Intelligent scheduling system of three-in-one unmanned aerial vehicle base station of communication relay, overcharge and hydrogenation

Abstract

The invention relates to the technical field of unmanned aerial vehicle base station scheduling, in particular to an intelligent scheduling system of a three-in-one unmanned aerial vehicle base station of communication relay, overcharge and hydrogenation. The node distance between the starting point and the end point of the unmanned aerial vehicle and the node distance between the base station are measured before a task is executed, signal attenuation records of a relay antenna array, energy consumption measurement of a battery heat management unit and the energy complementing rate of a fixed hydrogenation interface are combined, head and tail node closed loops are recognized, signal attenuation and energy consumption are accumulated, and total cost comparison is carried out; task segments are recombined on the basis of a topological sequence, a continuous execution chain without a closed loop is formed, multi-aircraft task conflicts and repeated flight are avoided, energy differences, signal differences and progress differences are comprehensively classified, sorted, combined and arranged through a decision tree, energy, signal and traffic states are compared with corresponding threshold values in the execution stage, and the execution efficiency is improved. And the abnormal task segment is quickly allocated to the nearest unmanned aerial vehicle in a normal state, and the execution sequence is dynamically adjusted.

Description

Technical Field

[0001] This invention relates to the field of drone base station scheduling technology, and in particular to an intelligent scheduling system for a drone base station that integrates "communication relay + supercharging + hydrogen refueling". Background Technology

[0002] The field of UAV base station scheduling technology aims to achieve continuous support and efficient execution of UAV missions by centralizing and intelligently managing energy supply, communication relay, take-off and landing control, station allocation, and multi-UAV collaborative operations between UAVs and ground base stations. Based on mission requirements and the real-time status of UAVs, it can rationally arrange return and refueling times, dynamically optimize mission allocation and path planning, improve operational coverage, and ensure the safety, stability, and resource utilization of UAV operations.

[0003] The purpose of an intelligent scheduling system for a three-in-one drone base station integrating communication relay, supercharging, and hydrogen refueling is to achieve integrated energy supply and communication support for drones in complex and remote operating environments, improve the continuity and coverage of drone operations, enhance mission execution efficiency and system resource utilization, including significantly extending the flight time of a single drone mission, reducing mission interruption rate, optimizing energy use structure, and achieving efficient, stable, and safe drone scheduling in multi-mission operations.

[0004] Existing UAV base station scheduling relies heavily on preset paths and static allocation modes. Route assessments are often based on single distances or flight durations, lacking comprehensive quantification of multi-dimensional parameters such as signal attenuation, recharging rate, and equipment energy consumption. This leads to discrepancies between actual path costs and planned costs in complex environments. High-cost closed loops formed by node connections are not identified and broken in advance, easily causing UAVs to fly repeatedly or conflict with other tasks. Although some status parameters can be collected during operation, segmented detection often replaces global comparison. Key factors such as energy consumption, communication attenuation, and remaining task time are not analyzed in conjunction, resulting in delayed risk identification, increased task interruption rates, and priority ranking based on single indicators, such as arranging tasks according to remaining time, without weighing the combined effects of multiple conditions such as energy, signal, and progress differences. This leads to uneven resource allocation. In the real-time scheduling stage, there are delays in handling abnormal flight segments, and the lack of spatial location sensitivity considerations when matching drones may result in excessively long maneuver paths for replacement UAVs, causing replacement delays and affecting the efficiency of multi-drone collaboration and the overall stability of the system. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an intelligent scheduling system for a three-in-one drone base station that integrates "communication relay + supercharging + hydrogen refueling".

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent scheduling system for a three-in-one unmanned aerial vehicle (UAV) base station integrating "communication relay + supercharging + hydrogen refueling" includes: Route energy consumption construction module: Based on the coordinates of the UAV mission start point, end point, and base station, measure the distance between nodes, record the attenuation of relay antenna array, energy consumption of battery thermal management unit, and energy replenishment rate of fixed hydrogen refueling interface, calculate the comprehensive cost, and generate a comprehensive cost route dataset. Closed-loop cost resolution module: Based on the comprehensive cost route dataset, check the node order to mark closed loops, sum the segment attenuation and energy consumption to obtain the total cost, compare it with the threshold to retain low-value paths, use topological sorting to remove and reorganize task segments, and generate a closed-loop-free route combination table. Mission interruption monitoring module: Based on the closed-loop route combination table, compare energy reserves with high-frequency DC charging pile demand, communication strength with signal switch stability value, execution time with automatic guidance track travel time, mark abnormal segments and generate a list of interruption risk segments; Layered guarantee scheduling module: Based on the list of interruption risk segments, the module uses a decision tree algorithm to sort, merge, and arrange the segments according to energy difference, signal difference, and schedule difference, determine the execution order and priority, and generate a layered sequential task table. Dynamic task adjustment module: Based on the hierarchical sequential task table, it compares energy changes, signal fluctuations, traffic changes with thresholds, assigns abnormal flight segments to the nearest normal UAV, updates the order, and generates an instant flight segment execution instruction set.

[0007] As a further aspect of the present invention, the comprehensive cost route dataset includes inter-node distance values, relay antenna array signal attenuation values, battery thermal management unit energy consumption values, and fixed hydrogen refueling interface refueling rates. The loopless route combination table includes low-cost segment node sequences, removed segment task segment numbers, and reorganized new segment node connection relationships. The interruption risk segment list includes abnormal segment numbers, risk classification labels, corresponding UAV numbers, energy difference values, communication signal difference values, and travel time difference values. The hierarchical sequential task table includes segment numbers, risk levels, energy difference sorting results, signal difference sorting results, progress difference sorting results, and task execution order. The real-time segment execution instruction set includes segment numbers, executing UAV numbers, task start nodes, task end nodes, and updated task execution order.

[0008] As a further aspect of the present invention, the route energy consumption construction module includes: Node distance calculation submodule: Based on the coordinates of the UAV mission start point, end point, and base station, measure the straight-line distance between the three points and record the value; obtain and record the signal attenuation value of the relay antenna array through a rangefinder; measure and record the energy consumption of the battery thermal management unit when it is running at full load using a power metering device; measure and record the time taken to replenish a unit amount of hydrogen at the fixed hydrogen refueling interface using a stopwatch; and generate a segment energy consumption attenuation record table. The comprehensive cost generation submodule: Based on the energy consumption attenuation record table of the flight segment, the distance value is multiplied by the energy consumption value to obtain the energy consumption amount. The result is weighted and accumulated with the signal attenuation value and the energy replenishment time value according to a set ratio. The comprehensive cost of each flight segment is summarized and archived to generate a comprehensive cost route dataset.

[0009] As a further aspect of the present invention, the closed-loop cost reduction module includes: Closed-loop calibration submodule: Based on the comprehensive cost route dataset, compare the node connection order one by one and record the path groups with consistent first and last nodes, accumulate the signal attenuation value and energy consumption value of each segment and record the sum, establish a correspondence table between the obtained total cost value and the path group number, and generate a total path cost list. Path filtering submodule: Based on the total path cost list, the total cost is compared with a preset threshold and paths exceeding the threshold are marked. The signal attenuation value and energy consumption value of each segment in the high-value path are comprehensively sorted and smaller items are selected to establish a list of retainable segment numbers and generate a set of low-cost segments. Task segment reorganization submodule: Based on the low-cost segment set, topological sorting is used to extract the task number and node location information of the non-reserved segments and divide them into units. These nodes are continuously combined with connectable nodes in the set and the connection order is updated to form a new segment node list and generate a loop-free route combination table.

[0010] As a further aspect of the present invention, the topology sorting involves establishing a directed connection table of all segment nodes based on the low-cost segment set, calculating the in-degree value of each node, placing nodes with an in-degree of zero into a sequential queue, sequentially removing nodes from the queue and recording them in an ordered list, decrementing the in-degree value of all outgoing edges of a node connected to the target node by one, and adding a target node to the sequential queue if its in-degree is reduced to zero. This process is repeated until the queue is empty or all nodes are recorded. The order of nodes in the ordered list is the task execution order that satisfies the directed connection relationship.

[0011] As a further aspect of the present invention, the task interruption monitoring module includes: Segment status comparison submodule: Based on the loopless route combination table, read the corresponding UAV energy reserve value and subtract it from the high-frequency DC charging pile demand value, read the communication signal strength and subtract it from the signal switch stable signal value, read the remaining execution time of the segment and subtract it from the available time of the automatic guidance track, record each comparison difference and abnormal mark, and generate an abnormal segment judgment table. Risk list generation submodule: Based on the abnormal flight segment judgment table, it filters flight segments with insufficient energy value, signal value below stable value, and execution time less than travel time, extracts flight segment number, risk category and corresponding UAV number, compiles risk mission segment data, and generates a list of interrupted risk flight segments.

[0012] As a further aspect of the present invention, the high-frequency DC charging pile demand value specifically refers to the amount of energy required for a certain UAV to complete the entire process within a specific mission segment. The difference is calculated by measuring the current energy reserve of the UAV and the estimated energy consumption corresponding to the mission segment, and the difference is converted according to the rated output power and charging amount per unit time of the high-frequency DC charging pile to obtain the total charging amount required to meet the execution of the mission segment. This amount is then directly compared with the energy reserve to determine the size of the energy replenishment demand.

[0013] As a further aspect of the present invention, the hierarchical guarantee scheduling module includes: Risk classification submodule: Based on the list of interruption risk segments, the decision tree algorithm is used to subtract the energy reserve value and demand value of each segment and determine it as energy priority; subtract the communication signal strength and stability value and determine it as communication priority; subtract the remaining execution time and travel time and determine it as schedule priority; establish an association table between the classification results and the segment number to generate a multi-category risk list. Differential Ranking Submodule: Based on the multi-category risk list, sort the energy priority segments in ascending order by energy difference, sort the communication priority segments in descending order by signal difference, and sort the schedule priority segments in ascending order by time difference. Merge the sorting results of each category in turn and retain the category labels to generate a risk ranking task set. Sequential generation submodule: Based on the risk sorting task set, integrate risky flight segments and risk-free flight segments, synchronously record priority tags, generate task execution order according to sorting position, establish numbering and sequence tables, and generate hierarchical sequential task tables.

[0014] As a further aspect of the present invention, the decision tree algorithm, based on the energy difference, communication signal difference, and time difference of each segment in the list of interruption risk segments, selects the difference parameter with the largest information gain as the root node attribute, splits all segments into several child node sets according to the threshold conditions of the attribute, calculates the information gain of the remaining difference parameters for each child node set in turn, and selects the largest one as the partitioning attribute of the current node, repeating the partitioning process until the segments in the child node sets belong to the same risk category or the difference parameters are exhausted, records the segment number and risk category label at each leaf node, forming a condition judgment path from the root node to the leaf node, so that the input segment data passes through the threshold conditions and sequentially descends to the corresponding leaf node and outputs the classification result of the segment.

[0015] As a further aspect of the present invention, the dynamic task adjustment module includes: Real-time flight segment detection submodule: Based on the hierarchical sequential task table, read the change value of UAV energy reserve and subtract it from the energy threshold, read the signal fluctuation value of signal switcher and subtract it from the signal stability threshold, read the change value of automatic guidance track travel time and subtract it from the travel time threshold, mark flight segments that exceed the threshold, and generate a cross-boundary flight segment screening table. The instruction generation submodule: Based on the out-of-bounds flight segment screening table, selects the UAV that is closest to the abnormal flight segment and whose energy value, signal value, and travel time are all within the normal range from the task table, assigns the flight segment to the UAV and adjusts the task order, records the updated flight segment number and order list, and generates an instant flight segment execution instruction set.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by measuring the node distances between the UAV's starting point, ending point, and base station before mission execution, and combining this with the signal attenuation records of the relay antenna array, the energy consumption measurement of the battery thermal management unit, and the refueling rate of the fixed hydrogen refueling interface, comprehensive route cost data is formed, thereby achieving unified quantification of path consumption and communication link status. In this invention, by identifying closed loops at the beginning and end nodes and accumulating the signal attenuation and energy consumption within them for total cost comparison, high-cost loops are eliminated and low-cost paths are retained. Then, based on the topological order, task segments are reorganized to form a continuous execution chain without closed loops, thus avoiding multi-aircraft task conflicts and repeated flights. In this invention, a decision tree is used to comprehensively classify, sort, and merge energy differences, signal differences, and schedule differences to obtain a priority task list. This enables the overall allocation of tasks under multiple constraints. During the execution phase, energy, signal, and traffic status are compared with corresponding thresholds to quickly assign abnormal task segments to the nearest and normal UAVs, dynamically adjusting the execution order. Attached Figure Description

[0017] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] Example 1 Please see Figure 1 This invention provides a technical solution: an intelligent scheduling system for a three-in-one drone base station integrating "communication relay + supercharging + hydrogen refueling" includes: Route energy consumption construction module: Based on the coordinates of the UAV mission start point, end point, and base station, measure the distance between nodes, record the attenuation of relay antenna array, energy consumption of battery thermal management unit, and energy replenishment rate of fixed hydrogen refueling interface, calculate the comprehensive cost, and generate a comprehensive cost route dataset. Closed-loop cost resolution module: Based on the comprehensive cost route dataset, it checks the node order to mark closed loops, sums up the segment attenuation and energy consumption to obtain the total cost, compares it with the threshold to retain low-value paths, uses topological sorting to remove and reorganize task segments, and generates a closed-loop route combination table. Mission interruption monitoring module: Based on the loopless route combination table, it compares energy reserves with the demand for high-frequency DC charging piles, communication strength with the stability value of signal switchers, execution time with the passage time of automatic guidance track, and marks abnormal segments to generate a list of interruption risk segments; Layered guarantee scheduling module: Based on the list of interruption risk segments, the module uses a decision tree algorithm to sort, merge, and arrange the segments according to energy difference, signal difference, and schedule difference, determine the execution order and priority, and generate a layered sequential task table. Dynamic task adjustment module: Based on a hierarchical sequential task table, it compares energy changes, signal fluctuations, traffic changes with thresholds, assigns abnormal flight segments to the nearest normal UAV, updates the order, and generates an instant flight segment execution instruction set.

[0020] The comprehensive cost route dataset includes inter-node distance values, relay antenna array signal attenuation values, battery thermal management unit energy consumption values, and fixed hydrogen refueling interface refueling rates. The loopless route combination table includes low-cost segment node sequences, removed segment task segment numbers, and reorganized new segment node connection relationships. The interruption risk segment list includes abnormal segment numbers, risk classification labels, corresponding UAV numbers, energy difference values, communication signal difference values, and travel time difference values. The hierarchical sequential task table includes segment numbers, risk levels, energy difference sorting results, signal difference sorting results, progress difference sorting results, and task execution order. The real-time segment execution instruction set includes segment numbers, executing UAV numbers, task start node, task end node, and updated task execution order.

[0021] The route energy consumption construction module includes: Node distance calculation submodule: Based on the coordinates of the UAV mission start point, end point, and base station, measure the straight-line distance between the three points and record the value; obtain and record the signal attenuation value of the relay antenna array through a rangefinder; measure and record the energy consumption of the battery thermal management unit when it is running at full load using a power metering device; measure and record the time taken to replenish a unit amount of hydrogen at the fixed hydrogen refueling interface using a stopwatch; and generate a segment energy consumption attenuation record table. The comprehensive cost generation submodule: Based on the segment energy consumption attenuation record table, the distance value is multiplied by the energy consumption value to obtain the energy consumption. The result is weighted and accumulated with the signal attenuation value and the energy replenishment time value according to the set ratio. The comprehensive cost of each segment is summarized and archived to generate a comprehensive cost route dataset. Node distance calculation submodule: Based on the coordinates of the UAV mission start point, end point, and base station, the three-dimensional Euclidean distance calculation method is used to calculate the coordinate difference of the three coordinate points. The x-axis difference is the sum of the x-coordinate of the end point and the x-coordinate of the start point, and the square. The y-axis difference is the sum of the y-coordinate of the end point and the y-coordinate of the start point, and the square. The z-axis difference is the sum of the z-coordinate of the end point and the z-coordinate of the start point, and the square. The three square values ​​are summed and the square root is taken and recorded. The signal strength value S1 of the relay antenna array and the signal source transmission power S0 are obtained through a rangefinder and the signal attenuation calculation method is used to perform the calculation. The ratio of S0 to S1 is taken, the logarithm to the base 10 is taken, and the result is multiplied by 20 to obtain the attenuation value and recorded. The output power P of the battery thermal management unit is collected under full load output state using a power metering device, and the running duration T is collected. P and T are multiplied to obtain the energy consumption value and recorded. The time is recorded and saved when 1 standard cubic meter of hydrogen is charged into the fixed hydrogen refueling interface using a stopwatch. A segment energy consumption attenuation record table is generated. The comprehensive cost generation submodule, based on the segment energy consumption attenuation record table, uses a weighted comprehensive cost calculation method to multiply the distance value and energy consumption value to obtain the energy consumption amount. A weighting coefficient W1 is assigned to the energy consumption amount with a weight of 0.5, signal attenuation value is assigned a weight of 0.3 with a weight of W2, and replenishment time value is assigned a weight of 0.2 with a weight of W3. The energy consumption amount is multiplied by W1 to obtain the weighted consumption value A1, the signal attenuation value is multiplied by W2 to obtain the weighted attenuation value A2, and the replenishment time value is multiplied by W3 to obtain the weighted time value A3. A1, A2, and A3 are summed to obtain the comprehensive cost value C. A key-value pair record is established between each segment number and its corresponding comprehensive cost value. The comprehensive cost of each segment is summarized and stored in the archived record, generating a comprehensive cost route dataset.

[0022] The closed-loop cost reduction module includes: Closed-loop calibration submodule: Based on the comprehensive cost route dataset, it compares the node connection order one by one and records the path groups with consistent first and last nodes, accumulates the signal attenuation value and energy consumption value of each segment and records the sum, establishes a correspondence table between the obtained total cost value and the path group number, and generates a list of total path costs. Path filtering submodule: Based on the total path cost list, the total cost is compared with the preset threshold and the path exceeding the threshold is marked. The signal attenuation value and energy consumption value of each segment in the high-value path are comprehensively sorted and the smaller item is selected to establish a list of retainable segment numbers and generate a set of low-cost segments. Task segment reorganization submodule: Based on the low-cost segment set, topological sorting is used to extract the task number and node location information of the non-reserved segments and divide them into units. These nodes are continuously combined with the connectable nodes in the set and the connection order is updated to form a new segment node list and generate a loop-free route combination table. Closed-loop calibration submodule: Based on the comprehensive cost route dataset, a list of all nodes and their connections is established using the strong connected component detection algorithm. Starting from the unvisited node, the nodes are visited sequentially according to the connection order, and the visited nodes are recorded. The visited nodes are temporarily stored in a stack. If the visit returns to the starting node that is already in the current visit path, it is determined to be a closed loop. All nodes in the closed loop are recorded as a path group. The signal attenuation value and energy consumption value of all segments in each path are calculated and totaled as the total cost of the path group. The path group number and the total cost are stored accordingly to generate a path total cost list. Path selection submodule: Based on the total path cost list, a cost-based sorting algorithm is used to arrange path groups in ascending order of total cost value. Path groups with total cost exceeding a preset threshold are marked as high-cost path groups. The signal attenuation value and energy consumption value of all segments in each high-cost path group are read, and the two data are added to obtain the single-segment cost value of the segment. The single-segment cost values ​​are sorted in ascending order, and the segments with lower cost values ​​are selected as reserved segments. The numbers of these segments are recorded to generate a reserved number list and generate a low-cost segment set. The task segment reorganization submodule, based on a low-cost segment set, uses a topology sorting algorithm to first count the number of incoming connections for each node. Nodes with zero incoming connections are added to an initial queue. Nodes are then sequentially removed from the queue and recorded in an ordered list. The number of incoming connections for all next-level nodes connected to that node is reduced by one. When the number of incoming connections for a node decreases to zero, it is added to the queue. This process continues until the queue is empty. The nodes in the ordered list are then sequentially combined into continuous segment paths, and connections are updated at the nodes that need to be connected. Finally, a loop-free path record table containing all nodes is formed, generating a loop-free route combination table.

[0023] Topological sorting involves establishing a directed connection table for all segment nodes based on the low-cost segment set. It calculates the in-degree of each node, adds nodes with an in-degree of zero to a sequential queue, and sequentially removes nodes from the queue and records them in an ordered list. It decrements the in-degree of all outgoing edges connecting a node to its target node by one. If the in-degree of a target node becomes zero, it is added to the sequential queue. This process is repeated until the queue is empty or all nodes are recorded. The order of nodes in the ordered list represents the task execution order that satisfies the directed connection relationships.

[0024] The task interruption monitoring module includes: Segment status comparison submodule: Based on the loopless route combination table, read the corresponding UAV energy reserve value and subtract it from the high-frequency DC charging pile demand value, read the communication signal strength and subtract it from the signal switch stable signal value, read the remaining execution time of the segment and subtract it from the available time of the automatic guidance track, record each comparison difference and abnormal mark, and generate an abnormal segment judgment table. Risk list generation submodule: Based on the abnormal flight segment judgment table, it filters flight segments with insufficient energy value, signal value below stable value, and execution time less than travel time, extracts flight segment number, risk category and corresponding UAV number, compiles risk task segment data, and generates a list of interrupted risk flight segments; The segment status comparison submodule, based on the loopless route combination table, uses a conditional difference calculation method to first read the energy reserve value (energyReserve) of the UAV corresponding to each segment, then read the energy replenishment demand value (energyNeed) of the high-frequency DC charging pile. The values ​​of energyReserve are subtracted from energyNeed in descending order, and the difference is recorded as energyDiff. Then, the communication signal strength (signalStrength) of the segment is read, followed by the stable signal value (signalStable) of the signal switcher. The value of signalStrength is subtracted from signalStable in the same steps, and the difference is recorded as signalDiff. Then, the remaining execution time (remainTime) of the segment is read, followed by the available time (trackTime) of the autoguided track. The remainingTime and trackTime are subtracted, and the difference is recorded as timeDiff. The module checks whether energyDiff, signalDiff, and timeDiff are lower than the set energy threshold, signal threshold, and time threshold, respectively. Segments lower than the threshold are marked as abnormal segments, and the abnormal markers and corresponding differences are written into the abnormal data record to generate an abnormal segment judgment table. The risk list generation submodule, based on the abnormal flight segment judgment table, uses a multi-condition rule filtering method to read the energyDiff, signalDiff, and timeDiff of each record. Segments with energyDiff below zero are marked as having insufficient energy, signalDiff as having insufficient communication, and timeDiff as having insufficient progress. For each marked segment, a unique segment ID (routeID), risk category (riskType), and corresponding UAV ID (uavID) are extracted. These data are then arranged in ascending order by segment ID to form a grouped data table. All records in the grouped data table are stored as a structured dataset in the original field order to generate a list of interruption risk flight segments.

[0025] The demand value of high-frequency DC charging piles specifically refers to the amount of energy required for a certain drone to complete the entire mission within a specific mission segment. It is calculated by measuring the difference between the current energy reserve of the drone and the estimated energy consumption of the mission segment, and then converting the difference according to the rated output power and charging amount per unit time of the high-frequency DC charging pile to obtain the total charging amount required to meet the mission segment. This amount is then directly compared with the energy reserve to determine the size of the energy replenishment demand.

[0026] The tiered guarantee scheduling module includes: Risk classification submodule: Based on the list of interruption risk flight segments, the decision tree algorithm is used to subtract the energy reserve value and demand value of each flight segment and determine it as energy priority; subtract the communication signal strength and stability value and determine it as communication priority; subtract the remaining execution time and travel time and determine it as schedule priority; establish an association table between the classification results and the flight segment number to generate a multi-category risk list. Differential Ranking Submodule: Based on a multi-category risk list, sort the energy priority segments in ascending order by energy difference, sort the communication priority segments in descending order by signal difference, and sort the schedule priority segments in ascending order by time difference. Merge the sorting results of each category in turn and retain the category labels to generate a risk ranking task set. Sequential generation submodule: Based on the risk-sorted task set, integrate risky segments and risk-free segments, synchronously record priority tags, generate task execution order according to sorting position, establish numbering and sequence tables, and generate hierarchical sequential task tables; Risk Classification Submodule: Based on the list of interruption risk flight segments, the module uses a decision tree algorithm to read the energy reserve value and energy demand value of each flight segment. The energy reserve value is subtracted from the energy demand value and recorded as the energy difference. Then, the communication signal strength and signal stability reference value of the flight segment are read, and the communication signal strength is subtracted from the signal stability reference value and recorded as the signal difference. Next, the remaining execution time and available travel time of the automatic guidance track are read, and the remaining execution time is subtracted from the available travel time and recorded as the time difference. According to the classification rules of the decision tree algorithm, the difference type with the highest distinguishability among energy difference, signal difference, and time difference is first determined. All flight segments are grouped according to the threshold of this difference type. Each group is then grouped again according to the remaining difference types until each group corresponds to only one risk category. The final classification results are then used to generate a correspondence table with the flight segment number to generate a multi-category risk list. The differential sorting submodule: Based on a multi-category risk list, a multi-key sorting algorithm is used to sort energy priority segments from smallest to largest energy difference, communication priority segments from largest to smallest signal difference, and schedule priority segments from smallest to largest time difference. The sorting results of the three categories are merged in the order of energy priority, communication priority, and schedule priority, and the category label of the segment is retained in each record. After merging, an ordered task data table including all segments is generated, and a risk sorting task set is generated. Sequential Generation Submodule: Based on the risk-sorted task set, a priority queue scheduling algorithm is used to put task data into the queue in order of priority label from smallest to largest. The tasks at the front of the queue are taken out in sequence to form an execution order list. When taking out each task, the task number and its execution order position are recorded in the task execution table. Risk category tasks and non-risk category tasks are sorted and stored together. The complete task scheduling list is formed by arranging them in the sequential list, generating a hierarchical sequential task table.

[0027] The decision tree algorithm, based on the energy difference, communication signal difference, and time difference values ​​of each segment in the list of interruption risk segments, selects the difference parameter with the largest information gain as the root node attribute. All segments are then split into several child node sets according to the attribute's threshold condition. For each child node set, the information gain of the remaining difference parameters is calculated sequentially, and the largest value is selected as the current node's splitting attribute. This splitting process is repeated until the segments in the child node sets belong to the same risk category or the difference parameter is exhausted. At each leaf node, the segment number and risk category label are recorded, forming a conditional judgment path from the root node to the leaf node. This allows the input segment data to sequentially pass through the threshold condition to the corresponding leaf node, outputting the segment classification result.

[0028] The dynamic task adjustment module includes: The real-time flight segment detection submodule reads the change value of the UAV's energy reserve based on the hierarchical sequential task table and subtracts it from the energy threshold; it reads the signal fluctuation value of the signal switcher and subtracts it from the signal stability threshold; it reads the change value of the automatic guidance track travel time and subtracts it from the travel time threshold; it marks flight segments that exceed the threshold and generates a flight segment over-boundary screening table. The instruction generation submodule: Based on the out-of-bounds flight segment screening table, selects the UAV that is closest to the abnormal flight segment and whose energy value, signal value, and passage time are all within the normal range from the task table, assigns the flight segment to the UAV and adjusts the task order, records the updated flight segment number and order list, and generates an instant flight segment execution instruction set; The real-time flight segment detection submodule, based on a hierarchical sequential task table, uses a conditional threshold comparison algorithm to read the energy reserve change value of the UAV corresponding to each flight segment from the task table one by one, and reads the energy threshold. It then directly compares the energy reserve change value with the energy threshold using natural language description and records the difference. Next, it reads the signal fluctuation value of the signal switcher corresponding to each flight segment and reads the signal stability threshold. It then compares the signal fluctuation value with the signal stability threshold and records the difference. Finally, it reads the change in the automatic guidance track travel time and reads the travel time threshold. It then compares the travel time change with the travel time threshold and records the difference. It compares energy, signal, and time in that order to determine whether they exceed their respective allowable ranges. Flight segments that exceed the allowable ranges are marked as out-of-bounds flight segments, and the flight segment number and corresponding difference data are recorded to generate an out-of-bounds flight segment screening table. The instruction generation submodule, based on the out-of-bounds flight segment screening table, uses a shortest distance matching algorithm to find all UAVs in normal status in the task table, obtains the current coordinates of each UAV, and obtains the starting position of the out-of-bounds flight segment. It directly compares the spatial distances of these two positions in the description, without using any mathematical functions, but instead using textual descriptions to compare the straight-line distances from each UAV to the target starting point. A list is then created in ascending order of distance, and the UAV with the shortest distance is selected. The out-of-bounds flight segment is assigned to this UAV, and the order of the UAV's task execution plan is modified. Simultaneously, the adjusted flight segment number and the new execution order are updated in the records, generating an instant flight segment execution instruction set.

[0029] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An intelligent scheduling system for a three-in-one drone base station integrating "communication relay + supercharging + hydrogen refueling", characterized in that, The system includes: Route energy consumption construction module: Based on the coordinates of the UAV mission start point, end point, and base station, measure the distance between nodes, record the attenuation of relay antenna array, energy consumption of battery thermal management unit, and energy replenishment rate of fixed hydrogen refueling interface, calculate the comprehensive cost, and generate a comprehensive cost route dataset. Closed-loop cost resolution module: Based on the comprehensive cost route dataset, check the node order to mark closed loops, sum the segment attenuation and energy consumption to obtain the total cost, compare it with the threshold to retain low-value paths, use topological sorting to remove and reorganize task segments, and generate a closed-loop-free route combination table. Mission interruption monitoring module: Based on the closed-loop route combination table, compare energy reserves with high-frequency DC charging pile demand, communication strength with signal switch stability value, execution time with automatic guidance track travel time, mark abnormal segments and generate a list of interruption risk segments; Layered guarantee scheduling module: Based on the list of interruption risk segments, the module uses a decision tree algorithm to sort, merge, and arrange the segments according to energy difference, signal difference, and schedule difference, determine the execution order and priority, and generate a layered sequential task table. Dynamic task adjustment module: Based on the hierarchical sequential task table, it compares energy changes, signal fluctuations, traffic changes with thresholds, assigns abnormal flight segments to the nearest normal UAV, updates the order, and generates an instant flight segment execution instruction set.

2. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station as described in claim 1, characterized in that, The comprehensive cost route dataset includes inter-node distance values, relay antenna array signal attenuation values, battery thermal management unit energy consumption values, and fixed hydrogen refueling interface refueling rates. The loopless route combination table includes low-cost segment node sequences, removed segment task segment numbers, and reorganized new segment node connection relationships. The interruption risk segment list includes abnormal segment numbers, risk classification labels, corresponding UAV numbers, energy difference values, communication signal difference values, and travel time difference values. The hierarchical sequential task table includes segment numbers, risk levels, energy difference sorting results, signal difference sorting results, progress difference sorting results, and task execution order. The real-time segment execution instruction set includes segment numbers, executing UAV numbers, task start nodes, task end nodes, and updated task execution order.

3. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 1, characterized in that, The route energy consumption construction module includes: Node distance calculation submodule: Based on the coordinates of the UAV mission start point, end point, and base station, measure the straight-line distance between the three points and record the value; obtain and record the signal attenuation value of the relay antenna array through a rangefinder; measure and record the energy consumption of the battery thermal management unit when it is running at full load using a power metering device; measure and record the time taken to replenish a unit amount of hydrogen at the fixed hydrogen refueling interface using a stopwatch; and generate a segment energy consumption attenuation record table. The comprehensive cost generation submodule: Based on the energy consumption attenuation record table of the flight segment, the distance value is multiplied by the energy consumption value to obtain the energy consumption amount. The result is weighted and accumulated with the signal attenuation value and the energy replenishment time value according to a set ratio. The comprehensive cost of each flight segment is summarized and archived to generate a comprehensive cost route dataset.

4. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 1, characterized in that, The closed-loop cost reduction module includes: Closed-loop calibration submodule: Based on the comprehensive cost route dataset, compare the node connection order one by one and record the path groups with consistent first and last nodes, accumulate the signal attenuation value and energy consumption value of each segment and record the sum, establish a correspondence table between the obtained total cost value and the path group number, and generate a total path cost list. Path filtering submodule: Based on the total path cost list, the total cost is compared with a preset threshold and paths exceeding the threshold are marked. The signal attenuation value and energy consumption value of each segment in the high-value path are comprehensively sorted and smaller items are selected to establish a list of retainable segment numbers and generate a set of low-cost segments. Task segment reorganization submodule: Based on the low-cost segment set, topological sorting is used to extract the task number and node location information of the non-reserved segments and divide them into units. These nodes are continuously combined with connectable nodes in the set and the connection order is updated to form a new segment node list and generate a loop-free route combination table.

5. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 4, characterized in that, The topology sorting process involves establishing a directed connection table for all segment nodes based on the low-cost segment set, calculating the in-degree value of each node, placing nodes with an in-degree of zero into a sequential queue, sequentially removing nodes from the queue and recording them in an ordered list, decrementing the in-degree value of all outgoing edges of a node connected to the target node by one, and adding a target node to the sequential queue if its in-degree is reduced to zero. This process is repeated until the queue is empty or all nodes are recorded. The order of nodes in the ordered list represents the task execution order that satisfies the directed connection relationship.

6. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 1, characterized in that, The task interruption monitoring module includes: Segment status comparison submodule: Based on the loopless route combination table, read the corresponding UAV energy reserve value and subtract it from the high-frequency DC charging pile demand value, read the communication signal strength and subtract it from the signal switch stable signal value, read the remaining execution time of the segment and subtract it from the available time of the automatic guidance track, record each comparison difference and abnormal mark, and generate an abnormal segment judgment table. Risk list generation submodule: Based on the abnormal flight segment judgment table, it filters flight segments with insufficient energy value, signal value below stable value, and execution time less than travel time, extracts flight segment number, risk category and corresponding UAV number, compiles risk mission segment data, and generates a list of interrupted risk flight segments.

7. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 1, characterized in that, The high-frequency DC charging pile demand value specifically refers to the amount of energy required for a certain UAV to complete the entire process within a specific mission segment. It is calculated by measuring the difference between the current energy reserve of the UAV and the estimated energy consumption of the mission segment, and then converting the difference according to the rated output power and charging amount per unit time of the high-frequency DC charging pile to obtain the total charging amount required to meet the execution of the mission segment. This amount is then directly compared with the energy reserve to determine the size of the energy replenishment demand.

8. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 1, characterized in that, The hierarchical guarantee scheduling module includes: Risk classification submodule: Based on the list of interruption risk segments, the decision tree algorithm is used to subtract the energy reserve value and demand value of each segment and determine it as energy priority; subtract the communication signal strength and stability value and determine it as communication priority; subtract the remaining execution time and travel time and determine it as schedule priority; establish an association table between the classification results and the segment number to generate a multi-category risk list. Differential Ranking Submodule: Based on the multi-category risk list, sort the energy priority segments in ascending order by energy difference, sort the communication priority segments in descending order by signal difference, and sort the schedule priority segments in ascending order by time difference. Merge the sorting results of each category in turn and retain the category labels to generate a risk ranking task set. Sequential generation submodule: Based on the risk sorting task set, integrate risky flight segments and risk-free flight segments, synchronously record priority tags, generate task execution order according to sorting position, establish numbering and sequence tables, and generate hierarchical sequential task tables.

9. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 8, characterized in that, The decision tree algorithm, based on the energy difference, communication signal difference, and time difference of each segment in the list of interruption risk segments, selects the difference parameter with the largest information gain as the root node attribute. It then splits all segments into several child node sets according to the threshold conditions of the attribute. For each child node set, it calculates the information gain of the remaining difference parameters and selects the largest one as the splitting attribute of the current node. The splitting process is repeated until the segments in the child node sets belong to the same risk category or the difference parameters are exhausted. At each leaf node, the segment number and risk category label are recorded, forming a condition judgment path from the root node to the leaf node. This allows the input segment data to pass through the threshold conditions and sequentially descend to the corresponding leaf node, outputting the segment classification result.

10. The intelligent scheduling system for the "communication relay + supercharging + hydrogen refueling" integrated drone base station according to claim 1, characterized in that, The dynamic task adjustment module includes: Real-time flight segment detection submodule: Based on the hierarchical sequential task table, read the change value of UAV energy reserve and subtract it from the energy threshold, read the signal fluctuation value of signal switcher and subtract it from the signal stability threshold, read the change value of automatic guidance track travel time and subtract it from the travel time threshold, mark flight segments that exceed the threshold, and generate a cross-boundary flight segment screening table. The instruction generation submodule: Based on the out-of-bounds flight segment screening table, selects the UAV that is closest to the abnormal flight segment and whose energy value, signal value, and travel time are all within the normal range from the task table, assigns the flight segment to the UAV and adjusts the task order, records the updated flight segment number and order list, and generates an instant flight segment execution instruction set.

Images