-PRS-ResourceSlotOffset.

[0111] The configuration is determined based on the positioning reference signal resource (Positioning SRS Resource) used for positioning, as shown in Table 1 below:

[0112] Table 1

[0113]

[0114] Here, INTEGER represents the integer data type, and ENUMERATED represents the enumeration data type.

[0115] Furthermore, based on the above-mentioned optional implementation methods, the first information may also include at least one of the following:

[0116] The format requirements for the second information; these format requirements are used to indicate the specific format of the second information, for example, indicating whether the second information is information generated based on sampling or information generated based on a path.

[0117] The second indication information is used to indicate the content carried by the second information. For example, the second indication information may indicate that the content carried by the second information includes at least one of time information (specifically, time information based on a reference time), power information, and phase information. Specifically, for example, the channel impulse response (CIR) carried by the second information may include time information, power information, and phase information; the power delay profile (PDP) may include time information and phase information; and the delay profile (DP) may include time information. In other words, the second indication information can specifically be used to indicate the time, power, and / or phase information included in each parameter carried in the second information (parameters obtained based on measurements of the reference signal).

[0118] The length of the measurement window; here, the length of the measurement window can be determined by the second device based on the specific scenario information, and the length of the measurement window will not affect the overhead of reporting the second information.

[0119] The selection rules for channel sampling information (the rules for selecting the reported sampling points from the sampling points); specifically, these selection rules include: within an Nt window (a measurement window of length Nt), selecting the Nt with the highest power. ’ Each sampling information, to Figure 2 or Figure 3 Taking the Nt window corresponding to TRP1 as an example, the information of the three sampling points with the strongest power (the sampling points represented by the dashed arrows in the Nt window) is selected for reporting, and / or, within the Nt window, the sampling information with the highest power and the Nt windows adjacent to the sampling information with the highest power are selected. ’ -1 sampling information; where, the Nt ’ -1 sample information is located on the same side as the sample information with the highest power (within) Figure 2 or Figure 3 Taking the Nt window corresponding to TRP2 as an example, select the sampling point with the highest power and the two sampling points to its right (the sampling points represented by the dashed arrows in the Nt window corresponding to TRP2) or both sides.

[0120] The number of channel sampling information reported, as mentioned above, is Nt. ’ .

[0121] The second piece of information is the time granularity; for example, the sampling time can be T=2. k ×Tc, where Tc is a reference parameter determined in the protocol, and k can be configured by the second device, for example, an integer belonging to [6, 6], but not limited thereto; wherein, the time granularity information includes at least one of the following: the sampling time T (1 / (system FFT points * subcarrier spacing)) corresponding to the system bandwidth, the time granularity configured according to the capability of the first device (for example, it can be less than T), and the time granularity configured according to the accuracy requirements of the AI ​​model of the second device (for example, it can be greater than T).

[0122] The second type of information is related to quality; for example, this quality-related information may be the Reference Signal Received Power (RSRP), the Signal to Noise Ratio (SNR), etc.

[0123] Reference time; here, as mentioned above, the reference time includes, for example, at least one of the following: the reference time of UL-RTOA, the reference time of the reference TRP, and the subframe boundary where the local timing reception reference signal is located.

[0124] The reporting method for the second information can be, for example, periodic reporting or non-periodic reporting.

[0125] As an optional implementation, step 102 includes at least one of the following:

[0126] (1) Determine the measurement start time based on the reference signal symbol determined by the reference signal configuration information; that is, obtain the measurement start time based on the position of the reference signal symbol, wherein the obtained measurement start time coincides with the time corresponding to the sampling point. For example, this step is as follows: First, determine the reference signal symbol based on the reference signal configuration information in the first information; second, determine the measurement start time based on the start time of the reference signal symbol, or determine the measurement start time based on the start time of the reference signal symbol and the first offset.

[0127] The first offset can be determined based on the time advance (TA), that is, the measurement start time is determined based on the first offset determined by the TA and the start time of the symbol where the reference signal is located. Specifically, the measurement start time can be obtained by shifting the start time of the symbol where the reference signal is located to the left or right by the first offset. The obtained measurement start time coincides with the time corresponding to the sampling point.

[0128] And / or, the first offset can be determined based on the CP, that is: the measurement start time is determined based on the symbol where the reference signal is located and the length of the CP. For example, the measurement start time is the start time of the symbol where the reference signal is located ± the first offset (the measurement start time is the start time of the symbol where the reference signal is located shifted forward by the first offset, or the measurement start time is the start time of the symbol where the reference signal is located shifted backward by the first offset), and the first offset is 0 to the length of the CP (e.g., Figure 4 (As shown). The obtained measurement start time coincides with the time corresponding to the sampling point.

[0129] (2) Determine the measurement start time based on the first offset information; that is, this step is: determine the first offset information carried in the first information as the measurement start time, wherein the first offset information is offset information relative to the reference time.

[0130] (3) Based on the target parameters, select the target information for determining the measurement start time from the first information, and determine the measurement start time based on the target information; wherein, the target parameters are parameters provided by the second device, pre-configured parameters, or pre-defined parameters.

[0131] In the above steps, the target parameters include, for example, the number of samples (Nt) included in the measurement window and the number of samples reported (Nt). ’ ), the number of TRPs for transmitting and receiving reference signals, etc. For example, when supporting the determination of the measurement start time using the above two methods (1) and (2), the implementation examples of the above step (3) include, but are not limited to:

[0132] Nt ’ If the value is greater than or equal to the first value, the measurement start time shall be determined by (1) above; otherwise, the measurement start time shall be determined by (2) above.

[0133] Nt is greater than or equal to the second value, and / or, Nt ’ If the value is greater than or equal to the third value, the measurement start time shall be determined by (1) above; otherwise, the measurement start time shall be determined by (2) above.

[0134] The number of TRPs is greater than the fourth value, and / or, Nt ’ If the value is greater than the fifth value, the measurement start time shall be determined by (1) above; otherwise, the measurement start time shall be determined by (2) above.

[0135] If the number of TRPs is greater than the sixth value and / or Nt is greater than the seventh value, the measurement start time shall be determined by (1) above; otherwise, the measurement start time shall be determined by (2) above.

[0136] (4) Determine the measurement start time using the measurement start time determination method indicated by the first indication information.

[0137] As a specific implementation, the measurement start time is determined using the measurement start time determination method indicated by the first indication information, including at least one of the following:

[0138] When the first indication information indicates that the measurement start time is determined based on the symbol start time, the measurement start time is determined according to the symbol offset and / or slot offset in the reference signal configuration information;

[0139] When the first indication information indicates that the measurement start time is determined based on the detected first path, path detection is performed, and the measurement start time is determined according to the time of the detected first path; wherein, the measurement start time is the time of the sampling point adjacent to the time of the first path detection (in the time domain, the sampling point where the time of the first path detection is located, and the sampling point before / after the time of the first path detection).

[0140] When the first indication information indicates that the measurement start time is determined based on the offset information configured by the second device, the measurement start time is determined according to the first offset information; wherein, the first offset time is a fixed offset time or / dynamic offset time configured by the second device based on a reference time;

[0141] When the first indication information indicates that the measurement start time is determined based on the symbol of the reference signal detected by the first device, the measurement start time is determined according to the boundary of the starting symbol of the reference signal detected by the first device.

[0142] When the first indication information indicates that the measurement start time is determined based on the detection path information, the time of the first sampling point in the detection path information is determined as the measurement start time.

[0143] As an optional implementation, the second information includes at least one of the following:

[0144] (1) AI model input related information, wherein, when the AI ​​model input related information includes time information, the time information is offset time information based on a reference time, such as: reference time + time offset. Specifically, firstly, the reference time includes, for example, at least one of the reference time of UL-RTOA, the reference time of the reference TRP, and the subframe boundary where the local timing reception reference signal is located; secondly, the time corresponding to this time information is at the sampling time point, and the sampling time granularity can be expressed as: T = 2. k ×Tc.

[0145] Here, the AI ​​model input information includes at least one of time information, power information, and phase information. Specifically, the AI ​​model input information includes CIR (including time information, power information, and phase information), PDP (including time information and phase information), and DP (including time information).

[0146] It should be noted that since the above time information is based on the time offset information of the reference time, the second information may also include the above reference time.

[0147] In addition, the relevant information input to the AI ​​model can be reported based on IE or its enhancements, such as the time difference between the receiver and transmitter of UL-RTOA and gNB (Rx-Tx Time difference), downlink reference signal time difference (DL-RSTD), time difference between the receiver and transmitter of UE (Rx-Tx Time difference), and reference signal received path power (RSRPP).

[0148] (2) Timestamp information, which is related to a reference signal; here, as mentioned above, the reference signal is, for example, DL PRS, UL-SRS, SL PRS, SRS-pos, or CSI-RS, etc., wherein the second information is obtained based on the measurement of the reference signal by the first device. Additionally, the timestamp information is, for example, the time at which the reference signal was measured; specifically, the timestamp information can be represented by the frame number and slot number of the reference signal.

[0149] (3) Quality indication information; for example, the quality indication information includes SNR and / or RSRP, etc.

[0150] Embodiments of this application also provide a method for determining AI model input information. This method is applied to a second device, such as a Location Management Function (LMF) network element / physical device. Figure 5 As shown, the method includes:

[0151] Step 501: Send first information to the first device, the first information being used to determine the measurement start time; here, the measurement start time is the start time of the measurement window of the first device, therefore, the measurement start time can also be described as "the first time, the first time being / used to indicate the start time of the measurement window of the first device".

[0152] Step 502: Receive second information sent by the first device. The second information is generated based on the measurement start time and the first information, and is used to generate AI model input information. Specifically, the AI ​​model input information can be AI model input information from at least one of the training, inference, and supervision phases.

[0153] It should be noted here that, as mentioned earlier, the second information is generated based on the sampling points selected from the sampling points that need to be reported. Here, the "sampling point" is obtained by sampling the measurement results of the reference signal detected by the first device within the measurement window; that is, the time information contained in the channel estimation carried in the second information is based on the time information of the sampling points.

[0154] In the method for determining AI model input information in this application embodiment, firstly, a second device sends first information to a first device, the first information being used to determine the measurement start time; so that the first device obtains second information for generating AI model input information based on the first information and the measurement start time, such that the second information can satisfy certain rules; secondly, the second device receives the second information sent by the first device, and generates AI model input information based on the second information, so that the generated AI model input information (which can be AI model input information from at least one of the training phase, inference phase, and supervision phase) all satisfy the same rules, that is, have good consistency, thereby helping to improve the performance of the AI ​​model.

[0155] As a specific implementation, the measurement start time includes at least one of the following:

[0156] The time determined based on the symbol of the reference signal; here, firstly, the symbol of the reference signal is, for example, the start time of the symbol of the reference signal. Secondly, the symbol of the reference signal is, for example, the start time of the starting symbol of the reference signal. Thirdly, the time determined based on the symbol of the reference signal specifically refers to: a time offset configured based on the reference time, wherein the time offset is the time offset between the symbol of the reference signal and the reference time, that is: configuring the measurement start time based on the time interval between the start time of the symbol of the reference signal and the reference time; wherein, this time interval may be different for different devices (such as TRP). Fourthly, the time determined based on the symbol of the reference signal can be implicitly included in the configuration information of the reference signal, such as being included in the time-frequency resources of the reference signal, which the TRP knows and maintains, and which time is at the sampling point of the time-domain channel estimation. That is to say, the measurement start time is obtained according to the position of the symbol of the reference signal, and the measurement start time is at / coincides with the sampling point. Fifthly, the time determined based on the symbol of the reference signal specifically refers to: the measurement start time determined based on the symbol of the reference signal and the length of the cyclic prefix (CP) of the symbol. The measurement start time determined in this step is a reference. Figure 2 .

[0157] The time of the sampling point adjacent to the time of the first path detected by the first device; here, the sampling point adjacent to the time of the detected first path is, for example, the time of the sampling point whose time overlaps with the time of the detected first path, the time of the previous sampling point adjacent to the time of the detected first path, or the time of the next sampling point adjacent to the time of the detected first path. Based on this, "the time of the sampling point adjacent to the time of the first path detected by the first device" can also be described as: the time obtained by rounding down the time of the first path of the wireless channel detected by the first device, where "rounding down" means: if the time of the detected path is not at a sampling point, then rounding down to the (previous) or (next) nearest sampling point, that is: using the nearest (closest) sampling point as the starting time. Specifically, a schematic diagram of the measurement start time (the start time of the measurement window) being the rounded time of the path detected by the first device is shown in the figure. Figure 3 As shown. Additionally, the aforementioned "first path" refers to the first path detected by the first device in time.

[0158] The measurement start time configured for the second device; here, the second device can configure a fixed offset time or a dynamic offset time based on the reference time. That is, the measurement start time is determined based on the reference time and the fixed or dynamic offset time configured by the second device. Specifically, the second device can configure this measurement start time (the start time of the measurement window) according to scene information, such as the distance between the transmitting and receiving reference signal devices in the scene. Furthermore, the second device can configure the same measurement start time for different first devices (such as TRPs) or configure different start times for different first devices.

[0159] The boundary of the starting symbol where the reference signal detected by the first device is located;

[0160] The time of the first sampling point in the detection path information.

[0161] As an optional implementation, the first information includes at least one of the following:

[0162] First offset information; here, the first offset information is, for example, the time offset configured by the second device based on the reference time, wherein the time offset is a fixed offset time or a dynamic offset time configured by the second device.

[0163] Reference signal configuration information; for example, the reference signal configuration information includes one or more of the following: reference signal resource information, reference signal identification (ID) information, and reference signal periodic information. As mentioned above, the reference signal here includes, for example, DL PRS, UL-SRS, SL PRS, SRS-pos, or CSI-RS.

[0164] The first indication information is used to indicate how the measurement start time is determined.

[0165] Furthermore, based on the above-mentioned optional implementation methods, the first information may also include at least one of the following:

[0166] The format requirements for the second information; these format requirements are used to indicate the specific format of the second information, for example, indicating whether the second information is information generated based on sampling or information generated based on a path.

[0167] The second indication information is used to indicate the content carried by the second information. For example, the second indication information may indicate that the content carried by the second information includes at least one of time information (specifically, time information based on a reference time), power information, and phase information. Specifically, for example, the channel impulse response (CIR) carried by the second information may include time information, power information, and phase information; the power delay profile (PDP) may include time information and phase information; and the delay profile (DP) may include time information. In other words, the second indication information can specifically be used to indicate the time, power, and / or phase information included in each parameter carried in the second information (parameters obtained based on measurements of the reference signal).

[0168] The length of the measurement window; here, the length of the measurement window can be determined by the second device based on the specific scenario information, and the length of the measurement window will not affect the overhead of reporting the second information.

[0169] The selection rules for channel sampling information (the rules for selecting the reported sampling points from the sampling points); specifically, these selection rules include: within an Nt window (a window containing Nt sampling points), selecting the Nt with the highest power. ’ Each sampling information, to Figure 2 or Figure 3 Taking TRP1 as an example, the information of the three sampling points with the strongest power (represented by the dashed arrows in the Nt window) is selected for reporting, and / or, within the Nt window (a window containing Nt sampling points), the sampling information with the highest power and the Nt adjacent to the sampling information with the highest power are selected. ’ -1 sampling information; where, the Nt ’ -1 sample information is located on the same side as the sample information with the highest power (within) Figure 2 or Figure 3 Taking TRP2 as an example, select the sampling point with the highest power and the two sampling points to its right (the sampling points represented by the dashed arrows in the Nt window corresponding to TRP2) or both sides.

[0170] The number of channel sampling information reported.

[0171] The second piece of information is the time granularity; for example, the sampling time can be T=2. k ×Tc; wherein the time granularity information includes at least one of the following: the sampling time T corresponding to the system bandwidth, the time granularity configured according to the capability of the first device (e.g., it can be less than T), and the time granularity configured according to the accuracy requirements of the AI ​​model by the second device (e.g., it can be greater than T).

[0172] The second piece of information is quality-related; for example, this quality-related information could be RSRP, SNR, etc.

[0173] Reference time; here, as mentioned above, the reference time includes, for example, at least one of the following: the reference time of UL-RTOA, the reference time of the reference TRP, and the subframe boundary where the local timing reception reference signal is located.

[0174] The reporting method for the second information can be, for example, periodic reporting or non-periodic reporting.

[0175] As an optional implementation, the second information includes at least one of the following:

[0176] The AI ​​model inputs relevant information, wherein, when the relevant information includes time information, the time information is offset time information based on a reference time, such as: reference time + time offset; wherein, firstly, the reference time includes at least one of the reference time of UL-RTOA, the reference time of reference TRP, and the subframe boundary where the local timing reception reference signal is located; secondly, the time corresponding to this time information is at the sampling time point, and the sampling time granularity can be expressed as: T = 2 k ×Tc.

[0177] Here, the AI ​​model input information includes at least one of time information, power information, and phase information. Specifically, the AI ​​model input information includes CIR (including time information, power information, and phase information), PDP (including time information and phase information), and DP (including time information).

[0178] It should be noted that since the above time information is based on the time offset information of the reference time, the second information may also include the above reference time.

[0179] In addition, relevant information input to the AI ​​model can be reported based on UL-RTOA, gNB's Rx-Tx Time difference, DL-RSTD, UE's Rx-Tx Time difference, RSRPP and other IEs or their enhancements.

[0180] The timestamp information is related to a reference signal; here, as mentioned above, the reference signal is, for example, DL PRS, UL-SRS, SL PRS, SRS-pos, or CSI-RS, etc., wherein the second information is obtained based on the measurement of the reference signal by the first device. Furthermore, the timestamp information is, for example, the time at which the reference signal was measured; specifically, the timestamp information can be represented by the frame number and slot number of the reference signal.

[0181] Quality indication information. For example, the quality indication information includes SNR and / or RSRP, etc.

[0182] As a specific implementation, the reference time includes at least one of the following:

[0183] UL-RTOA reference time;

[0184] Refer to the reference time of TRP;

[0185] The subframe boundary where the local timing reception reference signal is located.

[0186] It should be noted here that, as mentioned earlier, after receiving the second information, the second device will further obtain the AI ​​model input information based on the second information. The dimensionality of the AI ​​model input information depends on the second device; for example, the dimensionality of the AI ​​model input information may differ from the length (size Nt) of the measurement window configured for the first device. Specifically, for example:

[0187] Taking the first device as TRP and the second device as LMF as an example, the interpretation of Nt on the TRP and LMF sides may be inconsistent. Assuming Nt = 256 and Nt' = 32, the TRP detects and reports 32 sampled information points within a window length of 256 to the LMF. However, when the LMF obtains the input information for the AI ​​model based on these 32 sampled information points, the dimension of the AI ​​model's input information may be equal to or not equal to Nt. For example, the LMF's input may be related to Nt', depending on the LMF's implementation. It can be seen that the window length Nt configured by the LMF for the TRP does not affect the overhead of reporting Nt', nor does it affect the size of the AI ​​model's input obtained by the LMF. Therefore, when configuring Nt for the TRP, in order to cover more information, the LMF can configure the window length Nt as large as possible according to the size of the application scenario. For example, the window length could be the distance d from the UE to the TRP in the application scenario / time granularity T.

[0188] Furthermore, when the LMF receives AI model input information, it can adjust the timing information reported by different TRPs, such as performing consistency processing on the symbol / slot offsets of different TRPs. And / or, based on the synchronization information known to the LMF, it can adjust the timing information reported by different TRPs. This results in better consistency of the AI ​​model input received by the LMF.

[0189] Embodiments of this application provide an apparatus for determining AI model input information, applied to a first device, such as... Figure 6 As shown, the device includes:

[0190] Receiver module 601 is used to receive the first information;

[0191] The determining module 602 is used to determine the measurement start time based on the first information sent by the second device;

[0192] The sending module 603 is used to report second information to the second device based on the measurement start time and the first information, wherein the second information is used to generate AI model input information.

[0193] The measurement start time includes at least one of the following:

[0194] The time is determined based on the symbol of the reference signal;

[0195] The time of the sampling point closest to the time of the first path detected by the first device;

[0196] The measurement start time configured for the second device;

[0197] The boundary of the starting symbol where the reference signal detected by the first device is located;

[0198] The time of the first sampling point in the detection path information.

[0199] The first information includes at least one of the following:

[0200] First offset information;

[0201] Reference signal configuration information;

[0202] The first indication information is used to indicate how the measurement start time is determined.

[0203] The determining module 602 includes at least one of the following:

[0204] The first determining submodule is used to determine the measurement start time based on the symbol of the reference signal determined by the reference signal configuration information;

[0205] The second determining submodule is used to determine the measurement start time based on the first offset information;

[0206] The third determining submodule is used to select target information for determining the measurement start time from the first information according to the target parameters, and determine the measurement start time based on the target information; wherein, the target parameters are parameters provided by the second device, pre-configured parameters, or predefined parameters;

[0207] The fourth determining submodule is used to determine the measurement start time using the measurement start time determination method indicated by the first indication information.

[0208] The fourth determining submodule includes:

[0209] The first determining unit is configured to determine the measurement start time based on the symbol offset and / or time slot offset in the reference signal configuration information when the first indication information indicates that the measurement start time is determined based on the symbol start time.

[0210] The second determining unit is configured to perform path detection and determine the measurement start time based on the time of the detected first path when the first indication information indicates that the measurement start time is determined based on the first path detected.

[0211] The third determining unit is configured to determine the measurement start time based on the first offset information when the first indication information indicates that the measurement start time is determined based on the offset information configured by the second device.

[0212] The fourth determining unit is used to determine the measurement start time based on the boundary of the starting symbol of the reference signal detected by the first device when the first indication information indicates that the measurement start time is determined based on the symbol of the reference signal detected by the first device.

[0213] The fifth determining unit is used to determine the time of the first sampling point in the detection path information as the measurement start time when the first indication information indicates that the measurement start time is determined based on the detection path information.

[0214] The second information includes at least one of the following:

[0215] The AI ​​model is input with relevant information, wherein, when the relevant information input to the AI ​​model includes time information, the time information is offset time information based on a reference time;

[0216] Timestamp information, which is related to a reference signal;

[0217] Quality indication information.

[0218] The reference time includes at least one of the following:

[0219] UL-RTOA reference time;

[0220] Refer to the reference time of TRP;

[0221] The subframe boundary where the local timing reception reference signal is located.

[0222] It should be noted that the AI ​​model input information determination device provided in this application embodiment can implement all the method steps implemented in the above-mentioned AI model input information determination method embodiment applied to the first device, and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0223] Embodiments of this application provide an apparatus for determining AI model input information, applied to a second device, such as... Figure 7 As shown, the device includes:

[0224] The sending module 701 is used to send first information to the first device, the first information being used to determine the measurement start time;

[0225] The receiving module 702 is used to receive second information sent by the first device. The second information is generated based on the measurement start time and the first information, and the second information is used to generate AI model input information.

[0226] The measurement start time includes at least one of the following:

[0227] The time is determined based on the symbol of the reference signal;

[0228] The time of the sampling point closest to the time of the first path detected by the first device;

[0229] The measurement start time configured for the second device;

[0230] The boundary of the starting symbol where the reference signal detected by the first device is located;

[0231] The time of the first sampling point in the detection path information.

[0232] The first information includes at least one of the following:

[0233] First offset information;

[0234] Reference signal configuration information;

[0235] The first indication information is used to indicate how the measurement start time is determined.

[0236] The second information includes at least one of the following:

[0237] The AI ​​model is input with relevant information, wherein, when the relevant information input to the AI ​​model includes time information, the time information is offset time information based on a reference time;

[0238] Timestamp information, which is related to a reference signal;

[0239] Quality indication information.

[0240] The reference time includes at least one of the following:

[0241] UL-RTOA reference time;

[0242] Refer to the reference time of TRP;

[0243] The subframe boundary where the local timing reception reference signal is located.

[0244] It should be noted that the AI ​​model input information determination device provided in this application embodiment can implement all the method steps implemented in the above-mentioned AI model input information determination method embodiment applied to the second device, and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0245] Embodiments of this application also provide an AI model input information determination device, including a transceiver 810, a processor 800, a memory 820, and a program stored in the memory 820 and executable on the processor 800; wherein, when the processor 800 executes the program, it implements the AI ​​model input information determination method applied to a first device as described above, or implements the AI ​​model input information determination method applied to a second device as described above.

[0246] The transceiver 810 is used to receive and send data under the control of the processor 800.

[0247] Among them, Figure 8 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 800) and memory (memory 820). The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 810 can be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium.

[0248] The processor 800 is responsible for managing the bus architecture and general processing, while the memory 820 can store the data used by the processor 800 during operation.

[0249] It should be noted that the AI ​​model input information determination device provided in this application embodiment can implement all the method steps implemented in the above-mentioned method embodiment for determining AI model input information applied to the first device or the second device, and can achieve the same technical effect. Therefore, the parts that are the same as those in the method embodiment and the beneficial effects will not be described again here.

[0250] Embodiments of this application also provide a readable storage medium storing a program. When executed by a processor, this program implements the various processes of the above-described embodiment of the method for determining AI model input information applied to a first device, or implements the various processes of the above-described embodiment of the method for determining AI model input information applied to a second device, achieving the same technical effect. To avoid repetition, further details are omitted here. The readable storage medium may be, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0251] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, disk, optical disk) and includes several instructions for executing the methods described in the various embodiments of this application.

[0252] Therefore, embodiments of this application also provide a computer program product, including computer instructions, which, when executed by a processor, implement the steps in the method for determining AI model input information applied to a first device as described above, or implement the steps in the method for determining AI model input information applied to a second device as described above, and can achieve the same technical effect. To avoid repetition, further details are omitted here.

[0253] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0254] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for determining input information of an AI model, characterized in that, Applied to a first device, the method includes: Receive the first message; Based on the first information, determine the measurement start time; Based on the measurement start time and the first information, second information is reported to the second device; wherein, the second information is used to generate AI model input information.

2. The method according to claim 1, characterized in that, The measurement start time includes at least one of the following: The time is determined based on the symbol of the reference signal; The time of the sampling point closest to the time of the first path detected by the first device; The measurement start time configured for the second device; The boundary of the starting symbol where the reference signal detected by the first device is located; The time of the first sampling point in the detection path information.

3. The method according to claim 1, characterized in that, The first information includes at least one of the following: First offset information; Reference signal configuration information; The first indication information is used to indicate how the measurement start time is determined.

4. The method according to claim 3, characterized in that, Based on the first information, the measurement start time is determined, including at least one of the following: The measurement start time is determined based on the symbol of the reference signal as determined by the reference signal configuration information; The measurement start time is determined based on the first offset information; Based on the target parameters, target information for determining the measurement start time is selected from the first information, and the measurement start time is determined based on the target information; wherein, the target parameters are parameters provided by the second device, pre-configured parameters, or pre-defined parameters; The measurement start time is determined using the measurement start time determination method indicated by the first indication information.

5. The method according to claim 4, characterized in that, The measurement start time is determined using the measurement start time determination method indicated by the first indication information, including at least one of the following: When the first indication information indicates that the measurement start time is determined based on the symbol start time, the measurement start time is determined according to the symbol offset and / or time slot offset in the reference signal configuration information; When the first indication information indicates that the measurement start time is determined based on the detected first path, path detection is performed, and the measurement start time is determined according to the time of the detected first path; When the first indication information indicates that the measurement start time is determined based on the offset information configured by the second device, the measurement start time is determined according to the first offset information; When the first indication information indicates that the measurement start time is determined based on the symbol of the reference signal detected by the first device, the measurement start time is determined according to the boundary of the starting symbol of the reference signal detected by the first device. When the first indication information indicates that the measurement start time is determined based on the detection path information, the time of the first sampling point in the detection path information is determined as the measurement start time.

6. The method according to claim 1, characterized in that, The second information includes at least one of the following: The AI ​​model is input with relevant information, wherein, when the relevant information input to the AI ​​model includes time information, the time information is offset time information based on a reference time; Timestamp information, which is related to a reference signal; Quality indication information.

7. The method according to claim 6, characterized in that, The reference time includes at least one of the following: Reference time for uplink relative arrival time (UL-RTOA); Reference transmission and access point TRP reference time; The subframe boundary where the local timing reception reference signal is located.

8. A method for determining input information of an AI model, characterized in that, Applied to a second device, the method includes: Send first information to the first device, the first information being used to determine the measurement start time; The system receives second information sent by the first device. The second information is generated based on the measurement start time and the first information. The second information is used to generate AI model input information.

9. The method according to claim 8, characterized in that, The measurement start time includes at least one of the following: The time is determined based on the symbol of the reference signal; The time of the sampling point closest to the time of the first path detected by the first device; The measurement start time configured for the second device; The boundary of the starting symbol where the reference signal detected by the first device is located; The time of the first sampling point in the detection path information.

10. The method according to claim 8, characterized in that, The first information includes at least one of the following: First offset information; Reference signal configuration information; The first indication information is used to indicate how the measurement start time is determined.

11. The method according to claim 8, characterized in that, The second information includes at least one of the following: The AI ​​model is input with relevant information, wherein, when the relevant information input to the AI ​​model includes time information, the time information is offset time information based on a reference time; Timestamp information, which is related to a reference signal; Quality indication information.

12. The method according to claim 11, characterized in that, The reference time includes at least one of the following: UL-RTOA reference time; Refer to the reference time of TRP; The subframe boundary where the local timing reception reference signal is located.

13. A device for determining input information of an AI model, characterized in that, Applied to a first device, the device includes: The receiving module is used to receive the first information; The determination module is used to determine the measurement start time based on the first information sent by the second device; The sending module is used to report second information to the second device based on the measurement start time and the first information, wherein the second information is used to generate AI model input information.

14. A device for determining input information of an AI model, characterized in that, Applied to a second device, the device includes: The sending module is used to send first information to the first device, wherein the first information is used to determine the measurement start time; The receiving module is used to receive second information sent by the first device. The second information is generated based on the measurement start time and the first information, and the second information is used to generate AI model input information.

15. A device for determining input information of an AI model, comprising a transceiver, a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the method for determining AI model input information as described in any one of claims 1 to 6, or implements the method for determining AI model input information as described in any one of claims 7 to 12.

16. A readable storage medium having a program or instructions stored thereon, characterized in that, When the program or instructions are executed by the processor, they implement the method for determining AI model input information as described in any one of claims 1 to 6, or implement the method for determining AI model input information as described in any one of claims 7 to 12.

17. A computer program product, characterized in that, The method includes computer instructions that, when executed by a processor, implement the method for determining AI model input information as described in any one of claims 1 to 6, or implement the method for determining AI model input information as described in any one of claims 7 to 12.

Images