Large fixed-wing unmanned aerial vehicle operation management system

Abstract

The invention discloses a large fixed-wing unmanned aerial vehicle operation management system, and the system comprises a flight situation intelligent sensing module which is used for obtaining and processing the flight state data and external environment data of an unmanned aerial vehicle in real time through the fusion processing of a multi-source sensor; the self-adaptive flight path planning and risk avoiding module is used for carrying out dynamic risk assessment and generating or optimizing a flight path based on the output of the flight situation intelligent sensing module; the flight safety boundary dynamic calculation module is used for calculating a dynamic safety envelope evolved along with time according to the real-time state and the mission profile of the unmanned aerial vehicle; the autonomous intelligent decision and instruction generation module is used for synthesizing the dynamic safety envelope and the optimized flight path to generate a flight control instruction or alarm information; and the cloud collaborative management platform is in communication connection with the autonomous intelligent decision and instruction generation module and is used for collaborative management of multiple unmanned aerial vehicles. According to the invention, operation management of the fixed-wing unmanned aerial vehicle can be carried out more intelligently.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) management systems, and more specifically to a large fixed-wing UAV operation and management system. Background Technology

[0002] Compared to small or rotary-wing drones, large fixed-wing drones possess core advantages such as longer endurance, higher payload capacity, faster flight speed, and larger operational radius, making them widely used in complex scenarios such as long-range maritime monitoring, material transportation, and large-scale operations in urban and suburban areas. However, their technical complexity is significantly higher: on the one hand, precise control of flight status is required to avoid fuselage damage caused by aerodynamic parameter imbalances or excessive structural loads; on the other hand, they are more sensitive to external environment and energy consumption management. Any perception deviation, insufficient energy consumption, or untimely risk response could lead to more serious flight accidents than with small drones, thus imposing stringent requirements on the safety, accuracy, and dynamic adaptability of operational management. The current operation and management system for large fixed-wing UAVs has many shortcomings, including insufficient specific perception accuracy, one-sided and rigid flight path planning, and fixed and one-sided safety boundaries. The current operation and management system for large fixed-wing UAVs is difficult to adapt to the needs of complex flight scenarios. Therefore, a new operation and management system for large fixed-wing UAVs is proposed. Summary of the Invention

[0003] The present invention solves the above-mentioned technical problems through the following technical solutions, the present invention comprising: The flight status intelligent perception module is used to acquire and process the UAV's flight status data and external environment data in real time through multi-source sensor fusion processing. The adaptive trajectory planning and risk avoidance module is used to perform dynamic risk assessment and generate or optimize flight trajectories based on the output of the flight situation intelligent perception module. The flight safety boundary dynamic calculation module is used to calculate the dynamic safety envelope that evolves over time based on the real-time status of the UAV and the mission profile. The autonomous intelligent decision-making and command generation module is used to integrate the dynamic safety envelope and the optimized flight trajectory to generate flight control commands or alarm information; The cloud-based collaborative management platform communicates with the autonomous intelligent decision-making and instruction generation module and is used for collaborative management of multiple UAVs.

[0004] Furthermore, the multi-source sensor fusion processing in the intelligent flight situation perception module specifically includes: A three-dimensional situation field is constructed with UAVs at its core. The three-dimensional situation field is formed by integrating data from airborne inertial navigation systems, global positioning systems, weather radar, and air traffic control transponders. The fusion process employs an improved weighted distributed fusion algorithm, in which a signal integrity factor is introduced to weight the Global Positioning System data, and a meteorological unit threat level factor is introduced to weight the meteorological radar data. Finally, by calculating the fusion weight of each data source in the fusion center, high-confidence comprehensive UAV situational information is obtained; The calculation process for the fusion weight is as follows: First, the reliability index of each data source at the current moment is evaluated. The reliability index is obtained by exponentially smoothing the historical error sequence of the data source. The exponential smoothing process uses a single exponential smoothing method, and the smoothing coefficient is dynamically adjusted based on the characteristics of the data source. Then, the reliability index of each data source is normalized to obtain the fusion weight of each data source in the fusion center. The calculation formula is as follows: ; Among them, w i R represents the fusion weight of the i-th data source. i R represents the reliability index of the i-th data source after exponential smoothing, where n represents the total number of data sources participating in the fusion. j This represents the reliability index of the j-th data source after exponential smoothing.

[0005] Furthermore, the dynamic risk assessment in the adaptive trajectory planning and risk avoidance module specifically includes: Based on the three-dimensional situation field, identify static and dynamic obstacles ahead of the flight path; For dynamic obstacles, predict their trajectory and perform collision detection with the machine's predicted trajectory; The risk quantification value of the collision detection is obtained by calculating the total collision probability index between the machine and the dynamic obstacle; The calculation process of the total conflict probability index is as follows: First, based on the time series of the predicted trajectories of the two, the relative distance and approach rate at each future time point are calculated; Then, the relative distance and approach rate are input into a trained risk probability function to calculate the instantaneous conflict probability at that moment, where the instantaneous conflict probability function is defined as the product of a negative exponential function of the relative distance and a positive proportional function of the approach rate. Finally, by integrating the instantaneous conflict probabilities at all future time points, we obtain the total conflict probability index for a future period starting from the current time. The calculation formula is as follows: ; Among them, P c Let dt represent the total probability of conflict, t0 represent the current time, T represent the time window for future prediction, dt represent the relative distance at time t, and v represent the total probability of conflict. app t represents the approach rate at time t. Let α represent the instantaneous conflict probability function based on relative distance and approach rate, where α and β are positive coefficients obtained through training with historical conflict data.

[0006] Furthermore, the process by which the adaptive trajectory planning and risk avoidance module generates or optimizes the flight trajectory is as follows: The total conflict probability index, flight path length, and energy consumption are used as multi-objective optimization parameters for the comprehensive cost function; The comprehensive cost function is solved using a constrained optimization algorithm, whereby the constraints include performance constraints of the UAV and airspace rule constraints. After solving, one or more alternative optimized trajectories are generated, and the comprehensive cost function value of each alternative optimized trajectories is lower than a preset risk threshold.

[0007] Furthermore, the process by which the flight safety boundary dynamic calculation module calculates the dynamic safety envelope is as follows: The dynamic safety envelope is jointly defined by the flight state safety boundary and the mission energy consumption safety boundary; The flight state safety boundary is calculated by analyzing the real-time aerodynamic parameters, structural loads, and real-time wind field data obtained by the flight state intelligent perception module of the UAV. It represents the attitude, overload, and airspeed range in which the UAV can fly safely under the current state. The mission energy consumption safety boundary is calculated jointly by the UAV's remaining battery power, remaining mission range, and real-time energy consumption rate. The calculation process is as follows: First, the average energy consumption rate of future flight segments is predicted based on historical flight data and current flight status; then, the theoretical maximum remaining range is calculated based on the remaining battery power and the predicted average energy consumption rate; finally, the theoretical maximum remaining range is compared with the mission's remaining range, and combined with the preset safety margin, the battery power warning value at which the UAV must start returning to base or making an emergency landing is dynamically calculated.

[0008] Furthermore, the calculation process for the theoretical maximum remaining range in the mission energy consumption safety boundary is as follows: First, based on the UAV's flight status data over a fixed time window, including airspeed, rate of climb, and motor power, its average energy consumption rate is calculated. ; Then, using the drone's current remaining battery power... and the average energy consumption rate Estimate the theoretical maximum remaining flight time T max ; Finally, considering the ground speed V of the UAV in the current mission segment... g Calculate the theoretical maximum remaining range S max The calculation formula is as follows: ; The battery warning value E that necessitates initiating a return trip or forced landing alert Calculated using the following formula: ; Where k is a safety factor greater than 1, S return This returns the distance to the nearest available landing point from the current location.

[0009] Furthermore, the workflow of the autonomous intelligent decision-making and instruction generation module is as follows: The system receives dynamic safety envelope data from the flight safety boundary dynamic calculation module in real time. The dynamic safety envelope data includes flight state safety boundary and mission energy consumption safety boundary, as well as optimized trajectory from the adaptive trajectory planning and risk avoidance module. The optimized trajectory is compared with the dynamic safety envelope to determine whether the optimized trajectory is completely within the range defined by the flight state safety boundary and the mission energy consumption safety boundary. If the optimized trajectory is entirely within the dynamic safety envelope, the optimized trajectory is calculated into specific flight control commands and sent to the flight control system. If any part of the optimized trajectory exceeds the dynamic safety envelope, an alarm is triggered, and the alternative optimized trajectory that is completely within the dynamic safety envelope and has the lowest comprehensive cost function value is selected as the execution plan.

[0010] Furthermore, the cloud-based collaborative management platform is used to receive and record the flight decision data of a single drone, and to perform group collaborative management when multiple drones are operating in the same airspace; Group collaborative management includes: based on the optimized flight paths and dynamic safety envelopes reported by each UAV, performing conflict resolution calculations in the cloud, assigning passage sequences with time differences or adjusting horizontal and vertical intervals for UAV groups with potential conflicts, generating group collaborative flight commands, and issuing them to the corresponding UAVs.

[0011] Compared with existing technologies, this invention has the following advantages: This large fixed-wing UAV operation and management system constructs a three-dimensional situational field through multi-source sensor fusion. Combined with an improved weighted distributed fusion algorithm, it incorporates factors such as signal integrity and weather threat level to obtain high-confidence comprehensive situational information. The fusion weights are dynamically adjusted based on the real-time reliability of the data source, reducing the impact of single sensor errors and ensuring data accuracy. Dynamic risk assessment can identify static / dynamic obstacles and accurately quantify collision risks by predicting trajectories and calculating the total conflict probability. Path planning optimizes multiple objectives such as conflict probability, flight length, and energy consumption. Combined with UAV performance and airspace rule constraints, it generates low-risk, high-efficiency alternative paths. The dynamic safety envelope covers both flight status and mission energy consumption boundaries, adapting to real-time flight conditions and mission requirements. Dynamically calculating the power warning value and combining it with safety margins avoids insufficient energy consumption, provides early warning of return or emergency landing opportunities, reduces the risk of loss of control, and compares and optimizes the path and safety envelope in real time, automatically generating control commands or triggering alarms. It can quickly adapt to flight scenarios without manual intervention. Prioritize safe and lowest-cost alternative flight paths to enhance emergency response capabilities in extreme situations. The cloud platform coordinates flight data from multiple UAVs, calculates and assigns passage sequences and adjusts intervals through conflict resolution to avoid airspace conflicts, improve the coordination of multiple UAV operations in the same airspace, and enhance airspace utilization and overall operational efficiency. Attached Figure Description

[0012] Figure 1 This is a structural block diagram of the present invention. Detailed Implementation

[0013] The embodiments of the present invention are described in detail below. These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and specific operation processes. However, the scope of protection of the present invention is not limited to the following embodiments.

[0014] like Figure 1 As shown, this embodiment provides a technical solution: a large fixed-wing unmanned aerial vehicle (UAV) operation and management system, comprising: The flight status intelligent perception module is used to acquire and process the UAV's flight status data and external environment data in real time through multi-source sensor fusion processing. The adaptive trajectory planning and risk avoidance module is used to perform dynamic risk assessment and generate or optimize flight trajectories based on the output of the flight situation intelligent perception module. The flight safety boundary dynamic calculation module is used to calculate the dynamic safety envelope that evolves over time based on the real-time status of the UAV and the mission profile. The autonomous intelligent decision-making and command generation module is used to integrate the dynamic safety envelope and the optimized flight trajectory to generate flight control commands or alarm information; The cloud-based collaborative management platform communicates with the autonomous intelligent decision-making and instruction generation module and is used for collaborative management of multiple UAVs.

[0015] The multi-source sensor fusion processing in the intelligent flight situation perception module specifically involves: A three-dimensional situation field is constructed with UAVs at its core. The three-dimensional situation field is formed by integrating data from airborne inertial navigation systems, global positioning systems, weather radar, and air traffic control transponders. The fusion process employs an improved weighted distributed fusion algorithm, in which a signal integrity factor is introduced to weight the Global Positioning System data, and a meteorological unit threat level factor is introduced to weight the meteorological radar data. Finally, by calculating the fusion weight of each data source in the fusion center, high-confidence comprehensive UAV situational information is obtained; The calculation process for the fusion weight is as follows: First, the reliability index of each data source at the current moment is evaluated. The reliability index is obtained by exponentially smoothing the historical error sequence of the data source. The exponential smoothing process uses a single exponential smoothing method, and the smoothing coefficient is dynamically adjusted based on the characteristics of the data source. Then, the reliability index of each data source is normalized to obtain the fusion weight of each data source in the fusion center. The calculation formula is as follows: ; Among them, w i R represents the fusion weight of the i-th data source. i R represents the reliability index of the i-th data source after exponential smoothing, where n represents the total number of data sources participating in the fusion. j This represents the reliability index of the j-th data source after exponential smoothing. By fusing multi-source sensors to construct a three-dimensional situational awareness field centered on the UAV, and combining an improved weighted distributed fusion algorithm, and normalizing the fusion weights based on the real-time reliability of the data sources, this approach achieves comprehensive data coverage of both the UAV's own flight status and the external environment, avoiding the limitations of information from a single data source. Furthermore, it dynamically adapts weights according to data characteristics and real-time reliability, prioritizing high-value data with a greater impact on flight safety, reducing the influence of single sensor errors and environmental interference on the perception results. Ultimately, it outputs high-confidence comprehensive UAV situational awareness information, providing accurate and stable data support for downstream modules such as adaptive trajectory planning and dynamic risk assessment.

[0016] When drones perform material transport missions in hilly areas, information gaps are likely to occur due to a single data source: if relying solely on the Global Positioning System (GPS), the hills will block the GPS signal, making it impossible to obtain location data stably; if relying solely on the airborne inertial navigation system, although it can provide attitude (such as pitch angle and roll angle) and acceleration data in the short term, errors will accumulate over long-term flight; if relying solely on weather radar, it is impossible to know the dynamics of other aircraft in the vicinity.

[0017] The solution presented in this case, which integrates data from airborne inertial navigation systems, GPS, weather radar, and air traffic control transponders to construct a three-dimensional situational field, can comprehensively present multi-dimensional information: For example, at a certain moment, the system displays "The drone's current pitch angle is 2.5°, roll angle is 0°, and altitude is 650 meters (inertial navigation + GPS fusion). There is a crosswind with a wind speed of 12 m / s 1.8 kilometers ahead (weather radar). There is a light drone flying southwest 1.2 kilometers to the left front, with a flight altitude of 620 meters (air traffic control transponder)", which completely solves the problem of "unclear self-status and unclear environmental dynamics" caused by a single data source, and provides a comprehensive data foundation for subsequent track adjustments.

[0018] When a drone performs an inspection mission in the airspace within 5 kilometers of an airport, the complex electromagnetic environment at the airport easily interferes with GPS signals. In this situation, the GPS "signal integrity factor" drops from 0.9 in normal flight scenarios to 0.3. The improved weighted distributed fusion algorithm automatically reduces the fusion weight of GPS data based on this factor. Simultaneously, weather radar detects a small area of ​​severe convective weather (wind speeds reaching 18 m / s) east of the airport runway. The "meteorological unit threat level factor" for this area rises from the usual 0.2 to 0.9, and the algorithm synchronously increases the fusion weight of weather radar data. Without such dedicated weighting factors, the system might still accept the interfered GPS data with high weight, leading to location errors or ignoring high-threat weather data, increasing flight risks. However, the solution in this case prioritizes warnings of severe convective weather risks based on highly reliable weather radar data and supplements location information with inertial navigation data, significantly improving perception accuracy in special scenarios.

[0019] Assuming there are three data sources involved in the data fusion: Data Source 1 (GPS), Data Source 2 (Airborne Inertial Navigation System), and Data Source 3 (Weather Radar), the specific weight calculation process is as follows: First, the reliability indices of each data source are obtained through exponential smoothing: During normal flight, the GPS signal is stable, and its historical error sequence is processed by exponential smoothing (the smoothing coefficient is set to 0.6 based on the characteristic of GPS being "susceptible to obstruction and interference"), resulting in a reliability index R1=0.8; the inertial navigation system has no external interference and its historical error fluctuations are small, so the smoothing coefficient is set to 0.4, resulting in a reliability index R2=0.9; the weather radar has no complex weather within its detection range, so the smoothing coefficient is set to 0.5, resulting in a reliability index R3=0.7.

[0020] The fusion weights are calculated based on the normalization formula used in this case: ; Where n=3 (the total number of data sources participating in the fusion), substituting the data, we get: ; When the drone flew into the valley, GPS errors increased due to obstruction. After exponential smoothing, R1 decreased to 0.4, and the weights were recalculated. ; By dynamically adjusting the weights, the system automatically reduces the impact of interfered GPS data and increases the weight of inertial navigation data, ensuring that the fusion results are always based on highly reliable data sources and avoiding perception bias caused by fixed weights.

[0021] The dynamic risk assessment in the adaptive trajectory planning and risk avoidance module specifically includes: Based on the three-dimensional situation field, identify static and dynamic obstacles ahead of the flight path; For dynamic obstacles, predict their trajectory and perform collision detection with the machine's predicted trajectory; The risk quantification value of the collision detection is obtained by calculating the total collision probability index between the machine and the dynamic obstacle; The calculation process of the total conflict probability index is as follows: First, based on the time series of the predicted trajectories of the two, the relative distance and approach rate at each future time point are calculated; Then, the relative distance and approach rate are input into a trained risk probability function to calculate the instantaneous conflict probability at that moment, where the instantaneous conflict probability function is defined as the product of a negative exponential function of the relative distance and a positive proportional function of the approach rate. Finally, by integrating the instantaneous conflict probabilities at all future time points, we obtain the total conflict probability index for a future period starting from the current time. The calculation formula is as follows: ; Among them, P c Let dt represent the total probability of conflict, t0 represent the current time, T represent the time window for future prediction, dt represent the relative distance at time t, and v represent the total probability of conflict.app t represents the approach rate at time t. This represents the instantaneous conflict probability function based on relative distance and approach rate, where α and β are positive coefficients obtained through training with historical conflict data; By accurately identifying static and dynamic obstacles ahead of the flight path based on a three-dimensional situational field, the risk of missing a single type of obstacle is avoided. Flight trajectory prediction is performed for dynamic obstacles, and conflict detection is carried out with the aircraft's trajectory to solve the problem of lag in traditional static assessments that cannot cope with moving targets. At the same time, through the quantitative method of "calculating the instantaneous conflict probability - integrating to obtain the total conflict probability" (the instantaneous conflict probability combines the negative exponential characteristic of relative distance and the positive proportional characteristic of approach rate), the conflict risk is transformed from a qualitative judgment into a quantitative indicator, providing a clear and comparable risk basis for subsequent trajectory optimization, significantly improving the scientificity, foresight, and operability of risk assessment, and reducing the probability of decision-making errors caused by risk ambiguity.

[0022] For example, when a drone is conducting aerial photography in suburban airspace (flying at an altitude of 500 meters and a speed of 40 m / s), based on the solution in this case, the system can simultaneously identify two types of obstacles through a three-dimensional situational awareness field: one is a high-voltage transmission tower located 1.5 kilometers directly ahead of the flight path at an altitude of 480 meters (a static obstacle, with its position coordinates calibrated by multi-source sensing data with an error of ≤5 meters); the other is a light general aviation aircraft located 2 kilometers to the left front at an altitude of 520 meters (a dynamic obstacle, whose motion parameters of 30° heading and 60 m / s speed are acquired in real time through air traffic control transponder and radar data capture). If this solution is not adopted, relying solely on a single visual sensor may result in the failure to detect the high-voltage tower due to tree obstruction, or the neglect of general aviation aircraft due to the lack of specific identification of dynamic targets. However, the multi-type obstacle identification capability of this solution can simultaneously identify both static and dynamic risk sources, laying the foundation for subsequent risk assessment.

[0023] This system can predict dynamic obstacle trajectories and detect collisions, allowing for early identification of potential risks. The system performs trajectory prediction for the identified light general aviation aircraft (dynamic obstacle): Based on the aircraft's current heading of 30° and speed of 60m / s, combined with the airspace flight rules (general aviation aircraft in this area must maintain straight flight), the system predicts its trajectory for the next 60 seconds as "flying at a constant speed along a heading of 30°, with position coordinates every 10 seconds as (1800m, 1039m) → (1200m, 693m) → (600m, 346m)" (establishing a coordinate system with the UAV's current position as the origin); at the same time, the system predicts its own trajectory for the next 60 seconds as "flying at a constant speed along a heading of 0°, with position coordinates every 10 seconds as (400m, 0) → (800m, 0) → (1200m, 0)". Track comparison revealed that in the next 40 seconds, the horizontal distance between the two at the positions "(1200m,0)" and "(1200m,693m)" is only 693 meters, and the height difference is 20 meters (less than the safety interval of 100 meters), indicating a potential conflict. This conflict detection 40 seconds in advance avoids the passive response of the traditional "detection is proximity" and allows sufficient time for track adjustment.

[0024] Based on the suburban aerial photography scenario shown, the system in this case calculates the total probability of conflict between the aircraft and the light general aviation aircraft using the following method: Parameter settings: Current time t0 = 0 (unit: s), future prediction time window T = 40s (i.e., prediction up to t = 40s); based on the predicted trajectories, the formula for the relative distance d(t) changing with time is: ; (After simplification) Due to trajectory correction, the simplified measured value is used. (Unit: m); Approach speed v app t = 25 m / s (constant, calculated from the relative motion of the two); The instantaneous conflict probability function coefficients are α=0.002 and β=0.003, trained using historical conflict data.

[0025] Instantaneous conflict probability function: Substituting the parameters, we get: ; Total conflict probability calculation: Integrating over the interval [0, 40], we get: ; Simplify the calculation (let) ,but , (When t=0, u=-3; when t=40, u=0): ; Risk decision-making application: The system presets risk thresholds. Since the calculated Pc=0.633>0.5, it is determined that the current trajectory has a high risk of conflict, and the adaptive trajectory planning module is immediately triggered to optimize the trajectory (such as shifting to the right by 100 meters). If this quantitative scheme is not adopted, relying solely on the qualitative judgment of "potential conflict" may delay the adjustment time. The quantitative Pc value makes risk decision-making more accurate and executable.

[0026] The process by which the adaptive trajectory planning and risk avoidance module generates or optimizes the flight trajectory is as follows: The total conflict probability index, flight path length, and energy consumption are used as multi-objective optimization parameters for the comprehensive cost function; The comprehensive cost function is solved using a constrained optimization algorithm, whereby the constraints include performance constraints of the UAV and airspace rule constraints. After solving, one or more alternative optimized trajectories are generated, and the comprehensive cost function value of each alternative optimized trajectories is lower than a preset risk threshold. By optimizing parameters with total conflict probability, track length, and energy consumption as the comprehensive cost function, a triple balance between track planning safety, efficiency, and energy consumption is achieved, avoiding the loss of one aspect due to a single objective. At the same time, by combining UAV performance constraints and airspace rule constraints, the generated track is ensured to be practically executable. Furthermore, it can output multiple alternative optimized tracks below the risk threshold, providing flexible selection space for subsequent autonomous decision-making, and significantly improving the scientific nature, practicality, and emergency adaptability of track planning.

[0027] Taking a drone's mission of transporting supplies from warehouse A (coordinates: 30°15′N, 120°20′E) to community B (coordinates: 30°18′N, 120°25′E) as an example, the mission requires minimizing time and energy consumption while ensuring safety. If only the shortest flight path is considered, the planned direct flight path (length L1=8km) is the shortest, but it requires traversing an area with frequent bird activity, increasing the overall probability of conflict. (Higher than the preset threshold of 0.5), and requires frequent altitude adjustments to avoid birds during flight, energy consumption E1 = 2.5 kWh; if the sole objective is "low conflict", the planned detour path ( While safe, the current approach results in a flight path length of L2 = 12km, energy consumption of E2 = 3.8kWh, and a 40% increase in travel time. However, the multi-objective optimization scheme in this case comprehensively considers all three factors: the final optimized flight path (L3 = 9.5km)... (E3=2.8kWh) It satisfies the requirement that the probability of conflict is below the threshold, shortens the flight path by 2.5km and reduces energy consumption by 1.0kWh compared to a purely safety-oriented flight path, and is safer than a purely efficiency-oriented flight path, truly achieving a triple balance and avoiding the extreme drawbacks of a single objective.

[0028] In scenarios such as material transportation, the optimization algorithm in this case needs to simultaneously satisfy two types of constraints: one is the performance constraint of the drone (maximum climb rate H of the drone). max =3m / s, maximum level flight speed V max =50m / s); secondly, airspace rule constraints (there is an airport no-fly zone with a radius of 1km between warehouse A and community B, with an upper limit of 400m at the boundary of the no-fly zone, and the mission requires the flight altitude to be maintained between 450-500m). Without performance constraints, the algorithm may plan a trajectory that "climbs from 450m to 600m within 10 seconds" (climb rate 15m / s > 3m / s), exceeding the performance limit of the drone and causing loss of control; without airspace rule constraints, the algorithm may generate a trajectory that brushes past the no-fly zone (altitude 420m < 450m), violating airspace management regulations. In this case, the constrained optimization algorithm takes "climb rate ≤ 3m / s" and "altitude ≥ 450m and avoidance of the no-fly zone within 1km" as hard conditions when solving the problem. The final generated flight path is "take-off from warehouse A at an altitude of 450m, climb to 500m at a climb rate of 2m / s, fly around the outside of the no-fly zone for 1.2km, and fly at a speed of 45m / s < 50m / s throughout the flight path". This fully complies with the performance and airspace rules, ensuring that the flight path can be actually executed and avoiding the problem of "theoretically feasible but not actually flyable".

[0029] Based on the transportation scenario, the optimization algorithm in this case will generate three alternative optimized trajectories. A comprehensive cost function needs to be constructed for quantitative evaluation (with weights set). (Total conflict probability weight) (track length weight) (Energy-consuming masses) (Risk threshold) (Baseline flight length) (Baseline energy consumption), the comprehensive cost function formula is: ; Normalizing the conflict probability, the smaller the J value, the better the overall trajectory performance, and it must also satisfy... ; E represents real-time energy consumption, and L represents real-time flight length. Step 1: Calculate the combined cost of the three alternative routes: Alternative route 1: L1 = 9.2 km, E1 = 2.7 kWh; ; Alternative route 2: L2 = 8.8 km, E2 = 2.9 kWh; ; Alternative route 3: L3 = 9.8 km, E3 = 2.6 kWh; ; Step 2: Select the optimal flight path based on the scenario: If a crosswind is encountered during flight (potentially increasing energy consumption), the system predicts that the energy consumption of flight path 1 will rise to 2.7 × 1.1 = 2.97 kWh, and will recalculate. At this time, track 3 Even better, the system can quickly switch to track 3; If the no-fly zone is temporarily expanded (requiring flight path 3 to detour to 10.5km), then flight path 3... track 1 Back to optimal performance.

[0030] It is evident that multiple alternative tracks allow the system to switch between low-risk tracks that have already been generated when the scenario changes, without the need for replanning. This significantly improves emergency response efficiency. In contrast, if only one track is generated, the system may be left without any usable tracks in the event of an emergency.

[0031] The process by which the flight safety boundary dynamic calculation module calculates the dynamic safety envelope is as follows: The dynamic safety envelope is jointly defined by the flight state safety boundary and the mission energy consumption safety boundary; The flight state safety boundary is calculated by analyzing the real-time aerodynamic parameters, structural loads, and real-time wind field data obtained by the flight state intelligent perception module of the UAV. It represents the attitude, overload, and airspeed range in which the UAV can fly safely under the current state. The mission energy consumption safety boundary is calculated jointly by the UAV's remaining battery power, remaining mission range, and real-time energy consumption rate. The calculation process is as follows: First, the average energy consumption rate of future flight segments is predicted based on historical flight data and current flight status; then, the theoretical maximum remaining range is calculated based on the remaining battery power and the predicted average energy consumption rate; finally, the theoretical maximum remaining range is compared with the mission's remaining range, and combined with the preset safety margin, the battery power warning value at which the UAV must start returning to base or making an emergency landing is dynamically calculated.

[0032] The calculation process for the theoretical maximum remaining range in the mission energy consumption safety boundary is as follows: First, based on the UAV's flight status data over a fixed time window, including airspeed, rate of climb, and motor power, its average energy consumption rate is calculated. ; Then, using the drone's current remaining battery power... and the average energy consumption rate Estimate the theoretical maximum remaining flight time T max ; Finally, considering the ground speed V of the UAV in the current mission segment... gCalculate the theoretical maximum remaining range S max The calculation formula is as follows: ; The battery warning value E that necessitates initiating a return trip or forced landing alert Calculated using the following formula: ; Where k is a safety factor greater than 1, S return This returns the distance to the nearest available landing point from the current location. By constructing a dual-dimensional dynamic safety envelope encompassing both flight state safety boundaries and mission energy consumption safety boundaries, the system covers the safe range of UAV flight states such as real-time attitude, overload, and airspeed (calculated based on aerodynamic parameters, structural loads, and real-time wind field data), while also taking into account energy consumption safety related to remaining battery power, mission range, and energy consumption rate (dynamically predicting average energy consumption rate, theoretical maximum remaining range, and battery warning values). This avoids safety blind spots caused by a single safety boundary. Furthermore, the system dynamically adjusts the safety envelope based on real-time flight data (such as sudden wind fields and energy consumption fluctuations) rather than using fixed boundaries, adapting to complex and ever-changing flight conditions. It can also provide early warnings of return-to-home or emergency landing opportunities through dynamic battery warning values, preventing the risk of loss of control due to insufficient energy consumption. Ultimately, this provides clear and dynamic safety judgment standards for the autonomous intelligent decision-making module, ensuring that the optimized flight path is always within a safe range, and significantly improving the comprehensiveness and dynamic adaptability of UAV flight safety.

[0033] For example, in a drone transport mission from warehouse A (30°15′N, 120°20′E) to community B (30°18′N, 120°25′E) in the suburbs (with a payload of 30kg and a ground speed of 60km / h), the flight safety boundary dynamic calculation module simultaneously generates two types of safety boundaries: Flight safety boundaries: By analyzing the real-time aerodynamic parameters of the UAV (wing lift coefficient 1.2, drag coefficient 0.3) and structural load (maximum allowable overload of fuselage ±2.5g), combined with the real-time wind field data (wind speed 8m / s, wind direction angle 30°) obtained by the flight situation intelligent perception module, the current safe flight attitude range (pitch angle -5° to +7°, roll angle -6° to +6°), overload range (-1g to +2g), and airspeed range (75km / h to 115km / h) are calculated to avoid damage to the fuselage structure due to attitude exceeding limits or excessive overload; Task power consumption safety boundary: based on current remaining power. Remaining flight distance of the mission Based on the flight data from the past 20 minutes (airspeed 58–62 km / h, climb rate 1–2 m / s), the average energy consumption rate for the future flight segment is predicted. Calculate the theoretical maximum remaining range: (far greater than) (Preliminary assessment indicates sufficient energy consumption).

[0034] Relying solely on flight status safety boundaries may overlook the risk of a sudden surge in energy consumption due to strong winds in subsequent flight segments; conversely, relying solely on energy consumption safety boundaries may lead to excessive side loads due to unrestricted roll angles (such as accidental operation to +8°). The dynamic safety envelope defined by both boundaries comprehensively covers the two core risks of flight status and energy consumption, eliminating the safety blind spots of a single boundary.

[0035] Having undertaken the aforementioned transportation task, the drone flew to the midway point (20km from warehouse A, with remaining battery power). When the flight situation intelligent perception module suddenly detects a sudden strong crosswind (wind speed increases from 8 m / s to 18 m / s, and the angle between the wind direction and the heading changes to 60°), the flight safety boundary dynamic calculation module immediately adjusts the safety boundary based on the updated data. Flight safety boundary adjustment: Strong crosswinds will increase fuselage side loads by 30%, and the original safe roll angle range (-6° to +6°) will be reduced to (-4° to +4°) (to avoid side loads exceeding the fuselage structure's tolerance limits); at the same time, crosswinds will reduce airspeed stability, and the original airspeed range (75km / h to 115km / h) will be adjusted to (85km / h to 110km / h) (to prevent crosswinds from causing instability at low speeds and crosswind loads from exceeding limits at high speeds). Task energy consumption safety boundary adjustment: Strong crosswinds caused motor power to increase from 0.7 kWh / h to 0.9 kWh / h, and the average energy consumption rate was re-predicted. The theoretical maximum remaining range has been updated as follows: (Still satisfied) The remaining task requirements are met, but energy redundancy is reduced.

[0036] If a fixed safety boundary is adopted (such as always using the initial roll angle ±6° and airspeed 75~115km / h), when the UAV maintains a roll angle of +5° in strong crosswinds, the side load may exceed the structural limit and cause wing deformation; while a dynamically adjusted safety boundary adapts to sudden wind fields in real time, ensuring that the flight status and energy consumption safety are always within a controllable range.

[0037] Based on the aforementioned transportation mission, the flight safety boundary dynamic calculation module calculates the core indicators of the mission's energy consumption safety boundary, including the power warning value, according to the calculation process set in this case. The specific process is as follows: Predicted average energy consumption rate: Based on the UAV's flight status data over the past 30 minutes (fixed time window) (airspeed V=59km / h, rate of climb H=1.5m / s, motor power P=0.8~0.9kWh / h), the average energy consumption rate is calculated. ; Calculate the theoretical maximum remaining range: current remaining battery power , ground speed Theoretical maximum remaining flight time Theoretical maximum remaining range ; Dynamically calculates the battery warning value: The current location returns the distance to the nearest available landing point (Warehouse A). To avoid sudden increases in energy consumption, a safety factor k = 1.2 (k > 1) is set. The calculation logic for the power warning value is "the warning power must at least support the energy consumption of returning to the landing point, with a safety margin reserved," and the formula is: ; Substitute the data into the calculation: ; Risk warning application: If the remaining battery power drops to 0.38kWh during subsequent flights (close to...) If the remaining battery level drops below 0.34 kWh, the system will immediately trigger a "battery warning" alarm, indicating that "the remaining battery level is about to reach the warning value, and it is recommended to stop the mission and return to home". If the remaining battery level continues to drop below 0.34 kWh, the return to home procedure will be started automatically.

[0038] If not dynamically calculated A fixed warning based solely on "10% of total battery power" (e.g., a fixed warning value of 0.5kWh when the total battery power is 5kWh) might lead to an early warning (wasting mission time) when the return energy consumption is only 0.28kWh (without a safety margin), or an early warning (insufficient battery power to return) when the energy consumption suddenly increases to 1.0kWh / h. Dynamically calculated warnings, however, would be more effective. It can be precisely adapted to the current mission and flight conditions, balancing safety and mission efficiency.

[0039] The workflow of the autonomous intelligent decision-making and instruction generation module is as follows: The system receives dynamic safety envelope data from the flight safety boundary dynamic calculation module in real time. The dynamic safety envelope data includes flight state safety boundary and mission energy consumption safety boundary, as well as optimized trajectory from the adaptive trajectory planning and risk avoidance module. The optimized trajectory is compared with the dynamic safety envelope to determine whether the optimized trajectory is completely within the range defined by the flight state safety boundary and the mission energy consumption safety boundary. If the optimized trajectory is entirely within the dynamic safety envelope, the optimized trajectory is calculated into specific flight control commands and sent to the flight control system. If any part of the optimized trajectory exceeds the dynamic safety envelope, an alarm is triggered, and the alternative optimized trajectory that is completely within the dynamic safety envelope and has the lowest comprehensive cost function value is selected as the execution plan. By enabling the autonomous intelligent decision-making and command generation module to receive dynamic safety envelope and optimized trajectory data in real time, accurate real-time comparison of safety boundaries and optimized trajectories is achieved, avoiding the disconnect between trajectory planning and safety judgment. It can automatically determine whether the trajectory is within the safe range and quickly generate flight control commands or trigger alarms without human intervention, reducing the risks caused by delays in human decision-making. At the same time, when the trajectory exceeds the boundary, the safest alternative trajectory with the lowest overall cost is selected first, ensuring that there is still an optimal execution plan in emergency scenarios. This significantly improves the real-time performance, safety, and autonomy of UAV flight decisions, avoiding the problem of safe planned trajectories but no alternative plan when exceeding the boundary or encountering sudden risks.

[0040] For example, in the task of transporting supplies by drone from warehouse A (30°15′N, 120°20′E) in the suburbs to community B (30°18′N, 120°25′E), the autonomous intelligent decision-making and instruction generation module works according to the following process: The system synchronously acquires the dynamic safety envelope output by the flight safety boundary dynamic calculation module (flight state safety boundary: pitch angle -5° to +7°, roll angle -4° to +4°, airspeed 85km / h to 110km / h; mission energy consumption safety boundary: remaining power ≥ 0.34kWh), as well as the optimized trajectory generated by the adaptive trajectory planning module (track number T1, key parameters: pitch angle 1° to 3°, roll angle -2° to +2°, airspeed 90km / h to 95km / h, remaining power consumption expected to decrease from 3.4kWh to 2.1kWh, no over-boundary segments throughout the entire journey). The key parameters of T1's trajectory were compared one by one with the dynamic safety envelope: pitch angle (1°~3°) within -5°~+7°, roll angle (-2°~+2°) within -4°~+4°, airspeed (90~95km / h) within 85~110km / h, and remaining battery power (2.1kWh) ≥ 0.34kWh. Therefore, it was determined that T1 was completely within the safety envelope. The system automatically converts the T1 trajectory into specific flight control commands, such as "0-5 minutes: pitch angle 2°, roll angle 0°, airspeed 92km / h; 5-10 minutes: pitch angle 1.5°, roll angle +1°, airspeed 90km / h", and sends them to the flight control system in real time. This ensures the safe execution of the trajectory without manual operation and avoids the 10-15 minute delay caused by the traditional "plan first, then manual review" process.

[0041] If, when the drone is halfway through its flight (15km from warehouse A), the flight situation intelligent perception module detects a sudden gust of wind (wind speed increases from 18m / s to 25m / s), the adaptive trajectory planning module, due to data update delays, causes the temporarily generated optimized trajectory T2 to partially exceed the limits. The key parameters of the T2 track received by the module show that during the "12-13 minute" period, the roll angle needs to be adjusted to +5° and the airspeed reduced to 82km / h in order to counteract the gusts. At the same time, the flight safety boundary dynamic calculation module has updated the safety boundary (the roll angle has been adjusted to -3° to +3° due to the gusts, and the lower limit of airspeed has been increased to 90km / h). Comparison triggers alarm: Comparing T2 with the updated safety envelope, it is found that the roll angle +5° > +3° and the airspeed 82km / h < 90km / h are in the "12-13 minutes" period, and it is determined that T2 partially exceeds the dynamic safety envelope; The system immediately triggers a Level 3 audible and visual alarm (the cockpit / cloud simultaneously displays "T2 roll angle and airspeed exceed safety boundaries, suspend command issuance"), and automatically suspends T2 command generation, jumping to the alternative track selection process to avoid exceeding the limits of the fuselage side load due to executing an out-of-bounds track (roll angle +5° may cause side load to reach 3g, exceeding the structural limit of 2.5g) or causing instability due to excessively low airspeed.

[0042] After another out-of-bounds alarm occurs, the autonomous intelligent decision-making and command generation module obtains three alternative optimized tracks (T3, T4, T5) from the adaptive track planning module that are below the risk threshold (Pc < 0.5), and selects them according to the principle of being completely within the safety envelope and having the lowest overall cost: Initial screening of candidate tracks and safety envelopes: First, confirm that all 3 tracks meet the updated safety boundaries (roll angle -3° to +3°, airspeed 90 to 110 km / h, remaining battery power ≥ 0.34 kWh), and exclude tracks that exceed the boundaries; Set weights (Total conflict probability Pc, baseline probability) ), (Try track length L, reference track length) ), (Real-time energy consumption E, baseline energy consumption) Cost function formula: ; Substitute the parameters of the three alternative flight paths into the calculation: T3: ; ; T4:P c =0.41, L=10.2km, E=3.1kWh; ; T5:P c =0.42, L=9.6km, E=2.7kWh; ; Optimal flight path selection and execution: Since J5 < J3 < J4, T5 has the lowest overall cost. The module automatically converts T5 into control commands (such as "12-13 minutes: pitch angle 2.5°, roll angle +2°, airspeed 95km / h") and issues them. The entire process from selection to command generation takes only 2 seconds, ensuring continuous flight missions and avoiding mission interruptions or emergency landings due to the lack of alternative flight paths.

[0043] The cloud-based collaborative management platform is used to receive and record flight decision data of a single UAV, and to perform group collaborative management when multiple UAVs are operating in the same airspace. Group collaborative management includes: calculating in the cloud based on the optimized flight paths and dynamic safety envelopes reported by each UAV, assigning passage sequences with time differences or adjusting horizontal and vertical intervals for UAV groups with potential conflicts, generating group collaborative flight instructions, and issuing them to the corresponding UAVs.

[0044] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0045] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0046] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A large fixed-wing unmanned aerial vehicle (UAV) operation and management system, characterized in that, include: The flight status intelligent perception module is used to acquire and process the UAV's flight status data and external environment data in real time through multi-source sensor fusion processing. The adaptive trajectory planning and risk avoidance module is used to perform dynamic risk assessment and generate or optimize flight trajectories based on the output of the flight situation intelligent perception module. The flight safety boundary dynamic calculation module is used to calculate the dynamic safety envelope that evolves over time based on the real-time status of the UAV and the mission profile. The autonomous intelligent decision-making and command generation module is used to generate flight control commands or alarm information by integrating the dynamic safety envelope and the optimized flight trajectory. The cloud-based collaborative management platform communicates with the autonomous intelligent decision-making and command generation module for the collaborative management of multiple drones.

2. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 1, characterized in that: The multi-source sensor fusion processing in the intelligent flight situation perception module specifically involves: A three-dimensional situation field is constructed with UAVs at its core. The three-dimensional situation field is formed by integrating data from airborne inertial navigation systems, global positioning systems, weather radar, and air traffic control transponders. The fusion process employs an improved weighted distributed fusion algorithm, in which a signal integrity factor is introduced to weight the Global Positioning System data, and a meteorological unit threat level factor is introduced to weight the meteorological radar data. Finally, by calculating the fusion weight of each data source in the fusion center, high-confidence comprehensive UAV situational information is obtained; The calculation process of the fusion weight is as follows: First, the reliability index of each data source at the current moment is evaluated. The reliability index is obtained by exponentially smoothing the historical error sequence of the data source. The exponential smoothing process adopts a single exponential smoothing method, and its smoothing coefficient is dynamically adjusted based on the characteristics of the data source. Then, the reliability index of each data source is normalized to obtain the fusion weight of each data source in the fusion center.

3. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 2, characterized in that: The dynamic risk assessment in the adaptive trajectory planning and risk avoidance module specifically includes: Based on the three-dimensional situation field, identify static and dynamic obstacles ahead of the flight path; For dynamic obstacles, predict their trajectory and perform collision detection with the machine's predicted trajectory; The risk quantification value of the collision detection is obtained by calculating the total collision probability index between the machine and the dynamic obstacle; The calculation process of the total conflict probability index is as follows: First, based on the time series of the predicted trajectories of the two, the relative distance and approach rate at each future time point are calculated; Then, the relative distance and approach rate are input into a trained risk probability function to calculate the instantaneous conflict probability at that moment, where the instantaneous conflict probability function is defined as the product of a negative exponential function of the relative distance and a positive proportional function of the approach rate. Finally, the instantaneous conflict probabilities at all future time points are integrated to obtain the total conflict probability index for a future period starting from the current time.

4. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 3, characterized in that: The process by which the adaptive trajectory planning and risk avoidance module generates or optimizes the flight trajectory is as follows: The total conflict probability index, flight path length, and energy consumption are used as multi-objective optimization parameters for the comprehensive cost function; The comprehensive cost function is solved using a constrained optimization algorithm, with constraints including UAV performance constraints and airspace rule constraints. After solving, one or more alternative optimized paths are generated, and the comprehensive cost function value of the alternative optimized paths is lower than the preset risk threshold.

5. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 4, characterized in that: The process by which the flight safety boundary dynamic calculation module calculates the dynamic safety envelope is as follows: The dynamic safety envelope is jointly defined by the flight state safety boundary and the mission energy consumption safety boundary; The flight state safety boundary is calculated by analyzing the real-time aerodynamic parameters, structural loads, and real-time wind field data obtained by the flight state intelligent perception module of the UAV. It represents the attitude, overload, and airspeed range in which the UAV can fly safely under the current state. The mission energy consumption safety boundary is calculated jointly by the UAV's remaining battery power, remaining mission range, and real-time energy consumption rate. The calculation process is as follows: First, the average energy consumption rate of future flight segments is predicted based on historical flight data and current flight status; then, the theoretical maximum remaining range is calculated based on the remaining battery power and the predicted average energy consumption rate; finally, the theoretical maximum remaining range is compared with the mission's remaining range, and combined with the preset safety margin, the battery power warning value at which the UAV must start returning to base or making an emergency landing is dynamically calculated.

6. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 1, characterized in that: The calculation process for the theoretical maximum remaining range in the mission energy consumption safety boundary is as follows: First, based on the flight status data of the drone over a fixed time window in the past, including airspeed, rate of climb and motor power, its average energy consumption rate is calculated. Then, using the current remaining battery power and average energy consumption rate of the drone, the theoretical maximum remaining flight time is estimated; Finally, the theoretical maximum remaining range is calculated by combining the ground speed of the UAV in the current mission segment.

7. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 6, characterized in that: The workflow of the autonomous intelligent decision-making and instruction generation module is as follows: It receives dynamic safety envelope data in real time from the flight safety boundary dynamic calculation module. The dynamic safety envelope data includes the flight state safety boundary and the mission energy consumption safety boundary, as well as the optimized trajectory from the adaptive trajectory planning and risk avoidance module. The optimized trajectory is compared with the dynamic safety envelope to determine whether the optimized trajectory is completely within the range defined by the flight state safety boundary and the mission energy consumption safety boundary. If the optimized trajectory is entirely within the dynamic safety envelope, the optimized trajectory is calculated into specific flight control commands and sent to the flight control system. If any part of the optimized trajectory exceeds the dynamic safety envelope, an alarm is triggered, and the alternative optimized trajectory that is completely within the dynamic safety envelope and has the lowest comprehensive cost function value is selected as the execution plan.

8. The large fixed-wing unmanned aerial vehicle (UAV) operation and management system according to claim 7, characterized in that: The cloud-based collaborative management platform is used to receive and record flight decision data of a single UAV, and to perform group collaborative management when multiple UAVs are operating in the same airspace. Group collaborative management includes: based on the optimized flight paths and dynamic safety envelopes reported by each UAV, performing conflict resolution calculations in the cloud, assigning passage sequences with time differences or adjusting horizontal and vertical intervals for UAV groups with potential conflicts, generating group collaborative flight commands, and issuing them to the corresponding UAVs.

Images