Boundary range expansion method based on mobile communication

Abstract

The invention relates to the technical field of mobile communication, and discloses a boundary range expansion method based on mobile communication. The method comprises the following steps: determining a user position and an environment three-dimensional model of the position, and planning signal transfer nodes over a mountainous area so as to form a signal transfer chain; working time slots, communication wavebands and communication protocols adopted by the unmanned aerial vehicle corresponding to different signal transfer nodes are arranged; adjusting the position of the signal transfer node based on the position of the signal transfer node, the interference threshold distance of the unmanned aerial vehicle frequency band and the three-dimensional model of the environment; the unmanned aerial vehicle is dispatched to a signal transfer node, and communication connection among a base station, a signal transfer chain and a user is established; and the positions of the signal transfer node and the unmanned aerial vehicle are adjusted based on the real-time flight condition of the unmanned aerial vehicle. According to the invention, the defects of mountain terrain obstruction and traditional communication means can be overcome, and the boundary range of mobile communication is expanded.

Description

Technical Field

[0001] This application relates to the technical field of mobile communications, and in particular, to a method for expanding the boundary range based on mobile communications. Background Art

[0002] Mobile communication technology has gone through multiple rounds of updates and upgrades, and its coverage and service quality have been continuously improved. A relatively complete signal network has been achieved in cities and most plain areas. However, in mountainous areas with complex terrain, the number of base stations is limited and scattered, and traditional mobile communication methods cannot achieve comprehensive and stable coverage. Mountain blockages often cause wireless signal attenuation or inability to reach, posing great challenges to field communication.

[0003] For personnel engaged in scientific research, search and rescue, etc. in mountainous areas, real-time data information transmission is extremely important. They often need to upload a large amount of data (such as images and videos) collected at high bandwidth and high speed to the cloud for analysis using remotely deployed computing programs or large models. These large models have huge demands for computing resources and energy and cannot be carried to the field site. Only through data upload can real-time processing be achieved. However, traditional satellite communication solutions can provide wide-area coverage, but they have insufficient bandwidth and high costs in high-load data upload and are difficult to achieve efficient real-time interaction. Summary of the Invention

[0004] In order to overcome the deficiencies of mountain terrain barriers and traditional communication means, this application provides a method for expanding the boundary range based on mobile communications.

[0005] This application provides a method for expanding the boundary range based on mobile communications, adopting the following technical solutions:

[0006] A method for expanding the boundary range based on mobile communications includes the following steps:

[0007] S1. Obtain user location information, retrieve the three-dimensional model of the user's environment, and determine the user's position in the three-dimensional model of the environment;

[0008] S2. Based on the three-dimensional model of the environment and the user location information, plan signal relay nodes in the air over the mountainous area to form a signal relay chain, where the signal relay nodes are used for drones to hover;

[0009] S3. Based on the adjacent node relationship of the signal relay chain, arrange the working time slots, communication bands, and communication protocols adopted by the drones corresponding to different signal relay nodes, where the communication band of each signal relay node with the previous node is different from that with the next node;

[0010] S4. Adjust the position of the signal relay node based on the position of the signal relay node, the interference threshold distance of the UAV frequency band, and the three-dimensional model of the environment;

[0011] S5. Dispatch the UAV to the signal relay node to establish a communication connection between the base station, the signal relay chain, and the user;

[0012] S6. Adjust the positions of the signal relay node and the UAV based on the real-time flight situation of the UAV.

[0013] Optionally, S2 includes the following steps:

[0014] S21. Identify all candidate points that can satisfy the UAV hovering based on the terrain;

[0015] S22. For each pair of candidate nodes, detect whether there is a direct line of sight; if there is a mountain blockage, mark it as not directly connectable;

[0016] S23. For two visible points, calculate their three-dimensional distance and determine whether it exceeds the upper limit of the safe communication distance. If it exceeds, mark it as not directly connectable;

[0017] S24. Form a candidate link from the base station to the user with the candidate nodes that can be directly connected, and construct a node connection graph;

[0018] S25. Take the upper limit of the frequency band communication distance that alternates in sequence along the nodes as a constraint condition, and find the minimum node set through the minimum node coverage strategy.

[0019] Optionally, S3 includes the following steps:

[0020] S31. Determine frequency band one and frequency band two; among them, frequency band one and frequency band two correspond to the dual-mode communication module equipped on the UAV;

[0021] S32. Determine the working time slot one and working time slot two of the UAV;

[0022] S33. Determine the communication band and communication protocol adopted by the UAV corresponding to the signal relay node, so that among any adjacent five consecutive UAVs, the first UAV and the second UAV communicate through frequency band one in working time slot one, the second UAV and the third UAV communicate through frequency band two in time slot one, the third UAV and the fourth UAV communicate through frequency band one in time slot two, and the fourth UAV and the fifth UAV communicate through frequency band two in time slot two.

[0023] Optionally, S3 further includes the following steps:

[0024] Insert a protection interval of 0.1 - 0.5 ms between time slot one and time slot two.

[0025] Optionally, S4 includes the following steps:

[0026] S41. Take three consecutive signal relay nodes and calculate the geometric distance between the first signal relay node and the third signal relay node;

[0027] S42. Determine whether the geometric distance between the first signal relay node and the third signal relay node is greater than the signal interference distance threshold; if so, determine the positions of these three signal relay nodes, if not:

[0028] Based on the 3D environmental model, determine whether the area between the first signal relay node and the third signal relay node is blocked by a mountain; if so, determine the positions of these three signal relay nodes, if not, move the signal relay node farther from the base station on this link to increase the angle at the middle signal relay node of the connection broken line of these three signal relay nodes, and return to S41;

[0029] S43. Repeat the above step and perform hierarchical optimization starting from the near-base station section based on chain transmission.

[0030] Optionally, S4 includes the following steps:

[0031] S401. Take three consecutive signal relay nodes and calculate the geometric distance between the first signal relay node and the third signal relay node;

[0032] S402. Determine whether the geometric distance between the first signal relay node and the third signal relay node is greater than the signal interference distance threshold; if so, determine the positions of these three signal relay nodes, if not:

[0033] Based on the 3D environmental model, determine whether the area between the first signal relay node and the third signal relay node is blocked by a mountain; if so, determine the positions of these three signal relay nodes, if not, move the signal relay node farther from the base station on this link to increase the angle at the middle signal relay node of the connection broken line of these three signal relay nodes, and return to S401;

[0034] S403. Repeat the above step and perform hierarchical optimization starting from the near-base station section based on chain transmission.

[0035] Optionally, S4 includes the following steps:

[0036] S41. Evaluate multiple connectable candidate signal relay links to screen out the candidate links to be optimized;

[0037] S42. Perform local structure optimization on the positions of the signal relay nodes in the candidate links to be optimized to improve the link geometric shape and communication quality;

[0038] S43. Introduce a node perturbation template and a path switching mechanism during the local optimization process to improve the optimizability and convergence effect of the overall link.

[0039] Optionally, S41 includes the following steps:

[0040] S411. Based on the constructed node connection relationship graph, identify at least two visible link paths from the base station to the target user;

[0041] S412. For each link path, calculate the comprehensive score value corresponding to the path according to multiple preset performance indicators; wherein, the preset performance indicators include the number of relay nodes, path length, occlusion rate, and signal interference intensity between nodes;

[0042] S413. Based on the score value, select at least one path with a better score from all feasible paths as the candidate link to be optimized.

[0043] Optionally, S42 includes the following steps:

[0044] S421. Select three consecutive relay nodes in the candidate link in sequence to form a broken line segment triple;

[0045] S422. Calculate the included angle value of the broken line segment and the geometric distance between the first and last nodes respectively, and determine whether the set included angle threshold and interference distance threshold are satisfied;

[0046] S423. If the threshold requirements are not met, perform position perturbation on the intermediate node, including preset direction perturbation, distance perturbation, and height perturbation operations, generate multiple perturbation points and calculate the corresponding parameters;

[0047] S424. If there are perturbation points that meet the requirements of the broken line angle and interference limit, update the node position; if there are no effective perturbation points, expand the analysis range to a link segment including five consecutive nodes and continue the optimization judgment.

[0048] Optionally, S43 includes the following steps:

[0049] S431. When the node position perturbation attempt fails to obtain a satisfactory update result in multiple consecutive groups of broken line segments, trigger the preset perturbation template mechanism, and perform replacement tests on the node position according to the standard template stored in the perturbation library;

[0050] S432. If the current candidate link cannot be further optimized at multiple key node positions, switch to another candidate link path and return to S42;

[0051] S433. Compare all optimized candidate links based on a unified link performance evaluation criterion, and select the link with the highest evaluation score as the final signal relay communication path.

[0052] Optionally, the S5 includes the following steps:

[0053] For a drone that uses two frequency bands for information transmission and reception in one of two adjacent working time slots and does not perform information transmission and reception in the other working time slot, decode and re-encode the received signal during the non-information transmission and reception working time slot and the guard interval of the time slot, and send it out during the working time slot for information transmission and reception;

[0054] For a drone that performs information transmission and reception in both of two adjacent working time slots, perform gain forwarding on the received information.

[0055] Optionally, the S6 includes the following steps:

[0056] S61. Discretize the three-dimensional space of the user's environment into a three-dimensional grid and perform position encoding;

[0057] S62. Quantify the channel quality based on the received signal strength and the bit error rate;

[0058] S63. Standardize and classify the flight actions based on the flight adjustment orientation and the flight adjustment distance and number them;

[0059] S64. Set a multi-objective reward function based on the adjusted link throughput, the change amount of the distance to adjacent drones after adjustment, and the mobile energy consumption;

[0060] S65. Control the drone to perform different flight actions and reset, and collect data to substitute into the multi-objective reward function for calculation to obtain a local optimal solution;

[0061] S66. Repeat the above step, and perform hierarchical optimization starting from the near base station segment based on chain transfer.

[0062] In summary, the present application includes at least one of the following beneficial technical effects:

[0063] 1. By flexibly deploying signal relay nodes in a mountainous environment and using different frequency band and time slot configurations between adjacent nodes, the present application can significantly reduce the communication blind area caused by mountain blockage. Compared with the traditional single base station coverage method in high mountains or valleys, this solution can use the chain relay method of high-altitude drones to bypass mountain peaks and terrain obstacles, enabling remote users to still obtain good bandwidth and communication quality under mountainous terrain.

[0064] 2. By arranging adjacent nodes in different working frequency bands and time slots respectively, the present application significantly reduces the interference superposition generated in the same frequency band. At the same time, a protection interval is inserted between two working time slots. Combining the fine-tuning of the geographical positions between nodes and the evaluation of mountain occlusion can further reduce the influence of signal leakage from adjacent or non-adjacent nodes, thereby ensuring that the entire relay network still maintains relatively stable performance during high-concurrency or high-load transmission.

[0065] 3. From node planning to frequency band allocation, then to position optimization and flight action adjustment, the present application aims to reduce the total number of UAVs and energy consumption, while ensuring the necessary communication throughput and safety distance. Under the guidance of the multi-objective reward function, each UAV can complete the decoding, gain forwarding or re-encoding of the received signal within a limited time slot, realizing the refined utilization of frequency resources and hardware computing power, and enabling the entire communication system to have both efficient data backhaul and sustainable flight deployment in the complex mountain environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a flowchart of a method for expanding the boundary range based on mobile communication in an embodiment of the present application.

[0067] Figure 2 It is a flowchart of sub-step S1 in an embodiment of the present application.

[0068] Figure 3 It is a flowchart of sub-step S2 in an embodiment of the present application.

[0069] Figure 4 It is a flowchart of sub-step S3 in an embodiment of the present application.

[0070] Figure 5 It is a flowchart of sub-step S4 in an embodiment of the present application.

[0071] Figure 6 It is a flowchart of sub-step S41 in an embodiment of the present application.

[0072] Figure 7 It is a flowchart of sub-step S42 in an embodiment of the present application.

[0073] Figure 8 It is a flowchart of sub-step S43 in an embodiment of the present application.

[0074] Figure 9 It is a flowchart of sub-step S6 in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0075] The following details the embodiments of the present application, and the examples of the embodiments are shown in the drawings.

[0076] In the description of this specification, the descriptions referring to terms such as "certain embodiments", "one embodiment", "some embodiments", "schematic embodiments", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiments or examples. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0077] An embodiment of the present application discloses a method for expanding the boundary range based on mobile communication. Referring to Figure 1 , it includes the following steps S1 - S6.

[0078] S1. Obtain the user's location information, retrieve the three - dimensional model of the user's environment, and determine the user's position in the three - dimensional model of the environment.

[0079] The system first needs to obtain the user's location information and retrieve the corresponding three - dimensional environment model to accurately determine the user's specific position in the environment model. This process usually relies on the integration of geolocation technologies, such as GPS, Beidou, or other satellite navigation systems, to obtain the initial longitude and latitude coordinates of the area where the user is located.

[0080] Then, by mapping these coordinates to a high - precision three - dimensional geographic information database, the user's position can be accurately marked in a relatively complex terrain or urban scene. The three - dimensional model refers to the overall digital and visual representation of the target area in terms of the horizontal plane, vertical direction, and altitude changes. In different embodiments, it can be pre - entered into the database, or constructed based on aerial photography, satellite remote sensing data, and digital elevation models before the task starts, or more detailed spatial reconstruction can be carried out according to the lidar of the unmanned aerial vehicle.

[0081] In actual use, the user may be in a mountainous area with rough terrain and overlapping peaks. To make up for the lack of information in the height dimension of the conventional plane map, the three - dimensional environment model can reflect the true structures of different mountains, valleys, and slopes, facilitating the subsequent arrangement of relay nodes at the top or middle of the mountains. For example, if a scientific research team is conducting scientific research operations in a mountainous area at a high altitude, they only need to obtain the longitude, latitude, and altitude data through satellite positioning equipment in the wild and then import these data into the three - dimensional model to accurately estimate the relative positions of the user and the surrounding mountains and valleys. In this way, it is possible to initially judge whether there is a large - area mountain blockage between the user and the base station, and also to predict where it is more suitable to deploy unmanned aerial vehicles.

[0082] S2. Based on the 3D environmental model and the user's location information, plan signal relay nodes over the mountainous area to form a signal relay chain, where the signal relay nodes are used for the UAV to hover.

[0083] Based on the 3D environmental model and the specific location information of the user, the system plans signal relay nodes for UAV hovering over the mountainous area, thus constructing a communication chain that can achieve cross-mountain obstruction. Specifically, by setting a number of candidate points at positions with good line of sight in the air, and then according to terrain, interference and communication coverage requirements, these points are connected in series into one or more feasible candidate links. Candidate points refer to the set of coordinates that may become the hovering positions of UAVs in the 3D environment. The system will judge one by one whether there is mountain occlusion or excessive height difference between them to determine whether a direct connection can be established. If the spatial connection between two points is completely blocked by a mountain peak or cliff, it is considered non-line-of-sight and it is difficult to form a stable relay chain; if there is a basic line of sight despite the long distance and within the acceptable safe communication distance, they can be temporarily reserved as a feasible node combination.

[0084] Specifically, in one embodiment, S2 includes the following steps S21 - S25.

[0085] S21. Mark all candidate points that can meet the hovering requirements of the UAV based on the terrain.

[0086] In step S21, the system needs to mark all candidate points that can meet the hovering requirements of the UAV within the mountainous area based on the existing 3D terrain data. To achieve this goal, the target area can be discretized according to a certain spatial resolution, and then combined with terrain undulation, altitude distribution and meteorological conditions to judge which coordinate positions have enough airspace for hovering. In addition, the system will also consider parameters such as the power characteristics, maximum flight height and safety radius of different UAVs to ensure that the selected points are feasible in actual deployment.

[0087] As an example, in one embodiment, the system first analyzes the digital elevation model (DEM) or more detailed geographic information data, and marks areas with special terrain features such as ridges, valleys and steep slopes. For example, if the wind speed is too high in a certain mountaintop area all year round, it may be difficult to ensure the stable hovering of the UAV, and the system will exclude this area in the algorithm. On the other hand, in areas with relatively gentle slopes or open terrains, due to the relatively open field of vision and relatively small wind force, they are often more suitable for arranging aerial hovering points. In this process, the conditions that can meet the hovering requirements of the UAV often involve various considerations, including but not limited to: the terrain occlusion degree at the specified height, local wind field environment data, distribution of adjacent obstacles, and the impact of altitude on air density, etc. By quantifying these influencing factors one by one and writing them into the screening rules, the system can quickly exclude some areas that are not suitable for node layout.

[0088] S22. For each pair of candidate nodes, detect whether there is a direct line of sight; if there is mountain occlusion, mark it as not directly connectable.

[0089] In this step, the system will perform line-of-sight detection for each pair of candidate points screened in the previous step. The line of sight is whether the connection line from one spatial coordinate to another is completely blocked by mountains or other obstacles. Here, the digital elevation model or higher-resolution 3D geographic information data is usually combined to analyze the profile line between two points. Once it is found that the profile height exceeds the vertical height of the connection line at a certain point, it means there is occlusion, and thus it is determined that this pair of candidate points has no possibility of direct connection. Taking a point N1 on the mountainside and another point N2 on the mountaintop as an example, the system can sample the straight line from N1 to N2 in 3D space. If the altitude or mountain shape of any sampling point indicates that the connection line is substantially blocked, the potential link from N1 to N2 will be marked as "non-line-of-sight". Through this step, the subsequent interference assessment burden can be reduced because only node pairs with a line of sight are likely to perform wireless communication relaying, and there is no need to waste computing resources and flight time in areas with severe blockage.

[0090] S23. For two points with a line of sight, calculate their 3D distance and determine whether it exceeds the upper limit of the safe communication distance. If it exceeds, mark it as not directly connectable.

[0091] In this step, for those node pairs that have passed the line-of-sight detection, it is necessary to further verify whether their 3D spatial distance exceeds the upper limit of the safe communication distance. The safe communication distance here is selected as the maximum safe communication distance of all communication frequency bands used by the UAVs for mutual communication. This is because the maximum reliable transmission radius of different frequency bands (such as 2.4 GHz or 5 GHz) in the mountain environment is different. If the distance between a certain node pair exceeds the limit value of this frequency band, even if the line of sight is good, it cannot meet the requirements of stable communication. Therefore, a preliminary screening is carried out here, and then a secondary screening of the candidate links is performed according to the safe communication distance of different frequency bands in the subsequent steps. The system will uniformly store all visible and distance-compliant node pairs in the "feasible connection link" list and exclude the point pairs that exceed the distance threshold.

[0092] S24. Combine the candidate nodes that can be directly connected to form candidate links from the base station to the user, and construct a node connection diagram.

[0093] In this step, the system interconnects all candidate nodes that meet the line-of-sight condition and do not exceed the safe communication distance, thus forming one or more candidate links from the base station to the user. Here, a candidate link refers to a communication path that connects a certain node starting point at the base station end and a node ending point at the user end through several intermediate nodes. After integrating all possible candidate paths, a "node connection graph" covering all nodes and their connectable relationships can be constructed. This connection graph takes nodes as vertices and feasible connected links as edges, intuitively showing how different paths are intertwined in a three-dimensional terrain. For example, if three visible nodes N1, N2, and N3 are continuously distributed on a certain ridge, and the base station can first establish a connection with N1 and then transmit to the user through N2 and N3 in sequence, then the path N1, N2, N3 will be displayed as a complete link in the node connection graph.

[0094] S25. Take the upper limit of the communication distance of the frequency band that alternates in sequence along the nodes as a constraint condition, and find the minimum node set through the minimum node coverage strategy.

[0095] In this step, a minimum node coverage strategy is introduced to find one or several effective paths with as few nodes as possible but meeting the frequency band communication distance constraint from among many candidate links. Since in the communication process, adjacent nodes need to alternately use different frequency bands to avoid interference or conflicts, this will also become a constraint condition for screening paths. By taking the upper limit of the alternating frequency band communication distance as the screening condition for candidate links, the system will perform another optimization search on the node connection graph to eliminate unnecessary intermediate nodes, thereby reducing the requirements for the number of drones and flight energy consumption.

[0096] Since the maximum reliable transmission radius of different frequency bands (such as 2.4 GHz or 5 GHz) varies in a mountainous environment, if the distance between a certain node pair exceeds the limit value of this frequency band, stable communication requirements cannot be met even with good line of sight. For example, if the distance between N3 and N4 is 2.5 kilometers, but the best communication range of the drone in the 5 GHz frequency band is only about 2 kilometers, the system will automatically mark the N3 - N4 link as not establishable.

[0097] For example, in a long link of N1 → N2 → N3 → N4 → N5, if N2 and N4 can be directly connected and there is no insurmountable terrain obstruction in the middle, the system can ignore the intermediate node N3, thereby reducing the overall number of hops and resource occupancy.

[0098] S3. Based on the adjacent node relationship of the signal relay chain, arrange the working time slots, communication frequency bands, and communication protocols adopted by the drones corresponding to different signal relay nodes. Among them, the communication frequency band of each signal relay node is different from that of the previous node and the next node.

[0099] The system configures the communication band, working time slot, and specific communication protocol for each node in the signal relay chain formed in the previous stage, ensuring that the frequency bands used by adjacent nodes do not conflict. The signal relay chain mentioned here refers to the multi-hop communication path from the base station to the user, where each hop corresponds to the connection relationship between a drone node and its front and rear nodes. Since in mountainous environments, different frequency bands (such as 2.4 GHz and 5 GHz) have their own propagation characteristics and interference characteristics, to achieve continuous and stable chain connection, when the same drone transmits and receives with its previous node in a specific band, it needs to switch to another band when communicating with the subsequent node, thus avoiding mutual interference on the same frequency band.

[0100] In specific implementation, the system first identifies two available main frequency bands and allocates multi-mode communication modules to these drones to meet the technical requirements of simultaneously supporting two different bandwidths and frequencies. Immediately afterwards, it is also necessary to divide the entire transmission process into several time segments, that is, working time slots, so that each drone can communicate with its front and rear nodes in different time slots respectively, or perform signal decoding, caching, and power management during idle time slots.

[0101] Specifically, in one embodiment, S3 includes the following steps S31 - S33.

[0102] S31. Determine frequency band one and frequency band two; where frequency band one and frequency band two correspond to the dual-mode communication modules equipped on the drones.

[0103] S32. Determine the working time slot one and working time slot two of the drones.

[0104] The system first clarifies two different communication frequency bands. Usually, mainstream frequency bands such as 2.4 GHz and 5 GHz can be selected, or a more appropriate frequency range can be determined according to specific task requirements and equipment capabilities. The reason for using two frequency bands is to make full use of the characteristics of the dual-mode communication modules and reduce co-channel interference by alternately switching frequency bands between adjacent nodes. At the same time, the system also needs to distinguish between working time slot one and working time slot two, allowing each drone to perform transmission and reception tasks or perform temporary idle, data processing, etc. operations in different time slots. In this way, multi-reuse or isolation can be achieved at the time level and frequency level, reducing the risk of mutual interference between each link.

[0105] S33. Determine the communication band and communication protocol used by the drones corresponding to the signal relay nodes, so that among any adjacent five consecutive drones, the first drone and the second drone communicate through frequency band one in working time slot one, the second drone and the third drone communicate through frequency band two in time slot one, the third drone and the fourth drone communicate through frequency band one in time slot two, and the fourth drone and the fifth drone communicate through frequency band two in time slot two.

[0106] In the specific implementation process, the operator will number each UAV participating in the communication and arrange the frequency bands and time slots they use according to the order of the nodes in the relay link. For example, if there are a total of seven UAVs connected in sequence by number in the system, the first and the second UAVs can communicate using frequency band one during working time slot one, the second and the third UAVs communicate using frequency band two during the same time slot one, while the third and the fourth UAVs switch to working time slot two and then use frequency band one, the fourth and the fifth UAVs switch to frequency band two during the same time slot two, and so on. Through this way of "time slot interleaving + frequency band alternation", even if neighboring nodes all transmit signals in the near vicinity, the co-frequency interference can be significantly reduced. As shown in the table, it can be clearly seen which frequency band each UAV uses in each time slot, achieving an efficient arrangement of the communication link.

[0107]

[0108] Optionally, step S3 further includes the following steps:

[0109] S34. Insert a protection interval of 0.1 - 0.5 ms between time slot one and time slot two.

[0110] Inserting a protection interval of 0.1 - 0.5 ms between time slot one and time slot two mainly aims to reserve a buffer for the switching of the hardware state and the tail or pre-operation of short-time signals, rather than for completing the full decoding and encoding processes. Since in some scenarios, the UAV may need to perform communication tasks using different frequency bands in two adjacent time slots, the radio frequency front end or the baseband processing part often needs to adjust the transmission power, switch the local oscillator frequency, or refresh the transceiver cache, etc. Without a protection interval, these short-time hardware actions are likely to cause signal collisions or device mismatches at the junction of the two time slots. By setting such a very short protection interval, each UAV can have a necessary hardware idle cycle when switching the frequency band or working mode, thus maintaining the smoothness and stability of the link transmission.

[0111] S4. Adjust the position of the signal relay node based on the position of the signal relay node, the interference threshold distance of the UAV frequency band, and the three-dimensional model of the environment.

[0112] The system will further optimize and fine-tune the coordinates and layout of these relay nodes based on the preliminary position arrangement of the signal relay nodes, in combination with the interference threshold of the UAV frequency band and the three-dimensional model of the mountainous environment. The signal interference distance threshold described here is different from the above-mentioned secure communication distance. Here, the signal interference threshold is greater than the secure communication distance, which means that when the distance between two nodes operating in the same or adjacent frequency bands is too close, excessive interference may occur, thus affecting the communication quality. In a scenario with complex terrain and changing mountain blockages, it is also necessary to comprehensively consider the attenuation and scattering of electromagnetic waves by different altitudes and mountain terrains. Through the accurate measurement of the geometric distance between nodes and the cross-sectional analysis of terrain obstacles such as mountains and valleys, the system can determine whether it is necessary to slightly move the position of some relay nodes or adjust the angle of a certain connection, so that the nodes are neither too close nor too far apart, thereby minimizing the mutual interference risk while ensuring communication coverage.

[0113] Specifically, in one embodiment, S4 includes the following steps S401 - S403.

[0114] S401. Take three consecutive signal relay nodes and calculate the geometric distance between the first signal relay node and the third signal relay node.

[0115] S402. Determine whether the geometric distance between the first signal relay node and the third signal relay node is greater than the signal interference distance threshold; if so, determine the positions of these three signal relay nodes, if not:

[0116] Based on the three-dimensional environment model, determine whether the first signal relay node and the third signal relay node are blocked by mountains; if so, determine the positions of these three signal relay nodes, if not, move the position of the signal relay node farther from the base station on this link to increase the included angle of the connection broken line of these three signal relay nodes at the middle signal relay node, and return to S401.

[0117] S403. Repeat the above step and perform hierarchical optimization starting from the near base station section based on chain transmission.

[0118] In S401 to S403, the system takes three consecutive signal relay nodes as a processing unit and focuses on evaluating whether the distance between the first and the third nodes will cause excessive interference. First, calculate the three-dimensional geometric distance between these two nodes. If this distance exceeds the pre-set interference distance threshold, it means they are relatively dispersed and not likely to cause severe co-channel or adjacent-channel interference. At this time, the distribution positions of these three nodes can be directly determined without further modification. However, if the measured distance is too close, the system will use the three-dimensional environmental model to judge whether there is sufficient mountain or terrain occlusion to weaken the possible interference between the two. For example, a mountain depression between two peaks can effectively block most electromagnetic waves. If such a natural barrier does exist, the original layout can also be retained. But if there is neither sufficient distance nor terrain to utilize, the system will fine-tune the position of the node farthest from the base station on this link, by increasing the included angle of the broken line in the middle of the connection of the three nodes to widen the relative positions of the first and the third nodes. After the adjustment is completed, the system will check again whether the new node distribution meets the requirements of the interference threshold. If there are still problems, this cycle will be repeated until an optimal solution that takes into account both distance and terrain factors is found.

[0119] For example, assume that a user wants to continuously deploy multiple UAV nodes in a terrain with alternating ridges and valleys. If on a certain ridge, the horizontal span between node A and node C is only one kilometer and they may be too close under the set transmission power, the system will first observe whether there is a mountain peak spanning between the two to reduce the intensity of the direct signal. If the height of this mountain peak is insufficient or its position is deviated and cannot form an effective block, the system will automatically move the farther node C slightly to a higher ridge or a more lateral airspace, thus increasing the included angle of the connection of points A - B - C. In this way, not only can the interference risk between A and C be alleviated, but it can also ensure that node B continues to have line-of-sight with A and C. This process is passed step by step from the base station to deeper mountainous areas, ensuring that each group of adjacent three nodes undergoes the same distance verification, terrain verification, and position fine-tuning, and finally enabling the entire UAV relay link to maintain stable and efficient communication in complex terrains. Through such step-by-step iteration, the system can maximize the use of natural occlusion while avoiding link instability caused by signal coupling between nodes.

[0120] However, through the above simplified scheme, in complex terrains, the optimization between consecutive nodes is likely to form local extrema. Especially in areas with dense relay points, the impact of angle adjustment on overall interference may be limited. Therefore, in another embodiment of this application, S4 discloses the following steps S41 - S43 to reduce the occurrence of this problem.

[0121] S41. Evaluate multiple connectable candidate signal relay links to screen out candidate links to be optimized.

[0122] This step identifies a number of initially reasonable candidate links from a large number of redundant candidate nodes to narrow the search space for subsequent operations such as broken line optimization and perturbation optimization.

[0123] Specifically, S41 includes the following steps S411 - S413.

[0124] S411. Based on the constructed node connection relationship graph, identify at least two visible link paths from the base station to the target user.

[0125] Based on the node connection relationship graph constructed in step S2, enumerate all visible paths from the base station to the target user. This connection relationship graph is represented as an undirected graph structure, where the nodes of the graph correspond to geographical location points that can satisfy the UAV hovering condition, and the edges represent that there is a direct line of sight between nodes and the communication distance does not exceed the maximum allowable communication distance threshold.

[0126] Preferably, path enumeration uses a depth - first search (DFS) or an A* search algorithm with heuristic rules, aiming to traverse all simple paths starting from the base station and ending at the target user. The node sequence in the path needs to ensure that there is an effective communication condition between continuously adjacent nodes and does not contain a loop structure to meet the requirements of chain - type relay deployment.

[0127] S412. For each link path, calculate the corresponding comprehensive score value according to multiple preset performance indicators; where the preset performance indicators include the number of relay nodes, path length, occlusion rate, and signal interference strength between nodes.

[0128] For each enumerated candidate link path, calculate its path structure and communication attributes respectively, and then generate the corresponding comprehensive score value. The scoring is based on the following multiple preset indicators:

[0129] Number of relay nodes: The fewer the number of nodes, the lower the complexity of link deployment and the smaller the cumulative communication loss;

[0130] Total path length: The geometric total length of the link, which is used to measure the energy required for signal propagation and path loss;

[0131] Occlusion rate: The proportion of the edges with mountain occlusion in the path, which is used to reflect the visibility of the link and the anti - environmental interference ability;

[0132] Interference intensity: The spatial distribution density and frequency band reuse situation between adjacent relay nodes, which reflect the severity of potential electromagnetic interference.

[0133] Preferably, the scoring function can be defined in the following weighted form:

[0134] Score i= w1·N i + w2·L i + w3·O i + w4·I i where N i represents the number of relay nodes of the i-th path, L i represents the path length, O i represents the occlusion rate, I i represents the interference intensity, and w1 to w4 are empirically set weight coefficients for adjusting the evaluation weights of different factors according to the scenario requirements.

[0135] By performing scoring calculations on each path, the performance ranking results of all candidate paths are obtained.

[0136] S413. Based on the scoring values, select at least one path with a relatively better score from all feasible paths as the candidate link to be optimized.

[0137] After sorting the scoring results, select at least one path with the highest score as the candidate link to be optimized for the subsequent geometric structure optimization in step S42. Preferably, to enhance the path stability and robustness of the system, the first K (e.g., Top-3) paths with the highest scoring values can be retained to form a candidate link set for performing path switching operations in case of subsequent node perturbation failures or path breaks.

[0138] For example, in a mountainous area, the communication system base station is deployed at point A at the foot of the mountain, and the target user is located at point B on the opposite hillside. After analysis by the three-dimensional terrain model and processing in step S2, the system identifies 12 relay nodes with good hovering conditions. After generating a node connection relationship diagram, about 28 visible paths between A and B are formed.

[0139] After executing S411, the system enumerates these 28 paths through the DFS algorithm and calculates the comprehensive score for each path in S412:

[0140] Path 1 (A → P1 → P2 → B): The number of relay nodes is 2, the total length is 540 meters, the occlusion rate is 5%, the interference intensity is low, and the score is 78;

[0141] Path 2 (A → P3 → P4 → P5 → B): The number of relay nodes is 3, the total length is 460 meters, the occlusion rate is 0%, but the interference value is high due to the short distance between the middle nodes, and the score is 73;

[0142] Path 3 (A → P6 → P7 → P8 → P9 → B): The number of relay nodes is 4, the total length is 800 meters, the occlusion rate is 15%, and the score is 52;

[0143] Finally, the system retains path 1 and path 2 as candidate links to be optimized based on the scores, and enters the subsequent S42 node position disturbance and structure adjustment process. Path 3 is excluded due to severe occlusion and too many nodes.

[0144] S42. Perform local structural optimization on the signal relay node positions in the candidate links to be optimized to improve the link geometry and communication quality.

[0145] Specifically, S42 includes the following steps S421-S424. :

[0146] S421. Select three consecutive transit nodes in the candidate link in order to form a broken line segment triplet.

[0147] In a link composed of multiple signal relay nodes, three adjacent nodes can form a geometric polyline segment, which is recorded as a triple (N i , N i+1 , N i+2 ). In this triple, N i To N i+1 and N i+1 To N i+2 They form two direction vectors, and their angle determines the degree of tortuosity of the broken line.

[0148] By calculating the angle of the broken line segment, it is possible to determine whether the link has excessive return, direction deviation and other geometric structures that are not conducive to communication transmission. Especially in mountainous terrain, a too small broken line angle can easily cause: line of sight obstruction between adjacent nodes, electromagnetic interference caused by spatial overlap, increased energy consumption of the drone to adjust the flight direction at the node, and increased signal path reflection and attenuation.

[0149] Therefore, extracting and analyzing the triplet structure of broken line segments in the link is of great significance for determining whether it is necessary to perform position perturbations of intermediate nodes.

[0150] During the implementation process, the system calculates the node sequence P = {N1, N2, ..., N n}Execute the following processing:

[0151] Starting from the beginning of the sequence, extract the triples (N1, N2, N3), (N2, N3, N4), ..., (N n-2 ,N n-1 ,N n );

[0152] For each triple, the angle value θ is calculated, and the intermediate node numbers of the broken lines are recorded, marking each group of three-node structures to determine whether they need to enter the subsequent angle judgment and disturbance candidate process (S422).

[0153] In this step, each intermediate node N i+1 It will be used as a candidate disturbance node, and whether its position needs to be adjusted is determined by the subsequent angle and interference judgment.

[0154] Take a selected link path to be optimized as an example:

[0155] The link from base station A to user B passes through 7 signal transfer nodes, and the node sequence is as follows:

[0156] P={A,P1,P2,P3,P4,P5,P6,B}

[0157] According to the triple extraction principle, the system generates the following six groups of polyline segment triplets: (A, P1, P2), (P1, P2, P3), (P2, P3, P4), (P3, P4, P5), (P4, P5, P6), and (P5, P6, B).

[0158] For example, in the fourth group of triples, the angle of the broken line segment formed by nodes P3, P4, and P5 is only 38°, which is less than the set angle threshold (such as 65°). At the same time, the relative distance between the front and rear segments is less than 50 meters, and there is a possibility of signal crosstalk. Therefore, the system will try to disturb the position of node P4 in S422 to improve the geometric rationality and communication performance of the broken line structure.

[0159] For other triples such as (P1, P2, P3), the angle is 92°, and the broken line is close to a straight line segment, which is considered to be of reasonable structure. The perturbation processing can be skipped and the next group of judgments can be entered.

[0160] S422. Calculate the angle value of the broken line segment and the geometric distance between the first and last nodes respectively, and determine whether the set angle threshold and interference distance threshold are met.

[0161] For any set of three consecutive signal relay node triplets (N i ,N i+1 ,N i+2 ), this step first conducts a geometric analysis of its spatial structure and makes a comprehensive judgment based on the communication characteristics, mainly including the following two dimensions:

[0162] Calculation of the angle between the broken lines:

[0163] This angle is used to measure the turning range of the link in this section. Too small an angle (such as less than 60°) often means that the path has obvious turns, which may cause the following problems: the directions between relay nodes are inconsistent, the flight energy consumption is increased; the signal overlap area is formed near the middle node of the broken line, causing communication interference or reflection.

[0164] The angle of the broken line is calculated using the three-dimensional vector cosine formula. Assume:

[0165] The first vector The second vector

[0166] Then the included angle θ between the first vector and the second vector is:

[0167]

[0168] Geometric distance between nodes and interference prediction:

[0169] By calculating the straight-line distance d from N i to N i+2 and comparing it with the signal interference distance threshold D t preset by the system, it is determined whether there is a potential interference area in this section. If the distance between the two endpoints is too close, signal crosstalk between the front and rear sections may occur in the frequency band reuse scenario.

[0170] Generally speaking, when the included angle θ is less than the threshold θ t (such as 65°), and the distance d is less than the interference threshold D t (such as 80m), then it is determined that this polyline segment is an unreasonable structure segment, and the intermediate node N i+1 needs to perform a perturbation attempt in S423.

[0171] S423. If the threshold requirements are not met, position perturbations are performed on the intermediate node, including preset direction perturbations, distance perturbations, and height perturbation operations, generating multiple perturbed points and calculating the corresponding parameters.

[0172] In this step, the intermediate node (denoted as N i+1 ) in the polyline triple is subjected to coordinate perturbation to generate j perturbed candidate points N′ i+1,j with position offsets to construct a new polyline (N i , N′ i+1,j , N i+2 ) and evaluate its geometric and communication attributes. Through position perturbation means, it is expected to increase the polyline included angle, reduce the path reverse folding phenomenon, and widen the spatial distance from adjacent nodes without changing the path connectivity structure, thereby reducing the interference overlap risk.

[0173] The perturbation methods include but are not limited to the following three forms:

[0174] 1. Direction perturbation: Perform left and right angular offsets (such as ±15°, ±30°) along the normal direction of the current polyline direction.

[0175] 2. Distance perturbation: Advance or pull away in the above direction with different amplitudes (such as ±10m, ±20m, ±40m).

[0176] 3. Height perturbation: It performs vertical up and down offset (e.g., ±10m) to improve the line-of-sight relationship and electromagnetic propagation conditions.

[0177] The above perturbation operations can be controlled by the angle pair (θ, φ) and the perturbation radius r in the spherical coordinate system, and finally converted into a three-dimensional displacement vector in the Cartesian coordinate system. Among them, θ is the polar angle in the spherical coordinate system, and φ is the azimuth angle in the spherical coordinate system.

[0178] For each node N to be perturbed i+1 , the system sequentially executes the following steps:

[0179] S4231: Construct a perturbation template library, which contains several direction perturbation combinations (such as θ = ±15°, φ = ±30°) and amplitude sets (such as r = 10m, 20m, 30m, 50m);

[0180] S4232: Based on the coordinate position of the current node N i+1 , apply each set of perturbation parameters to generate the perturbed position point N' i+1,j ;

[0181] S4233: For each perturbed point, form a new polyline segment N i →N' i+1,j →N i+2 , calculate its included angle value and the distance between nodes, and compare them with the thresholds set by the system;

[0182] S4234: Consider all the perturbed points that meet the included angle threshold (such as ≥65°) and the interference distance threshold (such as ≥80m) as valid candidate points;

[0183] S4235: Record the parameters of all valid perturbed points for subsequent position update determination in S424.

[0184] For example, taking a polyline segment of three nodes (P2, P3, P4) as an example, P3 is the middle node, its original coordinates are (x = 320, y = 540, z = 880), the included angle of its polyline is only 43°, and the distance from P2 to P4 is 65 meters, which does not meet the system requirements.

[0185] The system enables the perturbation template set, selects the direction perturbations of ±15° and ±30°, the radius perturbations of 20 meters and 40 meters, and the height perturbation of ±10 meters, and combines them to form multiple perturbation configurations.

[0186] For example, a set of perturbation parameters is a direction offset of 15°, a distance of 20 meters, and a height of +10 meters, and the corresponding new coordinates after perturbation are:

[0187] N' P3 =(x = 337.5, y = 557.1, z = 890)

[0188] Calculate the included angle of 71.2° for the new broken line (P2, N′ P3 , P4). The side length from P2 to P4 is extended to 92 meters, meeting the threshold conditions set by the system. Therefore, this perturbation point is marked as a valid candidate position point.

[0189] By traversing all perturbation combinations, the system finally generates 6 perturbation positions that meet the conditions for subsequent selection of the best in the sub - steps.

[0190] S424. If there are perturbation positions that meet the requirements of the broken - line included angle and interference limit, update the position of this node; if there are no effective perturbation points, expand the analysis range to a segment containing five consecutive nodes and continue with the optimization judgment.

[0191] Specifically, this step includes the following two stages:

[0192] (1) Node position update determination

[0193] S4241: Traverse all valid perturbation positions N′ generated by S423 i+1,1 , N′ i+1,2 , …, N′ i+1,k ;

[0194] S4242: For each perturbation point, recalculate the included angle θ′ of the new broken - line segment (N i , N′ i+1,j , N i+2 ) with the geometric distances d1 and d2 between the front and rear nodes, and calculate its communication scoring function. For example:

[0195] Fitness j =w1·θ j ′ + w2·(d 1,j + d 2,j ) - w3I j

[0196] Where, I j is the estimated interference value between the perturbation position and the front and rear nodes; w1, w2, w3 are weighting coefficients.

[0197] S4243: Select the perturbation point N′ with the optimal scoring function value from them i+1,opt , and use its position as the updated position of node N i+1 ;

[0198] S4244: If the included angles of all perturbation points have not improved, or the values of the communication scoring function are all lower than the original structure score, the system rejects the position update and keeps the original position unchanged.

[0199] (2) Broken - line segment window expansion mechanism

[0200] S4245: If the current triple structure has not improved after being perturbed in S423, the analysis window is automatically expanded to five consecutive nodes, i.e., a four-segment broken-line structure is formed;

[0201] S4246: Identify the "main turning point" in the five-tuple. For example, select the one with the smallest included angle or the largest interference as the optimized central node;

[0202] S4247: Centered on the main turning point, re-execute S422 and S423 to form a new perturbation and update process until the structure meets the threshold requirements or cannot be further optimized.

[0203] For example, take a candidate link segment (P1, P2, P3, P4, P5) as an example. Among them, the included angle of the triple (P2, P3, P4) is judged to be 41° by S422. After being perturbed in S423, only 2 valid points are obtained:

[0204] Point A: The included angle is increased to 48°, but the distances to the front and rear nodes are close, and the interference score is relatively high;

[0205] Point B: The included angle is increased to 63°, and at the same time the distance is 95m, the interference intensity is small, and the communication score is the highest;

[0206] Based on this, the system updates the position of P3 to the coordinates of the perturbed point B, replaces the original node position, and enters the analysis of the next broken-line segment.

[0207] If there are no valid candidate points for another group of nodes (P4, P5, P6) after the perturbation attempt, the system automatically expands the analysis window to (P3, P4, P5, P6, P7), identifies P5 as the strongest turning point, and enters the included angle evaluation and perturbed point generation process again.

[0208] S43. Introduce a node perturbation template and a path switching mechanism during the local optimization process to improve the optimizability and convergence effect of the overall link.

[0209] Specifically, S43 includes the following steps S431 - S433.

[0210] S431. When the node position perturbation attempt fails to obtain a satisfactory update result in multiple consecutive broken-line segments, trigger the preset perturbation template mechanism, and perform a replacement test on the node position according to the standard template stored in the perturbation library.

[0211] The purpose of this step is to handle the situation where the relay node position update fails during the optimization process of multiple consecutive broken line segments. Specifically, when the system detects that multiple consecutive broken line segments are perturbed unsuccessfully, indicating that the local path structure has fallen into an optimization convergence bottleneck, the system will automatically trigger the "perturbation library mechanism" and call a set of preset standard perturbation templates to attempt to reconstruct the structure of the nodes in this position or area, in order to improve the adjustability and overall optimizability of the path structure. Essentially, this step is a strategy fallback mechanism and a standard structure repair method to address the following two types of situations:

[0212] 1. In local path optimization, multiple consecutive sets of broken line segments (e.g., more than three) do not achieve angle optimization or interference mitigation through conventional perturbation strategies;

[0213] 2. The current path structure is restricted by terrain, node density, or turning-back structure, resulting in a severely limited position perturbation search space, and traditional perturbation searches can no longer effectively improve link performance.

[0214] Specifically, S431 includes the following sub-steps:

[0215] S4311. Record the number of node segments with consecutive perturbation failures in S423 and S424. If the angle optimization or node update cannot be completed in three or more broken line segments, then determine that the current path segment is a structure convergence failure segment;

[0216] S4312. The system calls the preset perturbation template library, where each template contains a specific geometric structure and the corresponding node position reconstruction logic, for example:

[0217] L-shaped turning-back adjustment template: used to optimize sharp turn structures within 90°;

[0218] Dense node thinning template: used for scenarios where interference is too strong due to dense relay nodes;

[0219] Height difference compensation template: used for areas with significant path elevation fluctuations;

[0220] S4313. The system identifies the geometric pattern to which the current structure segment belongs, and generates node position replacement suggestions after matching the template;

[0221] S4314. Fuse and judge the node position information generated by the template with the original structure. If it meets communication requirements such as the angle threshold and interference distance threshold, then perform the replacement; otherwise, try the next template or enter the path switching logic (S432).

[0222] It should be noted that this mechanism can run nested in multiple structure segments, or it can perform template batch processing according to path segment blocks.

[0223] For example, assume a relay link from base station A to user B, and the node sequence is as follows:

[0224] P = {A, P1, P2, P3, P4, P5, P6, B}

[0225] When the system performs perturbation optimization on the following broken line segments (P1, P2, P3), (P2, P3, P4), and (P3, P4, P5) in sequence, no perturbation points that meet the requirements of angle increase or interference mitigation are obtained, and it is determined that this path structure has fallen into local convergence.

[0226] The system automatically triggers the perturbation library mechanism, identifies the current structure segment as an "L-shaped reverse fold segment", and calls the "L-shaped return adjustment template". The strategy defined by this template is as follows:

[0227] Shift the entire node P3 30 meters in the normal direction of the broken line;

[0228] At the same time, extend node P4 20 meters away from P2 and raise its height by 10 meters.

[0229] After the template is applied, the three sets of broken line angles are recalculated to be 72°, 89°, and 94° respectively, and the signal interference intensity is reduced to 65% of the original. The system determines that the updated plan is effective and then executes this structure replacement.

[0230] S432. If the current candidate link cannot be further optimized at multiple key node positions, switch to another candidate link path and return to S42.

[0231] This step is used to trigger the candidate path switching mechanism when the path structure still cannot obtain a structure adjustment result that meets the communication performance requirements after multiple rounds of local perturbation optimization (S42) and template perturbation replacement (S431). Switch from the candidate path set to another link path and re-enter the structure optimization process of S42. This mechanism, as a "path-level backoff logic" in the link planning process, is used to improve the global optimal solution ability and the ability to jump out of the search space of the full-link deployment strategy.

[0232] The specific implementation process of this step is as follows:

[0233] S4321. The system monitors the node optimization status in the current path. If it detects that more than the preset number of node segments cannot complete position perturbation or template replacement (such as ≥3 segments), and the affected paragraph distribution is relatively concentrated (such as continuously appearing in more than 60% of the path length), then mark the current path as an "optimization failure path";

[0234] S4322. Select the next sub-optimal path with the second-highest score from the candidate path set reserved in S41 as the new "path to be optimized";

[0235] S4323. Load the new path as the current working path and restart the triple extraction, angle judgment, perturbation generation and update mechanism of S42;

[0236] S4324. If the current path is the last one in the candidate path set and the path optimization still fails, mark the path optimization status as "undeployable" and enter the alarm or task abort process.

[0237] Taking a mountain area scenario as an example, the system retains three candidate paths in S41:

[0238] Path A (score 87): 6 nodes, the path is relatively straight but passes through a cliff section;

[0239] Path B (score 84): 8 nodes, located on the mountainside, and the electromagnetic environment is complex in some areas;

[0240] Path C (score 81): 10 nodes, the path is relatively winding but the area is relatively open.

[0241] The system first optimizes Path A. After executing S42, four consecutive groups of nodes (P2, P3, P4), (P3, P4, P5), (P4, P5, P6), (P5, P6, P7) cannot perform effective position perturbations. After executing S431, the template replacement also fails, and the system identifies Path A as a path with non-optimizable structure.

[0242] According to the rules of S432, the system switches to Path B and restarts the S42 optimization process. In Path B, although the interference index is slightly higher, all node segments can obtain an angle improvement through perturbation, and the path is finally successfully deployed.

[0243] S433. Compare all optimized candidate links based on a unified link performance evaluation criterion, and select the link with the highest evaluation score as the final signal relay communication path.

[0244] This step is used to uniformly evaluate the performance of all candidate links that have completed node position perturbation and structure optimization processing, and select the path with the best comprehensive performance as the final signal relay communication link based on the preset evaluation criterion for the execution phase of UAV deployment and link establishment.

[0245] This step includes the following specific processing procedures:

[0246] S433-1: Collect the structure parameters and performance data after optimizing all paths to form a candidate path set

[0247] S4332: For each path P i Calculate the following five evaluation indicators:

[0248] Number of nodes N i : The number of path relay nodes, the fewer the better;

[0249] Total path length L i : The total length of the relay chain from the base station to the user, which affects energy consumption and transmission delay;

[0250] Minimum included angle value θ min,i : The minimum included angle of the broken line in the path, which reflects whether there is an obvious turn-back in the geometric structure;

[0251] Average channel quality index Q i : Weighted estimation based on the signal-to-noise ratio (SNR) and bit error rate (BER) between nodes;

[0252] Energy consumption ratio per unit throughput E i : The energy consumption corresponding to the unit throughput of the path, which reflects the link efficiency;

[0253] S4333: Construct a comprehensive scoring function:

[0254] Score i = w1*f1(N i ) + w2*f2(L i ) + w3·f3(θ min,i ) + w4·f4(Q i ) + w5·f5(E i )

[0255] Among them, f k (·) is an index normalization function, w k is a weight coefficient (set according to the scenario), and score sorting is performed on all paths P i ;

[0256] S4334: Select the path P with the highest comprehensive score value opt as the final deployment path and output it to the control module for subsequent execution of step S5.

[0257] S5. Dispatch the UAV to the signal relay node to establish a communication connection between the base station, the signal relay chain and the user.

[0258] Specifically, in one embodiment, S5 includes the following steps:

[0259] For the UAV that uses two frequency bands for information transmission and reception in one of the two adjacent working time slots and does not perform information transmission and reception in the other working time slot, decode and re-encode the received signal during the working time slot and the guard interval when not performing information transmission and reception, and send it out during the working time slot when performing information transmission and reception;

[0260] For drones that send and receive information in two adjacent working time slots, the received information is forwarded with gain.

[0261] The system will dispatch and deploy the signal transfer nodes that have been planned in the early positions and frequency bands to the drones, thereby building a complete communication chain in the target mountain environment and connecting the base station to the user. Different drones will perform data transmission and reception according to the established time slot and frequency band allocation rules. Some drones will use two frequency bands to transmit and receive information in one time slot, while being temporarily idle in adjacent time slots. Such drones can use idle working time slots and time slot protection intervals to decode and re-encode received data packets, and then send the processed information together in the next time slot with a transmission task. In this way, not only can the drone achieve transmission and reception separation in the time domain, but it also provides a more ample processing window for possible high computing power requirements.

[0262] On the other hand, for those drones that need to send and receive information in two adjacent time slots, the system will set them to perform gain forwarding on the received signals. Gain forwarding refers to directly applying amplification, equalization and other processing to the input signal at the hardware level, so that the signal quality can be improved to a certain extent without taking up a lot of decoding time, and then immediately forwarded to the downstream node in the next time slot. For example, if a drone receives a data packet from the previous node via 2.4GHz in time slot one, it must use 5GHz to quickly send the data to the next node in time slot two. At this time, the forwarding can be completed after gain amplification and simple error correction processing, without the need for full-process decoding and re-encoding like in idle time slots. Therefore, this division of labor not only ensures that high-load data streams can be forwarded quickly and continuously in multi-hop links, but also makes full use of the computing resources of drones that do not need to perform sending and receiving tasks in all time slots, so that the entire system can achieve a more reasonable balance in terms of bandwidth usage and energy consumption.

[0263] S6. Adjust the positions of the signal relay node and the drone based on the real-time flight status of the drone.

[0264] Specifically, in a certain embodiment, S6 includes the following steps S61-S66.

[0265] S61. Discretize the three-dimensional space of the user's environment into a three-dimensional grid and perform position encoding.

[0266] In this step, the system discretizes the three-dimensional space of the user's environment and assigns corresponding position codes to each grid cell to more precisely evaluate flight and communication conditions in the entire mountainous area or variable terrain. The discretization of the three-dimensional space means dividing the originally continuous mountains, terrain, and airspace into several grid cells with fixed volumes or coordinate step sizes, just like building a grid paper under a three-dimensional coordinate system in a real mountain environment. Each grid cell can have its own unique number to identify the coordinate interval it is in, as well as corresponding information such as altitude, slope, or obstacle distribution.

[0267] In the implementation process, the system usually needs to first determine a global coordinate range. For example, it divides the horizontal and vertical directions from the foot of the mountain to the top of the mountain into several levels, and then sets an appropriate grid size according to device performance or accuracy requirements. For example, if the horizontal span of a certain mountain range area is about dozens of kilometers, the horizontal direction can be divided into grids of every 100 meters or dozens of meters; if higher accuracy requirements for height changes are needed, the vertical direction can also be divided more finely. After the division is completed, the system assigns a unique coding identifier to each grid cell, usually consisting of a set of coordinate indexes or a coding matrix.

[0268] S62. Quantify the channel quality based on the received signal strength and the bit error rate.

[0269] In this step, the system quantifies the wireless channel quality in the current environment based on the received signal strength and the bit error rate. The received signal strength is usually characterized by RSSI (Received Signal Strength Indicator) or more refined channel quality indicators (such as SNR, SINR, etc.), aiming to illustrate whether the current wireless link has sufficient power margin; while the bit error rate (BER) can reflect the probability of bit errors occurring at the receiving end when transmitting data on this link. Combining the two can give a more comprehensive channel quality value, which can not only measure whether the signal becomes weak due to attenuation, scattering, or multipath effects in the propagation path, but also test the actual performance of the coding method and modulation depth for anti-interference. Since in the mountain environment, signal propagation is often affected by complex terrain and meteorological conditions, the system needs to score each discretized grid cell based on real-time or historical measured RSSI and BER data.

[0270] In specific implementation, during flight tests, the drone may perform "exploratory transmission and reception" on each grid point, recording the corresponding signal strength and bit error rate. For example, if severe multipath interference is caused by a certain valley terrain, it may lead to a significant increase in BER; or in an open area on a ridge, although the signal strength may remain sufficient, due to the influence of instantaneous wind disturbances, multiple samples are required to obtain stable data. By collecting sufficient sample information and performing statistical or interpolation operations, the system can store the channel quality parameters of each grid cell in the database, providing an accurate reference for subsequent path planning or relay node deployment.

[0271] As an example, in one embodiment, it can be divided into levels every 5 dB (-90 dBm to -30 dBm), and the bit error rate is divided into a low bit error rate level and a high bit error rate level with a threshold of 10^(-4).

[0272] S63. Standardize and number the flight actions based on the flight adjustment orientation and flight adjustment distance.

[0273] In this step, the system standardizes and numbers various flight actions that may occur during the execution of the drone, so that different adjustment behaviors can be recognized and processed under a unified coordinate system and strategy framework. Flight adjustment orientation refers to the change in the flight path or attitude of the drone, such as turning at a certain angle, pitching, or yawing; while flight adjustment distance involves dividing the amplitude of advancing or retreating, ascending or descending on the flight path. By combining these two key dimensions, the system can define clear discrete labels for all common actions.

[0274] As an example, through some examples of the following action space design, the flight actions are standardized and numbered.

[0275] Action Number Adjustment Direction Adjustment Range 0 Horizontal Eastward Movement +5m (Positive X-axis Direction) 1 Horizontal Westward Movement -5m (Negative X-axis Direction) 2 Vertical Upward Movement +2m (Positive Z-axis Direction) 3 Maintain in Place 0m

[0276] S64. Set a multi-objective reward function based on the adjusted link throughput, the change amount of the distance to adjacent drones after adjustment, and the mobile energy consumption.

[0277] S65. Control the drone to perform different flight actions and reset, and collect data to substitute into the multi-objective reward function for calculation to obtain a local optimal solution.

[0278] In this step, the system sets a multi-objective reward function based on elements such as the adjusted link throughput, the change in distance between the UAV and adjacent nodes, and flight energy consumption. The purpose of this is to enable the system, when optimizing paths or strategies, to not only focus on the single metric of communication performance but also take into account the flight cost of the aircraft body and the safe distance between nodes. Multi-objective reward functions are common in reinforcement learning or other intelligent decision-making methods and are used to find a balance among multiple different and even potentially conflicting optimization goals. For example, when establishing a multi-hop network for users in mountainous areas, an increase in link throughput allows for faster video and data transmission. However, if the UAV frequently flies in high-wind areas in order to pursue the highest bandwidth, the energy consumption and safety risks will increase accordingly. Similarly, if only power saving is considered, it may result in a decline in communication quality.

[0279] In the actual implementation process, the system first defines a numerical expression method to map each metric to a positive or negative reward that can be accumulated. For example, an action that can stably provide high throughput will receive a positive score, while being too close to adjacent UAVs, which may pose interference or collision risks, will introduce a negative penalty, and excessive energy consumption will also result in corresponding deductions in the reward function. By continuously updating this reward function during training or online operation, the system can gradually learn or search for an optimal decision path in the comprehensive dimension. Specifically, for example, if the UAV finds that the link quality has improved significantly after making a fine-tuning action and the additional energy consumption remains within a reasonable range, then the comprehensive reward for this action is relatively high, and the system will be more inclined to repeat or generalize this action; conversely, if the gain is limited but a large amount of energy is consumed, it will also be deducted points and thus be eliminated in subsequent iterations.

[0280] As an example, in a certain embodiment, assume that a UAV b was originally 105 meters away from the upstream UAV a, and the link throughput was 20 Mbps. Due to terrain changes or interference, the communication quality fluctuated slightly. The system decided to move b 5 meters eastward in an attempt to increase the throughput and maintain a reasonable safety distance without significantly increasing the energy consumption. After the move, the actual distance between b and a became 110 meters, and the measured link throughput increased from 20 Mbps to 25 Mbps. If the maximum reference throughput C max set by the system in advance is in Mbps, the reference safe distance d safe is 100 meters, the energy consumption per unit distance of movement E move is 0.5 Wh, and this flight can be regarded as consuming a total of 0.5 Wh from the starting point to the end point, and the energy consumption upper limit E max allowed by the UAV battery and power system is 2 Wh, then the following multi-objective reward function can be used to evaluate this adjustment:

[0281]

[0282] C new Let \(C\) be the adjusted link throughput and \(\Delta d\) be the change in the distance to the adjacent UAV.

[0283] Then, \(R = 0.335\) is calculated.

[0284] From this, it can be obtained that the comprehensive benefit brought by this fine-tuning is \(0.335\), indicating that moderate movement improves the throughput without having too much negative impact on the safety distance or power consumption. In the actual environment, the system usually continuously performs similar action evaluations in multiple rounds of iteration, selects the action sequence with the highest reward value, and gradually optimizes the link throughput to a higher level while maintaining a reasonable node interval and energy consumption control.

[0285] S66. Repeat the previous step and perform hierarchical optimization starting from the near-base station segment based on chain transfer.

[0286] Start hierarchical optimization from the link closest to the base station to avoid link oscillation caused by moving multiple nodes simultaneously. For example, if the relative position from \(a\) to \(b\) has been determined to be relatively optimal, the coordinate of \(b\) will be fixed in the next step and the relationship between \(b\) and \(c\) will be considered, and the optimization will be passed to more distant nodes in this way. If conflicts such as target grid overlap occur during multi-aircraft cooperation, a set of priority rules can be set, such as allowing upstream nodes to select positions first, and then downstream nodes to make supplements or concessions.

[0287] At the same time, in terms of dynamic stability control, if the reward value gain brought by consecutive adjustments is less than \(5\%\), or the cumulative movement distance of the nodes exceeds 30 meters, resulting in excessive deviation from the initial plan, the system will consider it to have reached the convergence state and stop this round of iteration. If a certain adjustment causes a regression of more than \(10\%\) in the overall performance, the node positions can also be restored to the previous best solution through a rollback mechanism, and the relevant adjustment actions can be temporarily frozen for a period of time to prevent irreversible or repeated fluctuations caused by frequent trial and error.

[0288] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A boundary range expansion method based on mobile communication, characterized in that: The following steps are involved: S1. Obtain user location information, retrieve a three-dimensional model of the user's environment, and determine the user's position in the three-dimensional model of the environment; S2. Based on the three-dimensional model of the environment and the user's location information, signal transfer nodes are planned over the mountainous area to form a signal transfer chain, where the signal transfer nodes are used for the drone to hover; S3. Based on the relationship between adjacent nodes in the signal transfer chain, arrange the working time slots, communication bands and communication protocols used by the drones corresponding to different signal transfer nodes, wherein the communication band between each signal transfer node and the previous node is different from the communication band between each signal transfer node and the next node; S4. adjusting the position of the signal relay node based on the position of the signal relay node, the interference threshold distance of the UAV frequency band, and the three-dimensional model of the environment; S5. Send the drone to the signal transfer node to establish a communication connection between the base station, the signal transfer chain and the user; S6. Adjust the positions of the signal relay node and the drone based on the real-time flight status of the drone.

2. The method for extending the boundary range based on mobile communication according to claim 1, characterized in that: The S2 comprises the following steps: S21. Identify all candidate points that can satisfy the hovering of the UAV based on the terrain; S22. For each candidate node, check whether there is direct line of sight; if there is a mountain blocking, mark it as not directly connectable; S23. For two points that can be seen, calculate their three-dimensional distance and determine whether it exceeds the upper limit of the safe communication distance. If so, mark them as not directly connectable; S24. For the candidate nodes that can be directly connected, a candidate link from the base station to the user is formed, and a node connection graph is constructed; S25. Taking the upper limit of the communication distance of the frequency bands alternating along the node sequence as a constraint condition, the minimum node set is found through the minimum node coverage strategy.

3. The method for extending the boundary range based on mobile communication according to claim 2, characterized in that: The S3 comprises the following steps: S31. Determine frequency band one and frequency band two; wherein frequency band one and frequency band two correspond to the dual-mode communication module of the drone equipment; S32. Determine the working time slot 1 and working time slot 2 of the drone; S33. Determine the communication band and communication protocol used by the drone corresponding to the signal transfer node, so that among any five consecutive adjacent drones, the first drone and the second drone communicate through frequency band one in working time slot one, the second drone and the third drone communicate through frequency band two in time slot one, the third drone and the fourth drone communicate through frequency band one in time slot two, and the fourth drone and the fifth drone communicate through frequency band two in time slot two; S34. Insert a guard interval of 0.1-0.5 ms between time slot 1 and time slot 2.

4. The method for extending the boundary range based on mobile communication according to claim 3, characterized in that: The S4 comprises the following steps: S401. Take three consecutive signal transfer nodes and calculate the geometric distance between the first signal transfer node and the third signal transfer node; S402. Determine whether the geometric distance between the first signal transfer node and the third signal transfer node is greater than the signal interference distance threshold; if so, determine the positions of the three signal transfer nodes, if not: Based on the three-dimensional model of the environment, determine whether the first signal transfer node and the third signal transfer node are blocked by the mountain; if so, determine the positions of the three signal transfer nodes; if not, move the positions of the signal transfer nodes on the link away from the base station to increase the angle of the connection line of the three signal transfer nodes at the middle signal transfer node, and return to S401; S403. Repeat the previous step, and optimize step by step starting from the segment close to the base station based on chain transfer.

5. The method for extending the boundary range based on mobile communication according to claim 3, characterized in that: The S4 comprises the following steps: S41. Evaluate multiple connectable candidate signal relay links to screen candidate links to be optimized; S42. Perform local structural optimization on the signal transfer node positions in the candidate links to be optimized to improve the link geometry and communication quality; S43. Introduce node perturbation templates and path switching mechanisms in the local optimization process to improve the optimizability and convergence effect of the overall link.

6. The method for extending the boundary range based on mobile communication according to claim 5, characterized in that: The S41 comprises the following steps: S411. Based on the constructed node connection relationship diagram, identify at least two visible link paths from the base station to the target user; S412. For each link path, a comprehensive score corresponding to the path is calculated based on a plurality of preset performance indicators; wherein the preset performance indicators include the number of transit nodes, path length, occlusion rate, and signal interference intensity between nodes; S413. Based on the score value, select at least one path with a better score from all feasible paths as a candidate link to be optimized.

7. The method for extending the boundary range based on mobile communication according to claim 6, characterized in that: The S42 comprises the following steps: S421. Select three consecutive transit nodes in the candidate link in order to form a broken line segment triple; S422. Calculate the angle value of the polyline segment and the geometric distance between the first and last nodes respectively, and determine whether the set angle threshold and interference distance threshold are met; S423. If the threshold requirement is not met, the position disturbance is performed on the intermediate node, including the preset direction disturbance, distance disturbance and height disturbance operations, and multiple disturbance points are generated and the corresponding parameters are calculated; S424. If there is a disturbance point that meets the requirements of the broken line angle and the interference limit, the node position is updated; if there is no valid disturbance point, the analysis range is expanded to include a chain segment of five consecutive nodes to continue the optimization judgment.

8. The method for extending the boundary range based on mobile communication according to claim 5, characterized in that: The S43 comprises the following steps: S431. When the node position disturbance attempt fails to obtain an update result that satisfies the conditions in multiple consecutive groups of polyline segments, the preset disturbance template mechanism is triggered, and the node position is replaced according to the standard template stored in the disturbance library. S432. If the current candidate link cannot be further optimized at multiple key node positions, switch to another candidate link path and return to S42; S433. All optimized candidate links are compared based on a unified link performance evaluation standard, and the link with the highest evaluation score is selected as the final signal relay communication path.

9. The method for extending the boundary range based on mobile communication according to claim 8, characterized in that: The S5 comprises the following steps: For a drone that uses two frequency bands to transmit and receive information in one of two adjacent working time slots and does not transmit and receive information in the other working time slot, the received signal is decoded and re-encoded in the working time slot where no information is transmitted or received and the protection interval of the time slot, and sent in the working time slot where information is transmitted or received; For drones that send and receive information in two adjacent working time slots, the received information is forwarded with gain.

10. The method for extending the boundary range based on mobile communication according to claim 9, characterized in that: The S6 comprises the following steps: S61. Discretize the three-dimensional space of the user's environment into a three-dimensional grid and perform position encoding; S62. Quantify the channel quality based on received signal strength and bit error rate; S63. Based on the flight adjustment direction and flight adjustment distance, the flight actions are standardized and classified and numbered; S64. Based on the adjusted link throughput, the adjusted distance change to the adjacent UAVs and the movement energy consumption, a multi-objective reward function is set; S65. Control the drone to perform different flight actions and reset, and collect data to substitute into the multi-objective reward function for calculation to obtain a local optimal solution; S66. Repeat the previous step, and optimize step by step starting from the segment close to the base station based on chain transfer.

Images