Intelligent charging and endurance optimization method for unmanned aerial vehicle nest

Abstract

The invention relates to the technical field of unmanned aerial vehicles, and discloses an intelligent charging and endurance optimization method for an unmanned aerial vehicle nest, and the method comprises the steps: collecting flight, battery and environment data through an unmanned aerial vehicle, transmitting the data to a nest, and precisely evaluating the remaining endurance of the unmanned aerial vehicle through the nest in combination with historical tasks and geographic information; performing dynamic task planning based on reinforcement learning, and generating an optimal charging strategy by using a dynamic planning algorithm; by means of a multi-mode sensor and a self-adaptive charging control model, intelligent regulation and control of the charging process are achieved; and through solar auxiliary power supply and multi-machine cooperative charging, the endurance is enhanced. The problems of low charging efficiency, poor endurance management, insufficient system reliability and the like in a traditional method are effectively solved, the utilization rate of nest charging resources is remarkably improved, the service life of a battery is prolonged, the endurance of the unmanned aerial vehicle and the system stability are enhanced, and the method has important significance in promoting efficient application of the unmanned aerial vehicle in multiple fields.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicles, in particular to an intelligent charging and endurance optimization method for an unmanned aerial vehicle nest. BACKGROUND

[0002] In the current situation of rapid development and wide application of unmanned aerial vehicles, the unmanned aerial vehicle nest, as the energy supply station and task scheduling hub in the process of unmanned aerial vehicle task execution, undertakes the important functions of charging, storing and task planning for unmanned aerial vehicles. With the deep penetration of unmanned aerial vehicles in the fields of logistics distribution, environmental monitoring, emergency rescue and security inspection, the operation scene is becoming increasingly complex, and the task demand is becoming more diversified. The dependence of unmanned aerial vehicles on long-time endurance and efficient charging is increasing. The traditional charging and endurance management mode has been difficult to meet the needs of current high-density and high-intensity operation of unmanned aerial vehicles, and therefore an intelligent charging and endurance optimization method is urgently needed to improve the comprehensive performance of the unmanned aerial vehicle nest and ensure the stable and efficient operation of the unmanned aerial vehicle system.

[0003] Currently, the traditional charging and endurance management technology of the unmanned aerial vehicle nest mainly adopts fixed power charging and simple task planning mode. In terms of charging, the unmanned aerial vehicle returns to the nest for fixed power charging when the battery is insufficient, without fully considering the actual situation of the battery state, environmental factors and charging resources of the nest, resulting in low charging efficiency and accelerated aging of the battery due to unreasonable charging, which greatly shortens the service life. In terms of task planning and endurance management, there is a lack of dynamic analysis of task priority, environmental parameters and real-time state of the unmanned aerial vehicle, and it is difficult to adjust the flight path and charging plan according to the actual situation, which easily causes problems such as increased invalid energy consumption of the unmanned aerial vehicle, delayed or even failed tasks. The limitations of these traditional technologies seriously restrict the application range and work efficiency of the unmanned aerial vehicle system, and therefore an intelligent charging and endurance optimization method for the unmanned aerial vehicle nest is proposed. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides an intelligent charging and endurance optimization method for an unmanned aerial vehicle nest, which solves the above problems.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: an intelligent charging and endurance optimization method for an unmanned aerial vehicle nest, comprising the following steps:

[0006] S1 data acquisition and transmission: the unmanned aerial vehicle acquires real-time flight data, battery state data and environmental meteorological data, and sends the data to the nest;

[0007] S2 remaining endurance capability evaluation: the nest receives the data sent by the unmanned aerial vehicle, combines the historical task data and surrounding geographic information data stored by itself, and accurately evaluates the remaining endurance capability of the unmanned aerial vehicle based on a preset multi-source data fusion algorithm;

[0008] S3 Dynamic task planning: a task priority ranking algorithm based on reinforcement learning, combining task urgency, battery state data, environmental meteorological data, surrounding geographic information data, and real-time state of the UAV to generate an optimal task sequence, adjust the UAV flight trajectory, and reduce invalid energy consumption;

[0009] S4 Charging strategy generation: the nest generates a charging strategy for the UAV based on the remaining endurance capability evaluation results, combined with the current charging queue situation and the charging pile state, through a dynamic programming algorithm;

[0010] S5 Adaptive charging regulation: if the UAV needs to be charged, the nest controls the UAV to land at the designated charging pile for charging, and through the multi-modal sensor to build an adaptive charging control model to monitor the battery health state (SOH) and the remaining power (SOC) in real time, combined with the environmental temperature and the load power to dynamically adjust the charging current and voltage;

[0011] S6 Intelligent endurance enhancement: using the built-in solar auxiliary power system and energy storage unit in the nest, combined with the task path planning, through wireless communication between nests to realize power sharing, preferentially allocating high-priority charging positions for low-power UAVs, realizing dynamic energy distribution and multi-UAV cooperative charging;

[0012] S7 Fault prediction and fault-tolerant control: predict battery aging trend and charging system failure through machine learning algorithm, trigger redundant power switching or task redistribution;

[0013] S8 Model optimization update: the cloud management system collects the operation data of multiple nests and the task completion data of UAVs, and uses machine learning algorithm to optimize and update the multi-source data fusion algorithm, dynamic programming algorithm, and adaptive charging control model.

[0014] Preferably, the flight data includes flight speed, flight altitude, flight attitude, and remaining range;

[0015] The battery state data includes the current battery capacity, the number of charge and discharge cycles, and the battery health state;

[0016] The environmental meteorological data includes wind speed, wind direction, air temperature, and air pressure;

[0017] The historical task data includes the flight path and power consumption data of past similar tasks;

[0018] The surrounding geographic information data includes terrain height, obstacle distribution, and no-fly zone.

[0019] Preferably, the multi-source data fusion algorithm in the S2 remaining endurance assessment is specifically: a Bayesian network is used to construct a data fusion model, flight data, battery state data, environmental meteorological data, historical task data, and surrounding geographic information data are taken as input variables, the conditional probability between each variable is calculated to obtain the probability distribution of the remaining endurance of the unmanned aerial vehicle, and then the remaining endurance is determined.

[0020] Preferably, the task priority ranking algorithm based on reinforcement learning in the S3 dynamic task planning adopts a Q-learning algorithm, takes the task completion time, energy consumption cost, and weather influence factor as the state space, takes the task scheduling decision as the action space, and takes the maximization of the task completion value as the objective function for training and optimization.

[0021] Preferably, the process of generating a charging strategy by a dynamic programming algorithm in the S4 charging strategy generation includes: taking the minimization of the task delay time and the charging cost of the unmanned aerial vehicle as the objective function, taking the charging power limit of the charging pile, the charging capacity limit of the unmanned aerial vehicle battery, and the charging queue waiting time as the constraint conditions, and solving the optimal charging strategy by a dynamic programming algorithm;

[0022] The charging strategy includes whether to charge immediately, the charging time, and the charging power.

[0023] Preferably, the S5 adaptive charging regulation specifically includes: an adaptive charging control model is constructed by integrating a multi-modal sensor array of voltage / current sensors, infrared thermal imagers, and pressure sensors, the battery internal resistance, temperature, and swelling state are monitored in real time, and a segmented charging strategy is used to dynamically switch the constant current-constant voltage-pulse charging mode according to the SOC stage;

[0024] The adaptive charging control model is specifically: a relationship model between the battery temperature, the charging current, the charging voltage, the charging efficiency, and the battery life is established, the charging power and the charging time are adjusted by a fuzzy control algorithm according to the real-time monitoring data, so that the charging process reaches the optimal efficiency under the premise of ensuring the safety of the battery;

[0025] The SOC stage is divided into: the remaining power is between 0-20%, the remaining power is between 20-80%, and the remaining power is between 80-100%.

[0026] Preferably, the solar auxiliary power supply system in the S6 intelligent endurance enhancement adopts a flexible gallium arsenide (GaAs) photovoltaic panel and a maximum power point tracking (MPPT) controller, the flexible gallium arsenide (GaAs) photovoltaic panel is laid on the top and side of the nest, and a curved surface fitting design is adopted.

[0027] Preferably, the multi-robot cooperative charging strategy specifically comprises: when multiple robots simultaneously request charging, the nest calculates an optimal charging distribution scheme through a game theory model according to the SOC, SOH, task urgency and subsequent flight tasks of each robot, so as to maximize the overall energy efficiency of the system.

[0028] Preferably, the S7 fault prediction and fault-tolerant control specifically comprises a battery health prediction model based on an LSTM neural network, the input parameters of which include the number of charge-discharge cycles, temperature fluctuation range and internal resistance change rate, and the model is combined with a redundant power switching mechanism, so that when the main charging module fails, the standby energy storage unit or the solar power supply system is automatically enabled.

[0029] Preferably, the process of using the machine learning algorithm to optimize and update the model by the cloud management system comprises: using a reinforcement learning algorithm to adjust the parameters of the multi-source data fusion algorithm, dynamic programming algorithm and adaptive charging control model, taking the task completion efficiency, battery service life and charging cost as the reward function, so as to realize continuous optimization of the model.

[0030] Compared with the prior art, the present application provides an intelligent charging and endurance optimization method for a UAV nest, which has the following beneficial effects:

[0031] Precise endurance evaluation: through deep fusion of multi-source data such as flight state, battery parameters, environmental meteorology, historical tasks and geographic information, a data fusion model is constructed by using a Bayesian network algorithm to fully exploit the potential influence of each factor on the endurance of the UAV. Compared with the traditional single data or simple weighted evaluation method, this method can more accurately quantify the probability distribution of the remaining endurance, providing scientific and reliable data support for charging decision-making, avoiding risks such as task interruption and UAV disconnection caused by endurance misjudgment from the source, and significantly improving the accuracy and reliability of the decision-making.

[0032] Intelligent strategy generation and optimized resource allocation: based on the dynamic programming algorithm, multiple factors such as the remaining endurance of the UAV, the task urgency, the charging queue situation and the charging pile state are comprehensively considered to minimize the task delay time and the charging cost, and an optimal solution model under multiple constraints is constructed. This strategy can realize fine allocation of charging resources, efficiently coordinate the charging time sequence and power of multiple UAVs, significantly improve the utilization efficiency of the nest charging resources, effectively shorten the task execution period, reduce the operating cost, and ensure the efficient and economic operation of the UAV system.

[0033] Adaptive charging regulation, escort battery life: Relying on multi-modal sensor array to collect real-time battery resistance, temperature, expansion state and other key health indicators and remaining power data, build adaptive charging control model. Combined with the segmented charging strategy, according to the different state of charge (SOC) stage, intelligent switching constant current, constant voltage, pulse charging mode, and dynamically adjusting the charging parameters through fuzzy control algorithm. This mechanism ensures the safety and stability of the charging process, maximizes the charging efficiency, effectively inhibits battery polarization, overheating and other loss problems, significantly prolongs the service life of the battery, and reduces the maintenance cost of the equipment.

[0034] Intelligent endurance enhancement: With the built-in solar auxiliary power supply system in the nest, flexible gallium arsenide (GaAs) photovoltaic panels and maximum power point tracking (MPPT) controllers are used to achieve efficient conversion and utilization of solar energy; At the same time, based on the game theory model, a multi-machine cooperative charging strategy is constructed to realize the intelligent sharing and preferential allocation of power among multiple unmanned aerial vehicles. Through the full use of renewable energy and the dynamic optimization and allocation of energy, the endurance of unmanned aerial vehicles is effectively improved, the limitations of traditional energy supply are broken through, and the working radius and working time of unmanned aerial vehicles are greatly expanded, providing a solid guarantee for the application in complex task scenarios.

[0035] Intelligent fault prevention and control, strengthen system resilience: Based on LSTM neural network to build battery health prediction model, through deep analysis of parameters such as charge and discharge cycle number, temperature fluctuation range, internal resistance change rate, realize accurate prediction of battery aging trend and fault risk. Combined with the redundant power switching mechanism, when the main charging module fails, the system can automatically and quickly enable the standby energy storage unit or solar power supply system to ensure the continuity of the charging process. The fault prediction and fault tolerance control mechanism significantly enhances the anti-risk ability of the system, effectively reduces the equipment failure rate, ensures the stable operation of the unmanned aerial vehicle system in complex environment, and improves the overall reliability and stability.

[0036] Continuous intelligent evolution, improve system efficiency: Cloud management system collects multi-nest operation data and unmanned aerial vehicle task completion data, uses machine learning algorithm to iteratively optimize multi-source data fusion, dynamic planning, adaptive charging control and other core models. Based on reinforcement learning algorithm, taking task completion efficiency, battery service life, charging cost, etc. as optimization objectives, drive dynamic adjustment and algorithm upgrade of model parameters. This continuous optimization mechanism enables the system to quickly adapt to changes in task requirements and dynamic evolution of the environment, continuously improves the intelligent charging and endurance optimization performance of the unmanned aerial vehicle nest, and maintains the advanced nature of the technology and the adaptability of the application. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 The figure is a flowchart of the intelligent charging and endurance optimization method of the unmanned aerial vehicle nest. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0039] Please refer to Figure 1 An intelligent charging and endurance optimization method for a UAV nest, comprising the following steps:

[0040] S1 Data acquisition and transmission: the UAV is equipped with multiple sensors to collect real-time flight data, battery status data, and environmental meteorological data. The flight data is obtained through the flight control system of the UAV, including flight speed, flight altitude, flight attitude, and remaining range. The battery status data is provided by the battery management system, including the current battery capacity, the number of charge and discharge cycles, and the battery health status. The environmental meteorological data is collected by means of meteorological sensors, including wind speed, wind direction, air temperature, and air pressure. The UAV sends the collected data to the nest through a wireless communication module, providing a data basis for subsequent evaluation and decision-making.

[0041] S2 Remaining endurance capability evaluation: after the nest receives the data sent by the UAV, it combines the historical task data stored by itself, including the flight path and power consumption data of past similar tasks, and the surrounding geographic information data, including terrain height, obstacle distribution, and no-fly zones. A pre-set multi-source data fusion algorithm is used to accurately evaluate the remaining endurance capability of the UAV. The multi-source data fusion algorithm uses Bayesian network to construct a data fusion model, which can well handle the uncertainty and correlation between data. By calculating the conditional probability between variables, the probability distribution of the remaining endurance capability of the UAV is obtained, so as to more accurately determine the remaining endurance capability. For example, under different wind speed and wind direction conditions, combined with the power consumption data under the same or similar meteorological conditions in the historical task, as well as the flight speed and altitude of the current UAV, the remaining endurance mileage can be more accurately evaluated.

[0042] S3 dynamic task planning: based on task priority, environmental parameters and real-time state of the UAV, dynamically generate multi-objective optimization charging plan. Through the task priority ranking algorithm based on reinforcement learning, the task priority ranking algorithm adopts Q-learning algorithm, takes the task completion time, energy consumption cost and weather influence factor as the state space, takes the task scheduling decision as the action space, and takes the maximization of task completion value as the objective function to train and optimize. The optimal task sequence is generated in combination with task urgency, energy consumption and weather conditions; the dynamic path planning unit is used to adjust the UAV flight trajectory according to real-time wind speed, no-fly zone and obstacle data, and reduce invalid energy consumption. For example, in strong wind weather, the flight height and heading are dynamically adjusted, and the air flow is used to reduce energy consumption; for emergency tasks, the UAV with sufficient power is preferentially scheduled to execute.

[0043] S4 charging strategy generation: the nest generates the charging strategy of the UAV according to the remaining endurance capability evaluation result, considering the current charging queue situation and the charging pile state at the same time. The dynamic programming algorithm takes the minimization of UAV task delay time and charging cost as the objective function, and the charging cost includes the power consumption cost in the charging process and the cost caused by battery loss. Taking the charging power limit of the charging pile, the charging capacity limit of the UAV battery and the charging queue waiting time as the constraint conditions, the optimal charging strategy is solved through step-by-step analysis and optimization of different charging schemes, including whether to charge immediately, charging time and charging power. For example, if there are multiple UAVs waiting for charging, and the current charging pile power is limited, the dynamic programming algorithm will consider the task urgency, remaining endurance capability and other factors of each UAV, reasonably arrange the charging sequence and charging time, and ensure the efficient completion of the overall task.

[0044] S5 Adaptive charging regulation: If the UAV needs to be charged, the nest controls the UAV to land on the designated charging pile for charging. An adaptive charging control model is constructed by integrating a multi-modal sensor array of voltage / current sensors, infrared thermal imagers, and pressure sensors to monitor the state of health (SOH) and the state of charge (SOC) of the battery, including the internal resistance, temperature, and swelling state of the battery. A segmented charging strategy is adopted to dynamically adjust the charging current and voltage in combination with the environmental temperature and load power. The constant current-constant voltage-pulse charging mode is dynamically switched according to the SOC stage. The adaptive charging control model specifically establishes a relationship model between the battery temperature, charging current, charging voltage, and charging efficiency, battery life. Based on real-time monitoring data, the charging power and charging time are adjusted through a fuzzy control algorithm to achieve optimal efficiency in the premise of ensuring battery safety. The SOC stage is divided into: 0-20% of the remaining power, 20-80% of the remaining power, and 80-100% of the remaining power. For example, when the SOC is 0-20%, a large current constant current charging is adopted to quickly supplement the power; when it is 20-80%, it is switched to constant voltage charging to ensure charging efficiency; when it is 80-100%, pulse charging is adopted to reduce battery polarization and prolong battery life.

[0045] S6 Intelligent range enhancement: The built-in solar auxiliary power system and energy storage unit of the nest are used in combination with task path planning to achieve dynamic energy distribution and multi-UAV cooperative charging. When multiple UAVs request charging at the same time, the nest calculates the optimal charging distribution scheme through a game theory model based on the SOC, SOH, task urgency, and subsequent flight tasks of each UAV to maximize the overall energy efficiency of the system. The solar auxiliary power system uses flexible gallium arsenide (GaAs) photovoltaic panels and a maximum power point tracking (MPPT) controller. The flexible gallium arsenide (GaAs) photovoltaic panels are laid on the top and sides of the nest with a curved surface design to improve daytime charging efficiency. The multi-UAV cooperative charging strategy achieves power sharing through wireless communication between nests and prioritizes low-power UAVs for high-priority charging positions. For example, when the sunlight is sufficient, the solar auxiliary power system can provide additional power to the nest, reducing dependence on the power grid. When multiple UAVs request charging at the same time, the system will reasonably allocate charging resources based on the power conditions and task urgency of each UAV.

[0046] S7 Fault Prediction and Fault Tolerant Control: Predict battery aging trend and charging system failure through machine learning algorithm, trigger redundant power switching or task redistribution; Specifically, the battery health prediction model based on LSTM neural network, input parameters include charge-discharge cycle number, temperature fluctuation range and internal resistance change rate, can accurately predict the remaining service life and failure risk of the battery; When the main charging module fails, the redundant power switching mechanism automatically enables the standby energy storage unit or solar power supply system to ensure the continuity of the charging process. For example, when the health status of a certain battery is predicted to decline to a certain extent, the system will arrange the UAV for maintenance or battery replacement in advance; When the main charging module fails, the standby power supply can immediately take over the charging task to avoid affecting the normal operation of the UAV.

[0047] S8 Model Optimization Update: The cloud management system collects the operation data of multiple nests and the task completion data of the UAVs, and uses machine learning algorithms to optimize and update the multi-source data fusion algorithm, dynamic planning algorithm and adaptive charging control model. Reinforcement learning algorithm is adopted, and task completion efficiency, battery service life and charging cost are used as reward function, and parameters of the model are adjusted according to actual operation data. With the continuous accumulation of data and the learning and optimization of algorithm, the performance of each model will be continuously improved, so that the method can better adapt to different application scenarios and changing environmental conditions.

[0048] Example 1: In the emergency rescue scene of sudden mountain fire, a large area of mountain fire breaks out in a certain area, and the command center quickly starts the UAV emergency monitoring scheme and allocates 5 UAVs to perform continuous fire monitoring tasks. At this time, the initial remaining power (SOC) of the 5 UAVs is less than 30%, and they are facing the serious challenge of insufficient power, and the intelligent charging and endurance optimization method of the application plays a key role in this scene, and the specific implementation process is as follows:

[0049] S1 Data Collection and Transmission: After receiving the task instruction, the 5 UAVs take off immediately, and the various sensors carried by the UAVs start working. The flight control system collects flight data such as flight speed, flight altitude, flight attitude, and remaining range in real time; the battery management system synchronously obtains battery state data such as current battery power, charge-discharge cycle number, and battery health status; the weather sensor continuously monitors environmental meteorological data such as wind speed, wind direction, air temperature, and air pressure. These data are transmitted to the nearby emergency rescue UAV nest in a real-time and stable manner through the wireless communication module equipped on the UAV, providing a comprehensive and accurate data basis for subsequent accurate decision-making.

[0050] S2 Remaining endurance assessment: After receiving the data sent by the drone, the nest immediately retrieves the historical task data stored in itself, including the flight path, power consumption data, and other information of similar mountain fire monitoring tasks under similar weather conditions; at the same time, combined with the surrounding geographic information data, such as the terrain height of the mountain fire area, obstacle distribution, and no-fly zone information. Using the pre-set Bayesian network multi-source data fusion algorithm, the remaining endurance of each drone is accurately assessed. By calculating the conditional probability between each data variable, the probability distribution of the remaining endurance of the drone is obtained, so as to accurately determine the time length of each drone that can execute the task under the current power state, and provide a scientific basis for subsequent task planning and charging decision.

[0051] S3 Dynamic task planning: The emergency command center sets the mountain fire monitoring task as the highest priority, and the system trains and optimizes based on the task priority sorting algorithm of reinforcement learning (using Q-learning algorithm), taking the task completion time, energy consumption cost, and weather influence factor as the state space, the task scheduling decision as the action space, and the maximum task completion value as the objective function. Considering factors such as task urgency, distance between current position of the drone and fire site, and expected energy consumption, the optimal task execution sequence is quickly generated. At the same time, the dynamic path planning unit plans a flight trajectory for each drone that can avoid smoke areas, high-temperature zones, and efficiently cover the monitoring range based on real-time weather data (such as wind speed and direction of strong wind) and terrain data (such as mountain fire spread area and obstacle distribution). For example, when strong wind is detected to increase the energy consumption of the drone, the flight height and heading are dynamically adjusted to take advantage of air flow to reduce energy consumption, ensuring that the drone safely and efficiently executes the monitoring task.

[0052] S4 Charging strategy generation: Based on the remaining endurance assessment results, combined with the current charging queue situation (no other drones are charging), and the charging pile state (equipped with 4 charging piles, each with a maximum power of 15kW), the nest generates a charging strategy through a dynamic programming algorithm. Considering the need to ensure uninterrupted execution of the continuous monitoring task of the mountain fire, the nest prioritizes allocating charging resources to 3 drones. Taking the minimization of task delay time and charging cost of the drone as the objective function, and taking the charging power limit of the charging pile, the charging capacity limit of the drone battery, and the charging queue waiting time as the constraint conditions, it is calculated that: charging the 3 drones immediately, the charging time is set to 15 minutes, and the charging power is 12kW; the other 2 drones use the built-in energy storage unit of the nest for power supply, continue to execute the monitoring task, and realize the dynamic and reasonable allocation of energy.

[0053] S5 Adaptive charging regulation: After landing on the designated charging pile, the 3 drones assigned for charging start charging. The nest uses a multi-modal sensor array integrating voltage / current sensors, infrared thermal imagers, and pressure sensors to monitor key data such as battery internal resistance, temperature, and swelling state in real-time. Based on the segmented charging strategy, when the battery SOC is between 0-20%, a large current constant current charging of 12kW is used to quickly increase the power; when the SOC reaches 20-80%, the constant voltage charging mode is switched to ensure charging efficiency; when the SOC is between 80-100%, the pulse charging mode is enabled to reduce battery polarization. During the charging process, if the infrared thermal imager detects that the battery temperature is too high, the adaptive charging control model adjusts the charging power and time in time through the fuzzy control algorithm, such as reducing the charging power to 10kW, to ensure safe and efficient charging process and prolong the service life of the battery.

[0054] S6 Intelligent range enhancement: The solar auxiliary power system built-in the nest is in full operation. The flexible gallium arsenide (GaAs) photovoltaic panels are laid on the top and sides of the nest, which efficiently convert solar energy into electrical energy through curved surface design and maximum power point tracking (MPPT) controller, providing additional energy support for the charging process and reducing dependence on traditional power grids. At the same time, the multi-drone cooperative charging strategy plays a role, when more drones may need to be charged in the future, the nest can calculate the optimal charging allocation scheme through game theory model according to the SOC, SOH, task urgency and subsequent flight tasks of each drone, realize power sharing through wireless communication between nests, and preferentially allocate high-priority charging positions to low-power drones to maximize the overall energy efficiency of the system.

[0055] S7 Fault prediction and fault-tolerant control: The battery health prediction model based on LSTM neural network continuously monitors the batteries of 5 drones, inputting parameters such as charge / discharge cycle number, temperature fluctuation range, and internal resistance change rate. During task execution, the model detects that the internal resistance of one of the drones has abnormally increased, predicting a battery failure risk, and immediately triggers a fault warning. The system then starts the task redistribution mechanism, redistributes the monitoring tasks of the drone to other drones with sufficient power and good state, and arranges the faulty drone to return to the nest for repair or battery replacement, ensuring the entire forest fire monitoring task is not affected and ensuring continuous and stable data transmission.

[0056] S8 model optimization update: During the execution of this emergency rescue task, the cloud management system collects real-time operation data of the emergency rescue drone nest, including the execution of the charging strategy, the use state of the charging pile, and the task completion data of the five drones, such as flight trajectory, task execution time, etc. Using machine learning algorithms, reinforcement learning algorithms are used to adjust the parameters of multi-source data fusion algorithms, dynamic programming algorithms, and adaptive charging control models with task completion efficiency (monitoring coverage and time required by this task), battery life (evaluated through real-time monitoring of battery status data), and charging cost (including electricity consumption cost and potential battery loss cost) as the reward function. With the continuous accumulation of data and the learning and optimization of algorithms, these models can provide more accurate and efficient intelligent charging and endurance optimization solutions in future similar emergency rescue scenarios.

[0057] Finally, with the support of the method of the present application, the five drones successfully achieved continuous monitoring of the forest fire area through intelligent charging and endurance optimization, with a task completion rate of 100% and no task interruption due to insufficient power or equipment failure, providing stable and reliable real-time data support for forest fire rescue command, fully demonstrating the effectiveness and practicality of the method in emergency rescue scenarios

[0058] Although embodiments of the present application have been shown and described, it will be understood by those having ordinary skill in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent charging and endurance optimization method for a drone nest, characterized in that, The method comprises the following steps: S1 data acquisition and transmission: the unmanned aerial vehicle collects flight data, battery status data and environmental meteorological data in real time, and sends the data to the nest; S2 remaining endurance assessment: the nest receives the data sent by the unmanned aerial vehicle, combines historical task data and surrounding geographic information data stored by itself, and accurately assesses the remaining endurance of the unmanned aerial vehicle based on a preset multi-source data fusion algorithm; S3 dynamic task planning: based on a task priority sorting algorithm of reinforcement learning, an optimal task sequence is generated by combining task urgency, battery status data, environmental meteorological data, surrounding geographic information data and real-time state of the unmanned aerial vehicle, and the flight trajectory of the unmanned aerial vehicle is adjusted to reduce invalid energy consumption; S4 charging strategy generation: the nest generates a charging strategy for the unmanned aerial vehicle according to the remaining endurance assessment result, in combination with the current charging queue and charging pile state, through a dynamic programming algorithm; S5 adaptive charging regulation: if the unmanned aerial vehicle needs to be charged, the nest controls the unmanned aerial vehicle to land at a specified charging pile for charging, and an adaptive charging control model is constructed by using a multi-modal sensor to monitor the state of health (SOH) and the remaining capacity (SOC) of the battery in real time, and the charging current and voltage are dynamically adjusted in combination with the environmental temperature and load power; S6 intelligent endurance enhancement: the solar auxiliary power supply system and the energy storage unit built-in the nest are used to realize power sharing through wireless communication between nests, high-priority charging positions are preferentially allocated to unmanned aerial vehicles with low power, and energy dynamic allocation and multi-machine cooperative charging are realized; S7 fault prediction and fault-tolerant control: the battery aging trend and the charging system fault are predicted through a machine learning algorithm, and redundant power switching or task redistribution is triggered; S8 model optimization and update: the cloud management system collects operation data of multiple nests and task completion data of unmanned aerial vehicles, and optimizes and updates the multi-source data fusion algorithm, dynamic programming algorithm and adaptive charging control model by using a machine learning algorithm. 2.The method of claim 1, wherein, The flight data includes flight speed, flight height, flight attitude and remaining range; The battery status data includes current battery capacity, charge and discharge cycle number and battery health status; The environmental meteorological data includes wind speed, wind direction, air temperature and air pressure; The historical task data includes flight path and power consumption data of past same or similar tasks; The surrounding geographic information data includes terrain height, obstacle distribution and no-fly zone. 3.The method of claim 2, wherein, The multi-source data fusion algorithm in the S2 remaining endurance assessment is specifically: a Bayesian network is used to construct a data fusion model, flight data, battery status data, environmental meteorological data, historical task data and surrounding geographic information data are used as input variables, the conditional probability between each variable is calculated to obtain the probability distribution of the remaining endurance of the unmanned aerial vehicle, and then the remaining endurance is determined. 4.The method of claim 3, wherein, The task priority sorting algorithm based on reinforcement learning in the S3 dynamic task planning adopts a Q-learning algorithm, takes task completion time, energy consumption cost and weather influence factor as a state space, takes task scheduling decision as an action space, and trains and optimizes a target function to maximize task completion value. 5.The method of claim 1, wherein, The process of generating the charging strategy by the dynamic programming algorithm in the S4 charging strategy generation includes: taking minimization of the unmanned aerial vehicle task delay time and the charging cost as an objective function, taking the charging power limit of the charging pile, the charging capacity limit of the unmanned aerial vehicle battery and the charging queue waiting time as constraint conditions, and obtaining an optimal charging strategy by a dynamic programming algorithm. The charging strategy includes whether to charge immediately, a charging time length and a charging power.

6. The intelligent charging and endurance optimization method for a UAV nest according to claim 5, characterized in that, The S5 adaptive charging regulation specifically includes: constructing an adaptive charging control model by a multi-modal sensor array integrating a voltage / current sensor, an infrared thermal imager and a pressure sensor, monitoring a battery internal resistance, a temperature and an inflation state in real time, and a segmented charging strategy, dynamically switching a constant current-constant voltage-pulse charging mode according to an SOC stage; The adaptive charging control model specifically includes: establishing a relationship model among a battery temperature, a charging current, a charging voltage and a charging efficiency and a battery life, adjusting a charging power and a charging time according to real-time monitoring data by a fuzzy control algorithm, and making the charging process reach an optimal efficiency under the premise of ensuring the safety of the battery; The SOC stage is divided into: a remaining power of 0-20%, a remaining power of 20-80% and a remaining power of 80-100%.

7. The intelligent charging and endurance optimization method for a drone nest according to claim 1, characterized in that, The solar auxiliary power supply system in the S6 intelligent endurance enhancement adopts a flexible gallium arsenide (GaAs) photovoltaic panel and a maximum power point tracking (MPPT) controller, and the flexible gallium arsenide (GaAs) photovoltaic panel is laid on the top and side of the nest in a curved surface fitting design. 8.The method of claim 6, wherein, The multi-machine cooperative charging strategy specifically includes: when multiple unmanned aerial vehicles request charging at the same time, the nest calculates an optimal charging distribution scheme according to the SOC, SOH, task urgency and subsequent flight tasks of each unmanned aerial vehicle by a game theory model, and realizes maximization of the overall energy efficiency of the system. 9.The method of claim 6, wherein, The S7 fault prediction and fault-tolerant control specifically includes a battery health prediction model based on an LSTM neural network, and the input includes parameters such as the number of charge and discharge cycles, the temperature fluctuation range and the internal resistance change rate, and cooperates with a redundant power supply switching mechanism to automatically enable a backup energy storage unit or a solar power supply system when the main charging module fails. 10.The method of claim 1-9, wherein, The process of optimizing and updating the model by the cloud management system using a machine learning algorithm includes: adopting a reinforcement learning algorithm to adjust parameters of a multi-source data fusion algorithm, a dynamic programming algorithm and an adaptive charging control model, and realize continuous optimization of the model, taking the task completion efficiency, the battery service life and the charging cost as a reward function.

Images