Unmanned aerial vehicle and flight dynamic multi-dimensional safety management and cooperative response system thereof

Abstract

The invention provides an unmanned aerial vehicle and a flight dynamic multi-dimensional safety management and cooperative response system thereof, and relates to the technical field of unmanned aerial vehicle flight control. By constructing and cooperatively operating an environment stability index model, an electric quantity state safety early warning model, a communication link health degree model, an obstacle avoidance state health degree model and a flight speed safety evaluation model, real-time fusion evaluation is performed on parameters such as wind speed and obstacle avoidance of the unmanned aerial vehicle, and multi-dimensional dynamic safety situation awareness and monitoring processing are realized. A mixed monitoring period is configured for different parameters based on an event-driven and hierarchical response mechanism so as to optimize resources; and when the key parameters suddenly change severely, conventional polling is interrupted, and emergency evaluation is triggered. And processing the concurrent events according to a preset rule to ensure that the highest risk is preferentially handled. The upgrade from passive polling to active sensing response is realized, and the homeward voyage decision accuracy, the emergency response speed and the overall flight safety of the unmanned aerial vehicle in a complex environment are improved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) flight control technology, specifically to a UAV and its flight dynamic multi-dimensional safety management and collaborative response system. Background Technology

[0002] In recent years, the application of drone technology has deepened across various industries, and the management of its flight safety has become a core concern. Existing drone safety management solutions mostly rely on threshold monitoring of single or a few flight parameters and adopt a passive response mechanism of periodic polling, which exposes systemic defects in increasingly complex operating environments.

[0003] First, in terms of risk assessment, existing technologies typically monitor single parameters in isolation, such as remaining battery power and GPS positioning signal, and trigger alarms or return-to-home based on static thresholds (e.g., a fixed percentage of remaining battery power). However, drone flight safety is a complex state determined by multiple factors, including environment, equipment, communication, and mission. For example, in strong winds, even with sufficient battery power, the positioning signal may become unstable due to violent shaking of the drone; in complex electromagnetic environments, the interruption of a single communication link does not equate to complete loss of control, and safe flight can still be maintained if the backup link is functioning normally. This single-dimensional monitoring mode lacks a fusion perception and comprehensive assessment of the overall safety status of the aircraft, making it highly susceptible to misjudgments due to incomplete assessment dimensions, or overlooking potential risks caused by the coupling of multiple parameters, which is one of the important causes of drone crashes.

[0004] Secondly, the static strategy of using fixed thresholds in setting decision thresholds is severely out of touch with actual flight dynamics. Taking power management as an example, the traditional fixed percentage return-to-home strategy does not consider variables such as real-time headwind speed, remaining return-to-home distance, and climb power consumption. This may lead to two extreme consequences: first, premature return-to-home in tailwind, short-range scenarios, severely wasting operational time and endurance resources; second, decision delays in headwind, long-range scenarios, causing the drone to fail to return safely and resulting in property damage. Static thresholds cannot adapt to the ever-changing actual flight conditions, and their decision-making accuracy and reliability have fundamental bottlenecks.

[0005] Finally, regarding system response mechanisms, existing solutions generally rely on fixed-period polling checks (e.g., once per second). This mechanism, which checks at set intervals regardless of demand, inherently suffers from response delays in the face of emergencies such as sudden communication link interruptions, approaching obstacles, or sudden drops in GPS signal. A risk event may occur instantaneously after a single polling check, but the system must wait for the next detection cycle to detect it, potentially missing the optimal intervention window and making the incident unavoidable. This mechanism is inherently passive and lagging, failing to meet the urgent need for real-time, proactive responses to security events in highly dynamic environments.

[0006] In view of the above, this application is hereby submitted. Summary of the Invention

[0007] This invention provides a multi-dimensional safety management and collaborative response system for unmanned aerial vehicles (UAVs) and their flight dynamics, which can at least partially improve the above-mentioned problems.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A flight dynamic multi-dimensional safety management and collaborative response system is applied to unmanned aerial vehicles (UAVs). The system includes a multi-dimensional safety management hub, a multi-dimensional management model group, and an event response engine. The output of the multi-dimensional safety management hub is electrically connected to the input of the multi-dimensional management model group, and the output of the multi-dimensional management model group is electrically connected to the input of the event response engine. The multi-dimensional security management hub is configured to perform the following steps by executing a computer program stored internally: According to the current monitoring cycle, the environmental parameters collected by the acquisition components configured on the UAV are obtained, and the environmental parameters are transmitted to the multi-dimensional management model group for data monitoring and processing to generate multiple monitoring results. The multi-dimensional management model group includes an environmental stability index model, a power status safety early warning model, a communication link health model, an obstacle avoidance status health model, and a flight speed safety assessment model. The event response engine is invoked to process the monitoring results according to a preset priority order. The response processing includes pilot attention, forced return to home, manual takeover, and threshold adjustment. The environmental parameters collected in this round are compared with those collected in the previous round. If the change in the environmental parameter is greater than the first preset value but less than the second preset value, it indicates that the environmental parameter has changed significantly, and the current monitoring cycle is adjusted to a high-frequency cycle. If the change in the environmental parameter is not greater than the first preset value, it indicates that the environmental parameter has changed slowly, and the current monitoring cycle is maintained as a low-frequency cycle.

[0009] The present invention also provides an unmanned aerial vehicle (UAV), comprising: the UAV body and the flight dynamic multi-dimensional safety management and collaborative response system as described in any of the above.

[0010] In summary, this invention provides a novel intelligent management and collaborative response solution for unmanned aerial vehicles (UAVs) based on a multi-dimensional dynamic safety model. Compared with existing technologies, the core contribution of this invention lies in constructing a closed-loop safety management system that extends from multi-dimensional perception to intelligent decision-making and efficient execution. Firstly, through the collaborative work of an environmental stability index model, a dynamic power warning model, a dual-link communication evaluation model, a three-dimensional obstacle avoidance health model, and a dynamic speed risk assessment model, the system achieves, for the first time, a comprehensive and integrated real-time assessment of the flight environment, energy, communication, perception, and motion status, fundamentally overcoming the limitations and misjudgment risks of traditional single-parameter monitoring. Secondly, this invention abandons the inherent defects of static threshold decision-making and innovatively introduces a dynamic calculation algorithm, enabling key safety thresholds (such as return-to-home power) to be dynamically adjusted according to real-time variables such as flight distance and wind speed, greatly improving the accuracy and adaptability of decision-making. Finally, this invention designs a concurrent processing mechanism that combines hierarchical response and event-driven approaches. By configuring differentiated monitoring frequencies for different parameters to optimize resources and establishing an instant response channel for sudden changes in key parameters, supplemented by central priority arbitration logic, it ensures that the system can orderly and without delay prioritize the handling of the highest level threats when facing multiple concurrent risks, thus achieving a fundamental leap from passive polling to proactive perception and rapid intervention.

[0011] This system has achieved significant improvements in key indicators such as return-to-home decision accuracy, emergency response delay, and crash rate in complex environments. It not only provides a systematic solution for the safe and reliable operation of industrial-grade drones in complex scenarios, but its design principles of multi-model collaboration, dynamic thresholds, and intelligent response mechanisms also offer a beneficial technological paradigm for the broader field of intelligent mobile device security management. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the framework of the flight dynamic multi-dimensional safety management and collaborative response system provided in the first embodiment of the present invention.

[0013] Figure 2 This is a flowchart illustrating the environmental stability index model provided in this embodiment of the invention. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0015] refer to Figure 1As shown, the first embodiment of the present invention discloses a flight dynamic multi-dimensional safety management and collaborative response system, applied to unmanned aerial vehicles (UAVs), which includes: a multi-dimensional safety management center, a multi-dimensional management model group, and an event response engine. The output end of the multi-dimensional safety management center is electrically connected to the input end of the multi-dimensional management model group, and the output end of the multi-dimensional management model group is electrically connected to the input end of the event response engine. The multi-dimensional security management hub is configured to perform the following steps by executing a computer program stored internally: S1. According to the current monitoring cycle, acquire the environmental parameters collected by the acquisition components configured on the UAV, and transmit the environmental parameters to the multi-dimensional management model group for data monitoring and processing to generate multiple monitoring results. The multi-dimensional management model group includes an environmental stability index model, a power status safety early warning model, a communication link health model, an obstacle avoidance status health model, and a flight speed safety assessment model. S2, invoke the event response engine to process the monitoring results according to the preset priority order. The response processing includes pilot attention, forced return to home, manual takeover, and threshold adjustment. The priority order is: insufficient battery > obstacle avoidance failure > overspeed risk > communication anomaly.

[0016] Please see Figure 2 In this embodiment, preferably, according to the current monitoring cycle, the environmental parameters collected by the acquisition components configured on the UAV are obtained, and the environmental parameters are transmitted to the environmental stability index model of the multi-dimensional management model group for data monitoring and processing. Specifically, according to the current monitoring cycle, the environmental parameters collected by the acquisition components configured on the UAV are obtained, wherein the environmental parameters include the number of GPS satellites acquired, the number of RTK satellites acquired, wind speed, obstacle avoidance status data, current return distance, current speed, headwind speed, and signal quality. The positioning stability parameter Spos is calculated based on the number of GPS satellites acquired and the number of RTK satellites acquired. The formula is: Spos = 0.7 IGPS +0.3 IRTK, IGPS = GPS satellite search count / 20, IRTK = RTK satellite search count / 30; The three-dimensional obstacle avoidance stability parameter Sobstacle is calculated based on the obstacle avoidance state data. The formula is: Sobstacle = Horizontal obstacle avoidance state 0.5 + Upward obstacle avoidance status 0.3 + Downward obstacle avoidance state 0.2, wherein the obstacle avoidance status data includes horizontal obstacle avoidance status, upward obstacle avoidance status, and downward obstacle avoidance status; The wind speed risk factor Wrisk is calculated based on wind speed, with the formula: Wrisk = wind speed / 20. The environmental stability index ESI is then calculated based on the positioning stability parameter Spos, the three-dimensional obstacle avoidance stability parameter Sobstacle, and the wind speed risk factor Wrisk, with the formula: ESI = 0.4. Spos + 0.3 Sobstacle - 0.3 Wrisk, where the output range of the Environmental Stability Index (ESI) is specified as [0, 1]. When the Environmental Stability Index (ESI) is determined to be ≥0.7, the monitoring result output by the Environmental Stability Index model is considered to be that the environment is normal. When the Environmental Stability Index (ESI) is determined to be less than 0.7 (≤ 0.4), the monitoring result output by the Environmental Stability Index model is a Level II warning. When the Environmental Stability Index (ESI) is determined to be less than 0.4, the monitoring result output by the Environmental Stability Index model indicates an environmental emergency.

[0017] In this embodiment, various raw environmental parameters are continuously acquired from the data acquisition components configured on the UAV's onboard sensors and communication modules according to a preset current monitoring cycle. These parameters include, but are not limited to, the number of GPS satellites acquired, the number of RTK satellites acquired, real-time wind speed, horizontal / upward / downward obstacle avoidance status data from visual or infrared sensors, the current return distance calculated based on the current location and return point, the UAV's current speed, the headwind speed measured by meteorological data or sensors, and the 4G and SDR signal quality at both the air and ground ends. These real-time acquired environmental parameters are synchronously transmitted to a multi-dimensional management model group composed of five professional models for parallel data monitoring and processing, thereby generating multiple independent, quantitative monitoring results. This model group specifically includes an environmental stability index model, a battery status safety warning model, a communication link health model, an obstacle avoidance status health model, and a flight speed safety assessment model, which together constitute the technical foundation for comprehensive perception and assessment of flight safety status.

[0018] To achieve a comprehensive quantitative assessment of the flight environment, the model employs a weighted fusion algorithm. Specifically, during the execution of the Environmental Stability Index (ESI) model, after acquiring the relevant parameters according to the monitoring cycle, the positioning stability parameter Spos is first calculated based on the number of GPS and RTK satellite acquisitions. This weighted calculation method accurately and quantitatively reflects the signal quality of the fused positioning system. Subsequently, the overall health of the UAV obstacle avoidance system is dynamically characterized based on obstacle avoidance status data from three dimensions: horizontal, look-up, and look-down. Simultaneously, the wind speed risk factor Wrisk is introduced. Finally, the Environmental Stability Index (ESI) is calculated using the formula ESI = 0.4. Spos + 0.3 Sobstacle - 0.3 Wrisk calculates that its output value range is normalized to the interval [0, 1]. This design integrates the originally complex and independent positioning, obstacle avoidance, and meteorological parameters into an intuitive and comprehensive quantitative index, effectively overcoming the shortcomings of traditional methods that focus on only a single dimension and thus lead to misjudgments. It should be noted that the wind speed risk factor Wrisk can be replaced with a temperature influence factor or a pressure stability factor according to specific application scenarios (such as high altitude, chemical industrial parks), making the model adaptable to special meteorological monitoring needs.

[0019] The effectiveness of this model is directly reflected in its output monitoring results: when the calculated Environmental Stability Index (ESI) is ≥ 0.7, the model outputs a monitoring result indicating a normal environment, signifying overall stable flight conditions; when the ESI value falls between 0.4 and 0.7 (i.e., 0.4 ≤ ESI < 0.7), the model outputs a monitoring result of Level 2 warning, at which point the system will trigger a prompt alarm (such as a pop-up window) to remind the pilot to pay attention to potential environmental risks, achieving early warning of risks and providing valuable buffer time for manual intervention; and when the ESI is < 0.4, it indicates that the risk is too high, and the model outputs a monitoring result of the highest level of environmental emergency, at which point the system will automatically execute a forced return to base or require immediate manual takeover according to preset rules. This mechanism, as a global safety fallback strategy, can significantly reduce the risk of drone crashes in extremely complex environments.

[0020] In this embodiment, preferably, the environmental parameters are transmitted to the power status safety early warning model of the multi-dimensional management model group for data monitoring and processing. Specifically, the theoretical power required for return trip, Bneed, is calculated based on the current return distance D and the current speed. The safe return-to-base charge Bsafe is calculated using the formula Bsafe = Bneed + safety redundancy + distance correction + headwind compensation, where safety redundancy is 5%, and distance correction = D / 10000. 0.3%, headwind compensation = Vwind 0.5%, Vwind is the headwind speed; When the real-time battery level Bcurrent is less than the safe return battery level Bsafe, the monitoring result output by the battery status safety warning model is a warning or an emergency.

[0021] In this embodiment, to address the limitations of a fixed power threshold, this model pioneers a dynamic safety power calculation method. The power status safety warning model in the multi-dimensional management model group also synchronously monitors and processes data. The key environmental parameters it receives include the current return distance D provided by the flight control system or the mission planning module, the current speed, and the headwind speed Vwind calculated from meteorological sensors or flight control status data. The core innovation of the model is that it completely changes the traditional static return decision logic based on a fixed power percentage threshold.

[0022] Specifically, the model first calculates the minimum power Bneed theoretically required to complete this return mission based on the current return distance D and the current speed of the UAV (usually the preset safe return speed or the current cruising speed), using the built-in power consumption - distance - speed mapping algorithm or a look-up table method. Subsequently, based on the dynamic safety power calculation method, the model further calculates the safe return power Bsafe. Here, the safety redundancy is a preset empirical value, usually set at 5% of the total power, to cope with unforeseen fluctuations during the return journey; the distance correction term is calculated by the formula "D / 10000 × 0.3%", which means that for every additional 10,000 meters of return distance, the safe power threshold needs to be increased by 0.3% accordingly, thus quantifying the additional energy consumption risk brought by long-distance flight; the headwind compensation term is calculated by the formula "Vwind × 0.5%", meaning that for every additional unit of headwind speed, the safe power threshold needs to be increased by 0.5%, which scientifically reflects the physical fact that headwind flight will significantly increase energy consumption. By introducing these two real-time variables of distance and wind speed, the present invention for the first time realizes the dynamic and context-based calculation of the return power threshold.

[0023] When the model real-time obtains the remaining power Bcurrent of the UAV through sensors and determines that Bcurrent < Bsafe, it indicates that the current power has fallen below the dynamic safety line comprehensively considering the mission distance and environmental resistance. At this time, the power status safety warning model will output a monitoring result at the warning or emergency level. This monitoring result is then sent to the central priority arbitration module. Since power shortage is defined as the highest level in the response priority of the present invention (i.e., power shortage > obstacle avoidance failure > overspeed risk > communication anomaly), once this monitoring result is triggered, the central module will generate and issue an automatic return command immediately, prior to any other concurrent events. This forward-looking decision-making mechanism based on dynamic calculation effectively avoids the waste of operation time and endurance resources caused by premature return in downwind and short-distance scenarios using traditional methods. More importantly, it fundamentally prevents power exhaustion and crash accidents caused by the failure of static thresholds to give timely warnings under adverse conditions such as headwind and long distance, thus significantly improving the accuracy and reliability of the return decision.

[0024] In short, the power status safety early warning model introduces real-time variables, transforming the power management strategy from static to dynamic. This significantly improves the accuracy and reliability of return-to-base decisions and effectively avoids equipment losses caused by inaccurate power estimation.

[0025] In this embodiment, preferably, the environmental parameters are transmitted to the communication link health model of the multi-dimensional management model group for data monitoring and processing. Specifically, this involves calculating a first signal quality value H4G and a second signal quality value HSDR based on signal quality, using the formula: H4G = 0.6. Antenna 4G signal quality +0.4 Ground-side 4G signal quality, HSDR = mean of SDR signal quality; When H4G < 3 and HSDR < 3 or |H4G - HSDR| > 2, the monitoring result output by the communication link health model is a warning.

[0026] In this embodiment, the communication link health model in the multi-dimensional management model group also monitors and evaluates the status of the UAV communication system in real time. The environmental parameters it processes are mainly signal strength or signal-to-noise ratio data that characterize communication quality, specifically including the 4G signal quality from the UAV's onboard communication module, the 4G signal quality from the ground control station, and the signal quality of the software-defined radio (SDR) link. The core design idea of ​​this model is to abandon the absolute dependence on a single communication link and instead comprehensively judge the overall health of the communication system through dual-link evaluation and difference comparison, thereby improving the control reliability in complex electromagnetic environments. In specific implementation, the model first calculates the first signal quality value H4G based on the received raw signal quality data through weighted fusion. This weight allocation reflects the design consideration of giving higher attention to the overhead signal (which directly reflects the communication environment of the UAV's location). At the same time, the model calculates the second signal quality value HSDR, which is the average value of the SDR signal quality within a preset time window to smooth out instantaneous fluctuations and more stably reflect the performance of the backup link.

[0027] After completing the calculation, the model applies a composite judgment logic to generate monitoring results: when both H4G < 3 and HSDR < 3 are simultaneously determined, it indicates that both the primary 4G link and the backup SDR link are in a poor quality state, and the entire communication system faces a high risk of failure; or, when the absolute difference between the signal quality values ​​of the two links is too large, i.e., |H4G - HSDR| > 2, it indicates that there is a serious inconsistency in the status of the primary and backup links, which may indicate that the data of one of the links is unreliable, or that there are abnormal situations such as local strong interference. As long as either of the above conditions is met, the communication link health model outputs a warning-level monitoring result. This dual-link collaborative evaluation mechanism effectively solves the shortcomings of traditional solutions that cannot identify hidden risks such as inconsistencies in the status of links or simultaneous degradation due to monitoring only a single link. After the warning result is output, it will be sent to the central priority arbitration module. According to the preset hierarchical response mechanism, the priority of communication anomalies is set after insufficient power, obstacle avoidance failure, and overspeed risk. This means that when no higher-level risk events occur concurrently, the communication warning will trigger the system to suggest manual intervention by the pilot or initiate a conservative automatic return-to-home strategy. If higher-level events exist simultaneously, their processing order will be postponed, thus ensuring that system resources and response actions are always focused on the most pressing safety threats. By implementing this model, the control reliability and situational awareness integrity of the UAV system in complex electromagnetic environments or scenarios with fluctuating signals such as urban buildings and canyons have been significantly improved, providing a solid communication foundation for safe operations.

[0028] It should be noted that in the communication link health model, the absolute value comparison method (|H4G-HSDR|>2) for determining link differences can be replaced by the ratio method (such as H4G / HSDR>th threshold or vice versa). This method may have higher detection sensitivity for differences in the status of primary and backup links in some scenarios.

[0029] In this embodiment, preferably, the environmental parameters are transmitted to the obstacle avoidance status health model of the multi-dimensional management model group for data monitoring and processing. Specifically, the health status Hobstacle is calculated based on the three-dimensional obstacle avoidance stability parameter Sobstacle, and the calculation formula is: Hobstacle = Sobstacle 100% = (Horizontal obstacle avoidance state) 0.5 + Upward obstacle avoidance status 0.3 + Downward obstacle avoidance state 0.2) 100%; When the Hobstacle health level is determined to be less than 50%, the monitoring result output by the obstacle avoidance health level model is "emergency".

[0030] In this embodiment, the obstacle avoidance health model continuously monitors the reliability of the perception system. The environmental parameters processed by this model are directly related to the UAV's active obstacle avoidance capability, such as obstacle avoidance status data in the horizontal, upward, and downward directions provided by onboard visual, infrared, or ultrasonic sensors. In a preferred embodiment of the invention, the core processing logic of this model is highly integrated and efficient. Specifically, the model directly utilizes the three-dimensional obstacle avoidance stability parameter Sobstacle, which has already been calculated in the Environmental Stability Index (ESI) model. This parameter itself is a weighted fusion result of obstacle avoidance status data in the horizontal, upward, and downward dimensions. Based on this, the obstacle avoidance health model calculates the health Hobstacle of the obstacle avoidance system using a simple conversion formula: Hobstacle = Sobstacle 100%. This design cleverly reuses existing intermediate calculation results, avoiding redundant data processing and wasting computational resources, reflecting the overall optimization concept of the system design. The calculated health status (Hobstacle) is an intuitive percentage value that represents the overall integrity rate of the UAV's 3D obstacle avoidance function.

[0031] Subsequently, a threshold judgment is made on the health status: when the health status Hobstacle is less than 50%, it indicates that more than half of the core obstacle avoidance functions (weighted) are in an abnormal or failed state, and the drone's obstacle perception and avoidance capabilities in the corresponding direction are severely impaired. At this time, the obstacle avoidance health status model will output the highest level emergency monitoring result. After this emergency result is output, it will be immediately transmitted to the central priority arbitration module. According to the response priority order clearly defined in this invention, this result is triggered. Unless the system faces a more serious low battery alarm at the same time, the central module will prioritize handling this obstacle avoidance failure event. The usual approach is to immediately generate and execute an automatic emergency return command to allow the drone to quickly leave the complex airspace where there may be obstacles, thereby effectively preventing the collision risk caused by the simultaneous or successive failure of multiple sensors. This veto-based judgment based on the overall health status percentage, combined with the hierarchical response mechanism, adds a crucial layer of insurance to the drone's active safety system, significantly improving its operational safety in scenarios with extremely high obstacle avoidance requirements, such as near-ground, obstacle-penetrating, or indoor environments.

[0032] In this embodiment, preferably, the environmental parameters are transmitted to the flight speed safety assessment model of the multi-dimensional management model group for data monitoring and processing. Specifically, based on the current speed, the speed risk index Rspeed is calculated, where Rspeed = (current speed / preset speed). (1 + headwind compensation coefficient), and adjust the preset threshold based on the current positioning and obstacle avoidance correction parameters; When the speed risk index Rspeed is determined to be ≥1.3, the monitoring result output by the flight speed safety assessment model is a warning.

[0033] In this embodiment, the key environmental parameters processed by the model include the current speed provided by the flight control system, the preset safe flight speed (which can be set according to different flight modes or airspace regulations), the coefficient for wind condition compensation obtained from meteorological data or sensors, and the current state parameters from the positioning and obstacle avoidance system for dynamically adjusting the judgment benchmark. In specific implementation, the model first dynamically calculates the speed risk index Rspeed based on the ratio of the current speed to the preset speed, combined with environmental influencing factors. The introduction of the headwind compensation coefficient scientifically reflects the change in the relationship between the UAV's airspeed and ground speed in headwind conditions. The same indicated airspeed may correspond to a higher ground speed and kinetic energy, thus increasing the actual risk, making the risk assessment more consistent with actual flight mechanics.

[0034] Furthermore, the model does not use a fixed threshold for judgment. Instead, it dynamically fine-tunes the risk index threshold used to trigger warnings by combining current positioning stability parameters (such as the real-time value of the positioning stability parameter Spos) and obstacle avoidance health parameters (such as the real-time value of the health status Hobstacle). For example, when the positioning signal is good and the obstacle avoidance system is fully functional, the system can tolerate a slightly higher speed risk; conversely, when the positioning is unstable or the obstacle avoidance function is degraded, a more conservative speed threshold is adopted. This design reflects the intelligent concept of linking risk assessment with the overall flight situation.

[0035] After the model completes calculations and threshold adjustments, a risk assessment is performed: when the speed risk index Rspeed ≥ 1.3, it indicates that the drone's current speed has significantly exceeded the safe range dynamically adjusted based on wind conditions and system status. The flight speed safety assessment model then outputs a warning-level monitoring result. This speed hazard warning is sent to the central priority arbitration module. According to the hierarchical response mechanism explicitly defined in this invention, overspeed risk has a higher priority than communication anomalies but lower than obstacle avoidance failure and insufficient battery power. This means that when this warning is triggered and there are no higher-priority concurrent events, the system will immediately take measures such as displaying a speed hazard warning pop-up on the control interface and forcibly requiring manual takeover by the pilot. This ensures timely intervention on speed, a key parameter directly affecting collision energy and control margin, preventing accidents caused by excessive speed and inability to brake or avoid obstacles in sudden situations. By implementing this model, the system upgrades flight speed monitoring from static threshold monitoring to dynamic risk assessment, intelligently correcting based on environmental and system status, significantly improving the proactive control capability for high-speed flight risks, especially under complex weather and perception conditions.

[0036] It should be noted that, to ensure the system can respond in an orderly and efficient manner under multiple concurrent alarms, this solution defines a clear response priority order. The central priority arbitration module will process concurrent events according to this rule, ensuring that the highest-risk event receives the highest priority.

[0037] S3. Compare the environmental parameters collected in this round with those collected in the previous round. If the change in the environmental parameters is greater than the first preset value but less than the second preset value, it indicates that the environmental parameters have changed significantly, and the current monitoring cycle is adjusted to a high-frequency cycle. If the change in the environmental parameters is not greater than the first preset value, it indicates that the environmental parameters have changed slowly, and the current monitoring cycle is maintained as a low-frequency cycle.

[0038] In this embodiment, to ensure the system's real-time response capability while achieving efficient utilization of computing and communication resources, the system also includes an adaptive hybrid monitoring cycle dynamic adjustment mechanism. The core of this mechanism lies in intelligently switching the frequency of data acquisition and model evaluation based on the drastic changes in environmental parameters; that is, employing a hybrid monitoring cycle to optimize computing resources.

[0039] During system operation, the control module continuously records snapshots of environmental parameters acquired by the acquisition components in each monitoring cycle. Before starting a new monitoring cycle, the system compares the expected values ​​of the environmental parameters to be acquired in this cycle with the historical values ​​of the environmental parameters already acquired and stored in the previous cycle, calculating the changes in key parameters within adjacent cycles. These key parameters specifically refer to variables that change relatively rapidly and are sensitive to sudden risks, such as the number of GPS satellite acquisitions, the number of RTK satellite acquisitions, wind speed, and raw readings from obstacle avoidance sensors. The system presets a threshold value, or first preset value, for the changes in these parameters.

[0040] When the comparison determines that the change in any key parameter is greater than the first preset value but less than the second preset value, the system determines that the current flight environment or equipment status is changing rapidly, with large changes in environmental parameters and increased potential risks. To more closely capture these dynamic changes, the system will immediately adjust the current monitoring cycle from the conventional low-frequency cycle (e.g., 1 second) to a high-frequency cycle (e.g., 200 milliseconds). By increasing the sampling and evaluation frequency, the system can more promptly detect emergencies such as sudden signal drops, sudden wind speed changes, and sudden approach of obstacles, providing a more timely data foundation for the event-driven response mechanism, thereby effectively shortening the delay from the occurrence of an anomaly to its detection by the system.

[0041] Conversely, when the comparison determines that the changes in all key parameters are not greater than the first preset value, it indicates that the current environmental parameters are changing slowly and the flight status is relatively stable. In this case, the system determines that the necessity of maintaining high-frequency monitoring is reduced and will continue to maintain or switch back to the low-frequency normal monitoring cycle. This strategy of allocating monitoring resources on demand avoids the increase in processor load and energy consumption caused by unnecessary high-frequency calculations during stable flight phases, making the overall system operation more economical and efficient, and ensuring that the system can maintain high real-time performance under low power consumption. This hybrid monitoring cycle mechanism complements the event-driven channel: periodic comparison provides the basis for frequency adaptation under normal conditions, while event-driven operation ensures zero-wait response to sudden changes. The combination of the two achieves the optimal balance between low-power normal operation and high real-time emergency response, ensuring zero-delay processing of critical safety events.

[0042] In this embodiment, preferably, it further includes: when it is determined that the change of a key parameter in the environmental parameters is greater than a second preset value, it indicates that the environmental parameters have changed drastically, interrupts the regular polling, and directly triggers the relevant model corresponding to the parameter to perform emergency assessment and response.

[0043] In this embodiment, in addition to the hybrid monitoring cycle dynamic adjustment mechanism, the system also constructs an independent and higher-priority event-driven emergency response channel. This channel, as a fundamental supplement to and transcendence of periodic polling monitoring, focuses on handling sudden and dramatic changes in environmental parameters. Specifically, during operation, the system continuously monitors key parameters among the environmental parameters, such as the number of GPS satellites acquired, the number of RTK satellites acquired, and the strength of 4G or SDR signals. The system sets a more stringent judgment threshold, namely a second preset value, for the instantaneous change amplitude of these key parameters. This value is typically significantly greater than the first preset value used to trigger monitoring cycle adjustments. When the system determines in real time at any point that the change in any key parameter (e.g., between two adjacent sample values, or the deviation from a recent baseline value) exceeds the second preset value—for example, if the number of GPS satellites acquired decreases by ≥3 in a very short time—the system will immediately determine that a drastic change in environmental parameters has occurred, indicating a possible emergency situation such as a momentary signal blockage, strong electromagnetic interference, or encounter with a sudden obstacle. At this time, the system will no longer wait for the currently executing or upcoming regular polling monitoring cycle but will immediately interrupt the regular polling process. After the interruption, the system directly triggers the specific safety model associated with the drastic change in parameters to conduct an emergency assessment and response.

[0044] For example, if the number of GPS satellites acquired drops sharply, the Environmental Stability Index (ESI) model is immediately recalculated; if the communication signal strength drops precipitously, the communication link health model is immediately assessed. This direct response path completely bypasses fixed time scheduling, significantly reducing the response delay to sudden anomalies. The results of the emergency assessment (such as an ESI value dropping below 0.4) are fed directly into the central priority arbitration module, just like a highest-level alarm. This module, based on preset response priorities, immediately makes decisions and drives the actuators (such as the flight control system) to take the most decisive measures, such as forced return or switching control modes. The introduction of this event-driven mechanism enables the system to achieve a fundamental paradigm shift from passive, timed queries to proactive, instantaneous perception and response, qualitatively improving its ability to handle emergencies such as sudden signal drops and sudden obstacles. It significantly reduces the response delay to sudden anomalies, ensuring that the handling of the most dangerous situations, such as sudden signal drops and approaching obstacles, can be initiated at millisecond speeds.

[0045] In summary, this invention provides a systematic multi-dimensional safety management model and collaborative response scheme based on UAV flight dynamics. The core of this scheme lies in constructing a multi-dimensional management model group that collaboratively works with an environmental stability index model, a battery status safety early warning model, a communication link health model, an obstacle avoidance health model, and a flight speed safety assessment model. Through real-time acquisition and fusion processing of multi-source environmental parameters such as GPS, RTK, wind speed, obstacle avoidance status, battery level, communication signals, and flight speed, it achieves a leap from single-parameter monitoring to comprehensive dynamic assessment of UAV safety status.

[0046] Specifically, the environmental stability index model, through weighted fusion of positioning stability, three-dimensional obstacle avoidance stability, and wind speed risk, outputs a global environmental risk level, playing a crucial role in early warning and mandatory backup in extreme situations. The power status safety early warning model, through an innovative dynamic safe power calculation method, integrates return distance and headwind speed, transforming the return decision threshold from a static fixed value to a dynamic intelligent value, significantly improving decision accuracy. The communication link health model, through dual-link evaluation and difference comparison, effectively overcomes dependence on a single communication link. The obstacle avoidance health model and flight speed safety assessment model provide safety barriers in three-dimensional space from the perspectives of perception system integrity and motion status risk, respectively. More importantly, this invention designs a highly efficient and collaborative intelligent response mechanism. The system adopts a hybrid monitoring cycle strategy, adaptively adjusting the evaluation frequency according to the drastic changes in parameters to optimize resources; simultaneously, it creatively introduces an event-driven channel, immediately interrupting regular polling and triggering an emergency assessment when key parameters undergo drastic changes, achieving instantaneous response to sudden risks. All monitoring results output by the model are uniformly arbitrated and processed by a central priority arbitration module according to a clear hierarchical order of insufficient power > obstacle avoidance failure > overspeed risk > communication anomaly, ensuring orderly response and the principle of prioritizing the highest risk under multiple concurrent alarms.

[0047] Through the deep integration and synergistic operation of the aforementioned multi-dimensional dynamic evaluation model and hierarchical response mechanism, this invention successfully upgrades the UAV safety management model from a passive, lagging, and isolated traditional paradigm to a proactive, real-time, and systematic intelligent paradigm. The differences between traditional solutions and this system are shown in Table 1.

[0048] Table 1

[0049] Compared with existing technologies, the advantages of this invention are: 1. Multi-dimensional dynamic evaluation: By fusing real-time data from five major models, the blind spots of single-parameter decision-making are overcome, and the overall risk of system crash is significantly reduced. 2. Dynamic threshold calculation: Key decision thresholds such as power consumption can be dynamically adjusted according to real-time tasks and the environment, greatly improving the level of intelligence. 3. Concurrent event response optimization: Based on priority arbitration and hybrid triggering mechanisms, system resources are utilized efficiently, ensuring zero-delay processing of critical safety events.

[0050] A second embodiment of the present invention provides an unmanned aerial vehicle (UAV) including the UAV body and the flight dynamic multi-dimensional safety management and collaborative response system as described in any of the above embodiments.

[0051] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A flight dynamic multi-dimensional safety management and collaborative response system, applied to unmanned aerial vehicles (UAVs), characterized in that, include: The system comprises a multi-dimensional security management hub, a multi-dimensional management model group, and an event response engine. The output of the multi-dimensional security management hub is electrically connected to the input of the multi-dimensional management model group, and the output of the multi-dimensional management model group is electrically connected to the input of the event response engine. The multi-dimensional security management hub is configured to perform the following steps by executing a computer program stored internally: According to the current monitoring cycle, the environmental parameters collected by the acquisition components configured on the UAV are obtained, and the environmental parameters are transmitted to the multi-dimensional management model group for data monitoring and processing to generate multiple monitoring results. The multi-dimensional management model group includes an environmental stability index model, a power status safety early warning model, a communication link health model, an obstacle avoidance status health model, and a flight speed safety assessment model. The event response engine is invoked to process the monitoring results according to a preset priority order. The response processing includes pilot attention, forced return to home, manual takeover, and threshold adjustment. The environmental parameters collected in this round are compared with those collected in the previous round. If the change in the environmental parameter is greater than the first preset value but less than the second preset value, it indicates that the environmental parameter has changed significantly, and the current monitoring cycle is adjusted to a high-frequency cycle. If the change in the environmental parameter is not greater than the first preset value, it indicates that the environmental parameter has changed slowly, and the current monitoring cycle is maintained as a low-frequency cycle.

2. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 1, characterized in that, According to the current monitoring cycle, environmental parameters collected by the acquisition components configured on the drone are obtained, and these environmental parameters are transmitted to the environmental stability index model of the multi-dimensional management model group for data monitoring and processing. Specifically: According to the current monitoring cycle, the environmental parameters collected by the acquisition components configured on the drone are obtained. These environmental parameters include the number of GPS satellites acquired, the number of RTK satellites acquired, wind speed, obstacle avoidance status data, current return distance, current speed, headwind speed, and signal quality. The positioning stability parameter Spos is calculated based on the number of GPS satellites acquired and the number of RTK satellites acquired. The formula is: Spos = 0.7 IGPS +0.3 IRTK, IGPS = GPS satellite search count / 20, IRTK = RTK satellite search count / 30; The three-dimensional obstacle avoidance stability parameter Sobstacle is calculated based on the obstacle avoidance state data. The formula is: Sobstacle = Horizontal obstacle avoidance state 0.5 + Upward obstacle avoidance status 0.3 + Downward obstacle avoidance state 0.2, wherein the obstacle avoidance status data includes horizontal obstacle avoidance status, upward obstacle avoidance status, and downward obstacle avoidance status; The wind speed risk factor Wrisk is calculated based on wind speed, with the formula: Wrisk = wind speed / 20. The environmental stability index ESI is then calculated based on the positioning stability parameter Spos, the three-dimensional obstacle avoidance stability parameter Sobstacle, and the wind speed risk factor Wrisk, with the formula: ESI = 0.

4. Spos + 0.3 Sobstacle - 0.3 Wrisk, where the output range of the Environmental Stability Index (ESI) is specified as [0, 1]. When the Environmental Stability Index (ESI) is determined to be ≥0.7, the monitoring result output by the Environmental Stability Index model is considered to be that the environment is normal. When the Environmental Stability Index (ESI) is determined to be less than 0.7 (≤ 0.4), the monitoring result output by the Environmental Stability Index model is a Level II warning. When the Environmental Stability Index (ESI) is determined to be less than 0.4, the monitoring result output by the Environmental Stability Index model indicates an environmental emergency.

3. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 2, characterized in that, The environmental parameters are transmitted to the power status safety early warning model of the multi-dimensional management model group for data monitoring and processing, specifically as follows: Based on the current return distance D and the current speed, the theoretical power consumption Bneed required for the return trip is calculated. The safe return-to-base charge Bsafe is calculated using the formula Bsafe = Bneed + safety redundancy + distance correction + headwind compensation, where safety redundancy is 5%, and distance correction = D / 10000. 0.3%, headwind compensation = Vwind 0.5%, Vwind is the headwind speed; When the real-time battery level Bcurrent is less than the safe return battery level Bsafe, the monitoring result output by the battery status safety warning model is a warning or an emergency.

4. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 2, characterized in that, The environmental parameters are transmitted to the communication link health model of the multi-dimensional management model group for data monitoring and processing, specifically as follows: The first signal quality value H4G and the second signal quality value HSDR are calculated based on signal quality, using the formula: H4G = 0.6 Antenna 4G signal quality +0.4 Ground-side 4G signal quality, HSDR = mean of SDR signal quality; When H4G < 3 and HSDR < 3 or |H4G - HSDR| > 2, the monitoring result output by the communication link health model is a warning.

5. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 2, characterized in that, The environmental parameters are transmitted to the obstacle avoidance status health model of the multi-dimensional management model group for data monitoring and processing, specifically as follows: The health status Hobstacle is calculated based on the three-dimensional obstacle avoidance stability parameter Sobstacle, and the formula is: Hobstacle = Sobstacle 100% = (Horizontal obstacle avoidance state) 0.5 + Upward obstacle avoidance status 0.3 + Downward obstacle avoidance state 0.2) 100%; When the Hobstacle health level is determined to be less than 50%, the monitoring result output by the obstacle avoidance health level model is "emergency".

6. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 2, characterized in that, The environmental parameters are transmitted to the flight speed safety assessment model of the multi-dimensional management model group for data monitoring and processing, specifically as follows: Calculate the speed risk index Rspeed based on the current speed: Rspeed = (current speed / preset speed) (1 + headwind compensation coefficient), and adjust the preset threshold based on the current positioning and obstacle avoidance correction parameters; When the speed risk index Rspeed is determined to be ≥1.3, the monitoring result output by the flight speed safety assessment model is a warning.

7. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 1, characterized in that, The priority order is: insufficient battery power > obstacle avoidance failure > overspeed risk > communication anomaly.

8. The flight dynamic multi-dimensional safety management and collaborative response system according to claim 1, characterized in that, Also includes: When it is determined that the change of a key environmental parameter is greater than the second preset value, it indicates that the environmental parameter has changed drastically. The regular polling is interrupted, and the relevant model corresponding to the parameter is directly triggered to conduct an emergency assessment and response.

9. A drone, characterized in that, It includes the unmanned aerial vehicle (UAV) itself and the flight dynamic multi-dimensional safety management and collaborative response system as described in any one of claims 1 to 8.

Images