Dynamic threat assessment and control method for non-visual flight of unmanned aerial vehicle in unmanned area

Abstract

The invention discloses an unmanned area unmanned aerial vehicle non-visual flight dynamic threat assessment management and control method, and belongs to the technical field of air traffic control systems, and the method comprises the steps: obtaining the multi-source sensing data in the flight process of an unmanned aerial vehicle in real time, and generating the state reference parameters of the unmanned aerial vehicle; generating a multi-source quantitative threat data set based on the multi-source sensing data; the state reference parameters of the unmanned aerial vehicle and the multi-source quantitative threat data set are fused to generate a comprehensive flight risk assessment result; generating an adaptive flight management and control strategy based on the comprehensive flight risk assessment result; generating a dynamic adjustment control instruction set based on the adaptive flight control strategy; and performing feedback adjustment based on the execution effect of the dynamic adjustment control instruction set. The technical means of quantifying multi-source threats, carrying out comprehensive risk assessment, generating a self-adaptive control strategy and carrying out feedback optimization based on an execution effect are adopted, so that the autonomous flight safety and intelligent level of the unmanned aerial vehicle in a complex dynamic environment can be remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of air traffic control systems, and in particular to a method for assessing and managing the dynamic threats of unmanned aerial vehicles (UAVs) flying without visual contact in uninhabited areas. Background Technology

[0002] Unmanned aerial vehicles (UAVs), commonly known as drones, are now widely used in various fields such as surveying, inspection, logistics, and security. Their non-visual long-range flight capabilities, particularly in vast, inaccessible uninhabited areas, demonstrate enormous application potential. These missions typically require UAVs to autonomously complete long-distance, long-duration flights according to mission plans, without direct visual observation from an operator. This places extremely high demands on the autonomy and reliability of the flight control system.

[0003] Existing drone flight control technologies largely rely on pre-planned waypoint flight paths and integrate basic onboard sensing systems for simple obstacle avoidance. During non-visual flight, ground station operators remotely monitor basic flight telemetry data of the drone, such as position, altitude, speed, and battery level. In case of emergencies, operators typically need to make manual judgments and issue remote control commands, or the onboard simple obstacle avoidance logic will trigger preset hovering or return-to-home procedures.

[0004] However, the aforementioned existing technical solutions have significant drawbacks. Preset flight paths cannot cope with dynamically changing threats during flight, such as sudden weather changes, unexpected equipment malfunctions, or other aircraft temporarily entering the airspace. Remote manual intervention is unreliable in uninhabited environments with signal delays or interruptions, and cannot meet the needs of large-scale autonomous operations. Airborne, simple obstacle avoidance logic often targets only a single type of threat, lacking the ability to comprehensively assess multiple concurrent threats. Its risk assessment models typically use fixed thresholds, failing to consider the drone's real-time status, such as remaining battery power or communication quality, resulting in inaccurate risk assessments and a lack of targeted and flexible response strategies. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a method for assessing and managing dynamic threats to unmanned aerial vehicles (UAVs) flying without visual contact in unmanned areas. This method employs techniques such as real-time quantification of multi-source threats, integration of the UAV's own status for comprehensive risk assessment, generation of adaptive management strategies, and feedback optimization based on execution results. This approach can significantly improve the safety and intelligence of UAVs' autonomous flight in complex dynamic environments.

[0006] The above objectives can be achieved through the following approach: A method for assessing and managing non-visual flight dynamic threats of unmanned aerial vehicles (UAVs) in unmanned areas includes: acquiring multi-source sensor data during UAV flight in real time to generate UAV self-state baseline parameters; quantifying preset threat sources based on the multi-source sensor data, calculating the offset between the quantified preset threat sources and preset safety thresholds, and merging the offsets to generate a multi-source quantified threat dataset, wherein the multi-source quantified threat dataset includes meteorological threat values, equipment failure risk threat values, and airspace conflict risk threat values; inputting the UAV self-state baseline parameters and the multi-source quantified threat dataset into a preset weight allocation fusion model to generate a comprehensive flight risk assessment result; calculating the flight control target value of the corresponding threat item based on the comprehensive flight risk assessment result, and combining the flight control target values ​​to generate an adaptive flight control strategy; generating a dynamic adjustment control command set based on the adaptive flight control strategy and issuing it to the UAV, setting the UAV heading angle, UAV altitude value, and UAV speed value; and generating strategy optimization data based on the execution effect of the dynamic adjustment control command set, comparing the threat values ​​of the threat items before and after execution.

[0007] Optionally, generating the UAV's own state reference parameters includes: extracting the UAV's three-dimensional position, velocity, attitude, remaining battery power, and communication link quality from the multi-source sensor data; and combining the UAV's three-dimensional position, velocity, attitude, remaining battery power, and communication link quality to generate the UAV's own state reference parameters.

[0008] Optionally, generating the multi-source quantitative threat dataset includes: classifying the multi-source sensor data to generate environmental data, body data, and airspace data; conducting risk assessments on the environmental data, body data, and airspace data respectively to generate meteorological threat values, equipment failure risk values, and airspace conflict risk values; and merging the meteorological threat values, equipment failure risk values, and airspace conflict risk values ​​to generate the multi-source quantitative threat dataset.

[0009] Optionally, generating meteorological threat values, equipment failure risk values, and airspace conflict risk values ​​includes: comparing real-time parameters in the environmental data, aircraft data, and airspace data with preset environmental data safety thresholds, aircraft data safety thresholds, and airspace data safety thresholds, respectively; calculating the degree of deviation of the real-time parameters from the safety thresholds; and quantifying the degree of deviation to generate meteorological threat values, equipment failure risk values, and airspace conflict risk values.

[0010] Optionally, generating the comprehensive flight risk assessment result includes: taking the UAV's own state baseline parameters and the multi-source quantitative threat dataset as input items, inputting a preset weight allocation fusion model to dynamically allocate different weight coefficients to the meteorological threat value, the equipment failure risk threat value, and the airspace conflict risk threat value, and then performing weighted fusion to generate the comprehensive flight risk assessment result.

[0011] Optionally, the generation of the adaptive flight control strategy specifically includes: comparing the comprehensive flight risk assessment result with a preset risk level threshold to generate a risk handling level; when the risk handling level indicates that the flight status needs to be adjusted, identifying the specific threat source and its quantitative parameters from the original data of the comprehensive flight risk assessment result that caused the risk assessment result to exceed the corresponding threshold; calculating and generating the corresponding flight control target value based on the identified specific threat source and its quantitative parameters; and combining the flight control target values ​​to form an adaptive flight control strategy.

[0012] Optionally, generating the dynamic adjustment control instruction set includes: parsing the adaptive flight control strategy to extract flight control target values; generating basic control instructions to be executed based on the flight control target values, wherein the basic control instructions include the setting of target heading angle, target altitude value and target speed value; and combining the basic control instructions to form a dynamic adjustment control instruction set.

[0013] Optionally, the generation of strategy optimization data includes: parsing the dynamic adjustment control instruction set to obtain the specific threat item to be addressed and the threat value of the threat item before the execution of the dynamic adjustment control instruction set; after the execution of the dynamic adjustment control instruction set, generating a new threat value based on real-time acquired multi-source sensor data; calculating the change between the threat value of the threat item before the execution of the dynamic adjustment control instruction set and the new threat value, and generating strategy optimization data.

[0014] Optionally, the method further includes: adjusting the weight coefficients assigned by the preset weight allocation fusion model based on the strategy optimization data.

[0015] Based on the same inventive concept, this invention also provides a non-visual flight control system for unmanned aerial vehicles (UAVs) in uninhabited areas based on dynamic threat assessment, comprising: The status monitoring module is used to acquire multi-source sensor data during the drone's flight in real time and generate the drone's own status baseline parameters. The threat quantification module is used to quantify preset threat sources based on the multi-source sensor data, calculate the offset between the quantified preset threat source and the preset safety threshold, and merge the offsets to generate a multi-source quantified threat dataset, wherein the multi-source quantified threat dataset includes meteorological threat values, equipment failure risk threat values ​​and airspace conflict risk threat values. The risk fusion module is used to input the UAV's own state baseline parameters and the multi-source quantitative threat dataset into a preset weight allocation fusion model to generate a comprehensive flight risk assessment result. The strategy decision module is used to calculate the flight control target value of the corresponding threat item based on the comprehensive flight risk assessment results, and combine the flight control target values ​​to generate an adaptive flight control strategy. The instruction issuing module is used to generate a dynamic adjustment control instruction set based on the adaptive flight control strategy and issue it to the UAV, setting the UAV heading angle, UAV altitude value and UAV speed value; The execution feedback module is used to generate strategy optimization data based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution; The adaptive update module is used to generate strategy optimization data based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution.

[0016] Compared with the prior art, the present invention has the following advantages: 1. This invention constructs a comprehensive assessment model that dynamically integrates the UAV's own state with external multi-source threats, enabling the generation of the most targeted risk profile for the current flight scenario in real time. Compared with traditional methods that isolate the assessment of various threats or use static risk thresholds, this greatly improves the accuracy and authenticity of risk assessment, thereby providing a solid and reliable foundation for subsequent decision-making and comprehensively enhancing flight safety.

[0017] 2. This invention achieves adaptive and precise control strategies. Based on comprehensive risk assessment results, the system can retrospectively identify specific threat sources that lead to increased risk, and then generate flight control targets directly targeting those threat sources. This point-to-point approach avoids broad or ineffective flight adjustments, improves the efficiency and success rate of risk avoidance operations, and ensures that the UAV can effectively mitigate risks at minimal cost.

[0018] 3. This invention establishes a complete closed-loop feedback and self-optimization mechanism. By quantitatively evaluating the actual implementation effect of each control strategy, the system can learn from this and continuously iterate its internal threat assessment and decision-making models. This allows the intelligence level of the flight control system to continuously improve with the accumulation of flight experience, enhancing its robustness and environmental adaptability in the face of unknowns and emergencies.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 This is a schematic diagram of the structure of the non-visual flight dynamic threat assessment and control method for unmanned aerial vehicles in uninhabited areas according to an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the relationship between multi-source threat quantification and dynamic weight allocation in an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram of the execution effect curve of the adaptive control strategy in an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the structure of the non-visual flight dynamic threat assessment and control method for unmanned aerial vehicles in uninhabited areas according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Reference Figure 1 One embodiment of the present invention proposes a method for assessing and managing dynamic threats to unmanned aerial vehicles (UAVs) flying without visual range. This method employs techniques such as real-time quantification of multi-source threats, integration of the UAV's own state for comprehensive risk assessment, generation of adaptive management strategies, and feedback optimization based on execution results. This approach can significantly improve the safety and intelligence of UAVs' autonomous flight in complex dynamic environments.

[0027] The method described in this embodiment specifically includes: Real-time acquisition of multi-source sensor data during drone flight to generate drone's own state baseline parameters; Based on the multi-source sensor data, the preset threat sources are quantified, the offset between the quantified preset threat sources and the preset safety threshold is calculated, and the offsets are merged to generate a multi-source quantified threat dataset, which includes meteorological threat values, equipment failure risk threat values ​​and airspace conflict risk threat values. The UAV’s own state baseline parameters and the multi-source quantified threat dataset are input into a preset weight allocation fusion model to generate a comprehensive flight risk assessment result. Based on the comprehensive flight risk assessment results, the flight control target values ​​of the corresponding threat items are calculated, and the flight control target values ​​are combined to generate an adaptive flight control strategy. Based on the adaptive flight control strategy, a dynamic adjustment control command set is generated and sent to the UAV, setting the UAV heading angle, UAV altitude value and UAV speed value; Based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution are used to generate strategy optimization data.

[0028] Specifically, a complete, self-optimizing closed-loop flight management process was constructed. Its core lies in the dynamic fusion of the UAV's self-state awareness and external threat perception to achieve intelligent decision-making. This method first establishes a baseline of the UAV's own state and a quantitative threat model of the external environment, equipment, and airspace through multi-source sensor data synchronization. These two models are then fused to generate a comprehensive flight risk assessment result, reflecting not only the severity of the threat but also the UAV's current resilience. Based on this comprehensive risk assessment, the system autonomously generates a customized, adaptive flight management strategy and translates it into specific control commands for execution. A feedback mechanism is introduced to evaluate the actual threat mitigation effect after the control commands are executed, generating strategy optimization data. This data is then used to adjust and optimize the initial threat quantification assessment process, forming a continuously learning and iteratively evolving intelligent management closed loop. This significantly improves the autonomy, safety, and environmental adaptability of UAVs performing non-visual flight missions in complex, dynamic, and unpredictable environments such as uninhabited areas. Through dynamic threat assessment and adaptive control, this method enables UAVs to proactively and in real-time identify and respond to multi-dimensional potential risks during flight, rather than passively executing preset routes or relying on remote manual intervention. More importantly, its built-in feedback optimization closed-loop mechanism endows the flight control system with the ability to learn and continuously evolve, allowing the UAV's risk management strategies to be continuously improved through experience accumulation. This results in higher reliability and intelligence levels in long-term operation, ultimately ensuring that UAVs can complete their flight missions safely and efficiently.

[0029] Optionally, the generated UAV self-state reference parameters include: The three-dimensional position, velocity, attitude, remaining battery power, and communication link quality of the UAV are extracted from the multi-source sensor data. The combined three-dimensional position, velocity, attitude, remaining battery power, and communication link quality of the UAV are used to generate the UAV's own state baseline parameters.

[0030] Specifically, multi-source sensor data is collected and aggregated in real time through the multi-source sensor onboard sensors and airborne data link of the UAV, forming a multi-source sensor data stream. From this multi-source sensor data stream, several key parameters characterizing the core operational status of the UAV are precisely analyzed. These include three-dimensional position and velocity information, the aircraft's roll, pitch, and yaw angles, the current remaining battery percentage, and the communication link quality between the ground station and the UAV. Communication link quality is typically expressed as received signal strength or signal-to-noise ratio. The three-dimensional position and velocity information, the aircraft's roll, pitch, and yaw angles, the current remaining battery percentage, and the communication link quality between the ground station and the UAV are integrated into a structured dataset, namely the UAV's own state baseline parameters. ,have: , Wherein, P represents the UAV's three-dimensional position vector in the preset coordinate system; V represents its three-dimensional velocity vector; A represents the attitude vector composed of three Euler angles: roll, pitch, and yaw; B is a scalar representing the remaining battery power; and C is a scalar representing the communication link quality. These UAV self-state reference parameters fully describe the UAV's instantaneous physical and communication states at any given moment.

[0031] Optionally, generating the multi-source quantified threat dataset includes: The multi-source sensor data is classified to generate environmental data, organism data, and spatial data. Risk assessments are conducted on environmental data, aircraft data, and airspace data respectively, generating meteorological threat values, equipment failure risk values, and airspace conflict risk values. The numerical values ​​of meteorological threats, equipment failure risks, and airspace conflict risks are merged to generate a multi-source quantitative threat dataset.

[0032] Specifically, multi-source sensor data is categorized into three independent data subsets based on its source and physical meaning, generating environmental data, airframe data, and airspace data. Environmental data primarily includes information such as wind speed, wind direction, temperature, humidity, and atmospheric pressure obtained from meteorological sensors and other equipment. Airframe data mainly covers information from sensors within the UAV's internal systems, such as vibration data, voltage and current feedback from the battery management system, and the speed and temperature of the motors in the power system. Airspace data includes the position and speed information of other nearby aircraft, as well as airspace restriction information such as electronic fences and no-fly zones obtained from ground stations or air traffic control systems. Independent threat quantification assessments are then performed based on the environmental, airframe, and airspace data. For meteorological threats, real-time environmental data is compared with preset environmental data safety thresholds to generate a meteorological threat value. For equipment failure risks, real-time airframe data is compared with preset airframe data safety thresholds to generate an equipment failure risk value. For airspace conflict risks, real-time airspace data is compared with preset airspace data safety thresholds to generate an airspace conflict risk value. The numerical values ​​of meteorological threats, equipment failure risks, and airspace conflict risks are combined into a structured dataset to generate a multi-source quantified threat dataset. ,have: , in, Represents the numerical value of weather threat. This represents a numerical value indicating the risk of equipment failure. This represents a numerical value indicating the risk of airspace conflict. The magnitude of the threat value directly reflects the severity of the corresponding threat source.

[0033] Optionally, the generation of meteorological threat values, equipment failure risk threat values, and airspace conflict risk threat values ​​includes: The real-time parameters in the environmental data, body data, and airspace data are compared with preset environmental data security thresholds, body data security thresholds, and airspace data security thresholds, respectively. The deviation of real-time parameters from safety thresholds is calculated, and the deviation is quantified to generate meteorological threat values, equipment failure risk values, and airspace conflict risk values.

[0034] Specifically, a pre-defined safety threshold system is established, including environmental data safety thresholds, such as maximum permissible flight wind speed and upper and lower limits of operating temperature; airframe data safety thresholds, such as maximum permissible motor temperature, minimum safe battery voltage, and maximum airframe vibration amplitude; and airspace data safety thresholds, such as minimum safe separation distance from other aircraft and minimum buffer distance from the boundary of a no-fly zone. Each real-time parameter from the acquired environmental, airframe, and airspace data is compared one by one with its corresponding pre-defined safety threshold. A normalization function is used to calculate the degree of deviation of the real-time parameter from its safety threshold. , in, A numerical value representing a specific threat. These are the real-time parameter values ​​corresponding to the threat item, such as real-time wind speed or motor temperature, which are obtained through multi-source sensor data. This is the preset safety threshold for this parameter, that is, the upper or lower limit for normal operation. This is a preset critical threshold, representing the state where the system will be in an extremely dangerous condition once this value is reached or exceeded. Function f is a clamping function that ensures that when... Less than or equal to hour, =0; when Greater than or equal to hour, =1; when When it is in between, It grows between 0 and 1. This process is repeated on all key parameters in environmental, physical, and airspace data to generate meteorological threat values, equipment failure risk values, and airspace conflict risk values.

[0035] Optionally, the generation of the comprehensive flight risk assessment results includes: The UAV's own state baseline parameters and the multi-source quantified threat dataset are used as input items. A preset weight allocation fusion model is used to dynamically assign different weight coefficients to the meteorological threat value, the equipment failure risk threat value, and the airspace conflict risk threat value, and then the weighted fusion is performed to generate a comprehensive flight risk assessment result.

[0036] Specifically, the model standardizes the input data of the UAV's own state baseline parameters and the multi-source quantized threat dataset. Using a max-min normalization method, the data is uniformly mapped to a dimensionless interval of 0 to 1, forming a standardized state vector and threat matrix. The standardized UAV state baseline parameters are used to construct a query vector, representing the current situation and state of the UAV. The various threat features from the standardized multi-source quantized threat dataset are used to construct a key matrix, representing the identity of each threat. The quantified values ​​of each threat are used to construct a value matrix, representing the degree of danger of each threat. The model generates an attention score by calculating the correlation between the query vector and each threat identity in the key matrix. , in, To score attention, Let K be the query vector, K be the key matrix, and T represent the matrix transpose. The attention score is input into the Softmax function for normalization, generating assigned weight coefficients. After obtaining the dynamically assigned weight coefficients, a weighted sum is performed. , in, The final comprehensive flight risk assessment result is a dimensionless scalar whose value directly reflects the overall risk level currently faced by the drone. , and These represent the numerical values ​​of meteorological threats, equipment failure risks, and airspace conflict risks obtained from the multi-source quantitative threat dataset, respectively. , and These are weighted coefficients dynamically generated based on the UAV's own state baseline parameters, corresponding to the three types of threats mentioned above, and their sum is set to [value missing]. For example... Figure 2 The figure shown is a schematic diagram of the relationship between multi-source threat quantification and dynamic weight allocation.

[0037] Optionally, the generation of the adaptive flight control strategy specifically includes: The comprehensive flight risk assessment results are compared with preset risk level thresholds to generate a risk handling level. When the risk handling level indication requires adjustment of the flight status, the specific threat source and the specific threat source quantification parameters that caused the risk assessment result to exceed the corresponding threshold are identified from the raw data of the comprehensive flight risk assessment result. Based on the identified specific threat sources and their quantitative parameters, corresponding flight control target values ​​are calculated and generated; the flight control target values ​​are combined to form an adaptive flight control strategy.

[0038] Specifically, the system compares the comprehensive flight risk assessment results with a set of preset risk level thresholds. This threshold system divides continuous risk values ​​into several discrete levels, such as safe, concern, warning, and danger. Through comparison, the system can quickly determine the severity of the current risk and generate a corresponding risk handling level. When the risk handling level indicates that the flight status needs to be adjusted, i.e., when the comprehensive flight risk assessment results exceed the safety threshold, the system checks various threat values ​​in the multi-source quantified threat dataset, such as weather threat values, equipment failure risk threat values, and airspace conflict risk threat values, as well as their corresponding specific parameters, to identify the specific threat sources and their quantified parameters that have led to the increase in the overall risk level. The system then calculates and generates corresponding flight control target values ​​to avoid or mitigate the threat. For example, if the identified threat is excessive crosswind, the system calculates a new target heading angle to adjust the UAV's attitude to counteract the crosswind effect and calculates a new target speed value to ensure flight stability and track accuracy. If the threat is airspace conflict, the system calculates a target altitude or target heading angle that can increase the safety separation. These calculated target values, such as target heading angle, target altitude, and target speed, are combined to generate an adaptive flight control strategy.

[0039] Optionally, the generation of the dynamic adjustment control instruction set includes: The adaptive flight control strategy is analyzed to extract flight control target values; Based on the flight control target value, basic control commands to be executed are generated, including the setting of target heading angle, target altitude value and target speed value; Combine basic control commands to form a dynamic adjustment control command set.

[0040] Specifically, the adaptive flight control strategy is analyzed. The target heading angle, target altitude, and target velocity values ​​are extracted one by one from the data packet. Based on these values, corresponding basic control commands conforming to the UAV flight control system's communication protocol are generated. For example, if the extracted target heading angle is 90 degrees, the system generates a command to set the UAV's target heading angle to 90 degrees. All the basic control commands calculated for the current risk are combined to generate a complete and time-sequential dynamic adjustment control command set.

[0041] Optionally, the generated strategy optimization data includes: The dynamic adjustment control instruction set is analyzed to obtain the specific threat item to be dealt with and the threat value of the threat item before the execution of the dynamic adjustment control instruction set; After the dynamic adjustment control instruction set is executed, a new threat value is generated based on the real-time acquired multi-source sensor data; Calculate the change in threat value of the threat item before the execution of the dynamic adjustment control instruction set and the change in the new threat value, and generate strategy optimization data.

[0042] Specifically, by parsing the dynamic adjustment control command set, accessing and retrieving historical data archives, a multi-source quantitative threat dataset prior to decision-making is obtained. Threat values ​​corresponding to specific threat items are extracted from this dataset. After the UAV executes the dynamic adjustment control command set and reaches a new stable flight state, the latest multi-source sensor data acquired in real time is used to re-evaluate the risk quantification of the same specific threat item. The same safety threshold as before is used to generate a new threat value. This new value reflects the severity of the threat item after the implementation of the control strategy. The threat value of the threat item before command execution is compared with the new threat value generated after execution, and the change between the two is calculated. This change is defined as the strategy optimization data for this control action. Its calculation formula can be expressed as follows: , in, This refers to the generated strategy optimization data, which is a dimensionless scalar representing the effectiveness of the control strategy. This is the threat value of the threat item before the instruction is executed. This value is obtained from historical multi-source quantitative threat datasets. It is a new threat value for the same threat item calculated based on new multi-source sensor data after the instruction is executed. Because... and All are dimensionless values ​​normalized using the same method, therefore subtraction can be performed directly. A positive... The value indicates that the implemented strategy effectively reduced the threat; the higher the value, the better the effect. Conversely, a negative value or zero indicates that the strategy is ineffective or even has the opposite effect.

[0043] Optionally, the method further includes: Based on the optimized data according to the strategy, the weight coefficients of the preset weight allocation fusion model are adjusted.

[0044] Specifically, after a dynamic adjustment control command set is executed and corresponding strategy optimization data is generated, the specific threat type targeted by the adjustment is identified, such as a weather threat. Based on the values ​​of the strategy optimization data, the generation rules for the weight coefficients related to that threat type are adjusted. , in, These are new weighting coefficient values ​​that will be used when encountering similar drone states in the future. These are the old weighting coefficients actually used in this decision-making process. It is a preset learning rate, a small normal number, which determines the magnitude of each adjustment to ensure the stability of the learning process. This is strategy optimization data, specifically the change in the threat value of a threat item before and after the instruction is executed. It directly quantifies the effectiveness of the current strategy. After updating a weight coefficient, all weight coefficients need to be normalized to ensure that their sum is still 1, maintaining the logical consistency of the weighted fusion model.

[0045] Based on the same inventive concept, such as Figure 4 As shown, this invention also provides non-visual flight control of unmanned aerial vehicles (UAVs) in uninhabited areas based on dynamic threat assessment. The system includes: The status monitoring module is used to acquire multi-source sensor data during the drone's flight in real time and generate the drone's own status baseline parameters. The threat quantification module is used to quantify preset threat sources based on the multi-source sensor data, calculate the offset between the quantified preset threat source and the preset safety threshold, and merge the offsets to generate a multi-source quantified threat dataset, wherein the multi-source quantified threat dataset includes meteorological threat values, equipment failure risk threat values ​​and airspace conflict risk threat values. The risk fusion module is used to input the UAV's own state baseline parameters and the multi-source quantitative threat dataset into a preset weight allocation fusion model to generate a comprehensive flight risk assessment result. The strategy decision module is used to calculate the flight control target value of the corresponding threat item based on the comprehensive flight risk assessment results, and combine the flight control target values ​​to generate an adaptive flight control strategy. The instruction issuing module is used to generate a dynamic adjustment control instruction set based on the adaptive flight control strategy and issue it to the UAV, setting the UAV heading angle, UAV altitude value and UAV speed value; The execution feedback module is used to generate strategy optimization data based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution; The adaptive update module is used to generate strategy optimization data based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution.

[0046] Example 1: To verify the feasibility of this invention in practice, it was applied to a drone inspection mission along a high-voltage power transmission corridor in an uninhabited area. This mission required the drone to autonomously fly 50 kilometers along a predetermined route under non-visual conditions without human remote control, conducting inspections of the power facilities along the line. The inspection area was located in a mountainous region with variable weather and the potential presence of other unreported general aviation aircraft, posing a significant challenge to the drone's autonomous flight safety.

[0047] In this embodiment, a quadcopter UAV codenamed "Inspection-01" was equipped with the flight control system described in this invention. Between August and October 2023, the system performed multiple complete inspection missions. During flight, the system continuously implemented the control method described in this invention, including: real-time acquisition of multi-source sensor data; generation of the UAV's own state baseline parameters and a multi-source quantified threat dataset; fusion of the two to generate a comprehensive flight risk assessment result; generation and execution of an adaptive flight control strategy; and finally, generation of strategy optimization data based on the execution effect to provide feedback and adjust the threat assessment model.

[0048] To verify the beneficial effects of the present invention, the handling processes of several typical risk events were recorded and analyzed. For example, during a flight mission on September 5, 2023, when "Inspection-01" was flying midway through the mission, the UAV's own status baseline parameters showed that its remaining battery power was 45% and the communication link quality was good. At this time, the airborne meteorological sensor detected that the crosswind speed had increased sharply to 15 m / s, which exceeded the preset environmental data safety threshold of 10 m / s.

[0049] The system then performed a threat quantification assessment, quantifying the wind speed parameter and generating a high meteorological threat value. Simultaneously, because the UAV's baseline parameters indicated low remaining battery power, the system dynamically assigned a higher weighting coefficient to the meteorological threat value when generating the comprehensive flight risk assessment result. The resulting comprehensive flight risk assessment exceeded the risk threshold for the "warning" level. The system identified "excessive crosswind" as the specific threat source causing the increased risk and calculated new flight control target values, including deflecting the target heading angle 12 degrees towards the windward direction to counteract the crosswind's effects, and reducing the target speed from 10 m / s to 8 m / s to ensure flight stability. These flight control target values ​​constituted an adaptive flight management strategy and were parsed into a dynamically adjusted control command set, which was then issued to the UAV.

[0050] After executing the command, the drone's flight path stabilized. Based on the new sensor data, the system recalculated the weather threat value, which was significantly reduced. The system recorded the change in threat value before and after command execution as strategy optimization data. This data is used in subsequent tasks to adjust the weight allocation model, enabling the system to respond earlier and more decisively when encountering a situation where "low battery" and "strong winds" coexist.

[0051] Similarly, in another mission on September 28, 2023, the UAV's ADS-B receiver detected a rapidly approaching general aviation helicopter. The safe separation distance in the airspace data was less than the preset airspace data safety threshold. The system immediately generated a high airspace conflict risk threat value and fused it to generate a comprehensive flight risk assessment result that reached the "dangerous" level. After identifying the specific threat source, the system generated an adaptive flight control strategy that included changing the target's altitude (emergency descent of 30 meters) and the target's heading angle (yaw 90 degrees to the right), and successfully directed the UAV to perform evasive maneuvers, avoiding a potential aerial conflict.

[0052] Table 1. Dynamic Threat Assessment and Risk Level Determination Data Table Table 2. Data on the Generation and Execution Effects of Adaptive Control Strategies Table 3 Performance Comparison Data Before and After System Optimization As can be seen from the data in Tables 1 to 3 above, the method described in this invention demonstrates excellent autonomous risk management capabilities in practical applications.

[0053] Table 1 demonstrates the system's ability to accurately quantify and dynamically assess different types of threats. Whether facing weather threats or airspace conflicts, the system can generate reasonable threat values ​​based on real-time data and, through a dynamic weight allocation mechanism, combine the UAV's own status (such as battery level) to generate a more context-relevant comprehensive risk assessment result, thereby triggering the corresponding level of contingency plan.

[0054] Table 2 clearly records the decision-making closed loop from risk identification to strategy generation. The system can "prescribe the right medicine," calculating and generating adaptive control strategies containing clear flight control target values ​​for specific threat sources such as "excessive crosswinds" and "airspace conflicts." After implementing the strategies, the threat values ​​were significantly reduced, and the generated strategy optimization data objectively quantified the effectiveness of each intervention, providing a basis for subsequent optimization.

[0055] Table 3 provides strong evidence of the system evolution effect brought about by the feedback adjustment mechanism of this invention. After two months of operation and self-learning, the system's response speed and handling effectiveness in dealing with similar risk events have significantly improved, with the threat reduction rate increasing from 55% to 80%, and the accuracy of comprehensive risk assessment also significantly improved. These data indicate that this invention can not only achieve effective risk avoidance in a single instance, but also continuously improve the UAV's autonomous survival and mission execution capabilities in complex uninhabited areas through continuous learning, such as... Figure 3 The figure shown is a schematic diagram illustrating the effect of the adaptive control strategy.

[0056] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method is applicable to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0057] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A method for assessing and managing the dynamic threat of unmanned aerial vehicles (UAVs) flying non-visually in uninhabited areas, characterized in that: include: Real-time acquisition of multi-source sensor data during drone flight to generate drone's own state baseline parameters; Based on the multi-source sensor data, the preset threat sources are quantified, the offset between the quantified preset threat sources and the preset safety threshold is calculated, and the offsets are merged to generate a multi-source quantified threat dataset, which includes meteorological threat values, equipment failure risk threat values ​​and airspace conflict risk threat values. The UAV’s own state baseline parameters and the multi-source quantified threat dataset are input into a preset weight allocation fusion model to generate a comprehensive flight risk assessment result. Based on the comprehensive flight risk assessment results, the flight control target values ​​of the corresponding threat items are calculated, and the flight control target values ​​are combined to generate an adaptive flight control strategy. Based on the adaptive flight control strategy, a dynamic adjustment control command set is generated and sent to the UAV, setting the UAV heading angle, UAV altitude value and UAV speed value; Based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution are used to generate strategy optimization data.

2. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The generated UAV self-state reference parameters include: The three-dimensional position, velocity, attitude, remaining battery power, and communication link quality of the UAV are extracted from the multi-source sensor data. The combined three-dimensional position, velocity, attitude, remaining battery power, and communication link quality of the UAV are used to generate the UAV's own state baseline parameters.

3. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The generated multi-source quantified threat dataset includes: The multi-source sensor data is classified to generate environmental data, organism data, and spatial data. Risk assessments are conducted on environmental data, aircraft data, and airspace data respectively, generating meteorological threat values, equipment failure risk values, and airspace conflict risk values. The numerical values ​​of meteorological threats, equipment failure risks, and airspace conflict risks are merged to generate a multi-source quantitative threat dataset.

4. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 3, characterized in that, The generated meteorological threat values, equipment failure risk threat values, and airspace conflict risk threat values ​​include: The real-time parameters in the environmental data, body data, and airspace data are compared with preset environmental data security thresholds, body data security thresholds, and airspace data security thresholds, respectively. The deviation of real-time parameters from safety thresholds is calculated, and the deviation is quantified to generate meteorological threat values, equipment failure risk values, and airspace conflict risk values.

5. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The generated comprehensive flight risk assessment results include: The UAV's own state baseline parameters and the multi-source quantified threat dataset are used as input items. A preset weight allocation fusion model is used to dynamically assign different weight coefficients to the meteorological threat value, the equipment failure risk threat value, and the airspace conflict risk threat value, and then the weighted fusion is performed to generate a comprehensive flight risk assessment result.

6. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The generated adaptive flight control strategy specifically includes: The comprehensive flight risk assessment results are compared with preset risk level thresholds to generate a risk handling level. When the risk handling level indication requires adjustment of the flight status, the specific threat source and the specific threat source quantification parameters that caused the risk assessment result to exceed the corresponding threshold are identified from the raw data of the comprehensive flight risk assessment result. Based on the identified specific threat sources and their quantitative parameters, corresponding flight control target values ​​are calculated and generated; the flight control target values ​​are combined to form an adaptive flight control strategy.

7. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The generation of the dynamic adjustment control instruction set includes: The adaptive flight control strategy is analyzed to extract flight control target values; Based on the flight control target value, basic control commands to be executed are generated, including the setting of target heading angle, target altitude value and target speed value; Combine basic control commands to form a dynamic adjustment control command set.

8. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The generated strategy optimization data includes: The dynamic adjustment control instruction set is analyzed to obtain the specific threat item to be dealt with and the threat value of the threat item before the execution of the dynamic adjustment control instruction set; After the dynamic adjustment control instruction set is executed, a new threat value is generated based on the real-time acquired multi-source sensor data; Calculate the change in threat value of the threat item before the execution of the dynamic adjustment control instruction set and the change in the new threat value, and generate strategy optimization data.

9. The method for assessing and controlling the dynamic threat of unmanned aerial vehicles (UAVs) in non-visual flight as described in claim 1, characterized in that, The method also includes feedback adjustments to the generation process of the comprehensive flight risk assessment results: Based on the optimized data according to the strategy, the weight coefficients of the preset weight allocation fusion model are adjusted.

10. A dynamic threat assessment and control system for unmanned aerial vehicles (UAVs) in non-visual flight, applied to the dynamic threat assessment and control method for unmanned aerial vehicles in non-visual flight as described in any one of claims 1-9, characterized in that... The system includes: The status monitoring module is used to acquire multi-source sensor data during the drone's flight in real time and generate the drone's own status baseline parameters. The threat quantification module is used to quantify preset threat sources based on the multi-source sensor data, calculate the offset between the quantified preset threat source and the preset safety threshold, and merge the offsets to generate a multi-source quantified threat dataset, wherein the multi-source quantified threat dataset includes meteorological threat values, equipment failure risk threat values ​​and airspace conflict risk threat values. The risk fusion module is used to input the UAV's own state baseline parameters and the multi-source quantitative threat dataset into a preset weight allocation fusion model to generate a comprehensive flight risk assessment result. The strategy decision module is used to calculate the flight control target value of the corresponding threat item based on the comprehensive flight risk assessment results, and combine the flight control target values ​​to generate an adaptive flight control strategy. The instruction issuing module is used to generate a dynamic adjustment control instruction set based on the adaptive flight control strategy and issue it to the UAV, setting the UAV heading angle, UAV altitude value and UAV speed value; The execution feedback module is used to generate strategy optimization data based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution; The adaptive update module is used to generate strategy optimization data based on the execution effect of the dynamically adjusted control instruction set, the threat values ​​of threat items before and after execution.

Images