Fast beam alignment method based on location and historical information in high-speed railway scenarios

Abstract

The present invention discloses a fast beam alignment method based on position and historical information in a high-speed railway scenario. The method comprises: setting an interruption threshold and a collaboration threshold in advance, with the collaboration threshold being slightly higher than the interruption threshold, for determining whether RRU collaboration is required to provide services for MR; establishing a lookup table for recording the latest optimal beam pair corresponding to each position, executing a full-space beam training process, determining the optimal beam pair for each position, and initializing the lookup table; when the train passes the same position again, the MR queries the beam training result of the position at the previous moment from the lookup table; if the latest historical beam pair is one pair, executing a single RRU fast beam alignment scheme based on position and historical information to determine the optimal beam pair at that moment; if the latest historical beam pair is two pairs, executing an RRU collaborative fast beam alignment scheme based on position and historical information to determine the optimal beam pair at that moment; and finally updating the lookup table for subsequent beam training.

Description

Technical Field

[0001] The present invention belongs to the field of mobile communications and relates to a wireless communication method used in high-speed rail communication scenarios. Background Art

[0002] With the development of wireless communication technology, the fifth-generation mobile communication technology, which features ultra-low latency, ultra-high speed and massive connections, has become a research hotspot. High-speed railway scenarios, as a typical application scenario of 5G, are also one of the research focuses.

[0003] Currently, spectrum resources in low-frequency bands are becoming saturated and unable to meet the enormous capacity demands of high-speed rail communications. To meet the explosive growth in wireless data traffic in high-speed rail communications, a direct and effective approach is to utilize spectrum resources in high-frequency bands to improve system reliability and data transmission rates. Therefore, future 5G-based high-speed rail wireless networks will utilize millimeter wave bands, which have abundant available spectrum resources, to meet capacity demands. However, the propagation characteristics of millimeter waves also pose challenges to the implementation of millimeter wave technology. By simultaneously applying multiple communication technologies in HSR wireless communication systems, electromagnetic wave energy can be concentrated in the desired direction, leveraging the high gain of directional narrow beams to counteract millimeter wave propagation losses and thereby extend the signal propagation range in millimeter wave communications.

[0004] While beamforming technology can extend coverage, it also complicates control layer processes, particularly initial access. Initial access refers to the process by which a mobile user establishes a physical communication link with a base station. In frequency bands below 6 GHz, this is typically done omnidirectionally or using very wide beams. However, if this approach is directly applied to millimeter wave systems, the range within which users and base stations can discover each other will be far smaller than the range required for directional, high-speed communication due to high path loss. Therefore, in millimeter wave communication systems, a directional initial access (IA) process is required to establish a communication link.

[0005] During directional IA, beam alignment is crucial for determining the optimal beam direction for both the transmitter and receiver. Exhaustive search is the most commonly used beam alignment solution, but it introduces significant access latency. From a time perspective, the IA process consists of two phases: cell search and random access. Downlink synchronization is achieved during the CS phase, and uplink synchronization is achieved during the RA phase. During the CS phase, the RRU performs an omnidirectional beam scan to locate available RRUs. Simultaneously, the user scans all receive directions, measures the received signal quality, and records the index of the optimal beam pair. During the RA phase, the user transmits the index information using the optimal beam direction recorded during the CS phase, while the base station scans all directions to obtain this information. Due to the narrow beams in millimeter-wave communication systems and the large number of beams used by the transmitter and receiver, the exhaustive search method introduces excessive latency. In high-speed millimeter-wave communication scenarios, millimeter-wave communications are prone to congestion, necessitating beam re-alignment. Furthermore, due to the high speed of trains and the narrow beams, the optimal beam direction can change rapidly during train operation. Excessive IA latency can render previously recorded optimal beam pairs completely invalid. Therefore, it is necessary to design a fast and effective beam alignment scheme for the HSR communication system to improve the success probability of IA.

[0006] In wireless communications, the wireless channel is highly dependent on the propagation environment, which in turn is highly dependent on the user's location. In urban wireless communication environments, due to numerous obstructing buildings, the signal between the user and the base station undergoes multiple reflections and scattering, resulting in a complex signal model. However, in high-speed rail (HSR) scenarios, particularly in China, where high-speed rail is primarily located on elevated overpasses, the scattering environment is relatively simple, and there is a strong direct path between the train and the base station. Using beamforming technology, the non-line-of-sight (NLOS) path component, with its lower signal strength compared to the line-of-sight path, can be negligible. Furthermore, train trajectories in HSR scenarios are periodic and regular, and the train's position and speed can be estimated by the train control system, facilitating the optimization of beam alignment. In a relatively stable environment, the results of two consecutive beam alignment training runs at the same location will not differ significantly, meaning that historical information can provide a reference for the current beam training process. Therefore, in HSR scenarios, a fast beam alignment solution can be designed by leveraging both location and historical information. Furthermore, considering that in traditional cellular network communication scenarios, mobile users are typically served by a single base station, when users move to the cell edge, the increased path loss causes a degradation in received signal quality, thereby reducing the probability of successful IA. To address this issue, this paper considers extending the proposed scheme to non-cellular scenarios, implementing collaboration at the remote radio unit service edge to improve cell-edge IA performance. Therefore, designing a fast and effective beam alignment scheme based on location and historical information in HSR scenarios has both theoretical and practical significance. Summary of the Invention

[0007] Technical problem: The present invention provides a fast beam alignment method based on location and historical information in a high-speed railway scenario. During the initial access process, the MR and RRU use location and historical information to construct a new beam training set, which greatly reduces the beam training space and reduces the IA delay. At the same time, in order to address the problem of decreased received signal quality and reduced IA success probability due to high path loss at far RRUs, the present invention introduces an RRU collaboration mechanism, using multiple RRUs to provide services to the MR at the same time, thereby increasing the IA success probability. This method makes full use of location information, historical information, and the concept of collaboration, improves the IA success probability, reduces IA delay, and achieves fast access.

[0008] Technical Solution: The present invention provides a fast beam alignment method based on location and historical information in a high-speed railway scenario, including the following contents:

[0009] In high-speed railway (HSR) scenarios, a wireless network communication system with decoupled control and user planes is constructed. The network architecture includes two types of remote radio units (RRUs). One is a low-frequency RRU using an omnidirectional antenna to ensure system coverage; the other is a millimeter wave remote radio unit (RRU) (mmW-RRU), which uses directional beamforming technology to compensate for millimeter wave signal transmission losses. These two RRUs are geographically distributed and connected to the baseband unit pool via high-speed backhaul. Mobile relays (MRs) are installed on the roof of the train, forming a two-hop structure with the onboard access points, thereby alleviating penetration loss and group handoff issues. Before the MRs can communicate with the wireless network, they must perform initial access with the RRUs through beam alignment. The following method is used to perform fast beam alignment based on location and historical information:

[0010] Step 1: Set the interruption threshold and collaboration threshold in advance. The collaboration threshold is slightly higher than the interruption threshold to determine whether RRU collaboration is required to provide services to the MR.

[0011] Step 2: Create a lookup table to record the best beam pair for each position; perform a full-space beam training process to determine the best beam pair for each position, thereby initializing the lookup table;

[0012] Step 3: When the train passes the same location again, the MR queries the beam training result of the previous moment at that location from the lookup table, that is, the latest historical beam training result;

[0013] Step 4: If the latest historical beam pair is 1, execute the single RRU fast beam alignment solution based on location and historical information to determine the best beam pair at that moment;

[0014] Step 5: If the latest historical beam pair is 2, the RRU cooperative fast beam alignment solution based on location and historical information is executed to determine the optimal beam pair at that moment;

[0015] Step 6: Update the lookup table. Only the latest best beam pair is stored in the lookup table for subsequent beam training.

[0016] in,

[0017] The full-space beam training process described in step 2 specifically includes the following steps:

[0018] Step 2.1: Single RRU full spatial beam training process;

[0019] The MR selects the nearest RRU to perform the Initial Access (IA) process. The process begins with a Cell Search (CS) phase. The RRU traverses all transmit beams, while the MR traverses all receive beams and measures the received signal quality. If the received signal quality falls below the interruption threshold, the CS phase is considered a failure. If the received signal quality exceeds the coordination threshold, the single-RRU solution is implemented, meaning the MR only needs one RRU to provide service. The MR selects the beam with the strongest received signal, records the beam index, and calculates the probability of success in the CS phase.

[0020] Next, since the single frequency network is considered in the full spatial beam training process, the RRU needs to traverse the entire space to send the random access channel (RACH) scheduling information to ensure that the MR can obtain this information.

[0021] After receiving the RACH scheduling information, the MR enters the random access (RA) phase. The MR randomly selects a preamble from a set of mutually orthogonal preambles and sends it to the RRU using the optimal beam direction recorded by the MR in the CS phase. Simultaneously, the RRU traverses all receiving directions to receive the signal. When the received signal quality exceeds the interruption threshold, it selects the beam with the strongest received signal and records its index. The success probability of the RA phase is then calculated.

[0022] Thus, the MR and RRU can determine an optimal beam pair and initialize the lookup table;

[0023] Step 2.2, RRU collaborative full spatial beam training process;

[0024] The MR selects the nearest RRU to perform the IA process. During the CS phase, if the MR's received signal quality is between the interruption threshold and the cooperation threshold, cooperation is required. The RRU next closest to the MR is selected to join and provide joint service. After the first RRU's CS phase ends, the RRU that joins later will start the CS phase. After the CS phase of both RRUs ends, the RRUs traverse the entire space and send RACH scheduling information. After the MR obtains the resource, it enters the RA phase.

[0025] During the RA phase, the MR sends a preamble using the optimal beam direction trained in the CS phase. The wireless network receives the preamble information through two RRUs. If the received signal quality exceeds the interruption threshold, the RA process succeeds, the success probability is calculated, and the beam index is recorded. If both the CS and RA phases are successful, a communication link between the MR and the network is established. The MR can now communicate with the network through multiple RRUs, achieving higher data transmission rates.

[0026] Thus, the MR and RRU can determine two optimal beam pairs and initialize the lookup table;

[0027] The single RRU fast beam alignment solution based on location and historical information described in step 4 specifically includes the following steps:

[0028] Step 4.1, CS phase. For a certain location, in the CS phase, the MR and RRU first query the lookup table for the latest historical beam training results corresponding to that location, and select the best beam and several surrounding beams to form a new beam training set for this beam training.

[0029] In step 4.2, the wireless network communication system uses a decoupled architecture between the control plane and the user plane. Therefore, the MR feeds back the beam training results of the CS phase to the RRU via an omnidirectional link.

[0030] Step 4.3: Through step 4.2, the RRU can obtain the beam training results of the MR in the CS phase. Therefore, the RRU selects an appropriate beam based on the results and sends RACH resources in a directionally controlled manner.

[0031] In step 4.4, after receiving the RACH resource, the MR enters the RA phase. Considering the angle offset caused by train movement during the transition from the CS phase to the RA phase, the beam training set is adjusted according to the train's direction of travel. Specifically, several beams are added along the train's direction of travel and the same number of beams are removed in the opposite direction of travel to form a new beam training set to accommodate the train's movement. The MR sends a preamble in the optimal beam direction obtained through CS phase training. The RRU traverses and receives the beams in the new beam training set and records the optimal beam reception direction.

[0032] This completes the IA process for MR and RRU. In addition, due to the use of location information and historical information, the beam training space is greatly reduced, which can reduce the IA delay.

[0033] The RRU cooperative fast beam alignment solution based on location and historical information described in step 5 specifically includes the following steps:

[0034] Step 5.1, CS phase: During this IA process, the MR conducts CS with both RRUs simultaneously. The specific process is the same as the single RRU fast beam alignment process based on location and historical information. Refer to step 4.1.

[0035] In step 5.2, the wireless communication network system uses a decoupled architecture between the control plane and the user plane. Therefore, the MR can feed back the beam training results of the CS phase to the RRU via an omnidirectional link.

[0036] Step 5.3: The RRU sends RACH scheduling information in a targeted manner based on the feedback information obtained.

[0037] In step 5.4, after receiving the RACH resource, the MR initiates the RA phase: the MR sends a preamble to the two RRUs using the optimal beam direction obtained through CS phase training. The specific process is the same as step 4.4. The wireless network receives the signal through the two RRUs. If the received signal quality is higher than the interruption threshold, the RA phase is successful and the RRU records the optimal beam index.

[0038] This completes the RRU-coordinated IA process. This invention leverages location and historical information to significantly reduce the beam training space and reduce IA latency. Furthermore, activating the collaborative mechanism when received signal quality is low increases the probability of IA success at remote RRUs.

[0039] Beneficial Effects: This invention fully utilizes location information and historical information to reduce the beam training space, significantly reducing the beam training space and reducing IA latency. Furthermore, to address the issue of reduced received signal quality and reduced IA success probability caused by high path loss at remote RRUs, this invention introduces an RRU collaboration mechanism, utilizing multiple RRUs to simultaneously serve the MR, thereby increasing the IA success probability. This method fully utilizes location information, historical information, and the concept of collaboration to improve the IA success probability, reduce IA latency, and achieve rapid access. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a diagram of the RRU collaborative network architecture of the C / U decoupled high-speed railway wireless communication system of the present invention;

[0041] Figure 2 1 is a flow chart of a fast beam alignment solution for HSR scenarios according to the present invention;

[0042] Figure 3 This is the initial access process based on the exhaustive method of the single frequency network of the present invention;

[0043] Figure 4 This is the specific process of initial access based on historical information and location information of the present invention;

[0044] Figure 5 This is the initial access process of the dual-frequency network based on historical information of the present invention;

[0045] Figure 6 This is a comparison chart of CS success probability simulations for different solutions of the present invention;

[0046] Figure 7 This is a comparison chart of the CS success probability details of different schemes of the present invention;

[0047] Figure 8 This is a comparison chart of RA success probability simulations for different schemes of the present invention;

[0048] Figure 9 This is a comparison chart of the success probability simulation of different schemes IA of the present invention;

[0049] Figure 10 This is a comparison chart of the success probability details of different schemes IA of the present invention;

[0050] Figure 11 This is a simulation comparison chart of the average time consumption of the RRU cooperation solution IA of the present invention. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0052] Construct a wireless communication network system with decoupling of control plane and user plane in high-speed railway scenario. Figure 1 The network architecture includes two types of RRUs: a low-frequency remote radio unit using an omnidirectional antenna to ensure the coverage performance of the system; and a millimeter-wave remote radio unit that uses directional beamforming technology to compensate for the transmission loss of millimeter-wave signals. These two types of RRUs are deployed geographically and connected to the baseband unit pool via a high-speed backhaul. MR is installed on the top of the train to form a two-hop structure with the access point inside the train, thereby alleviating the problems of penetration loss and group switching. Before MR communicates with the wireless network, it is necessary to perform an initial access process with the RRU through beam alignment, such as Figure 2 As shown in Figure 2, the fast beam alignment method based on position and history information is as follows:

[0053] Step 1: Set the interruption threshold and collaboration threshold in advance. The collaboration threshold is slightly higher than the interruption threshold to determine whether RRU collaboration is required to provide services to the MR.

[0054] Step 2: Create a lookup table to record the best beam pair corresponding to each position. Figure 3 As shown, a full-space beam training process is performed to determine the best beam pair for each position, thereby initializing the lookup table;

[0055] Step 3: When the train passes the same location again, the MR queries the beam training result of the previous moment at that location from the lookup table, that is, the latest historical beam training result;

[0056] Step 4: If the latest historical beam pair is 1, execute the single RRU fast beam alignment solution based on location and historical information. Figure 4 As shown in Figure 1, a new beam training set is constructed based on the best beam pair at the previous moment, and the beam training space is reduced and adjusted. The entire IA process of the dual-frequency network based on historical information is shown in Figure 1. Figure 5 As shown, the best beam pair at this moment is determined;

[0057] Step 5: If the latest historical beam pair is 2, the RRU cooperative fast beam alignment solution based on location and historical information is executed to determine the optimal beam pair at that moment;

[0058] Step 6: Update the lookup table. Only the latest best beam pair is stored in the lookup table for subsequent beam training.

[0059] The method of establishing theoretical analysis is as follows:

[0060] Assuming that the train moves linearly along the track and the RRU is deployed along the track, the maximum beam scanning range of both MR and RRU is set to 180°. Assuming that the number of beams on the MR side and the RRU side are N respectively MR and N BS , assuming that the transmit and receive beams on each side are the same, the transmit (receive) beam codebooks of RRU and MR are expressed as

[0061]

[0062]

[0063] in, Represents the i-th transmit (receive) beam codebook on the RRU side; represents the i-th transmit (receive) beam codebook on the MR side. Correspondingly, the beam widths of RRU and MR are β BS =π / N BS , β MR =π / N MR .

[0064] The single RRU solution is a special case of the RRU cooperation solution. Therefore, the following mainly analyzes the situation where RRU cooperation is required at a certain position at a certain moment.

[0065] Assume that the positions of RRU1 and RRU2 are denoted as d AP1 and d AP2 , the MR position is denoted as d.

[0066] (1) Success probability in the CS stage

[0067] Assume that at time t-1, the optimal transmit beam codebook of the i-th RRU is The corresponding beam index is The optimal receiving beam codebook of MR for the i-th RRU is The corresponding beam index is

[0068] According to the latest beam training results in the lookup table, the best beam at the time t-1 and several surrounding beams at the location are selected as the new beam training set. Assume that the best transmitting beam at the previous moment and the surrounding beams are selected at the RRU. The beams are used as the new transmit beam training set, and the best receiving beam at the previous moment and the surrounding common beams are selected at MR. beams as the new receiving beam training set. Assume that and are all odd numbers. Then at time t, the beam set at the i-th RRU is expressed as

[0069]

[0070]

[0071] The beam training set of MR for the i-th RRU is

[0072]

[0073] In this scheme, MR conducts CS process with two RRUs at the same time. Let the path loss of MR for the i-th RRU be PL APi (d), in dB; the transmit beam gain of the i-th RRU is It represents the angle difference between the main lobe direction of the transmit beam of the i-th RRU and the actual direction, in dB; the receive beam gain corresponding to MR is Indicates the angle difference between the main lobe direction of the MR receive beam for the i-th RRU and the actual direction, in dB.

[0074] in, f cIndicates the carrier frequency, d min represents the vertical distance between the RRU and the track, and d represents the position of the train on the track; β represents the 3dB beamwidth, Δθ represents the angular difference between the main lobe direction and the actual direction of the beam, and η = 4lg2.

[0075] Assume that the received signal of MR is represented by y DL , which can be modeled as

[0076]

[0077] Among them, P t,APi represents the transmit power of the i-th RRU, x i represents the symbol sent, the average power is 1, n represents the noise, G t,APi represents the transmit beam gain of the i-th RRU, G r,MRi β represents the receiving beam gain of MR to the i-th RRU. i represents the path loss fading factor, α i Represents the shadow fading factor. The above parameters have the following conversion relationship:

[0078]

[0079]

[0080]

[0081] Assume i represents the shadow fading between MR and the i-th RRU, and α i and ξ i Has the following conversion relationship:

[0082]

[0083] The received SNR can be expressed as

[0084]

[0085] Among them, σ 2 Represents the noise power.

[0086] When the signal reception quality of the MR is greater than the set interruption threshold Γ, the CS stage is successful, and the success probability is modeled as

[0087]

[0088] (2) Success probability of RA stage

[0089] According to the train movement direction, the i-th RRU beam training space is updated as follows in the RA phase:

[0090]

[0091] Where ΔX RA Indicates the number of beams adjusted during the RA phase.

[0092] From the CS phase to the RA phase, due to the high speed of the train, a long distance offset will occur, resulting in an angle offset from the beam direction selected in the previous CS phase. The Angle-of-Departure (AOD) at the RRU in the CS phase is expressed as The angle of arrival (AOA) at MR is expressed as Then, in the RA phase, the actual AOA direction of the RRU is expressed as The actual AOD direction of MR is expressed as in, Indicates the angular offset.

[0093] If the train is approaching the RRU, Modeled as

[0094]

[0095] If the train is moving away from the RRU, then Modeled as

[0096]

[0097] Where Δx proAPs Indicates the distance offset of the train from the CS phase to the RA phase.

[0098] In this solution, the MR and two RRUs carry out the CS process simultaneously. After receiving the beam in the CS phase, the beam training results are fed back through the omnidirectional link. Therefore, the distance offset from the CS phase to the RA phase is

[0099]

[0100] Among them, v represents the train speed, τ represents the duration of one beam scan, Indicates the transmit beam space size of the i-th RRU, represents the receiving beam space size of MR for the i-th RRU, Indicates the optimal transmit beam index of the i-th RRU.

[0101] After obtaining the distance offset, the angle offset is calculated according to equations (12) and (13): and Assume that the path loss of MR for the i-th RRU is PL APi (d+Δx proAPs ), in dB; the receiving beam gain of the i-th RRU is Indicates the angle difference between the main lobe direction of the RRU receiving beam and the actual direction in the RA phase, in dB; the corresponding transmit beam gain of MR is The unit is dB.

[0102] Similarly, the success probability of the RA stage can be modeled as

[0103]

[0104] Among them, P t,MRi represents the transmit power allocated by MR to the i-th RRU, G r,APi represents the receiving beam gain of the i-th RRU, G t,MRi represents the transmit beam gain of MR for the i-th RRU, β i represents the path loss fading factor, α i represents the shadow fading factor. The parameter conversion relationship is similar to equations (5) to (8).

[0105] Therefore, the IA process success probability is modeled as

[0106]

[0107] Based on the theoretical analysis results of the IA success probability, we will now analyze the average time consumption of the IA process. This can be divided into the following two cases:

[0108] If a single RRU can meet the service requirements,

[0109]

[0110] in, and They represent the beam training space size of the RRU and MR selected in the CS phase and the beam training space size of the RRU in the RA phase, τ represents the duration of a beam scan, and τ fb represents the omnidirectional link feedback time, and T represents the beam training period.

[0111] If RRU collaboration is performed,

[0112]

[0113] in, and They respectively represent the beam training space size of the i-th RRU in the CS phase, the beam training space size of the MR for the i-th RRU, and the beam training space size of the i-th RRU in the RA phase.

[0114] Compared to exhaustive methods and cellular network architectures, the technical solution described in this invention can significantly improve the initial access success probability and reduce latency in high-speed rail wireless communication systems. A detailed performance analysis will be provided in the simulation results.

[0115] The simulation parameters are shown in Table I, and the simulation results are shown in Table Figure 5 、 Figure 6 、 Figure 7 、 Figure 8 、 Figure 9 、 Figure 10 and Figure 11 shown.

[0116] Table I Parameter settings

[0117]

[0118] Figure 6 The following chart compares the success probability of the CS phase for different schemes. In the single-RRU scheme, the CS success probability near the RRU area is 1. As the train approaches the cell edge, the CS success probability drops rapidly, reaching approximately 0.77 at 250 meters. In the RRU coordination scheme, the CS success probability is almost always 1.

[0119] Figure 7 The simulation details of the CS phase success probability at distances of 200m to 300m are presented. In the RRU collaboration scheme, the CS phase success probability decreases slightly at distant RRUs due to increased path loss. However, the simulation curve for the historical information-based scheme changes more gently, while the CS success probability curve for the exhaustive method-based scheme shows a significant drop at 250m. The specific reason for this is that in the exhaustive method-based RRU collaboration process, RRUs that join the collaboration later must wait until the RRU closest to the MR completes the CS process before entering the CS phase. During this waiting period, distance offsets occur due to train movement, and in the exhaustive method, longer beam search times correspond to larger distance offsets. When the train collaborates closer to RRU1 (i.e., before 250m), the path loss of RRU2 after the distance offset is lower than before the distance offset due to the train's movement toward RRU2. This increases β2, improves the MR's integrated received signal quality, and increases the CS success probability. When the train moves closer to RRU2 (i.e., after 250 m), the path loss of RRU1, which joins the collaboration later, increases due to the distance offset, i.e., β1 decreases, and the MR comprehensive received signal quality decreases, resulting in a lower probability of CS success.

[0120] Comprehensive analysis shows that in the single-RRU solution, the MR received signal quality degrades at the cell edge due to increased path loss, resulting in a rapid decrease in the CS success probability. In contrast, the RRU collaboration solution considers the application of a cell-free network architecture and collaborates at remote RRUs, thereby improving received signal quality and increasing the CS success probability. Therefore, the RRU collaboration solution can effectively improve CS performance at the cell edge compared to the single-RRU solution.

[0121] Figure 8 Simulation curves of RA success probabilities for different schemes are shown. In areas near the RRU, the RA success probability for the exhaustive search scheme is 0, while that for the historical information-based scheme is 1. This is because the exhaustive search method produces a large range offset and leads to significant angular offset near the RRU, invalidating the previous beam alignment results and causing RA failure. The historical information-based scheme, on the other hand, significantly reduces the beam training space. Shorter displacements correspond to smaller angular offsets, ensuring the validity of previous beam training results. Therefore, the historical information-based scheme can effectively improve RA performance near the RRU.

[0122] Figure 9 The simulation results of IA success probability of different schemes are shown. Figure 10 The simulation details of the collaborative area are shown. Since the IA success probability is jointly determined by the CS success probability and the RA success probability, Figure 9 and Figure 10 It has the following comprehensive characteristics:

[0123] (1) Compared with the exhaustive method-based scheme, the historical information-based scheme reduces the beam search space and improves the IA success probability near the RRU.

[0124] (2) The RRU collaboration scheme can significantly improve the success probability of IA at the edge of RRU coverage.

[0125] exist Figure 9 In the IA success probability of the RRU cooperation solution, the IA success probability decreases slightly at the edge of the RRU coverage due to the existence of high path loss, and the IA curve of the solution based on historical information is relatively flat. The IA curve based on the exhaustive method will show a significant decrease in the middle of the RRU geographical distribution. The specific reasons are as follows: Figure 7 similar.

[0126] Figure 11The simulation results of the average IA time consumption in the RRU cooperation scheme are shown. From formulas (17) and (18), it can be seen that the average IA time consumption is inversely correlated with the IA success probability. Since the IA success probability near the RRU is close to 0 in the exhaustive method, the IA time consumption at the corresponding position tends to infinity. When the IA success probability of the exhaustive method gradually returns to 1, its average IA time consumption also tends to stabilize at approximately 138.3ms. After RRU cooperation, since the two RRUs in the exhaustive method start the CS phase at different times, the total CS phase consumption will be longer than the single RRU solution, so the average IA time consumption increases slightly to approximately 279ms.

[0127] In the solution based on historical information, the IA success probability at distant RRUs (200 to 300 meters) decreased slightly, reaching a minimum of approximately 0.9953. Consequently, the average IA duration in this area increased, reaching a maximum of approximately 2 ms. At other locations, the IA success probability was close to 1, with an average IA duration of approximately 1.7 ms.

[0128] Comprehensive analysis shows that compared to the exhaustive approach, the historical information-based approach significantly reduces the beam training space. Consequently, the average IA time is significantly reduced to only one percent of the exhaustive approach's average IA time, achieving rapid beam alignment. Furthermore, RRU collaboration significantly improves the success rate of IA at the cell edge in the original cellular network.

Claims

1. A fast beam alignment method based on location and historical information in high-speed railway scenarios, characterized by: Includes the following: In high-speed railway (HSR) scenarios, a wireless network communication system with decoupled control and user planes is constructed. The network architecture includes two types of remote radio units (RRUs). One is a low-frequency RRU with an omnidirectional antenna to ensure system coverage; the other is a millimeter-wave (MW) remote radio unit (RRU), which uses directional beamforming technology to compensate for millimeter-wave signal transmission losses. These two RRUs are geographically distributed and connected to the baseband unit pool via a high-speed backhaul. A mobile relay (MR) is installed on the roof of the train, forming a two-hop structure with the onboard access point, thereby alleviating penetration loss and group handoff issues. Before the MR mobile relay communicates with the wireless network, it must perform an initial access process with the MW remote radio unit (RRU) through beam alignment. The following method is used to perform fast beam alignment based on location and historical information: Step 1: Set the interruption threshold and collaboration threshold in advance. The collaboration threshold is slightly higher than the interruption threshold to determine whether RRU collaboration is required to provide services to the MR. Step 2: Create a lookup table to record the best beam pair for each position; perform a full-space beam training process to determine the best beam pair for each position, thereby initializing the lookup table; Step 3: When the train passes the same location again, the MR queries the beam training result of the previous moment at that location from the lookup table, that is, the latest historical beam training result; Step 4: If the latest historical beam pair is 1, execute the single RRU fast beam alignment solution based on location and historical information to determine the best beam pair at that moment; Step 5: If the latest historical beam pair is 2, the RRU cooperative fast beam alignment solution based on location and historical information is executed to determine the optimal beam pair at that moment; Step 6: Update the lookup table. Only the latest best beam pair is stored in the lookup table for subsequent beam training.

2. The method for rapid beam alignment based on location and historical information in a high-speed railway scenario according to claim 1 is characterized in that: The full-space beam training process described in step 2 specifically includes the following steps: Step 2.1: Single RRU full spatial beam training process; The MR selects the nearest RRU to perform the initial access (IA) process. The process begins with a cell search (CS) phase. The RRU traverses all transmit beams, while the MR traverses all receive beams and measures the received signal quality. If the received signal quality falls below the interruption threshold, the CS phase is considered a failure. If the received signal quality exceeds the coordination threshold, the single-RRU solution is implemented, meaning that the MR only needs one RRU to provide service. The MR selects the beam with the strongest received signal, records the beam index, and calculates the probability of success in the CS phase. Next, since the single frequency network is considered in the full spatial beam training process, the RRU needs to traverse the entire space to send the random access channel RACH scheduling information to ensure that the MR can obtain this information; After receiving the RACH scheduling information, the MR enters the random access (RA) phase. The MR randomly selects a preamble from a set of mutually orthogonal preambles and sends it to the RRU using the optimal beam direction recorded by the MR in the CS phase. Simultaneously, the RRU traverses all receiving directions to receive the signal. When the received signal quality exceeds the interruption threshold, it selects the beam with the strongest received signal and records its index. The success probability of the RA phase is then calculated. Thus, the MR and RRU can determine an optimal beam pair and initialize the lookup table; Step 2.2, RRU collaborative full spatial beam training process; The MR selects the nearest RRU to perform the IA process. During the CS phase, if the MR's received signal quality is between the interruption threshold and the cooperation threshold, cooperation is required. The RRU next closest to the MR is selected to join and provide joint service. After the first RRU's CS phase ends, the RRU that joins later will start the CS phase. After the CS phase of both RRUs ends, the RRUs traverse the entire space and send RACH scheduling information. After the MR obtains the resource, it enters the RA phase. During the RA phase, the MR sends a preamble using the optimal beam direction trained in the CS phase. The wireless network receives the preamble information through two RRUs. If the received signal quality exceeds the interruption threshold, the RA process succeeds, the success probability is calculated, and the beam index is recorded. If both the CS and RA phases are successful, a communication link between the MR and the network is established. The MR can now communicate with the network through multiple RRUs, achieving higher data transmission rates. Thus, the MR and RRU can determine two optimal beam pairs and initialize the lookup table.

3. The method for rapid beam alignment based on location and historical information in a high-speed railway scenario according to claim 1 is characterized in that: The single RRU fast beam alignment solution based on location and historical information described in step 4 specifically includes the following steps: Step 4.1, CS phase: For a certain location, in the CS phase, the MR and RRU first query the lookup table for the latest historical beam training results corresponding to the location, and select the best beam and several surrounding beams to form a new beam training set for this beam training; In step 4.2, the wireless network communication system uses a decoupled architecture between the control plane and the user plane. Therefore, the MR feeds back the beam training results of the CS phase to the RRU via an omnidirectional link. Step 4.3: Through step 4.2, the RRU can obtain the beam training results of the MR in the CS phase. Therefore, the RRU selects an appropriate beam based on the results and sends RACH resources in a directionally controlled manner. In step 4.4, after receiving the RACH resource, the MR enters the RA phase. Considering the angle offset caused by train movement during the transition from the CS phase to the RA phase, the beam training set is adjusted according to the train's direction of travel. Specifically, several beams are added along the train's direction of travel and the same number of beams are removed in the opposite direction of travel to form a new beam training set to accommodate the train's movement. The MR sends a preamble in the optimal beam direction obtained through CS phase training. The RRU traverses and receives the beams in the new beam training set and records the optimal beam reception direction. This completes the IA process for MR and RRU. In addition, due to the use of location information and historical information, the beam training space is greatly reduced, which can reduce the IA delay.

4. The method for rapid beam alignment based on location and historical information in a high-speed railway scenario according to claim 1 is characterized in that: The RRU cooperative fast beam alignment solution based on location and historical information described in step 5 specifically includes the following steps: Step 5.1, CS phase: During this IA process, the MR conducts CS with both RRUs simultaneously. The specific process is the same as the single RRU fast beam alignment process based on location and historical information. Refer to step 4.

1. In step 5.2, the wireless network communication system uses a decoupled architecture between the control plane and the user plane. This allows the MR to feed back the beam training results of the CS phase to the RRU via an omnidirectional link. Step 5.3: The RRU sends RACH scheduling information in a targeted manner based on the feedback information obtained. In step 5.4, after receiving the RACH resource, the MR initiates the RA phase: the MR sends a preamble to the two RRUs using the optimal beam direction obtained through CS phase training. The specific process is the same as step 4.

4. The wireless network receives the signal through the two RRUs. If the received signal quality is higher than the interruption threshold, the RA phase is successful and the RRU records the optimal beam index.

Images