Dual-unmanned-aerial-vehicle system capable of improving cruising ability through aerial charging

Abstract

The invention belongs to the technical field of unmanned aerial vehicles, and particularly relates to a dual-unmanned aerial vehicle system for improving cruising ability through aerial charging, which comprises an energy management and dynamic charging cooperation module for firstly monitoring the battery state of dual unmanned aerial vehicles in real time, triggering a charging request and formulating a charging strategy based on monitoring data, and then, in the aspect of cooperative control, controlling the dual unmanned aerial vehicles according to the charging request. A charging strategy is dynamically adjusted according to the environment by adopting a dynamic charging protocol, and then dynamic charging alignment with centimeter-level precision is realized based on a near-field positioning technology of a convolutional neural network in order to cope with the influence of environment change on energy transmission. A dynamic charging protocol is adopted, the laser charging mode and the microwave charging mode can be flexibly switched according to weather conditions, the environmental adaptability and the energy transmission stability are greatly improved, and the charging interruption problem caused by weather or position deviation is avoided.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, specifically to a dual UAV system that uses in-flight charging to enhance endurance. Background Technology

[0002] Drones are increasingly used in agricultural inspections and logistics delivery, but insufficient battery life remains a major challenge. Increasing battery capacity reduces payload capacity, and the emergence of in-flight charging technology offers a new solution, with laser and microwave transmission technologies showing particularly strong performance. A dual-drone collaborative system allows the primary drone to perform the task while the secondary drone acts as a mobile charging platform, optimizing resource allocation. Research by relevant teams has also validated its effectiveness.

[0003] However, the existing dual-drone system has the following specific problems: 1. Energy transmission efficiency and environmental adaptability: Laser charging is greatly affected by weather, microwave charging efficiency decreases significantly with distance, and it is also susceptible to electromagnetic interference, such as the main drone having to return to base due to interruption of charging caused by rain during agricultural inspections.

[0004] 3. Task allocation and energy management: Existing collaborative algorithms suffer from uneven task allocation in large-scale scenarios. For example, in emergency rescue, the auxiliary drone may run out of power due to excessive charging tasks, affecting the operation of the main drone.

[0005] Based on the above, a dual-drone system for improving endurance through in-flight charging is invented. Summary of the Invention

[0006] To address the aforementioned technical problems, according to one aspect of the present invention, the present invention provides the following technical solution: A dual-drone system that enhances endurance through in-flight charging, comprising: The energy management and dynamic charging coordination module is used to first monitor the battery status of the two drones in real time, and can trigger charging requests and formulate charging strategies based on the monitoring data. Then, in terms of coordinated control, a dynamic charging protocol is adopted to dynamically adjust the charging strategy according to the environment. After that, in order to cope with the impact of environmental changes on energy transmission, near-field positioning technology based on convolutional neural networks is used to achieve dynamic charging alignment with centimeter-level accuracy. The flight control and cooperative navigation module first uses multimodal perception fusion to achieve centimeter-level positioning, providing accurate position information for the flight and cooperation of the two UAVs. At the same time, it combines reinforcement learning obstacle avoidance strategies to ensure safe flight in complex environments. Next, in terms of cooperative control, it integrates high-altitude global maps and low-altitude fine perception data to construct an environmental model in a unified coordinate system, realizing multimodal cooperative navigation. Then, when facing obstacles in a dynamic environment, it uses particle swarm optimization algorithm to optimize the path, improve control response speed, and reduce collision risk. The task planning and energy balance module is used to first generate optimal waypoints based on the AeroDuo framework trained on the HaL-13k dataset, combined with the target probability distribution map and the A* algorithm, and rationally plan the mission path of the main UAV. Then, in terms of cooperative control, it realizes the coordination of task allocation and energy management to avoid uneven task allocation.

[0007] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the energy management and dynamic charging coordination module includes: The real-time status monitoring and data acquisition module is used to continuously collect the core battery parameters through the onboard sensors of the main and auxiliary UAVs, while synchronizing the UAV's position coordinates and relative distance. Then, after the data is preprocessed by the edge computing unit, it can be transmitted to the ground station and the other UAV through the 5G link. The charging trigger condition judgment module is used to automatically trigger a charging request when the main drone meets any of the following conditions: Condition 1: SOC ≤ 30% or SOC ≤ 40%; Condition 2: The remaining battery power is not expected to complete the next flight segment; Condition 3: The auxiliary drone actively detects abnormal battery temperature of the main drone and triggers a cooling and charging mode.

[0008] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the energy management and dynamic charging coordination module further includes: The charging mode dynamic decision module is used to select the following mode after the assisted drone receives a request, based on environmental perception data and relative distance: Mode 1: Visibility ≥ 1km and no precipitation, select laser charging: activate the 650nm wavelength laser, and set the initial power to 50W; Mode 2: When visibility is <1km or there is precipitation, automatically switch to microwave charging: activate the 24GHz band beamforming system, and the transmission power is graded according to distance; Mode 3: Extreme Environment Trigger Backup Plan: The auxiliary drone flies to within 5 meters of the main drone and activates contact wireless charging; The high-precision alignment and energy transfer module enables the auxiliary drone to achieve dynamic alignment through coarse and fine alignment. During the transmission process, it monitors the intensity of reflected signals in real time. When the main drone deviates due to airflow, the auxiliary drone completes angle correction through second-order harmonic feedback signals to maintain charging efficiency.

[0009] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the energy management and dynamic charging coordination module further includes: The charging process monitoring and safety redundancy module is used to continuously monitor the charging process and set up a three-level protection mechanism; The charging termination and mission coordination reset module is used to send a termination signal when the main drone's SOC is ≥80% or when emergency power replenishment is completed. At this time, the main drone will continue to fly along the original mission path, and the auxiliary drone will calculate the optimal waiting point (within 50 meters of the next waypoint of the main drone) and enter a low-power hovering state to wait for the next charging request.

[0010] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the flight control and cooperative navigation module includes: The environmental perception and data fusion module enables the main UAV and the auxiliary UAV to collect environmental data through multimodal sensors. After the data is processed by the edge computing unit, it can achieve bidirectional synchronization through the 5GMesh network and be fused into an environmental model in a unified coordinate system. The global path planning and task decomposition module is used to generate the global task path of the main UAV based on the task objectives and the initial environment model using an improved A* algorithm. The auxiliary UAV plans the charging support path based on the path data of the main UAV using a hierarchical clustering algorithm. At the same time, it sets potential charging points between the waypoints of the main UAV and calculates the energy replenishment efficiency of each point.

[0011] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the flight control and cooperative navigation module further includes: The real-time positioning and cooperative obstacle avoidance module is used to first enable the main UAV to achieve centimeter-level positioning through visual SLAM technology, and combine IMU data to compensate for positioning drift when GNSS signals are missing; at the same time, it enables the auxiliary UAV to correct its own coordinates through differential GNSS and relative position data of the main UAV; then, when a sudden obstacle is detected, the two UAVs start a cooperative obstacle avoidance mechanism. The main UAV first determines the type and trajectory of the obstacle through a visual recognition system, and then the auxiliary UAV calculates the deviation of the flight path based on a global environment model and generates obstacle avoidance commands through a reinforcement learning algorithm. The dynamic trajectory tracking and attitude control module is used to first enable both the main UAV and the auxiliary UAV to adopt the PID+LQR composite control algorithm; then, during cooperative flight, the auxiliary UAV adjusts its own flight parameters by receiving the real-time velocity vector of the main UAV and using the model predictive control algorithm.

[0012] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the flight control and cooperative navigation module further includes: The path dynamic adjustment module during the mission is used to enable the main UAV to identify the type of environmental change through a visual language model and calculate the path correction requirements when the environmental model changes significantly. At the same time, after receiving the correction information, the auxiliary UAV re-evaluates the feasibility of the charging point, updates the support path through the particle swarm optimization algorithm, and pushes alternative adjustment schemes to the main UAV. The return-to-home coordination module is used to send a return-to-home command to the auxiliary UAV after the main UAV completes its mission. The two UAVs plan a coordinated return-to-home path based on the final position data. The auxiliary UAV prioritizes routes with lower energy consumption and calculates the best rendezvous point between the two UAVs.

[0013] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the mission planning and energy balancing module includes: The initial mission planning and waypoint generation module is used to first generate the initial optimal waypoints of the main UAV based on the AeroDuo framework trained on the mission requirements and the HaL-13k dataset, combined with the target probability distribution map, using the A* algorithm; then, based on the waypoint distribution of the main UAV, it plans the corresponding charging support waypoints for the auxiliary UAV. The task allocation and energy budget calculation module is used to first use an improved particle swarm optimization algorithm combined with the Hungarian algorithm to quantitatively allocate tasks for the main and auxiliary UAVs; then, iterative calculation is used to achieve task load balancing.

[0014] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the mission planning and energy balancing module further includes: The real-time energy monitoring and dynamic adjustment trigger module is used to first collect the power data of the two drones in real time, compare it with the initial energy consumption estimate, and trigger the dynamic adjustment mechanism under the following conditions: Scenario 1: The actual power consumption of the main drone exceeds the estimate by 15%, and the remaining power is insufficient to cover the next charging point; Scenario 2: The charging task for the auxiliary drone takes 20% longer than estimated, causing the subsequent charging point to be unable to arrive on time; Scenario 3: New temporary tasks are added, and task weights need to be reallocated; Once dynamic adjustment is triggered, multiple alternative solutions will be generated within a preset time, and the optimal solution will be selected through an energy consumption-efficiency evaluation model. The task path and energy allocation co-optimization module is used to optimize paths and energy in response to triggered adjustment requests through the following methods: Method 1, Path Replanning: The improved A* algorithm is used to shorten the detour distance of the main UAV, while the charging point position of the auxiliary UAV is adjusted simultaneously to ensure that the path deviation between the two is ≤30 meters; Method 2, Adjusting the charging strategy: If the main drone's battery is low, temporarily increase the charging power of the auxiliary drone and extend the charging time, while reducing the non-essential tasks of the auxiliary drone. Method 3, Emergency Energy Allocation: When the auxiliary drone's battery is low, the main drone's backup battery is activated to provide emergency charging, ensuring the completion of the core charging task.

[0015] As a preferred embodiment of the dual-UAV system for enhancing endurance through in-flight charging as described in this invention, the mission planning and energy balancing module further includes: The mission completion and energy review module is used to automatically generate an energy utilization report after the mission is completed. This report includes statistics on the mission energy consumption distribution of the main UAV, the charging efficiency curve of the auxiliary UAV, and the number and effect of dynamic adjustments. Then, based on the report data, the parameters of the mission planning model are updated through reinforcement learning to improve the accuracy of energy consumption prediction for the next mission. At the same time, the battery health loss of the two UAVs is calculated. If the SOH drops beyond the preset value, a battery maintenance reminder is triggered.

[0016] Compared with existing technologies: 1. The system boasts significant advantages in energy transmission efficiency and environmental adaptability. Addressing the issues of laser charging being highly susceptible to weather conditions and microwave charging efficiency decreasing with distance and being easily interfered with, the system employs a dynamic charging protocol that flexibly switches between laser and microwave charging modes based on weather conditions. In clear weather, laser charging is used with a high-precision tracking system to ensure efficiency, while in complex weather, microwave charging is switched to maintain energy transmission through beamforming technology and dynamic power adjustment. Simultaneously, near-field positioning technology based on convolutional neural networks achieves centimeter-level precision alignment, maintaining high charging efficiency even with deviations in the main UAV. This greatly improves environmental adaptability and energy transmission stability, avoiding charging interruptions caused by weather or positional shifts. For example, situations where charging is interrupted due to rainfall during agricultural inspections will be significantly reduced. 2. The system also performs well in addressing collaborative control and communication latency issues. To mitigate the challenges of synchronizing information between two drones, communication congestion, and collision risks from obstacles, the system constructs a 5G+Mesh hybrid network. Control commands are transmitted rapidly via 5G, while data is fed back through the Mesh network, reducing data congestion. Simultaneously, multimodal collaborative navigation integrates various data sources to build an environmental model. The primary and secondary drones collaboratively identify targets and generate paths, improving obstacle avoidance response speed and reducing collision risks. Issues such as missing charging windows due to communication delays in logistics scenarios are effectively resolved. 3. The system has significant advantages in task allocation and energy management. Addressing the issue of uneven task allocation in existing algorithms, the system's task planning and energy balancing modules, based on improved particle swarm optimization and Hungarian algorithms, achieve a balance between fairness and efficiency in task allocation, reducing the variance in task completion time and preventing auxiliary drones from running out of power due to excessive workload. Simultaneously, the inclusion of backup batteries provides safety redundancy, ensuring security in emergency situations and allowing the main drone to complete tasks more stably. For example, the phenomenon of auxiliary drones running out of power and affecting the main drone's operations during emergency rescue operations will be significantly reduced. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall framework of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0019] This invention provides a dual-drone system that enhances endurance through in-flight charging. Please refer to [link / reference]. Figure 1 ,include: The energy management and dynamic charging coordination module is used to first monitor the battery status of the two drones in real time, including information such as SOC (State of Charge), temperature, and SOH. It can trigger charging requests and formulate charging strategies based on the monitoring data. Then, in terms of coordinated control, a dynamic charging protocol is used to dynamically adjust the charging strategy according to the environment. After that, in order to cope with the impact of environmental changes on energy transmission, near-field positioning technology based on convolutional neural networks is used to achieve dynamic charging alignment with centimeter-level accuracy. The flight control and cooperative navigation module first uses multimodal perception fusion to achieve centimeter-level positioning, providing accurate position information for the flight and cooperation of the two UAVs. At the same time, it combines reinforcement learning obstacle avoidance strategies to ensure safe flight in complex environments. Next, in terms of cooperative control, it integrates high-altitude global maps and low-altitude fine perception data to construct an environmental model in a unified coordinate system, realizing multimodal cooperative navigation. Then, when facing obstacles in a dynamic environment, it uses particle swarm optimization algorithm to optimize the path, improve control response speed, and reduce collision risk. The task planning and energy balance module is used to first generate optimal waypoints based on the AeroDuo framework trained on the HaL-13k dataset, combined with the target probability distribution map and the A* algorithm, and rationally plan the mission path of the main UAV. Then, in terms of cooperative control, it realizes the coordination of task allocation and energy management to avoid uneven task allocation.

[0020] The energy management and dynamic charging coordination module includes: The real-time status monitoring and data acquisition module is used to continuously collect core battery parameters through the onboard sensors of the main and auxiliary UAVs. It records the SOC (accuracy ±1%), battery temperature (sampling range -20℃ to 60℃, accuracy ±0.5℃), and SOH (calibrated daily, error ≤3%) every 0.5 seconds. At the same time, it synchronizes the UAV's position coordinates (GPS + Beidou dual-mode positioning, refresh rate 10Hz) and relative distance (LiDAR measurement, range 0-500 meters, accuracy ±0.1 meters). Then, after the data is preprocessed by the edge computing unit, it can be transmitted to the ground station and the peer UAV through the 5G link. The charging trigger condition judgment module is used to automatically trigger a charging request when the main drone meets any of the following conditions (the request signal includes the real-time position of the main drone, the expected dwell time window (±10 seconds), and a priority charging mode suggestion (laser / microwave)): Condition 1: SOC ≤ 30% (for regular tasks) or SOC ≤ 40% (for high-priority tasks, such as emergency rescue). Condition 2: The remaining battery power is not expected to complete the next flight segment (the difference between the path energy consumption calculated based on the A* algorithm and the current battery power is ≥15% safety redundancy); Condition 3: The auxiliary drone actively detects abnormal battery temperature of the main drone (>55℃) and triggers cooling charging mode (reducing charging power to 30%). The charging mode dynamic decision module is used to select the following modes after the assisted drone receives a request, based on environmental perception data (visibility and precipitation intensity transmitted in real time by weather sensors) and relative distance: Mode 1: Visibility ≥ 1km and no precipitation, select laser charging: activate 650nm wavelength laser, initial power set to 50W (dynamically adjusted according to distance, linearly attenuating to 20W within 300 meters). Mode 2: When visibility is less than 1km or there is precipitation, microwave charging will be automatically switched: the 24GHz band beamforming system will be activated, and the transmission power will be graded according to distance (50W within 10 meters, 20W within 20 meters, and 5W within 50 meters). Mode 3: Extreme environment (such as rainstorm, sandstorm) trigger backup plan: The auxiliary drone flies within 5 meters of the main drone and activates contact wireless charging (Qi standard, 15W power). The high-precision alignment and energy transfer module enables the auxiliary drone to achieve dynamic alignment through coarse and fine alignment. During the transmission process, the intensity of the reflected signal is monitored in real time. When the main drone deviates due to airflow, the auxiliary drone completes the angle correction through the second harmonic feedback signal to maintain charging efficiency. The coarse alignment: adjusts the flight trajectory based on GPS coordinate differences to bring the main UAV into the field of view of the auxiliary UAV's onboard camera (field of view angle 60°×45°), taking ≤3 seconds; The precise alignment is achieved by capturing the infrared positioning targets (three of which are distributed in an equilateral triangle) on the main UAV body using the auxiliary UAV's visual recognition system. The target center offset is calculated using a convolutional neural network (accuracy ±3cm), and the gimbal angle is adjusted by PID control until the laser / microwave transmitter is aligned with the main UAV's receiving panel (deviation ≤5cm). The charging process monitoring and safety redundancy module is used to continuously monitor the charging process and set up a three-level protection mechanism, which includes a first-level early warning, a second-level early warning, and a third-level emergency stop. Level 1 Warning (Charging can continue): When the transmission efficiency is <30% and the main drone battery temperature is >50℃, the charging power will be automatically reduced by 20% and an alarm will be issued. Level 2 warning (suspension of charging): If an obstacle is detected in the laser path (such as a bird flying in, visual recognition response time <50ms) or microwave band interference intensity >-80dBm, the power transmission will be immediately cut off and the laser will be re-aligned after the interference is eliminated. Level 3 Emergency Stop: When the battery voltage of the main / auxiliary drone suddenly drops by more than 10% and there is a risk of collision (distance < 5 meters and relative speed > 3 m / s), the emergency stop procedure for both drones is triggered. The auxiliary drone switches to obstacle avoidance mode and the main drone uses the backup battery to continue its operation. The charging termination and mission coordination reset module is used to send a termination signal when the main UAV's SOC is ≥80% (for regular missions) or when emergency recharging is completed (to meet the requirements of the next flight segment). This causes the auxiliary UAV to gradually reduce its power to 0 (1 second cool-down time is required in laser mode) and perform the following: 1. Record the charging data (duration, transmitted energy, average efficiency) to the local log; 2. Update the charging strategy library based on reinforcement learning algorithms (if interference occurs frequently in a certain area, automatically adjust the offset of the next charging position). At this time, the main UAV will continue to fly along the original mission path, and the auxiliary UAV will calculate the optimal waiting point (within 50 meters of the next waypoint of the main UAV), enter a low-power hovering state, and wait for the next charging request.

[0021] The flight control and cooperative navigation module includes: The environmental perception and data fusion module enables the main UAV and the auxiliary UAV to collect environmental data through multimodal sensors. The main UAV's visual camera (12 megapixels, 30fps) captures detailed low-altitude images, and the lidar (10Hz scanning frequency, 0-200m ranging range) generates point clouds of surrounding obstacles. The auxiliary UAV's millimeter-wave radar (0-500m detection range, 1° angular resolution) acquires the global environmental contour at high altitude, and the GNSS module (GPS + BeiDou dual-mode, 1m positioning accuracy) records real-time location information. After the data is processed by the edge computing unit, it can achieve bidirectional synchronization through the 5GMesh network and be fused into an environmental model under a unified coordinate system. For example, the data of "10-meter-high power transmission tower" identified by the main UAV and "airflow disturbance within 5 meters around the tower" detected by the auxiliary UAV are superimposed to generate a comprehensive environmental map that includes static obstacles and dynamic interference, with an update frequency of 500ms / time. The global path planning and task decomposition module is used to generate the global task path of the main UAV based on the task objective (such as "inspecting 3 square kilometers of farmland") and the initial environment model using an improved A* algorithm. This path includes key waypoints (intervals of 500 meters) and expected arrival times. The auxiliary UAV plans charging support paths based on the path data of the main UAV using a hierarchical clustering algorithm. At the same time, potential charging points are set between the waypoints of the main UAV (such as setting one alternative point between every two waypoints), and the energy replenishment efficiency of each point is calculated (considering factors such as distance and environmental occlusion). The real-time positioning and cooperative obstacle avoidance module first enables the main UAV to achieve centimeter-level positioning using visual SLAM (Simultaneous Localization and Mapping) technology, and then compensates for positioning drift when GNSS signals are missing by combining IMU (Inertial Measurement Unit, 100Hz sampling rate) data. Simultaneously, the auxiliary UAV corrects its own coordinates using differential GNSS and relative position data (LiDAR measurement, accuracy ±0.5 meters) from the main UAV. Then, when a sudden obstacle is detected (such as a bird intrusion or a temporarily erected communication antenna), the two UAVs initiate a cooperative obstacle avoidance mechanism. The main UAV first determines the obstacle type and trajectory using a visual recognition system (response time <100ms). Then, the auxiliary UAV calculates the deviation of the flight path based on a global environment model and generates obstacle avoidance commands using a reinforcement learning algorithm (PPO strategy). For example, if the main UAV deviates 2 meters from its original path due to a sudden lateral airflow, the auxiliary UAV simultaneously adjusts 3 meters in the opposite direction, maintaining a safe relative distance (≥10 meters). The obstacle avoidance process takes ≤2 seconds. The dynamic trajectory tracking and attitude control module is used to first enable both the main UAV and the auxiliary UAV to adopt a PID+LQR (linear quadratic regulator) composite control algorithm. The PID controller is responsible for position closed-loop control (position error ≤ 0.5 meters), and the LQR algorithm optimizes attitude parameters (pitch angle and roll angle control accuracy ±1°) to ensure smooth trajectory tracking. Then, during cooperative flight, the auxiliary UAV receives the real-time velocity vector of the main UAV (update frequency 10Hz) and uses a model predictive control (MPC) algorithm to adjust its own flight parameters. For example, when the main UAV is flying straight at a speed of 15 m / s, the auxiliary UAV maintains a following distance of 50 meters behind, and controls the relative speed difference within ±1 m / s by adjusting the throttle opening (response time < 50 ms) to avoid charging alignment deviation caused by speed mismatch. The dynamic path adjustment module during the mission is used to enable the main UAV to identify the type of environmental change (such as "road water depth > 30cm") and calculate the path correction requirement when the environmental model changes significantly (such as sudden rainfall causing some areas to be impassable). At the same time, after receiving the correction information, the auxiliary UAV re-evaluates the feasibility of charging points, updates the support path through the particle swarm optimization algorithm, and pushes alternative adjustment plans to the main UAV (such as "detour through the farmland on the east side and add 1 temporary charging point"). The post-mission homing coordination module sends homing commands to the auxiliary UAV after the main UAV completes its mission. Both UAVs plan a coordinated return path based on their final position data. The auxiliary UAV prioritizes routes with lower energy consumption (such as utilizing high-altitude tailwinds) and calculates the optimal rendezvous point for both UAVs (usually a safe take-off and landing field at the edge of the mission area). During the homing process, the two UAVs maintain a 100-meter distance and confirm their remaining battery power through two-way communication: if the auxiliary UAV's battery power is below 20%, the main UAV adjusts its flight altitude to provide airflow cover (reducing energy consumption by 15%); if the main UAV's battery power is insufficient, the auxiliary UAV provides a final recharge during the homing journey (20W power, lasting 5 minutes) to ensure that both UAVs can safely return to the take-off and landing field.

[0022] The task planning and energy balance module includes: The initial task planning and waypoint generation module is used to generate the initial optimal waypoints for the main UAV based on the task requirements (such as the farmland area for agricultural plant protection, the disaster area for emergency rescue) and the AeroDuo framework trained on the HaL-13k dataset, combined with the target probability distribution map (marking high-priority task areas, such as high-incidence areas of pests and diseases in farmland, and densely populated areas in disaster areas) using the A* algorithm. Each waypoint includes coordinates, estimated dwell time (such as 30 seconds / point for pesticide spraying in plant protection scenarios, and 10 seconds / point for data collection in inspection scenarios), and energy consumption estimation (calculated based on distance and task type, with an accuracy of ±5%). Then, according to the waypoint distribution of the main UAV, it plans supporting charging waypoints for the auxiliary UAV, ensuring that there is at least one alternative charging point within 50 meters of each main UAV waypoint, and that the energy consumption of the path between charging points is less than 60% of the auxiliary UAV's single full-charge range (reserve emergency power). The task allocation and energy budget calculation module first uses an improved particle swarm optimization (IPSO) algorithm combined with the Hungarian algorithm to quantitatively allocate tasks to the primary and secondary drones. The task weight of the primary drone is assigned according to waypoint priority (high-priority waypoints have a weight of 1.2, and regular waypoints have a weight of 1.0), while the task weight of the secondary drone is calculated based on the distance between the charging point and the primary drone's waypoints (the closer the distance, the higher the weight, e.g., within 10 meters, the weight is 1.1, and within 30 meters, the weight is 1.0). Next, task load balancing is achieved through iterative calculation (the number of iterations ≤ 50). If the estimated energy consumption of a certain segment of the primary drone's path exceeds the average level by 20%, the tasks of adjacent low-priority waypoints are automatically split and assigned to the secondary drone (e.g., the secondary drone undertakes some simple observation tasks during charging intervals). If the total energy consumption of the secondary drone's charging tasks exceeds 70% of its own battery capacity, the number of charging points it is responsible for is reduced, and some tasks are transferred to segments where the primary drone has sufficient range. The final task allocation scheme must meet the following requirements: the time difference between the completion of tasks of the two drones is ≤ 10%, and the remaining battery power of the secondary drone is ≥ 25% (to ensure energy consumption for the return trip). The real-time energy monitoring and dynamic adjustment trigger module is used to first collect the power data of the two drones in real time (updated once per second), compare it with the initial energy consumption estimate, and trigger the dynamic adjustment mechanism under the following conditions: Scenario 1: The actual energy consumption of the main drone exceeds the estimate by 15% (e.g., due to strong headwinds increasing flight drag), and the remaining battery power is insufficient to cover the next charging point; Scenario 2: The charging task for the auxiliary drone takes 20% longer than estimated (e.g., due to weather conditions extending the charging time), causing the subsequent charging point to be unable to arrive on time; Scenario 3: New temporary tasks are added (such as ground station instructions to add emergency inspection points), and task weights need to be reallocated; Once dynamic adjustment is triggered, multiple alternative solutions will be generated within a preset time, and the optimal solution will be selected through an energy consumption-efficiency evaluation model (with task completion rate and energy utilization rate as the core indicators). The task path and energy allocation co-optimization module is used to optimize paths and energy in response to triggered adjustment requests through the following methods: Method 1, Path replanning: Use an improved A* algorithm to shorten the detour distance of the main UAV (e.g., delete low-priority waypoints or select a shorter straight route), while simultaneously adjusting the charging point location of the auxiliary UAV to ensure that the path deviation between the two is ≤30 meters. Method 2, Adjusting the charging strategy: If the main drone's battery is low, temporarily increase the charging power of the auxiliary drone (e.g., increase the microwave charging power within 10 meters from 50W to 60W) and extend the charging time (up to 50%), while reducing the non-essential tasks of the auxiliary drone (e.g., suspending environmental observation in low-priority areas). Method 3, emergency energy allocation: When the auxiliary drone's power is insufficient, the main drone's backup battery (capacity 15% of the main battery) is activated to provide emergency charging (through short-range microwave transmission with a transmission efficiency of ≥40%), ensuring the completion of the core charging task; The mission completion and energy review module automatically generates an energy utilization report after the mission ends. This report includes statistics on the main drone's mission energy consumption distribution (e.g., 60% flight energy consumption and 40% mission equipment energy consumption), the auxiliary drone's charging efficiency curve (average efficiency at different distances), and the number and effect of dynamic adjustments (e.g., an adjustment extends the main drone's flight time by 12 minutes). Based on the report data, the module updates the parameters of the mission planning model through reinforcement learning (e.g., correcting the energy consumption coefficient in headwind conditions and optimizing the charging point distance threshold) to improve the accuracy of energy consumption prediction for the next mission. Simultaneously, it calculates the battery health loss of both drones (based on charge / discharge cycles and temperature curves). If the state of health (SOH) drops beyond a preset value, a battery maintenance reminder is triggered.

[0023] In practical use, the specific steps are as follows: S1: The energy management and dynamic charging coordination module first monitors the battery status of the two drones in real time, and can trigger charging requests and formulate charging strategies based on the monitoring data. Then, in terms of collaborative control, a dynamic charging protocol is adopted to dynamically adjust the charging strategy according to the environment. After that, in order to cope with the impact of environmental changes on energy transmission, near-field positioning technology based on convolutional neural networks is used to achieve centimeter-level accuracy dynamic charging alignment. S2: The flight control and cooperative navigation module first uses multimodal perception fusion to achieve centimeter-level positioning, providing accurate position information for the flight and cooperation of the two UAVs. At the same time, it combines reinforcement learning obstacle avoidance strategy to ensure safe flight in complex environments. Then, in terms of cooperative control, it integrates high-altitude global map and low-altitude fine perception data to build an environmental model under a unified coordinate system to achieve multimodal cooperative navigation. After that, when facing obstacles in the dynamic environment, it uses particle swarm optimization algorithm to achieve path optimization, improve control response speed, and reduce collision risk. S3: The AeroDuo framework, trained on the HaL-13k dataset, is first used in the task planning and energy balance module. The optimal waypoints are generated by combining the target probability distribution map and the A* algorithm to rationally plan the mission path of the main UAV. Then, in terms of cooperative control, the coordination of task allocation and energy management is realized to avoid uneven task allocation.

[0024] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A dual-drone system for enhancing endurance through in-flight charging, characterized in that: include: The energy management and dynamic charging coordination module is used to first monitor the battery status of the two drones in real time, and can trigger charging requests and formulate charging strategies based on the monitoring data. Then, in terms of coordinated control, a dynamic charging protocol is adopted to dynamically adjust the charging strategy according to the environment. After that, in order to cope with the impact of environmental changes on energy transmission, near-field positioning technology based on convolutional neural networks is used to achieve dynamic charging alignment with centimeter-level accuracy. The flight control and cooperative navigation module first uses multimodal perception fusion to achieve centimeter-level positioning, providing accurate position information for the flight and cooperation of the two UAVs. At the same time, it combines reinforcement learning obstacle avoidance strategies to ensure safe flight in complex environments. Next, in terms of cooperative control, it integrates high-altitude global maps and low-altitude fine perception data to construct an environmental model in a unified coordinate system, realizing multimodal cooperative navigation. Then, when facing obstacles in a dynamic environment, it uses particle swarm optimization algorithm to optimize the path, improve control response speed, and reduce collision risk. The task planning and energy balance module is used to first generate optimal waypoints based on the AeroDuo framework trained on the HaL-13k dataset, combined with the target probability distribution map and the A* algorithm, and rationally plan the mission path of the main UAV. Then, in terms of cooperative control, it realizes the coordination of task allocation and energy management to avoid uneven task allocation.

2. The dual-UAV system for enhancing endurance through in-flight charging according to claim 1, characterized in that, The energy management and dynamic charging coordination module includes: The real-time status monitoring and data acquisition module is used to continuously collect the core battery parameters through the onboard sensors of the main and auxiliary UAVs, while synchronizing the UAV's position coordinates and relative distance. Then, after the data is preprocessed by the edge computing unit, it can be transmitted to the ground station and the other UAV through the 5G link. The charging trigger condition judgment module is used to automatically trigger a charging request when the main drone meets any of the following conditions: Condition 1: SOC ≤ 30% or SOC ≤ 40%; Condition 2: The remaining battery power is not expected to complete the next flight segment; Condition 3: The auxiliary drone actively detects abnormal battery temperature of the main drone and triggers a cooling and charging mode.

3. The dual-UAV system for enhancing endurance through in-flight charging according to claim 2, characterized in that, The energy management and dynamic charging coordination module also includes: The charging mode dynamic decision module is used to select the following mode after the assisted drone receives a request, based on environmental perception data and relative distance: Mode 1: Visibility ≥ 1km and no precipitation, select laser charging: activate the 650nm wavelength laser, and set the initial power to 50W; Mode 2: When visibility is <1km or there is precipitation, automatically switch to microwave charging: activate the 24GHz band beamforming system, and the transmission power is graded according to distance; Mode 3: Extreme Environment Trigger Backup Plan: The auxiliary drone flies to within 5 meters of the main drone and activates contact wireless charging; The high-precision alignment and energy transfer module enables the auxiliary drone to achieve dynamic alignment through coarse and fine alignment. During the transmission process, it monitors the intensity of reflected signals in real time. When the main drone deviates due to airflow, the auxiliary drone completes angle correction through second-order harmonic feedback signals to maintain charging efficiency.

4. The dual-UAV system for enhancing endurance through in-flight charging according to claim 3, characterized in that, The energy management and dynamic charging coordination module also includes: The charging process monitoring and safety redundancy module is used to continuously monitor the charging process and set up a three-level protection mechanism; The charging termination and mission coordination reset module is used to send a termination signal when the main drone's SOC is ≥80% or when emergency power replenishment is completed. At this time, the main drone will continue to fly along the original mission path, and the auxiliary drone will calculate the optimal waiting point (within 50 meters of the next waypoint of the main drone) and enter a low-power hovering state to wait for the next charging request.

5. The dual-UAV system for enhancing endurance through in-flight charging according to claim 1, characterized in that, The flight control and cooperative navigation module includes: The environmental perception and data fusion module enables the main UAV and the auxiliary UAV to collect environmental data through multimodal sensors. After the data is processed by the edge computing unit, it can achieve bidirectional synchronization through the 5GMesh network and be fused into an environmental model in a unified coordinate system. The global path planning and task decomposition module is used to generate the global task path of the main UAV based on the task objectives and the initial environment model using an improved A* algorithm. The auxiliary UAV plans the charging support path based on the path data of the main UAV using a hierarchical clustering algorithm. At the same time, it sets potential charging points between the waypoints of the main UAV and calculates the energy replenishment efficiency of each point.

6. The dual unmanned aerial vehicle system for enhancing endurance through in-flight charging according to claim 5, characterized in that, The flight control and cooperative navigation module also includes: The real-time positioning and cooperative obstacle avoidance module is used to first enable the main UAV to achieve centimeter-level positioning through visual SLAM technology, and combine IMU data to compensate for positioning drift when GNSS signals are missing; at the same time, it enables the auxiliary UAV to correct its own coordinates through differential GNSS and relative position data of the main UAV; then, when a sudden obstacle is detected, the two UAVs start a cooperative obstacle avoidance mechanism. The main UAV first determines the type and trajectory of the obstacle through a visual recognition system, and then the auxiliary UAV calculates the deviation of the flight path based on a global environment model and generates obstacle avoidance commands through a reinforcement learning algorithm. The dynamic trajectory tracking and attitude control module is used to first enable both the main UAV and the auxiliary UAV to adopt the PID+LQR composite control algorithm; then, during cooperative flight, the auxiliary UAV adjusts its own flight parameters by receiving the real-time velocity vector of the main UAV and using the model predictive control algorithm.

7. The dual unmanned aerial vehicle system for enhancing endurance through in-flight charging according to claim 6, characterized in that, The flight control and cooperative navigation module also includes: The path dynamic adjustment module during the mission is used to enable the main UAV to identify the type of environmental change through a visual language model and calculate the path correction requirements when the environmental model changes significantly. At the same time, after receiving the correction information, the auxiliary UAV re-evaluates the feasibility of the charging point, updates the support path through the particle swarm optimization algorithm, and pushes alternative adjustment schemes to the main UAV. The return-to-home coordination module is used to send a return-to-home command to the auxiliary UAV after the main UAV completes its mission. The two UAVs plan a coordinated return-to-home path based on the final position data. The auxiliary UAV prioritizes routes with lower energy consumption and calculates the best rendezvous point between the two UAVs.

8. The dual-UAV system for enhancing endurance through in-flight charging according to claim 1, characterized in that, The task planning and energy balance module includes: The initial mission planning and waypoint generation module is used to first generate the initial optimal waypoints of the main UAV based on the AeroDuo framework trained on the mission requirements and the HaL-13k dataset, combined with the target probability distribution map, using the A* algorithm; then, based on the waypoint distribution of the main UAV, it plans the corresponding charging support waypoints for the auxiliary UAV. The task allocation and energy budget calculation module is used to first use an improved particle swarm optimization algorithm combined with the Hungarian algorithm to quantitatively allocate tasks for the main and auxiliary UAVs; then, iterative calculation is used to achieve task load balancing.

9. The dual unmanned aerial vehicle system for enhancing endurance through in-flight charging according to claim 8, characterized in that, The task planning and energy balance module also includes: The real-time energy monitoring and dynamic adjustment trigger module is used to first collect the power data of the two drones in real time, compare it with the initial energy consumption estimate, and trigger the dynamic adjustment mechanism under the following conditions: Scenario 1: The actual power consumption of the main drone exceeds the estimate by 15%, and the remaining power is insufficient to cover the next charging point; Scenario 2: The charging task for the auxiliary drone takes 20% longer than estimated, causing the subsequent charging point to be unable to arrive on time; Scenario 3: New temporary tasks are added, and task weights need to be reallocated; Once dynamic adjustment is triggered, multiple alternative solutions will be generated within a preset time, and the optimal solution will be selected through an energy consumption-efficiency evaluation model. The task path and energy allocation co-optimization module is used to optimize paths and energy in response to triggered adjustment requests through the following methods: Method 1, Path Replanning: The improved A* algorithm is used to shorten the detour distance of the main UAV, while the charging point position of the auxiliary UAV is adjusted simultaneously to ensure that the path deviation between the two is ≤30 meters; Method 2, Adjusting the charging strategy: If the main drone's battery is low, temporarily increase the charging power of the auxiliary drone and extend the charging time, while reducing the non-essential tasks of the auxiliary drone. Method 3, Emergency Energy Allocation: When the auxiliary drone's battery is low, the main drone's backup battery is activated to provide emergency charging, ensuring the completion of the core charging task.

10. The dual unmanned aerial vehicle system for enhancing endurance through in-flight charging according to claim 9, characterized in that, The task planning and energy balance module also includes: The mission completion and energy review module is used to automatically generate an energy utilization report after the mission is completed. This report includes statistics on the mission energy consumption distribution of the main UAV, the charging efficiency curve of the auxiliary UAV, and the number and effect of dynamic adjustments. Then, based on the report data, the parameters of the mission planning model are updated through reinforcement learning to improve the accuracy of energy consumption prediction for the next mission. At the same time, the battery health loss of the two UAVs is calculated. If the SOH drops beyond the preset value, a battery maintenance reminder is triggered.

Images