Data processing method, device and equipment

Abstract

The embodiment of the invention provides a data processing method, device and equipment. In the embodiment of the invention, a prediction request message from a request node can be received; executing a prediction operation according to the prediction request message, and generating prediction information; wherein the prediction information comprises a plurality of prediction values corresponding to a plurality of different time intervals in the future; and sending a feedback message to the request node, wherein the feedback message carries the prediction information containing the plurality of prediction values.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to a data processing method, apparatus, and device. Background Technology

[0002] With the deep integration of fifth-generation mobile communication technology (5G) and artificial intelligence (AI), intelligent management and optimization of wireless networks has become an important development trend. In 3GPP R17 and subsequent standards, research has been conducted on... Figure 1 The AI ​​model shown is applied to the architecture of Radio Resource Management (RRM) to achieve precise optimization of network performance through data-driven prediction and decision-making. Under this architecture, nodes in the network (such as base stations or centralized units) can request and provide each other with predicted values ​​of future network conditions (such as load and number of users), thereby providing forward-looking data support for key functions such as handover and load balancing.

[0003] In related technologies, the process of exchanging predictive information between nodes is as follows: Figure 2 As shown, this interaction process typically follows a strict periodic pattern. Specifically, the requesting node (such as the serving base station) sends a prediction request to the prediction node (such as a neighboring base station), specifying a fixed period T. After accepting the request, the prediction node predicts the state value for the next period T and feeds it back to the requesting node; thereafter, every time period T elapses, the prediction node performs the prediction again and feeds back the predicted value for the new period T. This "one request, periodic feedback" pattern constitutes the basic interaction process for related technologies to achieve periodic prediction.

[0004] However, in related technologies, the prediction period and feedback period are completely synchronized, both being T. This synchronization mechanism results in a severe shortage of valid prediction data held by the requesting node at any non-feedback time. Therefore, there is an urgent need for a method that can ensure the continuous validity of prediction data to enhance the practical application value of periodic feedback predictions. Summary of the Invention

[0005] This application provides a data processing method, apparatus, and device to address the problem in related technologies that the effective prediction data held by requesting nodes at any non-feedback time is severely insufficient.

[0006] This application provides a data processing method, including: Receive prediction request messages from the requesting node; Based on the prediction request message, a prediction operation is performed, and prediction information is generated; wherein, the prediction information contains multiple prediction values ​​corresponding to multiple different time intervals in the future; A feedback message is sent to the requesting node, the feedback message carrying the prediction information containing the multiple prediction values.

[0007] This application also provides a data processing apparatus, including: The receiving module is used to receive prediction request messages from the requesting node; The generation module is used to perform a prediction operation and generate prediction information based on the prediction request message; wherein the prediction information includes multiple prediction values ​​corresponding to multiple different time intervals in the future. The sending module is used to send a feedback message to the requesting node, the feedback message carrying the prediction information containing the multiple prediction values.

[0008] This application also provides an electronic device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. The processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform steps in the data processing method provided in this application.

[0009] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the data processing method provided in this application.

[0010] This application also provides a computer program product that stores instructions that, when executed by a computer, cause the computer to perform the steps in the data processing method provided in this application.

[0011] The data processing method provided in this application can generate multiple predicted values ​​corresponding to multiple different time intervals in the future based on the predicted request information of the request node, and feed them back to the request node. This makes the predicted value of the request node a future predicted data sequence at any time. Therefore, when it needs to make a real-time decision at any time, it can directly call or calculate the predicted information of the target time period at the current time from the multiple predicted values ​​corresponding to multiple different time intervals in the future that are stored locally, without waiting for the next periodic feedback. This effectively increases the amount of effective predicted data held by the request node at any non-feedback time. Attached Figure Description

[0012] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1A schematic diagram of the basic functional architecture of AI; Figure 2 This is a schematic diagram illustrating the interaction process of prediction information between nodes in related technologies. Figure 3 This diagram illustrates the problems existing in the predictive information interaction process between nodes in related technologies. Figure 4 A flowchart illustrating a data processing method provided for an exemplary embodiment of this application; Figure 5 A schematic diagram illustrating the interactive process of a data processing method provided in an exemplary embodiment of this application; Figure 6 A schematic diagram illustrating another implementation interaction flow of the data processing method provided in an exemplary embodiment of this application; Figure 7 A schematic diagram of the structure of a data processing apparatus provided for an exemplary embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of this application. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] The following is a description of the terms used in this application: Radio Resource Management (RRM) is a core technology in mobile communication systems that dynamically allocates resources such as radio channels and power to ensure quality of service and improve spectrum efficiency. It encompasses basic modules such as power control, channel allocation, and handover control, and extends to novel methods such as dynamic spectrum sharing. This technology achieves collaborative management of hardware and software resources through space-time processing techniques such as smart antennas and joint detection, significantly improving code channel multiplexing in TD-SCDMA systems. Modern RRM, combined with dynamic EDCA parameter adjustment and intelligent roaming mechanisms, can automatically optimize channel contention parameters, implement terminal migration, and air interface slicing to ensure low-latency services.

[0015] TD-SCDMA is short for Time Division-Synchronous Code Division Multiple Access. It's an internationally recognized and widely accepted wireless communication standard, primarily based on Chinese intellectual property, and officially listed by the International Telecommunication Union (ITU) as one of the third-generation mobile communication air interface technology specifications. The "TD" in TD-SCDMA stands for Time Division Multiplexing, meaning that in a TD-SCDMA system, a single user can communicate bidirectionally (transmit and receive) simultaneously using TDD (Time Division Duplex). Different time slots are allocated within the same frequency band for uplink and downlink duplex communication, facilitating flexible switching between uplink and downlink. For example, the time slot allocation and switching points between uplink and downlink can be determined based on the different uplink and downlink resource requirements of different services, thereby achieving efficient support for all 3G symmetric and asymmetric services. Compared to FDD mode, TDD can operate on unpaired radio frequency spectrum, thus offering significant advantages in the current complex spectrum allocation situation. TD-SCDMA supports various 3G services, including voice, video calls, and the Internet, with speeds ranging from 8kb / s to 2Mb / s and even higher, through optimal adaptive resource allocation and optimal spectrum efficiency.

[0016] As described in the background section, the prediction period and feedback period in related technologies are completely synchronized, both being T. This synchronization mechanism results in a severe shortage of effective prediction data held by the requesting node at any non-feedback time. Specifically, at the beginning of each period T, the requesting node possesses the newly received prediction data for the current period; however, as time progresses, towards the end of period T, the prediction for the current period is about to expire, while the prediction for the next period has not yet arrived, resulting in a gap in the prediction data available for decision-making by the node. Figure 3 As shown, when node 1 makes an RRM decision at the highlighted time point, it can only obtain part of the valid prediction data, while the other part of the data needs to wait for subsequent message reports. Due to the highly real-time nature of network state changes, Figure 3 The RRM decisions shown (such as handover decisions and resource scheduling) cannot wait for the next prediction feedback and must be enforced even with incomplete data. This inevitably leads to a decrease in prediction accuracy, which in turn causes a series of network performance problems such as handover failure, resource allocation imbalance, interference management failure, and load balancing lag.

[0017] To address the aforementioned problems in related technologies, embodiments of this application provide a data processing method, apparatus, and device. Because it can generate multiple predicted values ​​corresponding to multiple different future time intervals based on the predicted request information of the requesting node, and feed these values ​​back to the requesting node, the predicted values ​​of the requesting node are always a future predicted data sequence. Therefore, when it needs to make real-time decisions at any time, it can directly call or calculate the predicted information for the currently required target time period from the multiple predicted values ​​corresponding to multiple different future time intervals stored locally, without waiting for the next periodic feedback. This effectively increases the amount of effective predicted data held by the requesting node at any non-feedback time.

[0018] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0019] Figure 4 This is a schematic flowchart illustrating a signal processing method provided for an exemplary embodiment of this application. Figure 4 As shown, the method includes: Step 110: Receive the prediction request message from the requesting node.

[0020] The prediction request message includes a parameter indicating a one-time feedback of prediction values ​​for the next n periods, where n is an integer greater than 1. This embodiment of the application, by introducing the parameter n, can transform the feedback mode of prediction information from a single-point + serial mode in related technologies to a sequential + parallel mode.

[0021] In some exemplary embodiments, this application provides two prediction time methods to address the differentiated needs for the "granularity" and "scope" of prediction information in different scenarios, thereby enhancing the adaptability of the solution and the effectiveness of decision support. Specifically, the prediction request message also includes a prediction time method selection parameter, which indicates that the prediction period corresponding to each of the multiple prediction values ​​can be defined using any of the following methods: The prediction period length corresponding to the i-th prediction value is i times the duration of a single period, and all prediction periods have the same start time. The prediction period length corresponding to the i-th prediction value is the duration of a single period, and the start time of the i-th prediction period is the end time of the (i-1)-th prediction period.

[0022] In some exemplary embodiments, the prediction request message includes a prediction period parameter T2 and a feedback period parameter T1, wherein T2 > T1.

[0023] For Method 1, the prediction period length corresponding to the i-th predicted value is i times the duration of a single period, and all prediction periods start at the same time. This method can be summarized as period accumulation with the same starting point, providing a "nested" macro-trend prediction view. This method is typically suitable for RRM decisions requiring long-term strategic planning. In this method, the first predicted value (Pre1) reflects the short-term (T) situation, the second predicted value (Pre2) reflects the average trend in the medium term (2T), and so on. This structure allows requesting nodes to clearly assess the short-term fluctuations and long-term evolution of the network state. For example, when making load balancing decisions, nodes are not only concerned with which cell has the lightest load in the next period (see Pre1), but also with which cell can maintain a stable low load for a longer period of time (such as 3T, 4T) (see Pre3, Pre4), thus making more durable and stable load migration decisions and avoiding the signaling overhead and instability caused by frequent switching.

[0024] For Method 2, the prediction period length corresponding to the i-th predicted value is the duration of a single period, and the start time of the i-th prediction period is the end time of the (i-1)-th prediction period. This method can be summarized as a fixed period with consecutive beginnings and end, providing a high-resolution continuous prediction view with "uniform slicing." This method is suitable for real-time or near-real-time decision-making scenarios with high continuity requirements and fine-grained management. In this method, each predicted value covers a standard duration T, and the periods are seamlessly connected, forming a continuous future timeline. This provides a perfect foundation for requesting nodes to perform calculations accurate to specific time points. As in the example in the disclosure document, when a node needs to know the predicted value for any time period from 0.8T to 1.3T, it can accurately find the two predicted values ​​Pre1 (covering 0T~1T) and Pre2 (covering 1T~2T) that overlap with this time period and perform high-precision weighted fusion. This method is very suitable for use in triggering strict switching decisions or avoiding transient congestion.

[0025] Figure 5 This is a schematic diagram illustrating an interactive flow of a data processing method provided as an exemplary embodiment of this application. Figure 5 In the prediction request message sent by node 1, a period T may be included, and at least one of the following parameters may be included: 1. Parameter n is used to indicate the parameter that provides a one-time feedback of the predicted values ​​for the next n periods, where n is an integer greater than 1.

[0026] 2. Predict the start time t pre This parameter is used to specify the start time of the prediction. If this parameter is not included, the current time is used as the prediction start time by default. This parameter can be represented in any of the following forms: timestamp, frame number, or subframe number.

[0027] 3. Feedback start time t feed This parameter specifies the start time for feeding back the predicted value. If this parameter is not included, it defaults to feeding back the predicted value immediately after it is generated. This parameter can be represented by any of the following forms: timestamp, frame number, or subframe number.

[0028] 4. Selection of prediction time method sel This includes two prediction methods, as detailed in Table 1. If this parameter is included, it specifies whether prediction time method 1 or 2 is used. If this parameter is not included, method 1 or method 2 is used according to the default rules of the sender and receiver.

[0029] Table 1 Forecasting Time Methods

[0030] The sending time of message 2 can be referenced from the predicted request message. Subsequent messages 3 and 4, etc., are sent sequentially at intervals T. The predicted value fed back in message 2 contains n future predicted values. : 1. The first predicted value is the predicted value for the time period T1; 2. The second predicted value may be a prediction for the period T2, or a prediction for the periods T1 and T2, depending on the prediction time method. 3. The third predicted value may be a prediction for the period T3, or it may be a prediction for the periods T1, T2 and T3, depending on the prediction time method. And so on.

[0031] The predicted value fed back in message 3 also includes the predicted values ​​for the next n periods. Similar to message 2, it is the predicted value information for the next n periods starting from the time of message 3.

[0032] Follow-up news 4 and its subsequent news can be deduced in the same way.

[0033] Step 120: Perform a prediction operation based on the prediction request message and generate prediction information; wherein the prediction information contains multiple prediction values ​​corresponding to multiple different time intervals in the future.

[0034] In some exemplary embodiments, provided that the prediction request message includes parameters for indicating a one-time feedback of prediction values ​​for the next n periods, the multiple prediction values ​​in the prediction information are prediction values ​​for each of the next n consecutive periods.

[0035] Specifically, firstly, by Figure 3It is known that under a single-cycle feedback mechanism, when the requesting node (Node 1) makes an RRM decision at the end of cycle T, it faces a "data vacuum period" where the current cycle's predicted value is about to expire and the next cycle's predicted value has not yet arrived, leading to insufficient decision-making basis. This application addresses this issue by acquiring the predicted values ​​for the next n cycles at once, ensuring that the requesting node immediately obtains a continuous sequence of predicted data covering the next n*T time period at any feedback time. Therefore, regardless of... Figure 3 As shown, if a decision needs to be made at any point within the cycle, node 1 can directly find or calculate the required prediction information from the locally stored continuous data sequence, completely eliminating decision delays and data gaps caused by waiting for the next feedback cycle.

[0036] Secondly, when a user is at the edge of the cell and needs an emergency handover, Figure 3 Node 1, as shown, no longer relies solely on a single, potentially expiring forecast value. Instead, it comprehensively examines the load and user number forecast trends of the target cell over multiple future periods (n periods). This allows the system to more accurately determine which target cell can provide more stable and high-quality service in the coming period, thereby reducing handover failures and unnecessary handovers. Moreover, Node 1 can perform proactive load scheduling based on the complete forecast sequence of multiple surrounding cells over the next n periods. It can identify cells that are about to experience continuous overload and preemptively distribute services to cells with lighter loads, avoiding the passive situation of discovering and addressing overload only after data lag, as seen in related technologies.

[0037] Finally, the sequence of predicted values ​​for n consecutive periods provided in this application embodiment is a prerequisite for the implementation of subsequent processing procedures. The requesting node can utilize this series of data to intelligently fuse a smoother and more reasonable "final predicted value" for an arbitrarily specified target time period by calculating the overlap with multiple prediction periods. This processing method reduces the absolute dependence on the accuracy of a single prediction data point, and through statistical averaging effects, makes the final value used for decision-making more robust and reliable.

[0038] In some exemplary embodiments, the prediction request message includes a prediction period parameter T2 and a feedback period parameter T1. When T2 > T1, the multiple prediction values ​​in the prediction information are prediction values ​​corresponding to M sub-intervals divided in time order within a prediction period of length T2, where M is an integer greater than 1.

[0039] By decoupling the prediction period T2 and feedback period T1 (T2>T1) using the above method, the real-time performance and accuracy issues of long-period predictions are resolved, achieving a balance between prediction coverage and feedback overhead. Specifically, by setting T2>T1, each requesting node receives a prediction covering a longer future period (T2) rather than just the next short period. This ensures that nodes always have sufficient periodic data to handle upcoming events, meeting the needs of long-term planning. This embodiment, by frequently feeding back with a shorter feedback period parameter T1, essentially performs rolling updates and corrections to the longer prediction period parameter T2. Each new feedback (message 3, message 4, etc.) is based on the latest network state, re-predicting the time intervals within the T2 period that have not yet arrived. This allows long-term predictions to be continuously replaced and optimized by predictions closer to real-time over time, effectively suppressing the accumulation of prediction errors.

[0040] Figure 6 This is a schematic diagram illustrating another implementation of the data processing method provided in an exemplary embodiment of this application. Figure 6 In the message sent by requesting node 1 to send a prediction request, at least one of the following parameters may be included: 1. Feedback period T1, used to indicate that node 2 periodically feeds back the predicted value according to the T1 time interval; 2. Prediction period T2, used to indicate the predicted value information of node 2 within the future time period T2; 3. Predict the start time t pre This parameter is used to specify the start time of the prediction; if this parameter is not included, the current time is used as the prediction start time by default; this parameter can be represented in any of the following forms: timestamp, frame number, or subframe number. 4. Feedback start time t feed This parameter specifies the start time for feeding back the predicted value; if this parameter is not included, it defaults to feeding back the predicted value immediately after it is generated; this parameter can be represented in any of the following forms: timestamp, frame number, or subframe number.

[0041] When the prediction period T2 is relatively long, given that the predicted values ​​are difficult to maintain stability over a long period, node 2 can provide multiple predicted values ​​within that prediction period. Each predicted value corresponds to a different time interval within the prediction period T2.

[0042] When the duration of the prediction period T2 is greater than the feedback period T1, the interaction process is as follows: Figure 6 As shown. Wherein: The sending time of message 2 is determined by the prediction request message. t feedOnce the parameters are determined, subsequent messages such as message 3 and message 4 can be sent sequentially according to the time interval of feedback period T1.

[0043] The prediction period start time for the predicted value reported in message 2 is determined by the parameters in the prediction request message. t pre It is determined that the prediction period termination time is the starting time point plus the prediction period T2.

[0044] The prediction period start time for the predicted value reported in message 3 is the time specified in the prediction request message. t pre Including the feedback period T1, the prediction period ends at the start time of the message plus the prediction period T2.

[0045] Similarly, the prediction period start time of the predicted value fed back by message n is the time specified in the prediction request message. t pre Adding a feedback period of (n-2) times T1, the prediction period ends when the start time of the message is added to the prediction period T2.

[0046] Node 1 can obtain the feedback at the time of feedback. t pre The predicted values ​​for the next time period T2 are given, where T2 contains n predicted values. Each predicted value corresponds to the prediction result for the time period T2 / n; in terms of time distribution, For the most recent time period, the predicted value Secondly, The predicted value corresponds to the furthest future time period.

[0047] Step 130: Send a feedback message to the requesting node. The feedback message carries prediction information containing multiple prediction values.

[0048] In some exemplary embodiments, to transform the received batch prediction data into effective information that can be used for accurate decision-making, after sending a feedback message to the requesting node, the method provided in this application embodiment further includes: Determine one or more predicted values ​​that overlap with the target time period by requesting nodes; Calculate the overlap duration between the prediction period and the target time period for each overlapping prediction value; Weights are assigned to the predicted values ​​of each overlap based on the overlap duration; The overlapping predicted values ​​are weighted and averaged to obtain the final predicted value.

[0049] First, the multiple predicted values ​​received by the requesting node, whether from method one or method two, are for specific, discrete time intervals. However, actual RRM decisions (such as...) Figure 3The target time period required for the decision on the red-marked moment is often arbitrary and may not perfectly align with the boundary of the received prediction period. This application's embodiments construct a continuous prediction function by determining overlapping prediction values, calculating the overlap duration, and applying a weighted average. This allows the requesting node to calculate a reasonable prediction value for any starting point and any target window length (e.g., the time period corresponding to the 0.8T to 1.3T time window), thereby achieving seamless stitching and on-demand customization of prediction data in the time dimension.

[0050] Secondly, the influence of a predicted value on any point in time within its corresponding period is not equal. For a target time period spanning multiple prediction periods, the contribution of each predicted value should be proportional to its time proportion. For example, when calculating the predicted value for the time period from 0.8T to 1.3T, it overlaps with the first prediction period (0T~1T) by 0.2T and with the second prediction period (1T~2T) by 0.3T. Based on this, the embodiments of this application assign a higher weight to the second predicted value (weight 0.6) than to the first predicted value (weight 0.4). This weighting method based on time contribution effectively smooths out data jumps between different prediction periods, making the final predicted value more accurately reflect the overall expected state of the target time period.

[0051] For example, Figure 5 Node 1, as shown, can obtain the predicted value for n periods before the time of receiving the message at any feedback time. Using the time-overlapping weighted prediction method, receiving node 1 can obtain the predicted value for the target time period, including: Step 1, Obtain the target time period Forecasted demand; Step 2: Identify all prediction periods that overlap with the target time period. ; Step 3: Calculate the overlap duration between each prediction period and the target time period. ; Step 4: Calculate the weight of each predicted value based on the overlap duration. ; Step 5: Calculate a weighted average of the overlapping predicted values: .

[0052] For example, Node 1, receiving data according to prediction time method two, needs to know the predicted values ​​for the time period from 0.8T to 1.3T. Based on the calculation steps above, there are two overlapping periods, T1 and T2, with overlap durations of 0.2 and 0.3 respectively. According to step 4, their weights are calculated to be 0.4 and 0.6 respectively. According to step 5, the predicted values ​​for the target time period can be obtained. Multiply by 0.4 and add Multiply by 0.6.

[0053] Figure 6 The time overlap prediction calculation method shown is the same as Figure 5 The same applies, so I won't go into details here.

[0054] In some exemplary embodiments, to further improve the reliability of the final predicted value, the feedback message also carries prediction accuracy information corresponding to each of the multiple predicted values; Based on the overlap duration, weights are assigned to the predicted values ​​of each overlap, including: Based on the overlap duration and the corresponding prediction accuracy information, weights are assigned to the prediction values ​​of each overlap, wherein the assigned weights are linearly correlated with the prediction accuracy information.

[0055] Since prediction accuracy typically decreases the further into the future, this application's embodiments introduce prediction accuracy information calculated from historical data (a i This problem can be solved by shifting the basis for weight allocation from a single "time correlation" (Δt) i ")" is transformed into "time relevance and historical credibility (a)". i The "comprehensive consideration" approach ensures that the influence of a prediction with a long overlap with the target time period but poor historical performance is reasonably suppressed, while the contribution of a prediction with a short overlap but extremely accurate historical performance is appropriately amplified. This dynamic, evidence-based weighting adjustment makes the final prediction calculated by weighted average more closely reflect the actual trend of network state changes, thereby directly improving the reliability of the final prediction.

[0056] Considering that the accuracy of predictions decreases as time goes further into the future, Figure 5 Node 2 shown can retain the actual observations for the most recent n prediction periods. Calculate the prediction accuracy over n prediction periods based on the feedback of n predictions. ,in The average relative error between the i-th predicted value and the corresponding actual predicted value in each feedback result is obtained, and the calculation formula is as follows:

[0057] in, Indicates the first The predicted value of the feedback is the first Each prediction result. The feedback message allows you to select the prediction accuracy. This is so that the receiving node can evaluate and use it.

[0058] After receiving the prediction accuracy, Node 1 adjusts the calculation formula in step 4 to allocate higher weights to periods with higher prediction accuracy. The formula in step 4 can be adjusted accordingly. .

[0059] For example: if we need to obtain the predicted values ​​for the time period from 0.8T to 1.3T, assuming... It is 0.95. If the value is 0.80, then the updated weights are 0.44 and 0.56 respectively. According to step 5, the predicted value for the target time period can be obtained. Multiply by 0.44 and add Multiply by 0.56 to get the predicted value for the time period from 0.8T to 1.3T.

[0060] The requesting node can obtain predicted information for the target time period using the method described above. It can further refine and optimize the predicted data using its AI processing capabilities to improve the accuracy of the predictions and thus enhance the reliability of the decision. Taking handover as an example, if a user in requesting node 1 moves to the cell edge, a handover needs to be triggered immediately. The selectable target cells A and B belong to nodes 2 and 3, respectively. Node 1 has received n predicted values ​​from nodes 2 and 3 for cells A and B, including the number of users and cell load. Requesting node 1 can process these n predicted values ​​according to the steps described above to obtain effective predicted values ​​for the target time period. Inputting these predicted values ​​into the handover decision algorithm module allows for more effective selection of suitable target cells.

[0061] The principle of load balancing decision is similar. It also involves obtaining the predicted values ​​of multiple surrounding cells, processing the n predicted values ​​of each corresponding cell according to the above steps, and then using them as input to the load balancing algorithm to select a more reasonable target cell.

[0062] Similarly, considering that the accuracy of predictions decreases as time goes further into the future. Figure 6 Node 2, as shown, can retain the actual observations and feedback predictions for the most recent k historical prediction periods. Each prediction period contains n actual observations. ) and n predicted values Based on the above data, the prediction accuracy can be calculated. ,in The average relative error between the i-th predicted value and the corresponding actual predicted value from k feedback results is obtained, and the calculation formula is as follows:

[0063] in, and Indicates the first l The predicted value of the feedback is the first i Each prediction result. The feedback message allows you to select the prediction accuracy. This is so that the receiving node can evaluate and use the data. After receiving the prediction accuracy, Node 1 adjusts the calculation formula in step 4 to allocate higher weights to periods with higher prediction accuracy. The adjusted formula is consistent with... Figure 5 The scheme shown is the same. If node 1 needs to evaluate the prediction results for a specific time period within time period T2, then according to... Figure 5 The steps described involve multiplying the predicted values ​​for the corresponding time period by their respective weights and then accumulating these values ​​to obtain the predicted value for that time period. Regarding the specific application scenarios of the predicted data, its implementation method is similar to... Figure 5 The switching decisions and load balancing application scenarios described in the previous section are similar, and will not be repeated here.

[0064] In some exemplary embodiments, to ensure the objectivity and authenticity of the prediction accuracy assessment and to provide a reliable basis for weight optimization, the prediction accuracy information in the embodiments of this application is calculated by the prediction node based on the error between the predicted value and the corresponding actual observation value in the historical feedback.

[0065] In some exemplary embodiments, to directly convert predicted data into key network performance gains, the method provided in this application further includes: After obtaining the final prediction value, the requesting node uses the final prediction value as input to apply it to the Radio Resource Management (RRM) decision; wherein the RRM decision includes at least one of handover decision, resource allocation and load balancing.

[0066] The data processing method provided in this application can generate multiple predicted values ​​corresponding to multiple different time intervals in the future based on the predicted request information of the request node, and feed them back to the request node. This makes the predicted value of the request node a future predicted data sequence at any time. Therefore, when it needs to make a real-time decision at any time, it can directly call or calculate the predicted information of the target time period at the current time from the multiple predicted values ​​corresponding to multiple different time intervals in the future that are stored locally, without waiting for the next periodic feedback. This effectively increases the amount of effective predicted data held by the request node at any non-feedback time.

[0067] Figure 7 This is a schematic diagram of the structure of a data processing apparatus 700 provided for an exemplary embodiment of this application. For example... Figure 7 As shown, the device 700 includes: a receiving module 710, a generating module 720, and a transmitting module 730, wherein: The receiving module 710 is used to receive prediction request messages from the requesting node; The generation module 720 is used to perform a prediction operation and generate prediction information according to the prediction request message; wherein the prediction information includes multiple prediction values ​​corresponding to multiple different time intervals in the future; The sending module 730 is used to send a feedback message to the requesting node, the feedback message carrying the prediction information containing the multiple prediction values.

[0068] The data processing device 700 provided in this application embodiment can generate multiple predicted values ​​corresponding to multiple different time intervals in the future based on the predicted request information of the requesting node, and feed them back to the requesting node. This makes the predicted value of the requesting node a future predicted data sequence at any time. Therefore, when it needs to make a real-time decision at any time, it can directly call or calculate the predicted information of the target time period required at the current time from the multiple predicted values ​​corresponding to multiple different time intervals in the future that are stored locally, without waiting for the next periodic feedback. This effectively increases the amount of effective predicted data held by the requesting node at any non-feedback time.

[0069] Optionally, the prediction request message includes a parameter for indicating a one-time feedback of prediction values ​​for the next n periods, where n is an integer greater than 1; The multiple predicted values ​​in the prediction information are the predicted values ​​for each of the next n consecutive periods.

[0070] Optionally, the prediction request message further includes a prediction time mode selection parameter, used to indicate that the prediction period corresponding to each of the plurality of predicted values ​​is defined using any of the following methods: The prediction period length corresponding to the i-th prediction value is i times the duration of a single period, and all prediction periods have the same start time. The prediction period length corresponding to the i-th prediction value is the duration of a single period, and the start time of the i-th prediction period is the end time of the (i-1)-th prediction period.

[0071] Optionally, the prediction request message includes a prediction period parameter T2 and a feedback period parameter T1, wherein T2 > T1; The multiple predicted values ​​in the prediction information are the predicted values ​​corresponding to M sub-intervals divided in time order within a prediction period of length T2, where M is an integer greater than 1.

[0072] Optionally, after the sending module 730 sends a feedback message to the requesting node, the device further includes a final prediction value determination module, used for: The request node determines one or more predicted values ​​that overlap with the target time period; Calculate the overlap duration between the prediction period corresponding to each of the overlapping prediction values ​​and the target time period; Based on the overlap duration, weights are assigned to the predicted values ​​of each overlap. The weighted average of the overlapping predicted values ​​is used to obtain the final predicted value.

[0073] Optionally, the feedback message also carries prediction accuracy information corresponding to the plurality of predicted values ​​respectively; When the final prediction value determination module assigns weights to each overlapping prediction value based on the overlap duration, it is specifically used for: Based on the overlap duration and the corresponding prediction accuracy information, weights are assigned to the prediction values ​​of each overlap, wherein the assigned weights are linearly correlated with the prediction accuracy information.

[0074] Optionally, the prediction accuracy information is calculated by the prediction node based on the error between the predicted value and the corresponding actual observed value from historical feedback.

[0075] Optionally, the device further includes an input module for: After obtaining the final prediction value, the requesting node uses the final prediction value as input and applies it to the Radio Resource Management (RRM) decision; wherein the RRM decision includes at least one of handover decision, resource allocation, and load balancing.

[0076] The data processing device 700 is capable of achieving Figures 4-6 For details of the method implementation examples, please refer to [link / reference]. Figures 4-6 The data processing method of the illustrated embodiment will not be described in detail.

[0077] Figure 8 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of this application. For example... Figure 8 As shown, the device includes a memory 81 and a processor 82.

[0078] Memory 81 is used to store computer programs and can be configured to store various other data to support operation on the computing device. Examples of this data include instructions for any application or method used to operate on the computing device, contact data, phone book data, messages, images, videos, etc.

[0079] Processor 82, coupled to memory 81, is configured to execute a computer program in memory 81 for: receiving a prediction request message from a requesting node; performing a prediction operation based on the prediction request message and generating prediction information; wherein the prediction information contains multiple prediction values ​​corresponding to multiple different future time intervals; and sending a feedback message to the requesting node, the feedback message carrying the prediction information containing the multiple prediction values.

[0080] The electronic device provided in this application embodiment can generate multiple predicted values ​​corresponding to multiple different time intervals in the future based on the predicted request information of the request node, and feed them back to the request node. This makes the predicted value of the request node a future predicted data sequence at any time. Therefore, when it needs to make a real-time decision at any time, it can directly call or calculate the predicted information of the target time period required at the current time from the multiple predicted values ​​corresponding to multiple different time intervals in the future that are stored locally, without waiting for the next periodic feedback. This effectively increases the amount of effective predicted data held by the request node at any non-feedback time.

[0081] Furthermore, such as Figure 8 As shown, the electronic device also includes other components such as a communication component 83, a display 84, a power supply component 85, and an audio component 86. Figure 8 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 8 The components shown. Additionally, depending on the implementation of the traffic playback device, Figure 8 The components within the dashed box are optional, not mandatory. For example, when an electronic device is implemented as a terminal device such as a smartphone, tablet, or desktop computer, it may include... Figure 8 The components within the dashed box; when the electronic device is implemented as a server-side device such as a conventional server, cloud server, data center, or server array, it may be excluded. Figure 8 The component within the dashed box.

[0082] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in the above-described data processing method embodiments.

[0083] Accordingly, this application also provides a computer program product, which stores instructions that, when executed by a computer, cause the computer to perform the steps in the data processing method embodiments provided in this application.

[0084] The above Figure 8 The communication component is configured to facilitate wired or wireless communication between the device containing the communication component and other devices. The device containing the communication component can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component may further include a Near Field Communication (NFC) module, Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, etc.

[0085] The above Figure 8 The memory in the memory can be implemented by any class of volatile or non-volatile storage devices or combinations thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0086] The above Figure 8 The display includes a screen, which may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of the touch or swipe action, but also the duration and pressure associated with the touch or swipe operation.

[0087] The above Figure 8 The power supply component provides power to the various components of the device in which it resides. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device in which it resides.

[0088] The above Figure 8 The audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.

[0089] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0094] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0095] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other classes of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0096] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0097] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A data processing method, characterized in that, include: Receive prediction request messages from the requesting node; Based on the prediction request message, a prediction operation is performed, and prediction information is generated; wherein, the prediction information contains multiple prediction values ​​corresponding to multiple different time intervals in the future; A feedback message is sent to the requesting node, the feedback message carrying the prediction information containing the multiple prediction values.

2. The method as described in claim 1, characterized in that, The prediction request message includes a parameter for indicating a one-time feedback of prediction values ​​for the next n periods, where n is an integer greater than 1; The multiple predicted values ​​in the prediction information are the predicted values ​​for each of the next n consecutive periods.

3. The method as described in claim 2, characterized in that, The prediction request message also includes a prediction time mode selection parameter, which indicates that the prediction period corresponding to each of the multiple prediction values ​​can be defined using any of the following methods: The prediction period length corresponding to the i-th prediction value is i times the duration of a single period, and all prediction periods have the same start time. The prediction period length corresponding to the i-th prediction value is the duration of a single period, and the start time of the i-th prediction period is the end time of the (i-1)-th prediction period.

4. The method as described in claim 1, characterized in that, The prediction request message includes a prediction period parameter T2 and a feedback period parameter T1, where T2 > T1; The multiple predicted values ​​in the prediction information are the predicted values ​​corresponding to M sub-intervals divided in time order within a prediction period of length T2, where M is an integer greater than 1.

5. The method according to any one of claims 1 to 4, characterized in that, After sending the feedback message to the requesting node, the method further includes: The request node determines one or more predicted values ​​that overlap with the target time period; Calculate the overlap duration between the prediction period corresponding to each of the overlapping prediction values ​​and the target time period; Based on the overlap duration, weights are assigned to the predicted values ​​of each overlap. The weighted average of the overlapping predicted values ​​is used to obtain the final predicted value.

6. The method as described in claim 5, characterized in that, The feedback message also carries prediction accuracy information corresponding to the multiple predicted values ​​respectively; The process of assigning weights to the predicted values ​​of each overlap based on the overlap duration includes: Based on the overlap duration and the corresponding prediction accuracy information, weights are assigned to the prediction values ​​of each overlap, wherein the assigned weights are linearly correlated with the prediction accuracy information.

7. The method as described in claim 6, characterized in that, The prediction accuracy information is calculated by the prediction node based on the error between the historically fed-back prediction values ​​and the corresponding actual observation values.

8. The method according to any one of claims 5 to 7, characterized in that, The method further includes: After obtaining the final prediction value, the requesting node uses the final prediction value as input and applies it to the Radio Resource Management (RRM) decision; wherein the RRM decision includes at least one of handover decision, resource allocation, and load balancing.

9. A data processing apparatus, characterized in that, include: The receiving module is used to receive prediction request messages from the requesting node; The generation module is used to perform a prediction operation and generate prediction information based on the prediction request message; wherein the prediction information includes multiple prediction values ​​corresponding to multiple different time intervals in the future. The sending module is used to send a feedback message to the requesting node, the feedback message carrying the prediction information containing the multiple prediction values.

10. An electronic device, characterized in that, include: One or more processors, memory, and one or more computer programs; The one or more computer programs are stored in the memory and configured to be executed by the one or more processors, the one or more computer programs including instructions for performing the method as described in any one of claims 1 to 8.

Images