Information processing device

Abstract

An information processing device according to an embodiment of the present invention comprises a control unit. A first channel estimation using AI is performed on a first symbol group in uplink data of one slot, and a second channel estimation using a method different from the first channel estimation is performed on a second symbol group.

Description

Information processing apparatus 【0001】 The disclosed embodiment relates to an information processing apparatus. 【0002】 In a wireless communication system, studies have been conducted on the mapping rules of demodulation reference signals (e.g., DMRS (DeModulation Reference Signal)) when bundling a plurality of slots in the time direction (see, for example, Patent Document 1). 【0003】 Japanese Patent Application Laid-Open No. 2023-123583 【0004】 The information processing apparatus according to the embodiment includes a control unit. The control unit performs first channel estimation using AI on a first symbol group among the uplink data of one slot, and performs second channel estimation using a method different from the first channel estimation on a second symbol group. 【0005】 FIG. 1 is a diagram showing an example of the schematic configuration of a communication system according to the embodiment. FIG. 2 is a diagram for explaining an example of channel estimation using a known signal. FIG. 3 is a diagram for explaining an example of AI technology in image processing. FIG. 4 is a diagram for explaining an example of channel estimation using an AI method performed by the information processing apparatus. FIG. 5 is a diagram for explaining the processing delay of channel estimation by the AI method. FIG. 6 is a flowchart showing an example of the information processing flow according to the embodiment. FIG. 7 is a diagram showing a configuration example of the information processing apparatus according to the embodiment. FIG. 8 is a diagram showing an example of a slot according to the embodiment. FIG. 9 is a diagram for explaining the timing of information processing according to the embodiment. FIG. 10 is a diagram showing an example of a resource block according to a modified example of the embodiment. FIG. 11 is a diagram for explaining the timing of information processing according to the embodiment. FIG. 12 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing apparatus. 【0006】The following describes in detail, with reference to the drawings, the embodiments for implementing the information processing device according to the present application (hereinafter referred to as "embodiments"). Note that these embodiments do not limit the information processing device according to the present application. Furthermore, the same parts are denoted by the same reference numerals in each of the following embodiments, and redundant explanations are omitted. Also, in the following, AI (artificial intelligence) models and machine learning models are treated as synonymous terms. 【0007】 (Embodiment) [1. Introduction] Figure 1 is a diagram showing an example of the schematic configuration of a communication system SYS1 according to an embodiment. The communication system SYS1 shown in Figure 1 comprises a terminal device 10 and a RAN (Radio Access Network) 20. 【0008】 RAN20, for example, operates cell C1 and performs wireless communication with terminal devices 10 within cell C1. RAN20 includes an information processing device 100 and a base station 200. For example, the information processing device 100 may be a server device that implements the functions of a RIC (RAN Intelligent Controller). The information processing device 100 may be a device that has the functions of AI-RAN, which has a computing platform and a learning platform among the functions of RAN20. "AI-RAN" is a technology that combines AI (Artificial Intelligence) and next-generation mobile networks. 【0009】 For example, in order to support a society where AI is evolving at an accelerating pace, it is necessary to build next-generation social infrastructure that can cope with the rapidly increasing demand for data processing and the electricity required for that processing. Therefore, there is the AI-RAN concept, which involves building large-scale server clusters in data centers for each region (for example, each wide-area), and simultaneously running and coordinating vRAN (virtual Radio Access Network), MEC (Multi-Access Edge Computing), and AI applications on these abundant computing resources. The information processing device 100 may be applied to this AI-RAN concept. 【0010】AI-RAN is an architecture that allows AI and RAN20 (base station 200) to coexist. By using AI, it is possible to maximize the performance of RAN20 and realize an ultra-low latency, high-security computing infrastructure for various AI applications at the regional level. 【0011】 The AI-RAN information processing device 100 is deployed in a distributed manner for each region. Thus, the AI-RAN information processing device 100 may be a cloud server distributed as an edge server (also known as a MEC server) near the terminal device 10. The information processing device 100 can achieve high speed, large capacity, and low latency by utilizing a closed network isolated from the internet. On the other hand, services that utilize the data collected by the information processing devices 100 in each region (for example, large-scale computations or learning that requires a lot of power) may be executed on the public cloud (not shown) via the internet. The information processing device 100 may be deployed, for example, one per multiple base stations 200. 【0012】 The base station 200 is, for example, a communication device equipped with the function of a Radio Unit (RU). The RU has the function of transmitting and receiving radio waves to and from the terminal device 10 and converting signals between analog and digital. 【0013】 The base station 200 may have at least some of the functions of a DU (Distributed Unit) and a CU (Central Unit) in addition to the functions of a RU. Furthermore, the information processing device 100 may have at least some of the functions of a DU and a CU. 【0014】 A DU (Digital Unit) is a network element that performs physical layer processing. For example, a DU performs signal modulation, signal demodulation, radio resource allocation, and RLC (Radio Link Control) such as retransmission control. A CU (Control Unit) performs processing such as PDCP (Packet Data Convergence Protocol), which includes packet encryption, and RRC (Radio Resource Control), which includes radio resource management. 【0015】For the sake of simplicity, in the following explanation, it will be assumed that the information processing device 100 possesses the functions of both DU and CU. 【0016】 Here, the information processing device 100 performs channel estimation between the base station 200 and the terminal device 10 and demodulates the signal. Various methods are known for channel estimation processing, such as conventional methods using conventional propagation models (e.g., regression models, etc.) and AI methods using AI models. 【0017】 Figure 2 illustrates an example of channel estimation using a conventional method. Figure 2 is a diagram illustrating an example of channel estimation using a known signal. 【0018】 Generally, a pilot signal (known signal) using a known pattern is used for channel estimation between the terminal device 10 and the base station 200. The pilot signal is discretely arranged with respect to time and frequency. 【0019】 For example, in Figure 2(a), pilot signals are assigned to the second symbol from the beginning and the second symbol from the end of a slot consisting of seven symbols on the time axis. Also, on the frequency axis, pilot signals are assigned to the 1st, 3rd, 5th, 7th, 9th, and 11th subcarriers from the low-frequency side of the 11 subcarriers that make up the slot. The position in the resource block where a pilot signal is assigned is also referred to as the pilot position. 【0020】 The pilot signal is a signal containing known pattern information placed at the pilot position shown in Figure 2(a). 【0021】 The information processing device 100 receives pilot signals from the terminal device 10 at pilot positions as shown in Figure 2(a), which include known pattern information. Using the received signals at each pilot position (also called pilot received signals) and the known pattern information, the information processing device 100 estimates channel information (for example, channel estimates) at each pilot position, as shown in Figure 2(b). 【0022】Next, the information processing device 100 interpolates unknown channel information in two dimensions using the channel information estimated at the pilot position. First, in the frequency axis, the information processing device 100 estimates the channel information of subcarriers where no pilot signal is placed, using the channel estimation results for the pilot signal. 【0023】 In the example shown in Figure 2(c), the information processing device 100 uses the channel estimation result of the subcarrier on which the pilot signal is placed to estimate the channel information of the subcarrier on which the pilot signal is not placed, at the second symbol from the beginning on which the pilot signal is placed. 【0024】 Next, the information processing device 100 uses the channel information estimated in the frequency direction for a predetermined symbol to perform two-dimensional interpolation of unknown channel information in the time axis direction. In the example shown in Figure 2(d), the information processing device 100 uses the channel information estimation result for the second symbol from the beginning to estimate the channel information for the first symbol from the beginning and for the third to seventh symbols. 【0025】 Thus, the information processing device 100 performs channel estimation using known pattern information, for example, at the pilot position. This estimation method will also be referred to as the first conventional estimation method. Furthermore, the information processing device 100 uses the results of channel estimation using this known pattern information to perform two-dimensional interpolation of channel information for unknown resources (in other words, resources other than the pilot signal). This estimation method will also be referred to as the second conventional estimation method. 【0026】 In other words, when estimating channel information in a predetermined resource block (RB) using a conventional method, the information processing device 100 estimates the channel information for the entire frequency domain and time domain of the RB using two types of estimation methods. 【0027】 One example of a conventional estimation method is the least squares method (LS). When using the least squares method, the information processing device 100 estimates the channel at each pilot position by finding a solution that minimizes the sum of the squares of the errors between the pilot received signal at each pilot position and known pattern information. 【0028】 This first conventional estimation method has the advantage of not requiring statistical information on the channels. 【0029】 A second conventional estimation method is, for example, the Minimum Mean Squares Error (MMSE) method. When using the Minimum Mean Squares Error method, the information processing device 100 estimates the channels by multiplying the channel estimates obtained by the first conventional estimation method by a filtering matrix. The information processing device 100 performs the second conventional estimation method twice, once in the time domain and once in the frequency domain. 【0030】 For example, the information processing device 100 estimates the channel in the frequency domain by multiplying the channel estimate obtained based on the first conventional estimation method by a filtering matrix. By multiplying this estimation result (the channel estimate obtained by the second conventional estimation method in the frequency domain) by a filtering matrix, the information processing device 100 estimates the channel in the time domain. 【0031】 In MMSE, the information processing device 100 performs channel estimation using channel statistics and noise variance. This allows the information processing device 100 to improve the performance of channel estimation. Furthermore, since the channel statistics are ideally known, the information processing device 100 may be able to achieve the highest performance (performance limit) in channel estimation. 【0032】 Furthermore, there is a method that uses AI for channel estimation between the terminal device 10 and the base station 200. One such method that uses AI (hereinafter also referred to as the AI ​​method) is a method that applies AI technology used in image processing. 【0033】 Figure 3 illustrates an example of AI technology in image processing. Here, we describe an example of an AI model that estimates a high-resolution image from a low-resolution image. 【0034】 As an image processing technique using AI technology, one image (Figure 3 low-resolution Image (I LRBy inputting a high-resolution image (Super-resolution Image (I) in Figure 3) into the AI ​​model, the AI ​​model can process high-resolution images (Super-resolution Image (I) in Figure 3). SR There is a task that estimates and outputs ). This task is called Single Image Super-Resolution, and for example, the Convolutional Neural Network (CNN) algorithm is used to generate the AI ​​model. 【0035】 Thus, in image processing, AI technology is known that takes into account low-resolution, noisy images and reproduces higher-resolution, less noisy images. 【0036】 In channel estimation between the terminal device 10 and the base station 200, the Single Image Super-Resolution technique in this image processing may be used. 【0037】 In this case, the information processing device 100 first estimates the channel response at the pilot position (corresponding to the channel estimate described above) using the first conventional estimation method described above. 【0038】 The information processing device 100 inputs a low-resolution image to the AI ​​model, where the pilot position is used as a pixel and the channel response at this pilot position is used as the pixel value. The information processing device 100 then acquires a high-resolution image as the output of the AI ​​model, where each resource of the RB is used as a pixel and the channel response at each resource is used as the pixel value. 【0039】 Figure 4 illustrates an example of channel estimation using an AI method performed by the information processing device 100. In Figure 4, 7 symbols * 11 subcarriers are treated as one RB, and the information processing device 100 performs channel estimation on the entire RB. 【0040】 Figure 4(a) shows an example of RB after channel estimation at the pilot position. The information processing device 100 calculates the estimated channel value at the pilot position shown by the diagonal lines in Figure 4(a) using, for example, the first conventional estimation method. 【0041】The information processing apparatus 100 estimates a high-resolution image from a low-resolution image that uses the pilot position as pixels and the channel response at this pilot position as pixel values, by using the Single Image Super-Resolution technique. This high-resolution image uses the entire RB shown in Fig. 4(b) as pixels and the channel response as pixel values. That is, the information processing apparatus 100 estimates the channel responses in the entire time domain and frequency domain in the RB indicated by the slashes in Fig. 4(b) from the channel response at the pilot position. 【0042】 Channel estimation using this AI method has been reported to have higher accuracy than conventional methods at the research level. 【0043】 However, channel estimation using the AI method has a problem that the processing time becomes longer than when using the conventional method. This problem will be described using Fig. 5. In the following, for the sake of simplifying the explanation, the description of the channel estimation process at the pilot position by the first conventional method may be omitted assuming that it is implemented. 【0044】 Fig. 5 is a diagram for explaining the processing delay of channel estimation by the AI method. As shown in Fig. 5(a), the terminal device 10 transmits a transmission signal at time t01. This transmission signal is a signal with 14 symbols per slot. The third symbol from the start and the third symbol from the end of this transmission signal are pilot positions, and known signals (pilot signals) are assigned to these pilot positions. 【0045】 A propagation delay occurs between the terminal device 10 and the base station 200. Furthermore, a propagation delay also occurs between the base station 200 and the information processing apparatus 100. Therefore, as shown in Fig. 5(b), the information processing apparatus 100 starts receiving the transmission signal at time t02, which is later than time t01. 【0046】Here, channel estimation by the AI method is performed on a slot-by-slot basis. Therefore, as shown in FIG. 5(c), for example, the information processing apparatus 100 starts uplink channel estimation (in FIG. 5, "AI UL Ch. Est.") by the AI method from the time t03 when all symbols of the transmission signal transmitted in slot units are received. 【0047】 After that, the information processing apparatus 100 executes the uplink HARQ process using the channel estimation result. After executing the HARQ process (for example, at time t04 in FIG. 5), the information processing apparatus 100 notifies the terminal apparatus 10 of the result of the HARQ process (uplink HARQ ACK / NACK). 【0048】 The L1 part of the RAN 20 is strict in processing time, and the notification timing of the uplink HARQ ACK / NACK is predetermined. Therefore, the information processing apparatus 100 is required to execute the HARQ process by this notification timing. That is, an allowable processing time (for example, time T01 in FIG. 5) for the HARQ process for the information processing apparatus 100 is provided. 【0049】 As described above, channel estimation using the AI method takes longer processing time than channel estimation using the conventional method. In addition, since channel estimation using the AI method is executed on a slot-by-slot basis, the channel estimation process starts after all symbols of the slot are received. 【0050】 Depending on the start timing of this channel estimation process and / or the processing time of the channel estimation, there is a risk that the information processing apparatus 100 cannot notify the HARQ ACK / NACK within the allowable processing time. 【0051】 For example, as a method for the information processing apparatus 100 to more surely end the HARQ process within the allowable processing time and notify the HARQ ACK / NACK, a method of shortening the processing time of channel estimation using the AI method can be considered. For example, the information processing apparatus 100 can shorten the processing time of channel estimation by reducing the size of the AI model used for channel estimation. On the other hand, if the size of the AI model is reduced, the accuracy of channel estimation may decrease. 【0052】 The information processing device 100 is required to provide HARQ ACK / NACK notifications within an acceptable processing time while preventing a decrease in the accuracy of channel estimation. In other words, a method is needed to reduce the processing delay of channel estimation using AI. 【0053】 Therefore, in this embodiment, the information processing device 100 does not process channel estimation on a slot-by-slot basis, but rather divides the slots into one or more symbols and performs channel estimation by combining the AI ​​method and the conventional method. As a result, the information processing device 100 can extend the processing time for channel estimation using the AI ​​method. 【0054】 Hereinafter, the channel estimation process using an AI method (an example of the first channel estimation) will be referred to as the first estimation process, and the channel estimation process using a method different from the AI ​​method, such as a conventional method (an example of the second channel estimation), will be referred to as the second estimation process. 【0055】 An example of information processing according to this embodiment will be explained using Figure 6. Figure 6 is a flowchart showing an example of the flow of information processing according to this embodiment. The information processing in Figure 6 is performed, for example, when the information processing device 100 receives a signal in slot units from the terminal device 10 and performs channel estimation. 【0056】 As shown in Figure 6, the information processing device 100 receives N1 symbols from the beginning of the slot from the received signal (step S101). 【0057】 Next, the information processing device 100 performs a second estimation process using the N1 symbols it has received (step S102). As described above, the second estimation process is a channel estimation process using a conventional method. Examples of the second estimation process include MMSE and linear interpolation. 【0058】 The information processing device 100 receives N2 symbols starting from the N1+1th position from the beginning of the slot (step S103). 【0059】The information processing device 100 performs a first estimation process using the N2 symbols it has received (step S104). As described above, the first estimation process is a channel estimation process using an AI method. 【0060】 As described above, the information processing device 100 according to this embodiment divides the slot into one or more symbols and performs channel estimation. In the example shown in Figure 6, the information processing device 100 divides the slot into N1 symbols and N2 symbols. The information processing device 100 performs a second estimation process using a conventional method for the N1 symbols and a first estimation process using an AI method for the N2 symbols. 【0061】 As a result, the information processing device 100 can reduce the number of symbols targeted for channel estimation using the AI ​​method compared to when it is performed on a slot-by-slot basis, and the processing time for the first estimation process can be shortened compared to when it is performed on a slot-by-slot basis. By performing channel estimation using the AI ​​method, the information processing device 100 can reduce the processing time by reducing the number of symbols targeted for channel estimation using the AI ​​method, while preventing a decrease in channel estimation accuracy. 【0062】 [2. Configuration of the Information Processing Device] Next, the configuration of the information processing device 100 according to the embodiment will be described with reference to Figure 7. Figure 7 is a diagram showing an example of the configuration of the information processing device 100 according to the embodiment. As shown in Figure 7, the information processing device 100 has a communication unit 110, a storage unit 120, and a control unit 130. The information processing device 100 may also have an input unit (for example, a keyboard or mouse) that receives various operations from the administrator of the information processing device 100, and a display unit (for example, a liquid crystal display) for displaying various information. 【0063】 (Communication Unit 110) The communication unit 110 is implemented by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to the network by wire or wireless and transmits and receives information to and from the base station 200 and other information processing devices 100 via the network. 【0064】(Storage Unit 120) The storage unit 120 is implemented by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or by storage devices such as hard disks and optical discs. 【0065】 (Control Unit 130) The control unit 130 is a controller and is realized by executing various programs stored in the memory device inside the information processing device 100 using RAM as the working area, using a CPU (Central Processing Unit) or MPU (Micro Processing Unit). The control unit 130 is also realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). 【0066】 As shown in Figure 7, the control unit 130 includes a first estimation unit 131, a second estimation unit 132, and a processing unit 133, and realizes or executes the information processing operations described below. Note that the internal configuration of the control unit 130 is not limited to the configuration shown in Figure 7, and other configurations are also acceptable as long as they perform the information processing described later. 【0067】 (First estimation unit 131) The first estimation unit 131 performs channel estimation using, for example, an AI method. An example of an AI method is the CNN algorithm method used in Single Image Super-Resolution. The first estimation unit 131 executes the first estimation process described above. 【0068】 The first estimation unit 131 uses the channel estimation result obtained using the received signal at the pilot position and known patterns to estimate the channel response of symbols other than the pilot position using an AI model. Channel estimation using the received signal at the pilot position and known patterns is performed, for example, using the first conventional estimation method described above. 【0069】More specifically, the first estimation unit 131 estimates the channel of a symbol that does not contain a DMRS, according to the channel estimation result using the DMRS included in the slot of the received signal. 【0070】 Figure 8 shows an example of a slot according to the embodiment. Here, one slot contains 14 symbols. 【0071】 In the slot shown in Figure 8, the third symbol from the beginning of the slot is the pilot position. For example, a DMRS is placed in this pilot position. 【0072】 The first estimation unit 131 performs a first estimation process using the channel estimation result using the DMRS (channel estimation value at the pilot position). As a result, the first estimation unit 131 estimates the channels in the symbols from the end of the slot up to the 10th symbol (in the example in Figure 8, symbols included in region sb12 (an example of the first symbol group)). This region sb12 does not include, for example, the DMRS (an example of a pilot signal). 【0073】 The first estimation unit 131 can estimate the channels in region sb12 with higher accuracy by performing a first estimation process using an AI model. 【0074】 (Second estimation unit 132) The second estimation unit 132 performs channel estimation using, for example, a conventional method. Examples of conventional methods include MMSE and linear interpolation. The second estimation unit 132 executes the second estimation process described above. 【0075】 The second estimation unit 132 estimates the channel response of a symbol including a DL HARQ Indication using the channel estimation result obtained using the received signal at the pilot position and a known pattern. Channel estimation using the received signal at the pilot position and a known pattern is performed, for example, using the first conventional estimation method described above. 【0076】 More specifically, the second estimation unit 132 estimates the channel of the symbol including the DL HARQ Indication according to the channel estimation result using the DMRS included in the slot of the received signal. 【0077】 As mentioned above, in the slot shown in Figure 8, the third symbol from the beginning is the pilot position. For example, a DMRS is placed in this pilot position. 【0078】 Furthermore, for example, there is a PUCCH (Physical Uplink Control Channel) used as a channel for uplink transmission from terminal device 10 to base station 200. In some cases, the PUCCH includes UCI (Uplink Control Information). In this case, the DL HARQ Indication is included as the fourth symbol from the beginning of the slot. 【0079】 The second estimation unit 132 uses the channel estimation result using the DMRS (channel estimation value at the pilot position) to perform a second estimation process on the symbols from the beginning to the fourth symbol of the slot. As a result, the second estimation unit 132 estimates the channels in the symbols from the beginning to the fourth symbol of the slot (in the example in Figure 8, symbols included in region sb11 (an example of the second symbol group)). This region sb11 includes, for example, the DMRS (an example of a pilot signal) and the DL HARQ Indication. 【0080】 The second estimation unit 132 performs a second estimation process on the first four symbols of the slot, thereby estimating the channel in the region sb11 containing the DL HARQ Indication before receiving the last symbol of the slot. 【0081】 (Processing Unit 133) The processing unit 133 executes the HARQ process and notifies the terminal device 10 of the UL HARQ ACK / NACK. The processing unit 133 executes the HARQ process using the DL HARQ Indication after channel estimation. 【0082】 As described above, for symbols containing DL HARQ Indication (the fourth symbol from the beginning of the slot), channel estimation is performed by the second estimation unit 132 at the time the symbol is received. 【0083】 The processing unit 133 can start the HARQ process when the second estimation process by the second estimation unit 132 is completed. As described above, the second estimation unit 132 starts the second estimation process earlier than when it receives the last symbol of the slot (in the example of Figure 8, when it receives the fourth symbol from the beginning of the slot). 【0084】 Therefore, the second estimation unit 132 can start channel estimation of symbols including DL HARQ Indication earlier than when channel estimation is started at the timing when all slots have been received. 【0085】 As a result, the processing unit 133 can start the HARQ process earlier than if it had started channel estimation after receiving all slots, and can notify the terminal device 10 of the UL HARQ ACK / NACK at an earlier time. 【0086】 Therefore, the information processing device 100 can complete the HARQ process earlier within the processing allowable time and notify the terminal device 10 of the UL HARQ ACK / NACK. This allows the information processing device 100 to further reduce the overall processing burden on itself. 【0087】 [3. Example of Information Processing] Next, an example of the timing of information processing performed by the information processing device 100 will be explained using Figure 9. Figure 9 is a diagram for explaining the timing of information processing according to the embodiment. 【0088】 As shown in Figure 9(a), the terminal device 10 transmits a transmission signal at time t11. This transmission signal consists of 14 symbols per slot. The third symbol from the beginning of this transmission signal is the pilot position, and a known signal (pilot signal) is assigned to this pilot position. In addition, the fourth symbol from the beginning of this transmission signal contains the DL HARQ Indication. 【0089】A propagation delay occurs between the terminal device 10 and the base station 200. Furthermore, a propagation delay also occurs between the base station 200 and the information processing device 100. Therefore, as shown in Figure 9(b), the information processing device 100 starts receiving the transmission signal at a time t12, which is later than time t11. 【0090】 At time t13, when the first four symbols have been received, the information processing device 100 starts the second estimation process (referred to as "Second Ch. Est." in Figure 9(d)). 【0091】 The information processing device 100 starts the HARQ process at time t14 when the second estimation process is completed. The information processing device 100 notifies the terminal device 10 of the UL HARQ ACK / NACK at time t16 when the HARQ process is completed. 【0092】 Furthermore, at time t15, when the information processing device 100 has received all of the transmitted signals, in other words, when it has received the 14th symbol from the beginning of the slot, it starts the first estimation process (in Figure 9(c), "AI UL Ch. Est."). 【0093】 Here, channel fluctuations often show similar trends in adjacent symbols. Therefore, it is thought that sufficient accuracy can be obtained even when channel estimation is performed using conventional methods near DMRS (an example of a pilot signal). 【0094】 Therefore, in this embodiment, when the information processing device 100 receives a region near the DMRS that includes a DL HARQ Indication, it starts a second estimation process. This allows the information processing device 100 to start the HARQ process at an earlier timing while ensuring sufficient accuracy in channel estimation. 【0095】 For example, if the information processing device 100 starts the HARQ process after completing the first estimation process, the timing of notifying the terminal device 10 of the UL HARQ ACK / NACK will be the time t17, when the HARQ process is completed after the first estimation process. 【0096】On the other hand, in this embodiment, the information processing device 100 can notify the terminal device 10 of the UL HARQ ACK / NACK at time t16 when the HARQ process is completed after the second estimation process, thereby shortening the notification timing of the UL HARQ ACK / NACK by time T11. 【0097】 [4. Modifications] In the information processing according to the above embodiment, the information processing device 100 performs a second estimation process on symbols in one region sb12 that includes DMRS and DL HARQ Indication. Alternatively, the information processing device 100 may perform a second estimation process on symbols in multiple regions that include DMRS. 【0098】 For example, the above-described embodiment described information processing when one DMRS is included in a frame on a single subcarrier, but the arrangement of DMRS is not limited to the examples of the above-described embodiment. For example, two or more DMRS may be included in one frame. 【0099】 Figure 10 shows an example of a resource block according to a modified embodiment. For example, in Figure 10, pilot signals are assigned to the second symbol from the beginning and the second symbol from the end of a slot consisting of seven symbols in the time axis. Also, in the frequency axis, pilot signals are assigned to the 1st, 3rd, 5th, 7th, 9th, and 11th subcarriers from the lowest frequencies among the 11 subcarriers that make up the slot. 【0100】 In this case, the information processing device 100 (more specifically, the second estimation unit 132) performs a second estimation process on the region sb21_1, which consists of the first two symbols in the time direction of RB, and the region sb21_2, which consists of the last two symbols. These regions sb21_1 and sb21_2 contain DMRS. The information processing device 100 then uses this DMRS to estimate the channels of regions sb21_1 and sb21_2 using conventional methods (for example, MMSE or linear interpolation). 【0101】Furthermore, the information processing device 100 (more specifically, the first estimation unit 131) performs a first estimation process on the region sb22 from the third to the fifth symbol from the beginning in the time direction of RB. The information processing device 100 uses the DMRS contained in symbols adjacent to region sb22 (for example, the second symbol from the beginning and the second symbol from the end) to estimate the channels of region sb22 according to the AI ​​method. 【0102】 In this way, the information processing device 100 performs a second estimation process in the region from the end of the RB in the time direction (or the slot described later) to the symbol containing the DMRS. Therefore, the information processing device 100 can start the first estimation process when it receives the symbol containing the DMRS (the second to last symbol in the example of Figure 10). 【0103】 Furthermore, the information processing device 100 does not perform the first estimation process on the entire RB, but rather on a portion of it (in the example of Figure 10, region sb22). Therefore, the information processing device 100 can further reduce the processing load of the first estimation process and further shorten the processing time of the first estimation process. 【0104】 As a result, the information processing device 100 can complete the first estimation process earlier, and can start the HARQ process using the channel estimation results from the first estimation process at an earlier time. 【0105】 For example, if the symbol targeted by the second estimation process does not include a DL HARQ Indication, the information processing device 100 can notify the L HARQ ACK / NACK at an earlier time by using the method of this modified example. In other words, even when the first estimation process is performed on a symbol that includes a DL HARQ Indication, the information processing device 100 can notify the L HARQ ACK / NACK at an earlier time. 【0106】Next, an example of the timing of information processing performed by the information processing device 100 will be explained using Figure 11. Figure 11 is a diagram illustrating the timing of information processing according to the embodiment. Here, we will explain the information processing when the information processing device 100 receives a transmission signal in which one slot contains 14 symbols. 【0107】 As shown in Figure 11(a), the terminal device 10 transmits a transmission signal at time t21. This transmission signal is a signal with 14 symbols per slot. The third symbol from the beginning of this transmission signal is the pilot position, and a known signal (pilot signal) is assigned to this pilot position. 【0108】 A propagation delay occurs between the terminal device 10 and the base station 200. Furthermore, a propagation delay also occurs between the base station 200 and the information processing device 100. Therefore, as shown in Figure 11(b), the information processing device 100 starts receiving the transmission signal at a time t22, which is later than time t21. 【0109】 At time t23, when the first three symbols have been received, the information processing device 100 starts a second estimation process (indicated as "Second Ch. Est." in Figure 11(c)). This second estimation process is for estimating the channel in region sb31_1 (from the first to the third symbol of the slot) of the received slot. 【0110】 Furthermore, at time t24, when the information processing device 100 has received all the pilot signals included in the slot, in other words, when it has received the third symbol from the end of the slot, it starts the first estimation process (in Figure 11(d), "AI UL Ch. Est."). This second estimation process is for estimating the channel in region sb32 (from the fourth symbol to the eleventh symbol of the slot) of the received slot. 【0111】Furthermore, at time t25, when the information processing device 100 has received all of the transmitted signals, in other words, when it has received the 14th symbol from the beginning of the slot, it starts the first estimation process (in Figure 11(e), "Second Ch. Est."). This second estimation process is for estimating the channel in region sb31_2 (from the end of the slot to the 3rd symbol) of the received slot. 【0112】 The information processing device 100 starts the HARQ process at time t26 when the first estimation process is completed. The information processing device 100 notifies the terminal device 10 of the UL HARQ ACK / NACK at time t27 when the HARQ process is completed. 【0113】 The timing at which the HARQ process is completed (time t27 in Figure 11) is earlier than the timing at which the allowable processing time ends (time t28 in Figure 11). Therefore, the information processing device 100 can complete the HARQ process at an earlier timing (time t27 in Figure 11) than the timing at which it is required to have completed the notification of UL HARQ ACK / NACK to the terminal device 10 (time t28 in Figure 11). 【0114】 As described above, when the information processing device 100 decodes the uplink data, it is assumed that there are two DMRS symbols, the 3rd and 12th symbols from the beginning of the slot. 【0115】 Channel fluctuations often show similar trends in adjacent symbols. Therefore, it is thought that sufficient estimation accuracy can be obtained by the information processing device 100 performing a second estimation process using a conventional channel estimation method in the vicinity of the DMRS (for example, regions sb31_1 and sb31_2 in Figure 11). 【0116】 On the other hand, in locations far from the DMRS, if channel fluctuations are large, conventional methods may not be able to obtain sufficient estimation accuracy. In such locations far from the DMRS (for example, region sb32 in Figure 11), there is a high possibility that sufficient estimation accuracy can be obtained by performing the first estimation process using an AI method. 【0117】Therefore, in this modified example, the information processing device 100 starts L1 processing (for example, the second estimation process) at the timing when it receives a symbol near the DMRS (in the example above, the third symbol from the beginning of the slot) (time t23 in Figure 11). Also, the information processing device 100 starts predicting the channel fluctuation of a symbol portion located away from the DMRS (for example, region sb32 in Figure 11) (for example, the first estimation process) at the timing when it receives both DMRS (time t24 in Figure 11). 【0118】 This allows the information processing device 100 to shorten the time it takes to notify UL HARQ ACK / NACK while securing more processing time to perform AI processing (e.g., the first estimation process). This allows the information processing device 100 to select an AI model more flexibly. In other words, the information processing device 100 can more easily ensure flexibility in the AI ​​model used in the first estimation process. 【0119】 In the embodiments described above, an example was shown in which one DMRS was included in one slot, and in this modification, an example was shown in which two DMRS were included in one slot. However, the number of DMRS included in one slot is not limited to one or two. 【0120】 The arrangement of DMRS in a single slot (or RB) can vary widely. The information processing device 100 according to this embodiment performs a first estimation process or a second estimation process depending on the position of the DMRS (or DL ​​HARQ Indication) in the time axis direction. 【0121】 Even focusing on the time axis, the placement of the DMRS (i.e., pilot signal) is diverse. For example, in the example above, the DMRS was placed in the third and / or twelfth slot from the beginning, but the DMRS may be placed elsewhere. For example, the DMRS may be placed in the third, eighth and twelfth slots from the beginning, or in the third, sixth, ninth and twelfth slots from the beginning. 【0122】A larger number of DMRS improves the accuracy of channel fluctuation estimation by the information processing device 100. On the other hand, a larger number of DMRS reduces the amount of data that can be transmitted from the terminal device 10 to the information processing device 100. To improve reception accuracy, a larger number of DMRS is desirable, while to improve transmission efficiency, a smaller number of DMRS is desirable. 【0123】 In this embodiment, by using an AI method for estimating channel fluctuations, it is expected that the accuracy of channel fluctuation estimation can be further improved even when the number of DMRS placed in a resource block (or slot) is small. Furthermore, by applying conventional channel estimation methods to some symbols, including DMRS, the information processing device 100 can further improve the accuracy of channel estimation while improving the overall performance of the device, in particular, reducing processing delays such as HARQ ACK / NACK. 【0124】 [5. Effects] As described above, the information processing device 100 according to the embodiment includes a control unit 130. The control unit 130 performs a first channel estimation using AI for a first group of symbols among the uplink data of one slot, and performs a second channel estimation using a method different from the first channel estimation for a second group of symbols. 【0125】 As a result, the information processing device 100 can shorten the processing time for the first channel estimation using AI. By using AI for the first channel estimation, the information processing device 100 can further shorten the AI ​​processing time while preventing a decrease in channel estimation accuracy. In addition, since the information processing device 100 can estimate channel fluctuations with greater accuracy and in a shorter time, it can contribute to achieving Sustainable Development Goal (SDG) 9, "Build resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation." 【0126】 Furthermore, the second set of symbols includes known signals used for channel estimation. 【0127】 As a result, the information processing device 100 can estimate the channel fluctuations of the second symbol group with greater accuracy. 【0128】Furthermore, the known signal is DMRS. 【0129】 As a result, the information processing device 100 can estimate the channel fluctuations of the second symbol group with higher accuracy using DMRS. 【0130】 Furthermore, the second set of symbols includes a predetermined number of symbols from the beginning of the slot. 【0131】 This allows the information processing device 100 to start the second channel estimation before the first channel estimation. 【0132】 Furthermore, the second set of symbols contains information used in HARQ processing. 【0133】 As a result, the information processing device 100 can perform HARQ processing using the results of the first channel estimation that was estimated earlier, and can perform HARQ processing more quickly. 【0134】 Furthermore, the second set of symbols includes the symbols that are included in the slot, excluding those included in the first set of symbols. 【0135】 This allows the information processing device 100 to reduce the processing time for the first channel estimation using AI. 【0136】 Furthermore, the second set of symbols includes a predetermined number of symbols from the end of the slot. 【0137】 This allows the information processing device 100 to start the first channel estimation using AI at an earlier time. 【0138】 The second channel estimation method is channel estimation using at least one of the following techniques: least mean squares error method and linear interpolation. 【0139】 This allows the information processing device 100 to reduce the processing load for second channel estimation. 【0140】 The first channel estimation method is channel estimation using the CNN algorithm. 【0141】As a result, the information processing device 100 can estimate channel fluctuations with higher accuracy. 【0142】 [6. Hardware Configuration] The information processing device 100 according to the above-described embodiment is also realized by a computer 1000 having a configuration such as that shown in Figure 12. Figure 12 is a hardware configuration diagram showing an example of a computer 1000 that realizes the functions of the information processing device 100. The computer 1000 includes a CPU 1100, RAM 1200, ROM 1300, HDD 1400, communication interface (I / F) 1500, input / output interface (I / F) 1600, and media interface (I / F) 1700. 【0143】 The CPU 1100 operates based on programs stored in the ROM 1300 or HDD 1400, and controls various parts. The ROM 1300 stores boot programs executed by the CPU 1100 when the computer 1000 starts up, as well as programs that depend on the computer 1000's hardware. 【0144】 The HDD 1400 stores programs executed by the CPU 1100, and data used by such programs. The communication interface 1500 receives data from other devices via a predetermined communication network and sends it to the CPU 1100, and transmits data generated by the CPU 1100 to other devices via the predetermined communication network. 【0145】 The CPU 1100 controls output devices such as displays and printers, and input devices such as keyboards and mice, via the input / output interface 1600. The CPU 1100 acquires data from input devices via the input / output interface 1600. The CPU 1100 also outputs the generated data to output devices via the input / output interface 1600. 【0146】The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads the program from the recording medium 1800 onto the RAM 1200 via the media interface 1700 and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase Change Rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. 【0147】 For example, when the computer 1000 functions as an information processing device 100 according to the embodiment, the CPU 1100 of the computer 1000 realizes the functions of the control unit 130 by executing a program loaded on the RAM 1200. The CPU 1100 of the computer 1000 reads and executes these programs from the recording medium 1800, but as another example, these programs may be obtained from other devices via a predetermined communication network. 【0148】 Although some embodiments of the present invention have been described in detail above with reference to the drawings, these are illustrative examples, and the present invention can be implemented in various other forms with modifications and improvements based on the knowledge of those skilled in the art, starting with the embodiments described in the disclosure section of the invention. 【0149】 [7. Others] Furthermore, among the processes described in the above embodiments and modifications, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by known methods. In addition, the processing procedures, specific names, and information including various data and parameters shown in the above document and drawings can be changed at will unless otherwise specified. For example, the various information shown in each figure is not limited to the information shown. 【0150】Furthermore, the components of each illustrated device are functionally conceptual and do not necessarily need to be physically configured as shown. In other words, the specific forms of distribution and integration of each device are not limited to those shown, and all or part of them can be functionally or physically distributed and integrated in any unit according to various loads and usage conditions. 【0151】 Furthermore, the embodiments and modifications described above can be combined as appropriate, provided that the processing content is not inconsistent. 【0152】 10 Terminal device 100 Information processing device 110 Communication unit 120 Storage unit 130 Control unit 131 First estimation unit 132 Second estimation unit 133 Processing unit 200 Base station

Claims

1. An information processing device having a control unit that performs a first channel estimation using AI for a first group of symbols from the uplink data of one slot, and a second channel estimation using a method different from the first channel estimation for a second group of symbols.

2. The information processing apparatus according to claim 1, wherein the second set of symbols includes known signals used for channel estimation.

3. The information processing apparatus according to claim 2, wherein the known signal is DMRS.

4. The information processing apparatus according to claim 1, wherein the second group of symbols includes a predetermined number of symbols from the beginning of the slot.

5. The information processing apparatus according to claim 1, wherein the second group of symbols includes information used for HARQ processing.

6. The information processing apparatus according to claim 1, wherein the second symbol group includes the symbols from among the plurality of symbols included in the slot, excluding the symbols included in the first symbol group.

7. The information processing apparatus according to claim 1, wherein the second group of symbols includes a predetermined number of symbols from the end of the slot.

8. The information processing apparatus according to claim 1, wherein the second channel estimation is channel estimation using at least one method of least mean squares error and linear interpolation.

9. The information processing apparatus according to claim 1, wherein the first channel estimation is channel estimation using a CNN algorithm.

Images