A Dynamic Decision-Making Method and System for Unmanned Aerial Vehicle (UAV) Security Strategies
By employing trend prediction and a two-tiered evaluation mechanism, the problem of lagging and rigid decision-making in UAV safety strategies has been solved, enabling proactive management of UAV systems and improving both safety and cost-effectiveness.
Patent Information
- Application Number
- CN202610108554.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-26
- Publication Date
- 2026-05-26
AI Technical Summary
Existing drone safety strategy decision-making methods suffer from decision lag, rigidity, and one-sided assessment, failing to effectively cope with complex and continuously changing operating environments, resulting in insufficient safety and reliability.
By employing a trend prediction and two-layer evaluation mechanism, the failure time is predicted by monitoring the historical data sequence of the UAV system, and the feasibility and cost-effectiveness are evaluated. The early warning and execution strategies are dynamically adjusted to achieve proactive management.
It improves the safety and intelligence of the drone system, reduces the probability of safety accidents, and enhances the system's flexibility and economy.
Smart Images
Figure CN122086040A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned aerial vehicle (UAV) system control technology, and in particular to a dynamic decision-making method and system for UAV safety strategies. Background Technology
[0002] With the widespread application of drone technology in logistics, surveying, security and other fields, its flight safety and mission reliability have become crucial. Modern drone systems are typically equipped with redundancy sensors and fault detection and diagnosis modules, which can trigger preset emergency strategies, such as automatic return to home, emergency landing or hovering for rescue, when component failure or abnormal status is detected.
[0003] Currently, the mainstream drone safety strategy decision-making methods in the industry are mainly based on predefined rule bases or finite state machines, and their core workflow can be summarized in the following three steps:
[0004] Fault detection and status assessment: The system monitors the operating parameters (such as signal strength, battery voltage, and number of GPS satellites) of each subsystem (such as communication link, power system, and navigation system) in real time and compares them with preset, fixed safety thresholds. Once a parameter exceeds the threshold, it is determined to be a fault or abnormal state, thereby triggering the corresponding alarm flag.
[0005] Static strategy mapping: The system maintains a "fault-response" mapping table internally. This mapping relationship is static and pre-defined, and a specific fault signal will directly correspond to one or a set of fixed handling strategies.
[0006] Strategy execution: Once a strategy is triggered, the flight control system will unconditionally execute the preset strategy process until the strategy is completed or a new fault occurs.
[0007] Although the above solutions are simple to implement and can handle some typical faults, they are difficult to cope with the complex and continuously changing operating environments in the real world. Summary of the Invention
[0008] Based on the above analysis, the present invention aims to provide a dynamic decision-making method and system for drone safety strategies, in order to solve the problems of technological lag, rigidity and one-sided evaluation in existing dynamic decision-making for drone safety strategies.
[0009] On one hand, embodiments of the present invention provide a dynamic decision-making method for unmanned aerial vehicle (UAV) safety strategies, the method comprising the following steps: Continuously monitor the historical data sequence of at least one state parameter in the unmanned aerial vehicle system; Based on the historical data sequence, predict the expected time to failure (TTF) when the state parameter reaches a preset critical value in the future; Before the predicted failure occurs, when the expected failure time TTF ≤ warning time threshold N, an early warning is triggered and a two-layer evaluation mechanism is used to evaluate the strategies in the candidate strategy set to obtain a recommended strategy list. When the expected failure time TTF is less than or equal to the pre-execution time threshold M, the highest priority strategy in the recommended strategy list is executed without external intervention. Wherein, the warning time threshold N is greater than the pre-execution time threshold M.
[0010] Based on the above solution, the present invention also makes the following improvements: Furthermore, the two-layer evaluation mechanism is used to evaluate the strategies in the candidate strategy set to obtain a recommended strategy list, including: A feasibility assessment is performed on the candidate strategy set. The feasibility assessment is used to verify whether the resources required to execute the strategy exceed the current available resources of the UAV, and to filter out strategies that do not meet the preset feasibility constraints. For the candidate strategy set that has passed the feasibility assessment, a cost-effectiveness ratio quantitative evaluation is performed to obtain the cost-effectiveness ratio of each strategy. Based on the cost-effectiveness ratio, K strategies are selected from the candidate strategy set and sorted by cost-effectiveness ratio to form a recommended strategy list, where K is less than the total number of strategies in the candidate strategy set that has passed the feasibility assessment.
[0011] Furthermore, the feasibility assessment includes at least verifying one or more of the following: the power required for strategy execution and the current remaining power of the UAV; the range required for strategy execution and the current endurance of the UAV; and the availability of the navigation source on which the strategy depends.
[0012] Furthermore, the cost-effectiveness ratio quantitative evaluation is performed by using a cost-effectiveness ratio calculation model to quantitatively evaluate the strategies in the candidate strategy set that have passed the feasibility evaluation.
[0013] Furthermore, the cost-effectiveness calculation model is as follows: CER = ( ×Safety Benefits+ ×Task Rewards) / ( ×Resource Costs+ ×Time Cost) in, These are the weighting coefficients for security benefits, task benefits, resource costs, and time costs, respectively.
[0014] Furthermore, the execution process of the trend prediction algorithm includes: Assign weights to each data point in the historical data sequence; obtain the difference sequence of the historical data sequence, and calculate the trend slope based on the weights of each data point and the difference sequence; calculate the expected failure time based on the parameter value, trend slope, and preset critical value of the last time point in the historical data sequence.
[0015] Furthermore, the formulas for calculating the weights and the trend slope are as follows:
[0016] In the formula, Represents the first in the historical data sequence The weight of each data point This represents the total length of the historical data sequence. Index representing a data point ;
[0017] In the formula, The slope of the trend, representing the overall rate of change of a historical data series. Represents the first value in the difference sequence. The data point and the first The difference between the data points.
[0018] Furthermore, the predicted time to failure (TTF) for the state parameter to reach a preset critical value in the future is achieved by using a trend prediction algorithm that combines exponentially weighted moving average and linear regression. The formula for calculating the expected failure time is as follows:
[0019] In the formula, This represents the parameter value at the last point in time in the historical data sequence. This indicates a preset critical value.
[0020] Furthermore, the warning threshold N and the pre-execution threshold M are dynamically adjusted according to the environmental complexity and the criticality of the task.
[0021] On the other hand, embodiments of the present invention also provide a dynamic decision-making system for unmanned aerial vehicle (UAV) safety strategies, the system comprising: The data acquisition module is used to continuously monitor the historical data sequence of at least one state parameter in the UAV system; The failure time prediction module predicts the expected failure time (TTF) when the state parameter reaches a preset critical value in the future, based on the historical data sequence. The strategy evaluation module is used to trigger an early warning before the predicted failure occurs, when the expected failure time TTF is less than or equal to the early warning time threshold N, and to evaluate the strategies in the candidate strategy set using a two-layer evaluation mechanism to obtain a recommended strategy list. The strategy execution module is used to execute the highest priority strategy in the recommended strategy list without external intervention when the expected failure time TTF ≤ the pre-execution time threshold M. Wherein, the warning time threshold N is greater than the pre-execution time threshold M.
[0022] The beneficial effects of this technical solution are: The dynamic decision-making method for UAV safety strategy provided by this invention can identify and intervene before a failure occurs through trend prediction, gaining valuable time for decision-making and handling, and reducing the probability of safety accidents. By introducing a two-layer evaluation model of feasibility assessment and cost-effectiveness assessment, this invention can automatically eliminate infeasible strategies and select the optimal solution with the lowest overall cost and highest benefit from feasible strategies. The warning threshold N and pre-execution threshold M used in this invention can be dynamically configured or adaptively adjusted according to factors such as environmental complexity and mission criticality, enabling the system to flexibly respond to diverse flight scenarios and reduce false alarm rate.
[0023] In summary, this invention effectively addresses the inherent shortcomings of existing technologies, such as lag, rigidity, and one-sided assessment, through trend prediction, two-layer evaluation, and early warning triggering. It achieves a leap from passive response to proactive management, significantly improving the safety, economy, and intelligence level of unmanned aerial vehicle (UAV) systems, and has significant practical application value.
[0024] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description
[0025] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.
[0026] Figure 1 This is a flowchart of the dynamic decision-making method for drone safety strategy provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the structure of the dynamic decision-making system for drone safety strategy provided in Embodiment 2 of the present invention. Detailed Implementation
[0027] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.
[0028] The mainstream drone safety strategy decision-making methods in the industry are mainly based on predefined rule bases or finite state machines, but they have inherent defects, which are analyzed and explained in detail below: 1. Decision-making is delayed, making risk prevention impossible.
[0029] Existing technologies employ a passive "failure-on-demand" approach. The system must wait until parameters actually exceed the fault threshold before initiating a response procedure, completely missing the "risk window" before the fault occurs. For example, during the process of communication signals gradually weakening from good to interrupted, the system will not take any preventative measures until communication is completely lost before initiating the return-to-base procedure. By this time, the best opportunity for coordinated decision-making with the ground station may have been lost, and the delayed response may even exacerbate the risk.
[0030] 2. The strategy evaluation has only one dimension and the basis for decision-making is insufficient.
[0031] Existing solutions select strategies solely based on the Boolean condition of "whether a certain type of failure occurs," completely ignoring the feasibility of the strategy's execution and its overall cost-effectiveness. The system may mechanically execute a "correct" but practically infeasible or costly strategy. A typical example is that when the return-to-home strategy is triggered, the system does not check if the return-to-home point is set incorrectly (e.g., thousands of kilometers away), nor does it calculate whether sufficient power is required for the return, potentially causing the drone to crash due to battery depletion during the return journey.
[0032] 3. The judgment threshold is rigid and lacks adaptability to dynamic changes in the environment.
[0033] Existing technologies rely on fixed thresholds for status determination, failing to dynamically adjust according to different mission phases and environmental complexity. This results in the system being overly sensitive in simple environments (generating false alarms) and unresponsive in complex environments (leading to missed alarms). For example, the requirements for positioning accuracy differ between the takeoff and landing phases and the cruise phase, but the system uses a uniform GPS loss-of-lock threshold, failing to achieve precise adaptive management.
[0034] Existing drone safety strategy decision-making technologies, due to their passive response mechanisms, one-sided evaluation systems, and rigid judgment thresholds, are unable to effectively cope with continuously changing risk situations, which is the main technical bottleneck restricting drones from achieving higher autonomy and reliability in complex application scenarios.
[0035] Based on the above analysis, the present invention provides a dynamic decision-making method and system for unmanned aerial vehicle (UAV) safety strategies.
[0036] Specific embodiment 1 of the present invention discloses a dynamic decision-making method for drone safety strategies, the process of which is as follows: Figure 1 As shown, the specific steps include: Step S1: Continuously monitor the historical data sequence of at least one state parameter in the UAV system; Step S2: Based on the historical data sequence, predict the expected time to failure (TTF) when the state parameter reaches a preset critical value in the future; Step S3: Before the predicted failure occurs, when the expected failure time TTF ≤ warning time threshold N, trigger the warning and use a two-layer evaluation mechanism to evaluate the strategies in the candidate strategy set and obtain a recommended strategy list. Step S4: When the expected failure time TTF ≤ the pre-execution time threshold M, execute the highest priority strategy in the recommended strategy list without external intervention; Wherein, the warning time threshold N is greater than the pre-execution time threshold M.
[0037] Specifically, the multi-source sensors in the UAV system continuously collect key data from the communication system, power system, navigation system, and environmental data. Therefore, the state parameters represent data from various sensors in the UAV system, and each state parameter forms a historical data sequence according to the chronological order of the collected data.
[0038] Examples of such status parameters include the received signal strength indicator and signal-to-noise ratio of the communication link; the battery voltage, current, and temperature of the power system; the number of GPS satellites and the horizontal accuracy factor of the navigation system; and meteorological and terrain information in environmental data.
[0039] Furthermore, the execution process of the trend prediction algorithm includes: assigning weights to each data point in the historical data sequence; obtaining the difference sequence of the historical data sequence, and calculating the trend slope based on the weights of each data point and the difference sequence; and calculating the expected failure time based on the parameter value of the last time point in the historical data sequence, the trend slope, and the preset critical value.
[0040] Specifically, a trend prediction algorithm combining exponentially weighted moving averages and linear regression is employed to transform discrete state parameters into continuous, quantifiable predictive indicators. These indicators include expected failure times and prediction confidence levels, used to analyze historical changes in state parameters to predict future trends. This trend prediction algorithm has low computational complexity, does not involve highly complex iterations or large-scale matrix operations, and can analyze parameter trends based on historical data sequences, providing stable long-term trend judgments. This makes the algorithm highly suitable for real-time operation in flight control computers or embedded systems with limited onboard computing resources for UAVs, meeting the feasibility requirements of engineering applications.
[0041] Furthermore, the formulas for calculating the weights and the trend slope are as follows:
[0042] In the formula, Represents the first in the historical data sequence The weight of each data point This represents the total length of the historical data sequence. Index representing a data point ;
[0043] In the formula, The slope of the trend, representing the overall rate of change of a historical data series. Represents the first value in the difference sequence. The data point and the first The difference between the data points.
[0044] Furthermore, the formula for calculating the estimated failure time is as follows:
[0045] In the formula, This represents the parameter value at the last point in time in the historical data sequence. This indicates a preset critical value.
[0046] Among them, when the absolute value of the trend slope is less than or equal to If the time frame is not met, it is determined that there is no risk of failure, and the expected failure time is set to infinity.
[0047] Furthermore, the two-layer evaluation mechanism is used to evaluate the strategies in the candidate strategy set to obtain a recommended strategy list, including: A feasibility assessment is performed on the candidate strategy set. The feasibility assessment is used to verify whether the resources required to execute the strategy exceed the current available resources of the UAV, and to filter out strategies that do not meet the preset feasibility constraints. For the candidate strategy set that has passed the feasibility assessment, a cost-effectiveness ratio quantitative evaluation is performed to obtain the cost-effectiveness ratio of each strategy. Based on the cost-effectiveness ratio, K strategies are selected from the candidate strategy set and sorted by cost-effectiveness ratio to form a recommended strategy list, where K is less than the total number of strategies in the candidate strategy set that has passed the feasibility assessment.
[0048] Specifically, when the expected failure time (TTF) is less than or equal to the warning time threshold (N), a warning is triggered indicating a risk. The strategies in the candidate strategy set generated by the system are then sequentially subjected to feasibility resource constraint checks and cost-effectiveness analysis based on cost-effectiveness quantification. Finally, a recommended strategy list is output, which is prioritized according to cost-effectiveness quantification, to ensure that the strategies are both feasible and optimal, providing a quantitative basis for decision-making.
[0049] Furthermore, the feasibility assessment includes at least verifying one or more of the following: the power required for strategy execution and the current remaining power of the UAV; the range required for strategy execution and the current endurance of the UAV; and the availability of the navigation source on which the strategy depends.
[0050] Specifically, feasibility assessments are used to eliminate strategies that are absolutely unfeasible, including energy feasibility, distance feasibility, navigation feasibility, and environmental feasibility.
[0051] For example, when the product of the drone's current remaining battery power and the safety factor is greater than the power required by the strategy, the energy is deemed feasible; when the strategy's range is less than the drone's maximum endurance, the distance is deemed feasible; by determining whether the navigation source (such as GPS) on which the strategy depends is available or switchable, the navigation is deemed feasible; by determining whether the strategy's execution altitude exceeds the restricted flight zone or is lower than the terrain altitude, the environment is deemed feasible.
[0052] Furthermore, the cost-effectiveness ratio quantitative evaluation is performed by using a cost-effectiveness ratio calculation model to quantitatively evaluate the strategies in the candidate strategy set that have passed the feasibility evaluation.
[0053] Furthermore, the cost-effectiveness calculation model is as follows: CER = ( ×Safety Benefits+ ×Task Rewards) / ( ×Resource Costs+ ×Time Cost) in, These are the weighting coefficients for security benefits, task benefits, resource costs, and time costs, respectively. In the actual implementation process, the weighting coefficients can be adjusted according to the task type.
[0054] Furthermore, since the warning time threshold N is greater than the pre-execution time threshold M, when the estimated time to failure (TTF) ≤ the pre-execution time threshold M, the highest priority strategy is selected from the obtained recommended strategy list. In this case, without external manual intervention, the highest priority strategy is automatically executed. When the estimated time to failure (TTF) is greater than the warning threshold N, it indicates no risk, and monitoring continues without taking any action.
[0055] Furthermore, after the UAV flight control system executes the corresponding strategy, it updates the status parameters. The new sensor data after the status parameter update re-enters the monitoring process, forming a closed-loop control.
[0056] Furthermore, the warning time threshold N and the pre-execution time threshold M are dynamically adjusted according to environmental complexity and task criticality.
[0057] Generally, if drones are operating in complex urban environments, the warning time threshold N can be appropriately increased to allow for a longer decision-making time.
[0058] Specific embodiment 2 of the present invention discloses a dynamic decision-making system for unmanned aerial vehicle (UAV) safety strategies. The structural diagram of the system is shown below. Figure 2 As shown, the system includes: The data acquisition module is used to continuously monitor the historical data sequence of at least one state parameter in the UAV system; The failure time prediction module predicts the expected failure time (TTF) when the state parameter reaches a preset critical value in the future, based on the historical data sequence. The strategy evaluation module is used to trigger an early warning before the predicted failure occurs, when the expected failure time TTF is less than or equal to the early warning time threshold N, and to evaluate the strategies in the candidate strategy set using a two-layer evaluation mechanism to obtain a recommended strategy list. The strategy execution module is used to execute the highest priority strategy in the recommended strategy list without external intervention when the expected failure time TTF ≤ the pre-execution time threshold M. Wherein, the warning time threshold N is greater than the pre-execution time threshold M.
[0059] The above-described method embodiments and system embodiments are based on the same principle, and their related aspects can be referenced from each other, and can achieve the same technical effect.
[0060] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.
[0061] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.
Claims
1. A dynamic decision-making method for unmanned aerial vehicle (UAV) safety strategies, characterized in that, The method includes the following steps: Continuously monitor the historical data sequence of at least one state parameter in the unmanned aerial vehicle system; Based on the historical data sequence, predict the expected time to failure (TTF) when the state parameter reaches a preset critical value in the future; Before the predicted failure occurs, when the expected failure time TTF ≤ warning time threshold N, an early warning is triggered and a two-layer evaluation mechanism is used to evaluate the strategies in the candidate strategy set to obtain a recommended strategy list. When the expected failure time TTF is less than or equal to the pre-execution time threshold M, the highest priority strategy in the recommended strategy list is executed without external intervention. Wherein, the warning time threshold N is greater than the pre-execution time threshold M.
2. The dynamic decision-making method for UAV security strategy according to claim 1, characterized in that, The method employs a two-layer evaluation mechanism to evaluate the strategies in the candidate strategy set and obtain a recommended strategy list, including: A feasibility assessment is performed on the candidate strategy set. The feasibility assessment is used to verify whether the resources required to execute the strategy exceed the current available resources of the UAV, and to filter out strategies that do not meet the preset feasibility constraints. For the candidate strategy set that has passed the feasibility assessment, a cost-effectiveness ratio quantitative evaluation is performed to obtain the cost-effectiveness ratio of each strategy. Based on the cost-effectiveness ratio, K strategies are selected from the candidate strategy set and sorted by cost-effectiveness ratio to form a recommended strategy list, where K is less than the total number of strategies in the candidate strategy set that has passed the feasibility assessment.
3. The dynamic decision-making method for UAV security strategy according to claim 2, characterized in that, The feasibility assessment includes at least verifying one or more of the following: the power required for strategy execution and the current remaining power of the UAV; the range required for strategy and the current endurance of the UAV; and the availability of the navigation source on which the strategy depends.
4. The dynamic decision-making method for UAV security strategy according to claim 2, characterized in that, The cost-effectiveness ratio quantitative evaluation uses a cost-effectiveness ratio calculation model to quantitatively evaluate the strategies in the candidate strategy set that have passed the feasibility evaluation.
5. The dynamic decision-making method for UAV security strategy according to claim 4, characterized in that, The cost-effectiveness ratio calculation model is as follows: CER = ( ×Safety Benefits+ ×Task Rewards) / ( ×Resource Costs+ ×Time Cost) in, These are the weighting coefficients for security benefits, task benefits, resource costs, and time costs, respectively.
6. The dynamic decision-making method for UAV safety strategy according to any one of claims 1-5, characterized in that, The execution process of the trend prediction algorithm includes: Assign weights to each data point in the historical data sequence; obtain the difference sequence of the historical data sequence, and calculate the trend slope based on the weights of each data point and the difference sequence; calculate the expected failure time based on the parameter value, trend slope, and preset critical value of the last time point in the historical data sequence.
7. The dynamic decision-making method for UAV security strategy according to claim 6, characterized in that, The formulas for calculating the weights and the trend slope are as follows: In the formula, Represents the first in the historical data sequence The weight of each data point This represents the total length of the historical data sequence. Index representing a data point ; In the formula, The slope of the trend, representing the overall rate of change of a historical data series. Represents the first value in the difference sequence. The data point and the first The difference between the data points.
8. The dynamic decision-making method for UAV security strategy according to claim 7, characterized in that, The predicted time to failure (TTF) for the state parameters to reach a preset critical value in the future is achieved by using a trend prediction algorithm that combines exponentially weighted moving average and linear regression. The formula for calculating the expected failure time is as follows: In the formula, This represents the parameter value at the last point in time in the historical data sequence. This indicates a preset critical value.
9. The dynamic decision-making method for UAV safety strategy based on predictive state and cost-effectiveness assessment according to claim 1, characterized in that, The warning threshold N and the pre-execution threshold M are dynamically adjusted according to the complexity of the environment and the criticality of the task.
10. A dynamic decision-making system for unmanned aerial vehicle (UAV) safety strategies, characterized in that, The system includes: The data acquisition module is used to continuously monitor the historical data sequence of at least one state parameter in the UAV system; The failure time prediction module predicts the expected failure time (TTF) when the state parameter reaches a preset critical value in the future, based on the historical data sequence. The strategy evaluation module is used to trigger an early warning before the predicted failure occurs, when the expected failure time TTF is less than or equal to the early warning time threshold N, and to evaluate the strategies in the candidate strategy set using a two-layer evaluation mechanism to obtain a recommended strategy list. The strategy execution module is used to execute the highest priority strategy in the recommended strategy list without external intervention when the expected failure time TTF ≤ the pre-execution time threshold M. Wherein, the warning time threshold N is greater than the pre-execution time threshold M.