Detection method of pdcch, drx configuration method, terminal, base station

Abstract

Provided are a PDCCH detection method, a DRX configuration method, a terminal, and a base station. The detection method can include: obtaining input data required for predicting aggregation level (AL) related information; predicting the AL related information using an AL prediction model based on the obtained input data; and performing physical downlink control channel (PDCCH) detection based on the predicted AL related information. According to the present disclosure, blind detection of the PDCCH can be reduced or avoided, and power consumption of the terminal can be reduced.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the wireless communication technology, and more particularly to a PDCCH detection method, a DRX configuration method, a terminal and a base station. BACKGROUND

[0002] In a Long Term Evolution (LTE) or New Radio (NR) system, a Physical Downlink Control Channel (PDCCH) carries control information, which can include but is not limited to: transmission format, resource allocation information, uplink scheduling grant, power control information, and uplink retransmission information. According to the link direction, the information carried by the PDCCH can include uplink control information and downlink control information (DCI); according to the scope (for example, the search space defines the starting position of blind detection and the channel search method), the information carried by the PDCCH can include common control information and dedicated control information.

[0003] In the communication process, PDCCH detection is needed, but in the prior art, there is a technical problem that the number of PDCCH detection is large, which leads to long PDCCH detection time and large terminal power consumption. SUMMARY

[0004] The present disclosure predicts the aggregation level (AL) related information of the PDCCH based on a prediction model, and performs PDCCH detection based on the predicted AL related information to replace PDCCH blind detection, so as to reduce the number of PDCCH detection to reduce the power consumption of the terminal and reduce the delay.

[0005] According to an exemplary embodiment of the present disclosure, a PDCCH detection method is provided, wherein the detection method comprises: obtaining input data required for predicting aggregation level (AL) related information; predicting AL related information using an AL prediction model based on the obtained input data; and performing PDCCH detection based on the predicted AL related information.

[0006] Optionally, the input data required for predicting the AL related information includes link information; and / or the AL related information includes at least one of the following: detection order for a plurality of ALs, probability of each AL in the plurality of ALs, and the AL with the maximum probability in the plurality of ALs.

[0007] Optionally, the link information comprises at least one of the following: reference signal received power (RSRP), channel quality indicator (CQI), signal to noise ratio (SNR), downlink control information (DCI) payload, and PDCCH slot index.

[0008] Optionally, if the AL-related information comprises a most probable AL among a plurality of ALs, performing PDCCH detection based on the predicted AL-related information comprises: performing PDCCH detection based on the predicted most probable AL; and if the detection fails, performing PDCCH blind detection based on other ALs.

[0009] Optionally, performing PDCCH detection based on the predicted AL-related information comprises: determining a PDCCH detection order based on the predicted AL-related information; and performing PDCCH detection based on the determined PDCCH detection order.

[0010] Optionally, predicting the AL-related information using the AL prediction model comprises: searching for an AL prediction model corresponding to the cell to which the terminal is attached according to a cell identifier of the cell and a correspondence between the AL prediction model and the cell identifier; and predicting the AL-related information using the searched AL prediction model based on the obtained input data.

[0011] Optionally, the cell identifier comprises a new radio cell global identifier (NCGI), and the NCGI comprises at least one of the following: mobile country code (MCC), mobile network code (MNC), and next generation node (gNB) identifier.

[0012] Optionally, before predicting the AL-related information using the searched AL prediction model, the method further comprises: determining that the searched AL prediction model is valid based on a prediction accuracy of the searched AL prediction model.

[0013] Optionally, searching for the AL prediction model corresponding to the cell to which the terminal is attached comprises: searching for the AL prediction model corresponding to the cell to which the terminal is attached from the terminal locally or from a server.

[0014] Optionally, the method further comprises: determining whether to merge at least two AL prediction models according to at least one of the following: a message indicating that AL prediction model merging is to be performed, similarity between outputs of the at least two AL prediction models, coincidence degree of a cell identifier list, and similarity between inputs of the at least two AL prediction models; and modifying the correspondence between the cell identifier and the AL prediction model, so that each cell identifier corresponding to the merged AL prediction model corresponds to the same AL prediction model.

[0015] Optionally, the detection method further comprises: obtaining the constructed AL prediction model from the terminal locally or from a server; and training the obtained AL prediction model based on historical data of the terminal, wherein the historical data of the terminal comprises the input data reported by the terminal to the base station and the AL information of the PDCCH detected by the terminal.

[0016] Optionally, the detection method further comprises: reporting PDCCH detection capability information to the base station, wherein the PDCCH detection capability information is capable of indicating whether the terminal is capable of performing PDCCH detection based on the predicted AL-related information.

[0017] Optionally, the detection method further comprises: obtaining a discontinuous reception (DRX) cycle configured by the base station, wherein the DRX cycle is configured by the base station based on the PDCCH detection capability information reported by the terminal, and wherein the DRX cycle configured by the base station for the terminal capable of performing PDCCH detection based on the predicted AL-related information is shorter than the DRX cycle configured by the base station for the terminal incapable of performing PDCCH detection based on the predicted AL-related information; and performing PDCCH detection based on the predicted AL-related information comprises performing PDCCH detection based on the predicted AL-related information and the obtained DRX cycle.

[0018] Optionally, the step of reporting the capability information to the base station comprises reporting the capability information through an information element (IE) in radio resource control (RRC) signaling.

[0019] Optionally, predicting the AL-related information using the AL prediction model comprises: determining whether the input data required for the currently obtained predicted AL-related information has changed compared with historical input data; predicting the AL-related information using the AL prediction model when it is determined that the input data has changed; and the detection method further comprises: performing PDCCH detection using the AL-related information predicted based on the historical input data when it is determined that the input data has not changed.

[0020] According to another example embodiment of the present disclosure, a training method of an AL prediction model is provided, wherein the training method comprises: receiving physical downlink control channel (PDCCH) detection-related data reported by a terminal; training an AL prediction model based on the PDCCH detection-related data reported by the terminal; and sending the trained AL prediction model to the terminal.

[0021] Optionally, the PDCCH detection-related data reported by the terminal comprises link information between the terminal and a base station and AL information of a PDCCH detected by the terminal, and the step of training the AL prediction model comprises: training one AL prediction model based on the reported PDCCH detection-related data corresponding to multiple base stations of the same type.

[0022] According to another example embodiment of the present disclosure, a discontinuous reception (DRX) configuration method is provided, wherein the configuration method comprises: receiving physical downlink control channel (PDCCH) detection capability information from a terminal; and configuring a DRX cycle according to the received capability information, wherein the PDCCH detection capability information is capable of representing whether the terminal is capable of performing PDCCH detection based on predicted AL-related information.

[0023] Optionally, the capability information is carried by an information element (IE) on radio resource control (RRC) signaling, and / or configuring the DRX cycle according to the received capability information comprises: configuring a shorter DRX cycle for the terminal capable of performing PDCCH detection based on predicted AL-related information than for the terminal incapable of performing PDCCH detection based on predicted AL-related information.

[0024] According to another example embodiment of the present disclosure, a terminal is provided, wherein the terminal comprises: a data acquisition unit configured to acquire input data required for predicting aggregation level (AL)-related information; a prediction unit configured to predict the AL-related information using an AL prediction model based on the acquired input data; and a detection unit configured to perform physical downlink control channel (PDCCH) detection based on the predicted AL-related information.

[0025] Optionally, the input data required for predicting the AL-related information comprises link information, and / or the AL-related information comprises at least one of the following: detection order for multiple ALs, probability of each of the multiple ALs, and the AL with the highest probability among the multiple ALs.

[0026] Optionally, the link information comprises at least one of the following: reference signal received power (RSRP), channel quality indicator (CQI), signal-to-noise ratio (SNR), downlink control information (DCI) payload, and PDCCH slot index.

[0027] Optionally, the detection unit is configured to: if the AL-related information comprises the AL with the highest probability among the multiple ALs, perform PDCCH detection based on the predicted AL with the highest probability; and if the detection fails, perform PDCCH blind detection based on other ALs.

[0028] Optionally, the detection unit is configured to: determine a PDCCH detection order based on the predicted AL-related information; and perform PDCCH detection based on the determined PDCCH detection order.

[0029] Optionally, the prediction unit is configured to: according to a cell identifier of the cell to which the terminal is attached and a correspondence between the AL prediction model and the cell identifier, find the AL prediction model corresponding to the cell to which the terminal is attached; and based on the obtained input data, predict the AL-related information using the found AL prediction model.

[0030] Optionally, the cell identifier comprises a new radio cell global identifier (NCGI), and the NCGI comprises at least one of the following: a mobile country code (MCC), a mobile network code (MNC), and a next generation node (gNB) identifier.

[0031] Optionally, the terminal further comprises a judgment unit configured to determine, before predicting the AL-related information using the found AL prediction model, whether the found AL prediction model is valid based on a prediction accuracy of the found AL prediction model.

[0032] Optionally, the prediction unit is configured to find the AL prediction model corresponding to the cell to which the terminal is attached from the terminal locally or from a server.

[0033] Optionally, the terminal further comprises a merging unit configured to determine whether to merge at least two AL prediction models according to at least one of the following: a message indicating that the AL prediction model is to be merged, a similarity between outputs of the at least two AL prediction models, a coincidence degree of a cell identifier list, and a similarity between inputs of the at least two AL prediction models; and modify the correspondence between the cell identifier and the AL prediction model, so that each cell identifier corresponding to the merged AL prediction model corresponds to the same AL prediction model.

[0034] Optionally, the terminal further comprises a training unit and a model acquisition unit, the model acquisition unit is configured to acquire the constructed AL prediction model from the terminal locally or from a server, and the training unit is configured to train the acquired AL prediction model based on historical data of the terminal, wherein the historical data of the terminal comprises the input data reported by the terminal to the base station and the AL information of the PDCCH detected by the terminal.

[0035] Optionally, the terminal further comprises a reporting unit configured to report PDCCH detection capability information to the base station, wherein the PDCCH detection capability information is capable of representing whether the terminal is capable of performing PDCCH detection based on the predicted AL-related information.

[0036] Optionally, the terminal further comprises a receiving unit configured to acquire a discontinuous reception (DRX) cycle configured by the base station, wherein the DRX cycle is configured by the base station according to the PDCCH detection capability information reported by the terminal, and wherein the DRX cycle configured for the terminal capable of performing PDCCH detection based on predicted AL-related information is shorter than the DRX cycle configured for the terminal incapable of performing PDCCH detection based on predicted AL-related information; and a detecting unit configured to perform PDCCH detection based on the predicted AL-related information and the acquired DRX cycle.

[0037] Optionally, the reporting unit is configured to report the capability information through an information element (IE) in RRC signaling.

[0038] Optionally, the predicting unit is configured to determine whether input data required for the currently acquired predicted AL-related information has changed compared with historical input data, and predict the AL-related information using an AL prediction model when it is determined that the input data has changed; and the detecting unit is configured to perform PDCCH detection using the AL-related information predicted based on the historical input data when it is determined that the input data has not changed.

[0039] According to another example embodiment of the present disclosure, a terminal is provided, wherein the terminal comprises: a processor; and a memory configured to store machine-readable instructions, which, when executed by the processor, cause the processor to perform the above-mentioned method for detecting a physical downlink control channel (PDCCH).

[0040] According to another example embodiment of the present disclosure, a base station is provided, wherein the base station comprises: a communication unit configured to receive physical downlink control channel (PDCCH) detection capability information from a terminal; and a configuration unit configured to configure a DRX cycle according to the received capability information, wherein the PDCCH detection capability information is capable of representing whether the terminal is capable of performing PDCCH detection based on predicted AL-related information.

[0041] Optionally, the capability information is carried through an information element (IE) in RRC signaling, and / or the configuration unit is configured to configure a DRX cycle for the terminal capable of performing PDCCH detection based on predicted AL-related information to be shorter than a DRX cycle for the terminal incapable of performing PDCCH detection based on predicted AL-related information.

[0042] According to another example embodiment of the present disclosure, a base station is provided, wherein the base station comprises: a processor; and a memory configured to store machine-readable instructions, which, when executed by the processor, cause the processor to perform the above-mentioned method for configuring a discontinuous reception (DRX).

[0043] According to another example embodiment of the present disclosure, a server is provided, wherein the server comprises: a receiving unit configured to receive data related to PDCCH detection reported by a terminal; a training unit configured to train an AL prediction model based on the data related to PDCCH detection reported by the terminal; and a sending unit configured to send the trained AL prediction model to the terminal.

[0044] Optionally, the data related to PDCCH detection reported by the terminal comprises link information between the terminal and a base station and AL information of a detected PDCCH, and the training unit is configured to train one AL prediction model based on reported data related to PDCCH detection corresponding to a plurality of base stations of the same type.

[0045] According to another example embodiment of the present disclosure, a computer readable storage medium storing instructions is provided, wherein when the instructions are run by at least one computing device, the at least one computing device is caused to perform the above method.

[0046] The present disclosure can effectively utilize prior information to reduce the number of PDCCH detections in the PDCCH detection process, so as to achieve the goal of saving terminal power consumption.

[0047] In addition, if the terminal can complete PDCCH detection with less power consumption, a shorter DRX period (cycle) can be set for the terminal. The terminal can perform detection based on the predicted information to complete PDCCH detection in a shorter time than PDCCH blind detection, so as not to consume more terminal power, and a long non-DRX cycle can increase the scheduling opportunity of the terminal to improve transmission efficiency (for example, improve the amount of received and transmitted data of Massive MIMO).

[0048] Additional aspects and / or advantages of the general inventive concept will be set forth in part in the description which follows, and in part will be obvious from the description, or can be learned by practice of the general inventive concept. BRIEF DESCRIPTION OF DRAWINGS

[0049] The above and other objects and features of the present exemplary embodiments of the present disclosure will become more apparent from the following description of the exemplary embodiments of the present disclosure given for the purpose of illustrations only, wherein:

[0050] Figure 1 A schematic diagram of an algorithm for a base station to determine AL according to an exemplary embodiment of the present disclosure is shown;

[0051] Figure 2 A schematic diagram of an algorithm for predicting AL related information of PDCCH based on an AL prediction model according to an exemplary embodiment of the present disclosure is shown;

[0052] Figure 3 a flowchart illustrating a PDCCH detection method according to an example embodiment of the present disclosure;

[0053] Figure 4 a flowchart illustrating a PDCCH detection method according to another example embodiment of the present disclosure;

[0054] Figure 5 a structural schematic diagram of a network side according to an example embodiment of the present disclosure;

[0055] Figure 6 a flowchart illustrating a PDCCH detection method according to another example embodiment of the present disclosure;

[0056] Figure 7 a flowchart illustrating AI detection according to an example embodiment of the present disclosure;

[0057] Figure 8 a flowchart illustrating searching for an AI model according to an example embodiment of the present disclosure

[0058] Figure 9 a flowchart illustrating AI detection according to an example embodiment of the present disclosure;

[0059] Figure 10 interaction between a terminal and an AI server according to an example embodiment of the present disclosure;

[0060] Figure 11 single NCGI training and multi NCGI training according to an example embodiment of the present disclosure;

[0061] Figure 12 a flowchart illustrating AI detection according to an example embodiment of the present disclosure;

[0062] Figure 13 a flowchart illustrating base station and AI server joint training according to an example embodiment of the present disclosure. DETAILED DESCRIPTION

[0063] Reference will now be made in detail embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like elements throughout. The embodiments will be explained by referring to the drawings in detail.

[0064] The DCI includes Physical Uplink Shared Channel (PUSCH) DCI and Physical Downlink Shared Channel (PDSCH) DCI. The PDCCH can be identified or detected with a Radio Network Temporary Identity (RNTI), for example, a Cyclic Redundancy Check (CRC) of the DCI can be scrambled with the RNTI to identify the DCI.

[0065] Generally, there can be multiple PDCCHs in one Transmission Time Interval (TTI). A User Equipment (UE) needs to demodulate the DCI in the PDCCH, and then demodulate the PDSCH belonging to the UE, which includes but is not limited to broadcast messages, paging, and UE data.

[0066] Multiple PDCCHs can be transmitted simultaneously within the PDCCH transmission bandwidth. To configure PDCCH and other time-frequency resources more efficiently, Long Term Evolution (LTE) and New Radio (NR) protocols define Control Channel Elements (CCEs). The downlink channel environment where the UE is located affects the number of CCEs occupied by the PDCCH, and the base station can allocate fewer CCEs (e.g., one CCE) to the UE in a better downlink channel environment and more CCEs (e.g., 16 CCEs) to the UE in a worse downlink channel environment. The number of CCEs contained or occupied by the PDCCH can be referred to as the Aggregation Level (AL). To simplify the complexity of PDCCH detection, LTE and NR protocols also stipulate that the CCE number corresponding to the starting position of the PDCCH is an integer multiple of the AL.

[0067] In LTE and NR networks, the base station can send the PDCCH to the UE, but the base station does not inform the UE of the AL of the PDCCH and the CCE number corresponding to the starting position of the PDCCH. The UE needs to perform PDCCH detection for multiple ALs, and such detection can be referred to as blind detection. When the AL is determined through blind detection, the CCEs that transmit the PDCCH can be determined.

[0068] According to the Third Generation Partnerships Project (3GPP) standardization research and terminal development results, the increase of PDCCH detection will increase the power consumption of the terminal.

[0069] In an aspect, within a TTI, the UE detects PDCCH according to a configured Discontinuous Reception (DRX) cycle until one of the following two results is reached: in one result, the UE detects PDCCH, receives DCI carried by PDCCH, and stops PDCCH detection; in another result, the UE does not detect PDCCH until a defined maximum number of PDCCH detections for a search space is reached, and ends PDCCH detection.

[0070] In the process of PDCCH detection, channel decoding needs to be performed for each candidate PDCCH corresponding to each AL in multiple ALs. Since the completion of the channel decoding process consumes operations, and the increase of the number of channel decoding can lead to the increase of consumed operations, therefore, with the increase of the number of ALs for which detection is performed, the number of channel decoding increases, and the consumed operations also increase, thus causing the increase of power consumption.

[0071] In the traditional PDCCH detection process, blind detection can be performed according to a fixed AL order, for example, an AL with a size of 1 is taken as a starting AL and detection is started, and then the AL is gradually increased to try to detect a PDCCH corresponding to a specific AL until the PDCCH is detected or the maximum AL is reached; or blind detection can be performed according to limited prior information, for example, the AL at the end of the last blind detection is taken as a starting AL for the current blind detection, and then the AL is increased to try to detect a PDCCH corresponding to a specific AL.

[0072] As described above, the traditional PDCCH detection process does not utilize prior information or utilizes limited prior information, and the terminal needs to determine the AL of the PDCCH by trying, so the traditional PDCCH detection is called blind detection. Blind detection usually cannot directly perform PDCCH detection on the actual AL (the AL of the current PDCCH), and the trying before the actual AL is found will consume the power of the terminal.

[0073] Third Generation (3G), Fourth Generation (4G) and Fifth Generation (5G) systems need to consider the power consumption of terminals. 5G protocols will consider energy efficiency as a key performance indicator, so the power consumption of UEs needs to be considered more in the design process.

[0074] In another aspect, to reduce the power consumption of the terminal, the 5G standard makes changes to DRX and the like, for example, by increasing the DRX cycle to reduce the opportunity for the UE to detect the PDCCH (for example, the number of TTIs in which PDCCH detection is performed), that is, to reduce the number of PDCCH detections. The reduction of the detection opportunity of the PDCCH will result in a reduction of the opportunity for the UE to be scheduled, so that the latency of the UE is increased. In addition, the detection of the PDCCH often needs to go through multiple rounds of detection, and during the detection of the PDCCH, the signals for assisting scheduling and measurement (for example, Sounding Reference Signal (SRS)) cannot be transmitted. The 5G core technology includes the Massive Multi-Input Multi-Output (Massive MIMO) technology. The Massive MIMO technology needs to use SRS for channel measurement, so the non-transmission of SRS during the detection of the PDCCH can affect the performance of the Massive MIMO.

[0075] The basic idea of the present disclosure includes that the terminal performs PDCCH detection based on an AL prediction model, wherein the prior information that can be embodied by the AL prediction model can be used to predict the AL related information (for example, the order of the AL to which the PDCCH detection is directed), and the PDCCH detection is performed according to the predicted AL related information. The base station can calculate the AL of the PDCCH according to the link information (for example, applying the link quality information or channel quality information between the base station and the terminal to a PDCCH AL adaptation algorithm to calculate the AL of the PDCCH), and the terminal can use historical data (which can be referred to as training data) to learn the way in which the base station calculates the AL to form an AL prediction model, and the historical data includes the data (which can be referred to as PDCCH AL adaptation algorithm parameters) used by the base station when performing the PDCCH AL adaptation algorithm and the AL of the PDCCH that has been detected by the terminal. The order of the AL to which the detection is directed can be determined according to the information predicted by the AL prediction model, and the PDCCH detection is performed according to such order rather than blindly. In this case, the number of detections in the PDCCH detection process can be reduced to reduce power consumption, and more opportunities can be provided for the transmission of SRS and the like to improve the response speed and other performances of the terminal. In another aspect, the base station can configure the DRX and the like parameters to ensure the transmission of SRS and the like signals and the scheduling based on the PDCCH can be performed according to whether the terminal can perform PDCCH detection based on the AL prediction model, so that the throughput is improved and the latency is reduced.

[0076] In the exemplary embodiments of the present disclosure, the terminal (may be referred to as UE), the base station, and the server can support at least one of 3G protocol, 4G protocol, and 5G protocol, respectively, of course, this is not used for limitation, for example, the possibility of supporting a new generation protocol or an old protocol is not excluded.

[0077] Figure 1 A schematic diagram of an algorithm for a base station to determine an AL according to an exemplary embodiment of the present disclosure is shown.

[0078] As shown in Figure 1 , the terminal can report input data (for example, link information between a next generation node (gNB) and the terminal, which can be link quality information, channel quality information, etc.) of a PDCCH AL adaptive algorithm to the base station. On the network side, the base station determines the channel quality of the PDCCH according to the parameters reported by the user, and determines the AL of the PDCCH. The link information reported by the terminal (input of the PDCCH AL adaptive algorithm) is known information of the terminal, and through PDCCH detection, the terminal can know the AL of the PDCCH (i.e. the AL calculated according to the PDCCH AL adaptive algorithm and configured), for example, the AL of uplink (UL) and the AL of downlink (DL). That is, the terminal can know the input and output of the PDCCH AL adaptive algorithm on the base station side, so that the UE can learn the PDCCH AL adaptive algorithm on the base station side, to avoid blind detection, which provides the possibility for the terminal or a device other than the terminal (for example, a server) to learn the PDCCH AL adaptive algorithm of the base station.

[0079] Figure 2 A schematic diagram of an algorithm for predicting AL related information of a PDCCH based on an AL prediction model according to an exemplary embodiment of the present disclosure is shown.

[0080] As shown in Figure 2 , the input data is input to an AL prediction model (may be an artificial intelligence (AI) model or a non-AI model), and the AL prediction model predicts the output based on the input data. As Figure 1 the input data of the PDCCH AL adaptive algorithm in Figure 2The link information of the input data of the AL prediction model in the PDCCH AL adaptive algorithm can include at least one of the following: a downlink condition indicator of the terminal, including at least one of the following: a reference signal receiving power (RSRP), a channel quality indicator (CQI), a signal-to-noise ratio (SNR), a downlink control information (DCI) payload (i.e., the byte size of the DCI), a PDCCH slot index, and a transmission direction (including at least one of downlink (DL) and uplink (UL)). Figure 1 The output of the PDCCH AL adaptive algorithm in the PDCCH AL adaptive algorithm can include at least one of the following: an AL related information of the DL and an AL related information of the UL. Figure 2 The output of the AL prediction model in the PDCCH AL adaptive algorithm can include at least one of the following: an AL related information of the DL and an AL related information of the UL.

[0081] As an example, the input data can be collected by the terminal of an evolved node (eNB), a next generation node (gNB), etc. with reference to the input data required by the PDCCH AL adaptive algorithm, and the collected input data is input to the AL prediction model to predict the AL related information (such as the AL related information at the current detection time). According to the AL related information predicted by the AL prediction model, the terminal can determine the PDCCH detection order (for example, determine the order of the ALs to which the PDCCH detection is directed), and perform the PDCCH detection according to the determined detection order rather than the default detection order. Since the algorithm based on the AL prediction model utilizes prior information (for example, accurate ALs enable the PDCCH to be detected earlier than blind detection of the ALs), the number of detections (for example, blind detections) can be reduced, the detection time of the PDCCH can be accelerated, the power consumption caused by the detection can be reduced, and the purpose of power saving of the terminal can be achieved.

[0082] Figure 3 A flowchart of a PDCCH detection method according to an example embodiment of the present disclosure is shown.

[0083] As shown in Figure 3 The PDCCH detection method according to an example embodiment of the present disclosure can include steps 101 to 103. In step 101, input data required for predicting AL related information is obtained; in step 102, AL related information is predicted using an AL prediction model based on the obtained input data; and in step 103, PDCCH detection is performed based on the predicted AL related information. For example, Figure 3The illustrated detection method can be performed by the terminal, and the predicted AL-related information can be an order of ALs. Since the starting position of the PDCCH is related to the AL (for example, the starting position is an integer multiple of the AL), the starting position of the PDCCH is also determined to some extent when the AL is determined. In the case of scrambling by RNTI, it can be determined whether the PDCCH exists using the RNTI from the starting position. When the predicted AL-related information includes the order of the ALs, PDCCH detection can be performed based on the predicted order.

[0084] As an example, the input data required for predicting the AL-related information includes: link information; and / or, the AL-related information includes at least one of: a detection order for multiple ALs, a probability of each of the multiple ALs, and a most probable AL in the multiple ALs.

[0085] As an example, the link information includes at least one of: RSRP, CQI, SNR, DCI payload, and PDCCH slot index.

[0086] As an example, if the AL-related information includes a most probable AL in the multiple ALs, performing PDCCH detection based on the predicted AL-related information includes: performing PDCCH detection based on the predicted most probable AL; and if the detection fails, performing PDCCH blind detection based on other ALs.

[0087] As an example, performing PDCCH detection based on the predicted AL-related information includes: determining a PDCCH detection order based on the predicted AL-related information; and performing PDCCH detection based on the determined PDCCH detection order.

[0088] As an example, predicting the AL-related information using an AL prediction model includes: looking up an AL prediction model corresponding to the cell to which the terminal is attached according to a cell identifier of the cell and a correspondence relationship between the AL prediction model and the cell identifier; and predicting the AL-related information using the looked-up AL prediction model based on the obtained input data.

[0089] As an example, the cell identifier includes a New Radio Cell Global Identifier (NCGI), and the NCGI includes at least one of: a Mobile Country Code (MCC), a Mobile Network Code (MNC), and a next generation Node B (gNB) identifier.

[0090] As an example, before predicting the AL related information using the found AL prediction model, the method further comprises: determining whether the found AL prediction model is valid based on a prediction accuracy of the found AL prediction model.

[0091] As an example, finding the AL prediction model corresponding to the cell to which the terminal is attached comprises: finding the AL prediction model corresponding to the cell to which the terminal is attached from the terminal locally or from a server.

[0092] Optionally, the detection method further comprises: determining whether to combine the at least two AL prediction models according to at least one of the following: a message indicating that the AL prediction model combination is performed, a similarity between outputs of the at least two AL prediction models, a coincidence degree of a cell identifier list, and a similarity between inputs of the at least two AL prediction models; and modifying a correspondence between the cell identifier and the AL prediction model, so that each cell identifier corresponding to the combined AL prediction model corresponds to the same AL prediction model.

[0093] Optionally, the detection method further comprises: obtaining the constructed AL prediction model from the terminal locally or from a server; and training the obtained AL prediction model based on historical data of the terminal, wherein the historical data of the terminal comprises the input data reported by the terminal to the base station and the AL information of the PDCCH detected by the terminal.

[0094] Optionally, the detection method further comprises: reporting PDCCH detection capability information to the base station, wherein the PDCCH detection capability information is capable of representing whether the terminal is capable of performing PDCCH detection based on the predicted AL related information.

[0095] Optionally, the detection method further comprises: obtaining a discontinuous reception (DRX) cycle configured by the base station, wherein the DRX cycle is configured by the base station according to the PDCCH detection capability information reported by the terminal, and wherein the DRX cycle configured by the base station for the terminal capable of performing PDCCH detection based on the predicted AL related information is shorter than the DRX cycle configured by the base station for the terminal incapable of performing PDCCH detection based on the predicted AL related information; and performing PDCCH detection based on the predicted AL related information comprises: performing PDCCH detection based on the predicted AL related information and the obtained DRX cycle.

[0096] Optionally, the step of reporting the capability information to the base station comprises: reporting the capability information through an information element (IE).

[0097] Optionally, the AL related information is predicted using the AL prediction model, including: judging whether the input data required for the currently acquired prediction AL related information changes compared with the historical input data; when judging that the change occurs, using the AL prediction model to predict the AL related information; the detection method further includes: when judging that no change occurs, using the AL related information predicted based on the historical input data to perform PDCCH detection.

[0098] Figure 4 A flow chart of a PDCCH detection method according to another example embodiment of the present disclosure is shown.

[0099] As shown in Figure 4 In the case of starting PDCCH detection, if the AI detection condition is met (for example, the terminal can predict the AL related information based on the AI model and can perform PDCCH detection based on the AL related information), AI detection can be performed, and if the AI detection condition is not met, blind detection of PDCCH is performed.

[0100] In the example embodiments of the present disclosure, the process of predicting the AL related information using the AI model can also be referred to as AI detection. AI detection and prediction based on the AI model are only exemplary and do not limit the protection scope of the present disclosure. AI detection can be replaced by non-AI detection, or replaced by AL prediction including AI detection and non-AI detection, and the AI model can be replaced by a non-AI model or replaced by an AL prediction model including the AI model and the non-AI model. The AI model can be, but is not limited to, a deep learning model, a deep neural network model, a deep network model, a traditional three-layer neural network model, etc.

[0101] Steps 211 and 212 relate to blind detection. In step 211, the candidate PDCCH is determined based on the fixed (default) AL order. For example, the multiple ALs have a default order, and each AL can correspond to a candidate PDCCH. For another example, the default order of the ALs with sizes of 1, 2, 4, and 8 can be 1->2->4->8. In step 212, since the order of the candidate PDCCH is consistent with the order of the AL, the candidate PDCCH can be detected based on the order of the AL.

[0102] Steps 221 to 224 relate to AI detection. In step 221, the AL related information of the PDCCH can be predicted based on the AI model (in Figure 4The simplified AL is referred to as AL). In step 222, the detection order can be determined based on the predicted AL-related information. In step 223, based on the determined detection order, candidate PDCCHs can be determined. For example, the order of AL can be determined as 2->4->1->8. In step 224, detection is performed on the candidate PDCCHs, and data (which can also be referred to as terminal data) in the case of detecting a PDCCH (including input data (such as link information) required for predicting AL-related information and detected AL information) is recorded for subsequent processing (for example, detecting the accuracy of an AI model and / or training an AI model).

[0103] As an example, the PDCCH detection method of this embodiment is applicable to a gNB network, and the terminal can be a 5G terminal. The AI model can simulate the PDCCH AL adaptive algorithm of the gNB network to predict the AL-related information of the PDCCH. However, this is only an example provided for description and does not limit the protection scope of the present disclosure, and the method of the present disclosure is also applicable to terminals, base stations, servers or networks of 3G or 4G.

[0104] Corresponding to the detection method, the training method and the DRX configuration method described below can be implemented.

[0105] According to another exemplary embodiment of the present disclosure, a training method of an AL prediction model is provided, wherein the training method comprises: receiving related data of PDCCH detection reported by a terminal; training an AL prediction model based on the related data of PDCCH detection reported by the terminal; and sending the trained AL prediction model to the terminal.

[0106] As an example, the related data of PDCCH detection reported by the terminal includes link information between the terminal and the base station and AL information of the PDCCH detected (blind detection or AI detection) by the terminal, wherein the step of training the AL prediction model comprises: training one AL prediction model based on the reported related data of PDCCH detection corresponding to multiple base stations of the same type, and / or training the AL prediction model for each base station respectively (training one AL prediction model based on the reported related data of PDCCH detection corresponding to each base station), and / or training one AL prediction model for all base stations (training one AL prediction model based on all reported related data of PDCCH detection).

[0107] According to another exemplary embodiment of the present disclosure, a DRX configuration method is provided, wherein the configuration method comprises: receiving physical downlink control channel (PDCCH) detection capability information from a terminal; and configuring a DRX cycle according to the received capability information, wherein the PDCCH detection capability information can represent whether the terminal can perform PDCCH detection based on predicted AL-related information.

[0108] As an example, the capability information is carried by an information element (IE), and / or the DRX cycle is configured according to the received capability information, including: a DRX cycle configured for a terminal capable of PDCCH detection based on predicted AL-related information is shorter than a DRX cycle configured for a terminal incapable of PDCCH detection based on predicted AL-related information.

[0109] The training method and the DRX configuration method according to the example embodiments of the present disclosure can be implemented at a network side, where the training method can be implemented by a server, and the DRX configuration method can be implemented by a base station. In addition, the training method can also be implemented at a terminal.

[0110] Figure 5 A structure diagram of a network side according to an example embodiment of the present disclosure is shown.

[0111] The network side according to an example embodiment of the present disclosure can include a base station or include a base station and a server. The base station can be configured to receive PDCCH detection capability information reported by a terminal, and configure a short DRX cycle for a terminal with AL detection capability (e.g., capable of PDCCH detection based on predicted AL-related information), which can also be referred to as a terminal with AL intelligent detection capability, and can also be referred to as a terminal supporting AI detection, and configure a long DRX cycle for a terminal without AL detection capability.

[0112] According to an example embodiment of the present disclosure, the base station includes: a communication unit configured to receive physical downlink control channel (PDCCH) detection capability information from a terminal; and a configuration unit configured to configure a DRX cycle according to the received capability information, where the PDCCH detection capability information can represent whether the terminal is capable of PDCCH detection based on predicted AL-related information.

[0113] As an example, the capability information is carried by an information element (IE), and / or the DRX cycle is configured according to the received capability information, including: a DRX cycle configured for a terminal capable of PDCCH detection based on predicted AL-related information is shorter than a DRX cycle configured for a terminal incapable of PDCCH detection based on predicted AL-related information.

[0114] According to another example embodiment of the present disclosure, the base station includes: a processor; and a memory configured to store machine-readable instructions, which, when executed by the processor, cause the processor to perform the above-mentioned discontinuous reception (DRX) configuration method.

[0115] According to another example embodiment of the present disclosure, a server comprises: a receiving unit configured to receive relevant data of physical downlink control channel (PDCCH) detection reported by a terminal; a training unit configured to train an AL prediction model based on the relevant data of PDCCH detection reported by the terminal; and a sending unit configured to send the trained AL prediction model to the terminal.

[0116] As an example, the relevant data of PDCCH detection reported by the terminal comprises link information between the terminal and a base station and AL information of a PDCCH that has been detected by the terminal, wherein the training unit is configured to train one AL prediction model based on the reported relevant data of PDCCH detection corresponding to multiple base stations of the same type, and / or train the AL prediction model for each base station respectively, and / or train one AL prediction model for all base stations.

[0117] As an example, terminals with AL detection capability (e.g., supporting AI detection) and terminals without AL detection capability can be deployed within the coverage of a base station. The base station can send PDCCHs to these terminals regardless of whether the terminals have AL detection capability. Terminals with AL detection capability can have intelligent power saving functions and can detect PDCCHs more quickly than terminals without AL detection capability, thereby reducing the power consumed for PDCCH detection, so terminals with AL detection capability consume less power than terminals without AL detection capability.

[0118] As an example, an interface can be configured between the network side (e.g., a base station or an AI server) and a terminal to transmit capability information. This interface can be implemented as a standardized interface, for example, a new information element (IE) can be added on the radio resource control (RRC) signaling. The IE can be used to transmit capability information (e.g., the terminal reports capability information to the base station, the capability information can represent whether the terminal can perform PDCCH detection based on predicted AL related information) and can also be used to transmit an AL prediction model (e.g., the terminal receives or downloads a model trained by the server from the server, and the terminal submits terminal data to the server).

[0119] As an example, the capability information and / or the reported terminal data can be reported in real time or non-real time. In the case of non-real-time reporting, the trigger conditions for the terminal to report capability information and / or reported terminal data include but are not limited to at least one of the following: the data volume reaches a predetermined amount, the time reaches a reporting time (reaches a time specified by a reporting period, the reporting time can be agreed or agreed by agreement, in units of days, weeks, or months), the storage amount of the storage of the terminal reaches a predetermined storage amount (e.g., the storage is full), and the prediction accuracy of the existing AL prediction model is lower than a predetermined accuracy.

[0120] Since the capacity of the memory of the terminal is limited, the terminal data can be reported when the data volume of the collected terminal data reaches a predetermined value, and the stored data can be deleted after reporting to store new data. The data volume can be measured by bytes, for example, when the data volume occupies a predetermined number of bytes. The memory of the terminal stores the terminal data, and the storage capacity can also be measured by units such as bytes. The objects targeted by the trigger conditions based on the data volume and the trigger conditions based on the storage volume are data and memory, respectively. When the AL prediction accuracy of the terminal using the existing AI model cannot reach the threshold, it is considered that the base station updates the PDCCH aggregation level adaptive algorithm, and the input and output information of the AI model needs to be updated, so the terminal can upload the terminal data.

[0121] As an example, the server can be separate from the terminal or the base station, or can be included in the base station or the terminal. The server can train the AL prediction model based on the terminal data submitted by the terminal. In the case where the terminal includes the server, the server can be a unit of the terminal, so that the terminal can locally train the AL prediction model.

[0122] As an example, the base station can configure a shorter DRX cycle for the terminal with AL detection capability. A short DRX cycle means a short sleep period and a long scheduling time. During the time outside the DRX cycle, the terminal can perform PDCCH detection (for example, PDCCH detection based on the AL prediction model, when the AL prediction model is an AI model, the PDCCH detection can be referred to as PDCCH intelligent detection). However, since the PDCCH intelligent detection can be completed more quickly (with fewer detection times), it will not increase the terminal power consumption, on the contrary, it can increase the opportunity of the terminal being scheduled, improve the terminal response speed and throughput (for example, increase the performance of the terminal supporting Massive MIMO), improve the data transmission efficiency, and reduce the latency, which can meet the energy efficiency target required by the standard. The base station can configure a longer DRX cycle for the terminal without AL detection capability, so as to ensure that such a terminal does not consume too much power due to PDCCH blind detection, but the response speed of such a terminal is slow, the latency is long, and the throughput is low.

[0123] Figure 6 A flowchart of PDCCH detection according to another example embodiment of the present disclosure is shown.

[0124] The PDCCH detection method of the present example embodiment can be implemented in a terminal, and the detection method can include: a terminal attached to a cell determining the NCGI of the cell; searching for a valid AL prediction model (for example, an AI model) according to the NCGI; if a valid AI model is found, performing AI detection, and if no valid AI model is found, performing blind detection or updating the AI model from the server.

[0125] Specifically, the terminal can attach to a cell (e.g., a cell of a 5G base station) and obtain an NCGI of the cell, where the NCGI can be used to uniquely identify the cell globally and can include an MCC, an MNC, an ID of a gNB, and an ID of the cell. The terminal can look up a valid AI model according to the NCGI (e.g., locally or remotely), perform AI detection if a valid AI model is found, or perform blind detection or attempt to obtain an AI model from a server if a valid AI model is not found; after AI detection or blind detection of PDCCH, it can be determined whether a cell change occurs, and if a cell change occurs, the terminal obtains a new NCGI and performs the steps of looking up a valid AI model and the subsequent steps.

[0126] After obtaining (e.g., finding) an AI model, the valid AI model can be determined by a prediction accuracy, and the step of AI detection can also be performed. During AI detection, terminal data (e.g., input data of the AI model described above) can be input into the AI model to obtain an output of the AI model. The output of the AI model can include an AL order (e.g., an AL order of 2->4->1->8), and each AL can correspond to a candidate PDCCH. Each candidate PDCCH can be detected based on the AL order to determine a PDCCH configured by the base station for the terminal. After determining the PDCCH configured by the base station for the terminal, the terminal data and the AL of the detected PDCCH can also be recorded. During blind detection, detection is performed based on a fixed AL order (e.g., an AL order of 1->2->4->8), each AL can correspond to a candidate PDCCH, and the PDCCH configured by the base station for the terminal is found from each candidate PDCCH based on the fixed AL order.

[0127] As an example, the NCGI here can be applicable to a 5G cell, and can be replaced by a cell identifier applicable to a 3G or 4G cell. In this way, the step of looking up an AI model can include looking up an AI model corresponding to a cell according to a correspondence between the AI model and a cell identifier of a cell to which the terminal is attached, and determining whether a prediction accuracy of the AI model meets a requirement, and when the prediction accuracy meets the requirement, determining that the AI model is valid, where the step of predicting AL-related information can include predicting AL-related information based on the valid AI model.

[0128] After AI detection or blind detection, it can be determined whether a cell change occurs, and if a cell change occurs, the steps of determining an NCGI of a cell and the subsequent steps are performed.

[0129] A terminal according to an example embodiment of the present disclosure can include: a data obtaining unit configured to obtain input data required for predicting AL-related information; a predicting unit configured to predict the AL-related information using an AL prediction model based on the obtained input data; and a detecting unit configured to perform PDCCH detection based on the predicted AL-related information.

[0130] As an example, the input data required for predicting the AL-related information includes: link information; and / or, the AL-related information includes at least one of: a detection order for a plurality of ALs, a probability of each of the plurality of ALs, and an AL with the highest probability among the plurality of ALs.

[0131] As an example, the link information includes at least one of: RSRP, CQI, SNR, DCI payload, and PDCCH slot index.

[0132] As an example, the detecting unit is configured to: if the AL-related information includes an AL with the highest probability among the plurality of ALs, perform PDCCH detection based on the predicted AL with the highest probability; and if the detection fails, perform PDCCH blind detection based on other ALs.

[0133] As an example, the detecting unit is configured to: determine a PDCCH detection order based on the predicted AL-related information; and perform PDCCH detection based on the determined PDCCH detection order.

[0134] As an example, the predicting unit is configured to: find an AL prediction model corresponding to a cell to which the terminal is attached, according to a cell identifier of the cell and a correspondence relationship between the AL prediction model and the cell identifier; and predict the AL-related information using the found AL prediction model based on the obtained input data.

[0135] As an example, the cell identifier includes a New Radio Cell Global Identifier (NCGI), and the NCGI includes at least one of: a Mobile Country Code (MCC), a Mobile Network Code (MNC), and a Next Generation Node (gNB) identity.

[0136] As an example, the terminal further includes a judging unit configured to determine that the found AL prediction model is valid based on a prediction accuracy of the found AL prediction model, before predicting the AL-related information using the found AL prediction model.

[0137] As an example, the predicting unit is configured to find the AL prediction model corresponding to the cell to which the terminal is attached from the terminal locally or from a server.

[0138] As an example, the terminal further comprises a merging unit configured to determine whether to merge at least two AL prediction models according to at least one of the following: a message indicating to merge the AL prediction models, a similarity between outputs of the at least two AL prediction models, a coincidence degree of a cell identifier list, and a similarity between inputs of the at least two AL prediction models; and modify a correspondence between cell identifiers and AL prediction models, such that each cell identifier corresponding to a merged AL prediction model corresponds to a same AL prediction model.

[0139] As an example, the terminal further comprises a training unit and a model obtaining unit, the model obtaining unit being configured to obtain the constructed AL prediction model from a terminal locally or from a server; and the training unit being configured to train the obtained AL prediction model based on historical data of the terminal; wherein the historical data of the terminal comprises the input data reported by the terminal to a base station and AL related information of a PDCCH detected by the terminal.

[0140] As an example, the terminal further comprises a reporting unit configured to report PDCCH detection capability information to a base station, wherein the PDCCH detection capability information is capable of representing whether the terminal is capable of performing PDCCH detection based on predicted AL related information.

[0141] As an example, the terminal further comprises a receiving unit configured to obtain a discontinuous reception (DRX) cycle configured by a base station, wherein the DRX cycle is configured by the base station according to PDCCH detection capability information reported by the terminal, and wherein the base station configures a shorter DRX cycle for a terminal capable of performing PDCCH detection based on predicted AL related information than for a terminal incapable of performing PDCCH detection based on predicted AL related information; and a detecting unit configured to perform PDCCH detection based on predicted AL related information and the obtained DRX cycle.

[0142] As an example, the reporting unit is configured to report the capability information via an information element (IE).

[0143] As an example, the predicting unit is configured to determine whether input data required for currently obtained predicted AL related information has changed compared to historical input data; and when it is determined that there is a change, predict the AL related information using an AL prediction model; and the detecting unit is configured to, when it is determined that there is no change, perform PDCCH detection using AL related information predicted based on the historical input data.

[0144] A UE or terminal according to another example embodiment of the present disclosure can comprise: a processor; and a memory configured to store machine-readable instructions which, when executed by the processor, cause the processor to perform the above-described detection method.

[0145] Figure 7 A flowchart of AI detection according to an example embodiment of the present disclosure is shown. The difference between AI detection of PDCCH and blind detection of PDCCH includes that AI detection can perform PDCCH detection and extraction based on predicted AL order, while blind detection performs PDCCH detection and extraction based on fixed (e.g., default) AL order. Prediction can be made with an AI model, input data of the AI model can be collected by the terminal, and the input data can be related parameters of PDCCH AL adaptive algorithm of gNB network, including at least one of the following: reference signal receiving power (RSRP), received signal strength indication (RSSI), channel quality indicator (CQI), signal-to-noise ratio (SNR), byte size of downlink control information (DCI), and slot index of PDCCH. The output of the AI model can include the probability of each AL of uplink (UL) and downlink (DL), for example, the probability of each AL of DL and / or the probability of each AL of UL.

[0146] In an example embodiment of the present disclosure, according to the correlation between the AL of DL and the AL of UL, the same AL model can be used to predict the AL related information of UL and the AL related information of DL. Of course, it is not excluded that two AL models are used to predict the AL related information of UL and the AL related information of DL respectively.

[0147] In an example embodiment of the present disclosure, the input data and output data related to PDCCH detection of a specific base station can be collected by the terminal, and the association between the input data and the output data can reflect the PDCCH AL adaptive algorithm of the specific base station, which can be the form of the algorithm in a specific scenario. Therefore, the AI model can be obtained by learning the collected data, and the AL related information (e.g., the probability of AL, etc.) corresponding to the new input data can be predicted based on the obtained AI model.

[0148] As an example, according to the prediction result, the probability of AL with size 1 is 6%, the probability of AL with size 2 is 85%, the probability of AL with size 4 is 7%, and the probability of AL with size 8 is 2%, the AL order can be determined according to such prediction result. In the order of probability from large to small, the AL order can be determined as 2->4->1->8. The terminal can first determine the PDCCH with length of 2 CCEs at the starting position corresponding to AL=2, then determine the PDCCH with length of 4 CCEs at the starting position corresponding to AL=4, and so on, until the PDCCH is found or the predetermined detection number (round) is reached, wherein each AL can correspond to one detection. Assuming that the actual AL=2, if the blind detection starts from AL=1, two detections are needed to find the PDCCH, while the AI detection starting from AL=2 only needs one detection to find the PDCCH, so that the AI detection can reduce the number of detection rounds. Compared with blind detection, the number of PDCCH detections is reduced, the power consumption of PDCCH detection is reduced, and the purpose of power saving of the terminal (especially 5G terminal) is achieved.

[0149] In addition, when applied to a network based on an NR protocol, the AL can be 16. For example, when the probability of AL=16 is 0, the order of AL=16 is arranged after AL=8.

[0150] Figure 8 A flowchart of finding an AI model according to an example embodiment of the present disclosure is shown. Referring to Figure 8 , the process of finding an AI model can include the following steps: finding an AI model based on NCGI; determining whether the accuracy of the AI model meets the requirements; if the requirements are met, it means that an effective AI model is found, if the requirements are not met, it means that an effective AI model is not found, and the found AI model can be deleted.

[0151] Specifically, when the terminal is attached to a cell and registers a network service, the NCGI of the cell can be obtained. There can be a correspondence between NCGI and AI model, so the AI model can be found based on NCGI. The probability of successful prediction (i.e. prediction accuracy, or detection accuracy) can be counted when the number of detections reaches a predetermined number or the detection time reaches a predetermined time, and the AI model is determined to be effective when the prediction accuracy reaches a predetermined value.

[0152] As an example, the prediction accuracy can be the sum of the probabilities of successful detection of each round. Based on the trade-off between the amount of calculation and the statistical effect (accuracy of statistical results), it can be determined that the AI model is valid if the probability of successful detection of each of the first N (for example, N = 2) rounds reaches a predetermined value. Based on the statistical effect, the probabilities of successful detection of multiple rounds can be compared with a predetermined threshold (for example, the sum of the probabilities of the first three rounds is compared with the predetermined threshold, and the sum of the probabilities of the first four rounds is compared with the predetermined threshold), and if it is greater than or equal to the predetermined threshold, it is determined that the AI model is valid.

[0153] Whether the AI model is valid (whether a valid AI model is found) can be notified to the terminal. When the acquired AI model is not a valid AI model, the terminal deletes the acquired AI model or deletes the correspondence between the cell identifier (for example, NCGI) and the acquired AI model.

[0154] In order to quickly find an AI model based on NCGI at the terminal, a mapping table can be established and maintained to represent the correspondence between NCGI and AI model. The initial mapping table can be acquired (downloaded) from the server based on country, location, etc. information. The mapping table can record NCGI and AI model (model name, ID, etc. information) corresponding to NCGI. Table 1 shows a mapping table according to an example embodiment of the present disclosure.

[0155] Table 1

[0156] AI model list NCGI list AI model 1 NCGI 001, NCGI 003, NCGI 005,... AI model 2 NCGI 002, NCGI 028, NCGI 008,... …… …… AI model N NCGI 015,...

[0157] As shown in Table 1, each AI model can correspond to one or more NCGI, N is a natural number, and each row NCGI in the NCGI list can correspond to a series of base stations using the same PDCCH AL adaptive algorithm.

[0158] As an example, the terminal can manage the relationship between AI model and NCGI (for example, the above mapping table), and the management operation can include: finding the corresponding AI model based on NCGI; adding an item on the mapping table for recording the AI model and the corresponding NCGI list; when it is found that 2 or more AI models can be merged, merging (for example, merging according to categories).

[0159] As an example, the trigger conditions for merging include at least one of the following: the AI server pushes a message to the terminal, the message indicating that the AI model can be merged; evaluating the AI model using specific test data, the similarity between the outputs of multiple AI models exceeding a specific threshold; the coincidence degree of NCGI (the values of each item included in NCGI) (for example, the values of multiple items belonging to different NCGI are the same, indicating that the multiple items coincide) exceeds a predetermined threshold; the similarity of the input data of the AI model exceeds a predetermined threshold.

[0160] As an example, the merging manner can include deleting a specific AI model and adding the NCGI of the deleted AI model to the NCGI list corresponding to another AI model.

[0161] As an example, when an AI model is not applicable to the current network, for example, when a base station is found to be no longer applicable to a certain AI model, the NCGI corresponding to the AI model can be deleted from the NCGI list. The case that an AI model is not applicable to the current network includes that the prediction result of the AI model and the newly collected data conflict (for example, the predicted AL or the probability of the AL is different from the AL or the probability of the AL actually detected). The conflict can mean that a new base station is discovered (or connected to) or the configuration of the base station is changed.

[0162] As an example, the newly collected terminal data conflicts with the historical terminal data, which means that the PDCCH AL adaptive algorithm of the base station can be upgraded or modified, or the base station switches the PDCCH AL adaptive algorithm. In this case, the AI model is not applicable to the current network, and the corresponding AI model can also be deleted.

[0163] As an example, the above-mentioned deletion and merging operations can be respectively implemented by a deletion unit and a merging unit of the terminal or the UE.

[0164] In the example embodiments of the present disclosure, each AI model can learn the PDCCH AL adaptive algorithm of the base station in the gNB network, which can be bound to the type of base station equipment and the specific software version. In a real network environment, when the base stations of a certain operator in a certain area use the same base station equipment and the same software version, the AI models corresponding to these base stations are the same or can be merged. Considering that each base station can be uniformly identified by the NCGI, one row in the NCGI list is used to represent this series of base stations, and this series of base stations corresponds to one AI model, that is, this series of base stations uses the same PDCCH AL adaptive algorithm.

[0165] Figure 9A flow chart of AI detection according to the exemplary embodiments of the present disclosure is shown. The AI model can be referred to as a complete AI model, which can predict the probability of each AL and the like. After finding an effective and complete AI model, the terminal performs AI detection based on the found AI model, avoids PDCCH blind detection, and achieves the purpose of power saving. The process of AI detection can include: predicting the probability of AL, determining the AL order used in the detection process based on the predicted probability, and performing PDCCH detection according to the determined AL order. The AI detection described in the above embodiments is not repeated here. The present exemplary embodiments improve the AI detection to avoid repeated operations and further save power.

[0166] As an example, the terminal has found and obtained a complete and effective AI model suitable for the network in which the terminal is currently registered through NCGI. The process of AI detection can include: detecting whether the input data of the AI model (such as the input data described in the above embodiments, which can be obtained by the terminal) has changed, if the input data has changed, the AI model predicts the probability of Al and the like AL related information based on the new input data (changed input data), and determines the AL order used according to the predicted AL related information, if the input data has not changed, the previously used AL order is used to perform PDCCH detection without the need to perform the prediction operation, thereby avoiding repeated prediction.

[0167] As an example, the case where the input data changes includes but is not limited to: a) when base station switching occurs, the unique identifier (for example, NCGI) of the base station will change, and the input data associated with the NCGI changes; b) when the terminal changes or the network environment between the terminal and the base station changes, although the NCGI does not change, one or more of the input data changes, for example, at least one of RSRP, RSSI, CQI, SNR, DIC byte size, PDCCH slot index changes.

[0168] As an example, the terminal can store the value or probability of AL or store the AL order through the storage unit, and upload the stored content to the AI server for AI model training and / or judgment of the prediction accuracy of the AI model. Subsequently, it can be judged whether to enter the next round of detection, if yes (the prediction accuracy is high enough, for example, higher than a predetermined value), the step of judging whether the input data has changed is performed.

[0169] Figure 10Interaction between a terminal and an AI server according to an example embodiment of the present disclosure is shown. Embodiments of the present disclosure are described by taking an AI server as an example, but the AI server can also be replaced by any server, for example, a non-AI server. Example embodiments of the present disclosure are described for an example in which NCGI is used as a cell identifier, but this is not intended to be limiting, for example, other CGI can also be used to identify a cell.

[0170] According to example embodiments of the present disclosure, when a terminal cannot find a valid AI model locally, or the terminal finds that the local AI model does not meet the requirements (for example, the prediction accuracy does not meet the requirements), the terminal can try to obtain a suitable AI model from the AI server, and the terminal can also try to collect data related to PDCCH detection (data related to PDCCH AL adaptive algorithm) and upload it to the AI server (which can be located on the network side or set in the terminal). In this case, the terminal can update the AI model as needed, and the terminal can also upload the collected data (which the AI server may not have obtained) to the AI server in a timely manner.

[0171] As an example, the terminal has found an AI model through NCGI, but has not obtained a valid AI model suitable for the network in which the terminal is currently registered. When the terminal cannot find a valid AI model, the terminal can perform PDCCH blind detection. The terminal can try to collect data related to PDCCH detection (terminal data), and upload the collected data to the AI server after a certain threshold (such as 100) is reached. The terminal can also download and manage the AI model corresponding to the NCGI of the network in which the terminal is currently registered from the AI server based on the NCGI.

[0172] As an example, the AI server can be configured to: collect and store terminal data submitted by each terminal; train an AI model based on the collected terminal data; and provide each terminal with the trained AI model.

[0173] As an example, the terminal can upload the terminal data by at least one of the following: a) periodically submitting terminal data; b) submitting terminal data when the amount of collected terminal data exceeds a threshold; c) submitting terminal data according to a strategy negotiated with the AI server.

[0174] As an example, terminal data includes two parts, input X and output Y, where input X includes, but is not limited to, at least one of RSRP, RSSI, CQI, SNR, DCI payload, and time slot index; output Y includes DL AL related information and UL AL related information. Input X and output Y are one-to-one correspondence, a set of terminal data can be represented as (X, Y), and the base station to which the set of terminal data corresponds can be identified by a cell identifier (e.g., NCGI). Because the cell identifier (e.g., NCGI) can uniquely identify the base station, the AI server uses NCGI to distinguish terminal data belonging to different base stations. In addition, the AI server can use the unique identifier of the terminal and the time information to mark the source of the terminal data. Based on this, a series of terminal data marked with the same NCGI belonging to a specific base station can be represented as (X1, Y1), (X2, Y2), …, X N ,Y N ). Because the source of the terminal data does not need to be considered when training the AI model, the same NCGI is used to identify the terminal data (X1, Y1), (X2, Y2), …, X N ,Y N ) during AI training.

[0175] When the amount of terminal data collected by the AI server reaches a certain threshold (e.g., 10,000), the AI server will start training the AI model. In the AI server, the terminal data can be identified using a cell identifier (e.g., NCGI). The AI server can collect all terminal data belonging to a specific NCGI and train the AI model of the base station identified by the specific NCGI based on the collected data. This training of terminal data based on a single NCGI can be referred to as single-NCGI training.

[0176] In addition, since multiple base stations can use the same network equipment and the same software version, which means that a series of NCGIs can correspond to one AI model, the AI server can train the AI model based on terminal data corresponding to the series of NCGIs, so that the multiple base stations share one AI model. This training using terminal data corresponding to the series of NCGIs is referred to as multi-NCGI training.

[0177] Figure 11 Single-NCGI training and multi-NCGI training according to an example embodiment of the present disclosure are shown.

[0178] According to an example embodiment of the present disclosure, the AI model is trained when the terminal data corresponding to a specific base station or the terminal data corresponding to a series of base stations collected by the AI server accumulates to a certain extent.

[0179] As an example, a series of terminal data marked with the same NCGI belonging to a specific base station can be represented as (X1, Y1), (X2, Y2), …, XN ,Y N This represents terminal data corresponding to a specific NCGI. The numbers 1, 2, ..., N represent the sequence number of the terminal data. Because the AI ​​server can use the terminal identifier and time information to mark the source of the terminal data, the numbers 1, 2, ..., N can be associated with the terminal identifier and time information. Of course, the numbers 1, 2, ..., N may not represent the above associations.

[0180] As an example, (X1,Y1), (X2,Y2), ..., X N ,Y N Historical data can be used for NCGI training. Since historical data is related to the PDCCH AL adaptive algorithm, the output Y can be predicted using new input data X, for example, the probability of output AL, denoted as the probability function Y = P(Y|X), which simulates the PDCCH AL adaptive algorithm. As an example, the random forest algorithm can be used during training, but this disclosure is not limited to this; other algorithms are also feasible.

[0181] As an example, when the similarity between the trained AI model and a known AI model reaches a predetermined level, the NCGI corresponding to the trained AI model is added to the list of NCGI corresponding to the known AI model.

[0182] As an example, the merging conditions for AI models include at least one of the following: the similarity of the outputs of multiple AI models exceeds a threshold; the overlap of NCGI lists (the degree of repetition of NGCIs in the NCGI lists corresponding to multiple AI models) exceeds a threshold; and the similarity of the terminal data corresponding to NCGIs exceeds a threshold.

[0183] According to the example embodiments of this disclosure, multiple base stations may use the same network equipment and the same software version, which means that a series of NCGIs corresponding to the multiple base stations can correspond to one AI model, that is, the multiple base stations share one AI model.

[0184] As an example, the multi-NCGI training process includes using terminal data corresponding to the multiple base stations to train the AI ​​model. The AI ​​model can still be represented as a probability function: Y = P(Y|X).

[0185] As an example, the multi-NCGI training process includes: training an AI model corresponding to each base station using terminal data corresponding to each base station, and integrating the various AI models. Integration methods can include averaging (averaging the outputs of each AI model to arrive at the final output) or voting (relative majority voting, absolute majority voting, weighted majority voting).

[0186] As an example, when the trained AI model is different from the known AI model, the AI model and the corresponding NCGI can be added to the mapping table.

[0187] As an example, when the base stations to which the plurality of terminals belong use the same PDCCH AL adaptive algorithm, the plurality of terminals can use the same AI model, and the data of the plurality of terminals can be used together for training during training, so that the speed of collecting training data is faster, the data coverage is wider, and the accuracy of the trained AI model is higher.

[0188] If the AI model trained using the multi-NCGI is used, when the terminal performs network switching between base stations running the same algorithm, the terminal only needs to use one trained AI model, without switching.

[0189] Figure 12 A flowchart of AI detection according to an example embodiment of the present disclosure is shown.

[0190] According to an example embodiment of the present disclosure, the terminal has found and obtained the AI model of the network to which the terminal is currently registered through the NCGI. In the process of AI detection, it is detected whether the input data of the AI model has changed compared with the previous input data, if the input data has changed, the AI model is used to predict the AL with the greatest possibility (the AL with the greatest probability); if the input data has not changed, the previously predicted AL is used (for example, PDCCH detection is performed using the previous detection order). In the case where the input data has changed, PDCCH detection is performed based on the predicted AL with the greatest possibility. Subsequently, it is judged whether the PDCCH detection is successful, if the detection fails (PDCCH is not detected), PDCCH blind detection is performed, if the detection is successful (PDCCH is detected), it is judged whether to enter the next round of detection, if yes, the next round of PDCCH detection is entered.

[0191] As an example, the success rate of PDCCH detection based on the AL with the greatest possibility is about 80%, which means that using the AI model can reduce 80% or even more PDCCH blind detection. At the same time, the AI model only predicts the AL with the greatest possibility and does not predict all ALs, such an AI model can be called a simplified AI model. Such an AI model gives up the probability prediction of ALs other than the AL with the greatest possibility, has lower requirements for computing power and storage space, and can be applied to smart watches, bracelets and other Internet of Things (IOT) devices with weak computing power, and has higher universality.

[0192] According to the exemplary embodiments of the present disclosure, the training of the AI model can be selected to be jointly performed by the terminal and the AI server. When the terminal jointly trains the AI model with the AI server, the terminal can first download a general or basic version of the AI model suitable for the current specific base station from the AI server. When the terminal accumulates corresponding terminal data in use, and the accumulated data exceeds a certain threshold, the terminal re-trains the AI model based on the downloaded AI model and the collected local data. The new AI model obtained by the training can be uploaded to the AI server. When the AI server receives AI models from the terminal exceeding a certain threshold, the server re-trains based on the collected AI models to obtain the latest AI model.

[0193] Figure 13 A flowchart showing the joint training process of the base station and the AI server according to the exemplary embodiments of the present disclosure.

[0194] According to the exemplary embodiments of the present disclosure, the terminal has found and obtained the AI model of the network currently registered by the terminal through the NCGI. The terminal downloads a general or basic version of the AI model from the AI server based on the NCGI of the currently registered network. The general or basic version of the AI model can be shared by users, and can be trained based on historical data of PDCCH detection corresponding to one or a series of cells (which can be determined by a cell identifier such as NCGI). Because of the influence of data or the coverage of the AI model, the general or basic version of the AI model may not be able to perform the prediction well, and therefore needs to be retrained.

[0195] As an example, when the terminal data accumulated by the terminal exceeds a certain threshold (such as more than 1000 data), the terminal re-trains the AI model based on the downloaded AI model and the collected data.

[0196] The terminal re-trains the downloaded AI model using the personalized terminal data collected locally based on the downloaded general or basic version of the AI model. That is, the training is divided into two parts, first, the general or basic version of the AI model is obtained by training on the AI server side, and then the local collected data is used to re-train on the terminal side, so that a more suitable AI model for the user's daily use scenario can be obtained. The traditional AI training is mostly to collect data by the terminal, upload the data to the AI server, and train uniformly by the AI server.

[0197] As an example, the terminal can choose to upload the trained AI model to the server. When the server receives AI models corresponding to a certain NCGI exceeding a certain threshold (such as more than 50), the server re-trains based on the collected AI models to obtain the latest AI model.

[0198] If the user agrees to upload the AI model trained by the terminal to the AI server, the AI server can collect a large number of AI models obtained by retraining based on the personalized data of the user's resident base station. These models contain personalized data of the user's resident base station, and if the server performs iterative training of the AI model based on the set of all terminal-trained AI models, the server-side AI model will also be able to obtain the benefits of personalized data. Make the prediction accuracy of the server-side AI model higher.

[0199] Due to the use of joint training, the corresponding input is changed from terminal data to AI model uploaded by the terminal. The server performs iterative training of the AI model on the network side based on the terminal-trained AI model, thereby continuously improving the prediction accuracy of the AI model. The terminal can train locally without uploading user data to the server and upload the trained model, which can reduce the leakage of user privacy and reduce the operating cost of protecting user privacy. Because the AI model trained by the local data implies the user's personalized data, it can better provide personalized services for users. In addition, only the AI model needs to be uploaded instead of the original user data, which can reduce the amount of data uploaded and thus reduce resource and cost consumption. The joint training of the terminal and the AI server can be completed using federated learning.

[0200] In order to verify the real effect of the exemplary embodiments of the present disclosure, an AI model is obtained based on a random forest algorithm. The AI model based on random deep forest is trained by terminal data in an external field, and the AI model is tested using terminal data in another external field. The test results show that the prediction accuracy of the results determined using the AI model is more than 80%.

[0201] For the AI model, one training and multiple predictions can be performed; on the other hand, model training can be completed by the server rather than the terminal, and the computational complexity of the terminal can be negligible.

[0202] Using PDCCH intelligent detection requires considering the increased operation of using the AI model process (prediction based on the AI model). The complexity of the use process of the AI model trained by the random forest is proportional to the product of K and P, where K represents the number of trees and P represents the depth of the tree. Compared with the complexity of channel decoding, this complexity can be almost negligible.

[0203] Assuming that in the PDCCH configuration, the probability of occurrence of the five ALs (AL=1, AL=2, AL=4, AL=8, and AL=16) is uniformly distributed, and the complexity required for detecting one PDCCH candidate of different AL levels is proportional to the PDCCH length, i.e., the PDCCH AL level, the complexity of PDCCH blind detection is approximately 67.6 units. Assuming that the PDCCH AI model can accurately find the PDCCH AL level, the complexity of PDCCH detection is approximately 11.6 units, and the detection complexity of PDCCH can be reduced by approximately 83%. If the PDCCH AI model can accurately find the PDCCH AL level with a probability of 80%, the complexity of PDCCH detection is approximately 22.8 units, and the detection complexity of PDCCH can be reduced by approximately 66%.

[0204] It can be seen that the method in the present disclosure can significantly reduce the detection complexity of PDCCH, thereby effectively reducing the power consumption in the PDCCH detection process.

[0205] Meanwhile, based on the interface signaling between the base station and the terminal, the base station confirms the capability information of the terminal, configures different DRX parameters for terminals with different capability information, and can further improve the system transmission throughput and reduce the delay.

[0206] According to another exemplary embodiment of the present disclosure, a computer readable storage medium storing instructions is provided, wherein when the instructions are run by at least one computing device, the at least one computing device is caused to perform the detection method and / or the configuration method as described above.

[0207] The above describes embodiments according to the concept of the present disclosure, and features in various embodiments can be combined without departing from the scope of protection of the present disclosure, and such combinations will also fall within the scope of protection of the present disclosure.

[0208] The computer readable storage medium is any data storage device that can store data read by a computer system. Examples of the computer readable storage medium include a read-only memory, a random access memory, a read-only optical disc, a magnetic tape, a floppy disc, an optical data storage device, and a carrier wave (such as data transmission through the Internet via a wired or wireless transmission path).

[0209] In addition, it should be understood that each unit of the terminal and the base station according to the exemplary embodiments of the present disclosure can be implemented by hardware components and / or software components. A person skilled in the art can implement each unit, for example, using a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC), according to the processing performed by each unit as defined.

[0210] Further, a method according to an exemplary embodiment of the present disclosure can be implemented as computer code in a computer readable storage medium. A person skilled in the art can implement the computer code according to the description of the above-described method. The above-described method of the present disclosure is implemented when the computer code is executed in a computer.

[0211] While some exemplary embodiments of the present disclosure have been expressed and described, it should be understood by those skilled in the art that modifications can be made to the embodiments without departing from the principles and spirit of the present disclosure, the scope of which is defined by the claims and their equivalents.

Claims

1. A method performed by a terminal in a wireless communication system, wherein, The method comprises: determining a prediction accuracy of an aggregation level AL prediction model, wherein the prediction accuracy of the AL prediction model is determined based on a detection success probability calculated from each round of one or more rounds; determining whether the AL prediction model is valid based on the prediction accuracy of the AL prediction model and a preset value, wherein the AL prediction model is determined to be valid based on the prediction accuracy of the AL prediction model being equal to or greater than the preset value; in a case where the AL prediction model is determined to be valid, predicting AL-related information based on the AL prediction model, wherein the AL-related information comprises information about an AL of a control channel element CCE related to a physical downlink control channel PDCCH; determining a PDCCH detection order based on the predicted AL-related information; performing PDCCH detection based on the determined PDCCH detection order.

2. The method of claim 1, further comprising: obtaining input data required for predicting the AL-related information; predicting the AL-related information based on the AL prediction model, comprising: predicting the AL-related information based on the obtained input data and the AL prediction model.

3. The method of claim 2, wherein the input data is related to report data for determining the AL of the CCE sent to the base station.

4. The method of claim 2 or 3, wherein, The input data required for predicting the AL-related information comprises: link information; and / or The AL-related information comprises at least one of the following: a detection order for a plurality of ALs, a probability of each of the plurality of ALs, and an AL with the highest probability in the plurality of ALs.

5. The method of claim 4, wherein, The link information comprises at least one of the following: a reference signal received power RSRP, a channel quality indicator CQI, a signal-to-noise ratio SNR, a downlink control information DCI payload, and a PDCCH slot index.

6. The method of claim 4 or 5, wherein, If the AL-related information comprises an AL with the highest probability in the plurality of ALs, performing PDCCH detection based on the predicted AL-related information comprises: performing PDCCH detection based on the predicted AL with the highest probability; if the detection fails, performing PDCCH blind detection based on other ALs.

7. The method of claim 1, wherein, Predicting the AL-related information based on the AL prediction model comprises: determining a probability of detecting a PDCCH for each of a plurality of ALs; wherein determining the PDCCH detection order based on the predicted AL-related information comprises: determining the PDCCH detection order based on the probability for each of the plurality of ALs.

8. The method of claim 7, wherein, The PDCCH detection order is determined to be detecting a CCE corresponding to an AL with the largest probability value in the plurality of ALs first.

9. The method of any one of claims 1 to 8, further comprising: obtaining a cell identifier of a cell to which the terminal is connected; and determining the AL prediction model based on the cell identifier. Determining the AL prediction model based on the cell identifier comprises:

10. The method of claim 9, wherein, looking up the AL prediction model corresponding to the cell to which the terminal is connected based on the cell identifier and a correspondence relationship between the AL prediction model and the cell identifier. ​ predicting AL-related information based on the AL prediction model, including: predicting the AL-related information based on the found AL prediction model.

11. The method of claim 10, wherein, The method further includes: determining whether to merge at least two AL prediction models according to at least one of the following: a message indicating to merge AL prediction models, similarity between outputs of the at least two AL prediction models, coincidence degree of a cell identifier list, and similarity between inputs of the at least two AL prediction models; modifying a correspondence between cell identifiers and AL prediction models, so that each cell identifier corresponding to a merged AL prediction model corresponds to the same AL prediction model.

12. The method of any one of claims 9-11, wherein, The cell identifier includes a New Radio Cell Global Identifier (NCGI), and the NCGI includes at least one of the following: a Mobile Country Code (MCC), a Mobile Network Code (MNC), and a Next Generation Node (gNB) identifier.

13. The method of any one of claims 1 to 12, wherein, The method further includes: acquiring the constructed AL prediction model from the terminal locally or from a server; updating the acquired AL prediction model based on report data for determining AL of a CCE sent to a base station and result information detected by the PDCCH.

14. The method of any one of claims 1 to 13, wherein, The method further includes: reporting PDCCH detection capability information to a base station, wherein the PDCCH detection capability information can represent whether the terminal can perform PDCCH detection based on predicted AL-related information.

15. The method of claim 14, wherein, The PDCCH detection capability information is transmitted using Radio Resource Control (RRC) signaling.

16. The method of claim 14 or 15, wherein, The method further includes: receiving information about a Discontinuous Reception (DRX) cycle from a base station, wherein the DRX cycle is determined according to PDCCH detection capability information; performing PDCCH detection based on predicted AL-related information, including: performing PDCCH detection based on the acquired DRX cycle.

17. The method of claim 16, wherein, The DRX cycle configured by the base station for a terminal that can perform PDCCH detection based on predicted AL-related information is shorter than the DRX cycle configured for a terminal that cannot perform PDCCH detection based on predicted AL-related information.

18. The method of claim 16, wherein, The DRX cycle of the terminal is shorter than the DRX cycle of another terminal that performs blind detection on a PDCCH.

19. The method of any one of claims 1 to 7, wherein, Performing PDCCH detection based on predicted AL-related information, including: determining whether input data required for currently acquired predicted AL-related information has changed compared to historical input data; when it is determined that there has been a change, predicting the AL-related information using the AL prediction model; The method further includes: when it is determined that there has been no change, performing PDCCH detection using AL-related information predicted based on historical input data.

20. A method performed by a base station in a wireless communication system, wherein, The method includes: receiving Physical Downlink Control Channel (PDCCH) detection capability information from a terminal; configuring a Discontinuous Reception (DRX) cycle according to the received PDCCH detection capability information, wherein the PDCCH detection capability information can represent whether the terminal can perform PDCCH detection based on predicted AL-related information, and wherein the AL-related information includes information about AL of a Control Channel Element (CCE) related to a PDCCH, wherein, in a case that the PDCCH detection capability information represents that the terminal is capable of performing PDCCH detection based on predicted AL-related information, the base station configures a DRX cycle for the terminal capable of performing PDCCH detection based on predicted AL-related information to be shorter than a DRX cycle configured for a terminal incapable of performing PDCCH detection based on predicted AL-related information, wherein, in a case that the terminal determines that an AL prediction model for predicting AL-related information is valid according to a prediction accuracy of the AL prediction model, the PDCCH detection capability information represents that the terminal is capable of performing PDCCH detection based on predicted AL-related information, wherein, the prediction accuracy of the AL prediction model is determined based on a detection success probability calculated from each of one or more rounds, wherein, the AL prediction model is determined to be valid based on the prediction accuracy of the AL prediction model being equal to or greater than a preset value, wherein, the predicted AL-related information is used to determine a PDCCH detection order, and PDCCH detection is performed based on the determined PDCCH detection order.

21. The method of claim 20, wherein, the capability information is carried by an information element (IE) on radio resource control (RRC) signaling.

22. A terminal, wherein, the terminal comprises: a prediction unit configured to determine a prediction accuracy of an aggregation level (AL) prediction model, wherein the prediction accuracy of the AL prediction model is determined based on a detection success probability calculated from each of one or more rounds, and determine whether the AL prediction model is valid based on the prediction accuracy of the AL prediction model and a preset value, wherein the AL prediction model is determined to be valid based on the prediction accuracy of the AL prediction model being equal to or greater than the preset value, and predict AL-related information based on the AL prediction model in a case that the AL prediction model is determined to be valid, wherein the AL-related information comprises information about an AL of a control channel element (CCE) related to a physical downlink control channel (PDCCH); a detection unit configured to determine a PDCCH detection order based on predicted AL-related information, and perform PDCCH detection based on the determined PDCCH detection order.

23. A terminal, wherein, the terminal comprises a processor; and a memory configured to store machine-readable instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-19.

24. A base station, wherein, the base station comprises: a communication unit configured to receive physical downlink control channel (PDCCH) detection capability information from a terminal; a configuration unit configured to configure a discontinuous reception (DRX) cycle according to the received PDCCH detection capability information, wherein, the PDCCH detection capability information is capable of representing whether the terminal is capable of performing PDCCH detection based on predicted AL-related information, wherein the AL-related information comprises information about an AL of a control channel element (CCE) related to a physical downlink control channel (PDCCH), wherein, in a case that the PDCCH detection capability information represents that the terminal is capable of performing PDCCH detection based on predicted AL-related information, the configuration unit configures a DRX cycle for the terminal capable of performing PDCCH detection based on predicted AL-related information to be shorter than a DRX cycle configured for a terminal incapable of performing PDCCH detection based on predicted AL-related information, wherein, in a case that the terminal determines that the AL prediction model for predicting AL-related information is valid according to a prediction accuracy of the AL prediction model, the PDCCH detection capability information represents that the terminal is capable of performing PDCCH detection based on predicted AL-related information, wherein, the prediction accuracy of the AL prediction model is determined based on a detection success probability calculated from each round of one or more rounds, wherein, it is determined that the AL prediction model is valid based on the prediction accuracy of the AL prediction model being equal to or greater than a preset value, wherein, the predicted AL-related information is used to determine a PDCCH detection order, and PDCCH detection is performed based on the determined PDCCH detection order.

25. A base station, wherein, The base station comprises: a processor; and a memory configured to store machine-readable instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 20-21.

26. A computer-readable storage medium storing instructions, wherein, The instructions, when executed by at least one computing device, cause the at least one computing device to perform the method of any one of claims 1-21.

Images