System for preventing autonomous flight of unmanned aerial vehicle and decoy method

Abstract

The invention relates to the technical field of unmanned aerial vehicle decoy, and discloses a system for preventing autonomous flight of an unmanned aerial vehicle and a decoy method. The method comprises the steps of continuously monitoring a flight state data stream of an unmanned aerial vehicle to identify a current flight mode type; dynamically extracting a control parameter group from an unmanned aerial vehicle navigation system according to the identified flight mode type; constructing an interference activation signal sequence by adopting the control parameter group, and performing time sequence analysis on the interference activation signal sequence to position a parameter change event point; dividing the parameter change event points into intra-mode event points and inter-mode event points so as to synthesize flight path deviation parameters; and generating a decoy control parameter based on the flight path deviation parameter, and executing the decoy control parameter to complete a decoy action. According to the method, the flight mode of the unmanned aerial vehicle is accurately identified, and the decoy control parameters are specifically generated, so that effective interference and path deviation control of autonomous flight of the unmanned aerial vehicle are realized, and a new technical means is provided for unmanned aerial vehicle management and control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle decoy, in particular to a system for preventing autonomous flight of unmanned aerial vehicle and a decoy method. BACKGROUND

[0002] The popularization and application field expansion of unmanned aerial vehicle technology, while bringing convenience, have also caused problems such as unauthorized intrusion into no-fly zones, invasion of privacy, and even potential security threats. How to effectively deal with non-cooperative unmanned aerial vehicles, especially their autonomous flight mode, has become a challenge in the current security field. Traditional unmanned aerial vehicle countermeasures mainly include physical destruction, radio signal jamming, and GPS signal jamming. The physical destruction method may cause secondary damage from debris and is limited in application scenarios. Wideband radio jamming blocks the communication link between the unmanned aerial vehicle and the operator by emitting high-power jamming signals, forcing the unmanned aerial vehicle to land or return. This method is direct, but its non-selective interference may affect the normal operation of other legal electronic devices in the surrounding area, and for unmanned aerial vehicles with strong anti-interference capability or pre-set autonomous emergency procedures, the effect may be poor or even ineffective.

[0003] GPS signal jamming blocks the GPS positioning reception of the unmanned aerial vehicle by emitting a specific frequency signal, forcing the unmanned aerial vehicle to lose precise positioning signals. However, this simple jamming method can only cut off the positioning source and cannot stop the unmanned aerial vehicle from continuing to fly - most unmanned aerial vehicles are equipped with inertial navigation, visual navigation, and other backup navigation modules, and after losing GPS, they can still rely on backup systems to maintain autonomous flight and even complete tasks according to pre-set routes or emergency procedures, greatly reducing the effectiveness of countermeasures. In addition, the above methods are mostly passive responses or "hard kill", lacking in-depth understanding of the current behavior intention of the unmanned aerial vehicle and the targeted use of its flight control logic.

[0004] When an unmanned aerial vehicle is in autonomous flight, it will follow control laws and modes to maintain stability and complete tasks. Its navigation system continuously adjusts control parameters based on sensor feedback and internal and external instructions. If the current flight mode and the change rule of the control parameters can be deeply understood, it is possible to find specific "opportunities" or "nodes" in its control logic, and by applying carefully calculated and small misleading signals, rather than brute-force full-band interference, to induce its own flight decision deviation, so as to achieve the purpose of preventing it from continuing its original flight. This method requires the ability to analyze real-time flight state data of the unmanned aerial vehicle, accurately identify the flight mode, dynamically extract the control parameters, and accurately grasp the interference opportunity.

[0005] Existing technologies often lack exploration of this fine "soft" decoy strategy based on flight mode understanding and control parameter analysis, making it difficult to achieve an efficient, low-impact, and highly targeted new countermeasure. SUMMARY

[0006] The present application aims to provide a system and a deception method for preventing unmanned aerial vehicles from autonomous flight, so as to solve the problems in the background art.

[0007] To achieve the above-mentioned purpose, the present application provides a deception method for preventing unmanned aerial vehicles from autonomous flight, comprising: continuously monitoring flight state data stream of the unmanned aerial vehicle to identify current flight mode type; dynamically extracting control parameter group from the unmanned aerial vehicle navigation system according to the identified flight mode type; constructing interference activation signal sequence by using the control parameter group, and performing time sequence analysis on the interference activation signal sequence to locate parameter change event point; dividing the parameter change event point into intra-mode event point and inter-mode event point, thereby synthesizing flight path deviation parameter; generating deception control parameter based on the flight path deviation parameter, and executing the deception control parameter to complete deception action.

[0008] Preferably, the dynamically extracting control parameter group from the unmanned aerial vehicle navigation system according to the identified flight mode type comprises: capturing control reference parameter set and its duration window according to state change curve of flight mode type; capturing control update parameter set and its effective time interval according to transition trigger point of flight mode type; applying parameter standardization program to process the control update parameter set to output standardized update parameter set; and fusing the control reference parameter set and the standardized update parameter set to form the control parameter group.

[0009] Preferably, the applying parameter standardization program to process the control update parameter set to output standardized update parameter set comprises: calculating parameter difference measure of each state stage by deviation evaluation algorithm, wherein the deviation evaluation algorithm adopts state influence factor, time decay coefficient, time offset between state transition trigger point and parameter update time, and reference parameter setting time; and adjusting parameter value based on proportion of parameter difference measure in total difference by parameter normalization algorithm to obtain standardized update parameter set.

[0010] Preferably, the constructing interference activation signal sequence by using the control parameter group comprises: screening update parameter subset from the control parameter group to generate interference activation signal based on the update parameter subset; extracting standardized update parameter set from the control parameter group to generate interference activation signal based on the standardized update parameter set; comparing the interference activation signal based on the update parameter subset with the interference activation signal based on the standardized update parameter set to obtain activation signal mode of the same flight mode type and different flight mode types in state stage; and optimizing generation time point of the interference activation signal sequence according to the activation signal mode.

[0011] Preferably, the time sequence analysis on the interference activation signal sequence to locate the parameter change event point comprises: calculating a matching degree index of the interference activation signal sequence and a reference time sequence corresponding to the historical parameter change event point, and if the matching degree index exceeds a preset matching threshold, identifying it as a consistent type parameter change event point as an intra-mode parameter change event point; calculating a matching degree index of the interference activation signal sequence and a reference time sequence corresponding to the historical parameter reset event point, and if the matching degree index is lower than the preset matching threshold, identifying it as a conflict type parameter reset event point as an inter-mode parameter reset event point.

[0012] Preferably, the calculation of the matching degree index of the interference activation signal sequence and the reference time sequence corresponding to the historical parameter reset event point comprises: calculating a covariance value of the time data of the interference activation signal sequence and the reference time sequence, and calculating a standard deviation of the time data of the interference activation signal sequence and a standard deviation of the reference time sequence, and dividing the covariance value by the product of the standard deviations to obtain the matching degree index.

[0013] Preferably, the optimization of the generation time point of the interference activation signal sequence according to the activation signal mode comprises: if the activation signal mode is greater than a preset mode threshold, triggering the generation of the interference activation signal sequence in advance; if the activation signal mode is less than the preset mode threshold, delaying the generation of the interference activation signal sequence.

[0014] Preferably, the division of the parameter change event point into intra-mode event points and inter-mode event points to synthesize the flight path deviation parameter comprises: extracting parameter change event points caused by environmental disturbances within the same flight mode type and labeling them as intra-mode parameter change event points; extracting parameter reset event points triggered when the flight mode type is switched and labeling them as inter-mode parameter change event points; calculating the parameter variation amplitude within the same flight mode type based on the intra-mode parameter change event points as an internal path deviation parameter; calculating the parameter variation amplitude when the flight mode type is converted based on the inter-mode parameter change event points as a switching path deviation parameter; and combining the internal path deviation parameter and the switching path deviation parameter as the flight path deviation parameter.

[0015] Preferably, the generation of the decoy control parameter based on the flight path deviation parameter comprises: taking the flight state stage as the time reference, analyzing the common mode in the flight path deviation parameter, extracting the parameter group common to all flight mode types and marking it as a shared parameter; analyzing the specific mode in the flight path deviation parameter, extracting the parameter group corresponding only to a specific flight mode type and marking it as an exclusive parameter; applying the shared parameter as a global decoy reference to the decoy process of all flight mode types; and adaptively adjusting the decoy parameter of a specific flight mode type according to the exclusive parameter. The execution of the decoy control parameter to complete the decoy action includes: adopting a shared parameter to establish a basic decoy framework, and integrating a dedicated parameter to adjust the specific value of the decoy parameter, so as to realize dynamic decoy control.

[0016] Preferably, the present application also includes a system for preventing an unmanned aerial vehicle from autonomous flight, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the decoy method for preventing an unmanned aerial vehicle from autonomous flight when executing the computer program.

[0017] Compared with the prior art, the present application has the following beneficial effects: The method can master the flight characteristics and behavior patterns of the unmanned aerial vehicle in real time by continuously monitoring the flight state data stream of the unmanned aerial vehicle. This continuous monitoring mechanism ensures comprehensive perception of the flight state of the unmanned aerial vehicle by the system, providing sufficient data support for subsequent flight pattern recognition. Continuous collection of flight state data stream enables the system to capture the dynamic change process of the unmanned aerial vehicle flight, thereby more accurately understanding its flight intention and behavior rules.

[0018] Flight pattern type recognition based on the flight state data stream can effectively distinguish the flight characteristics of the unmanned aerial vehicle in different scenarios. This recognition capability enables the system to take targeted measures according to the specific flight pattern, improving the accuracy and effectiveness of the decoy operation. Through accurate recognition of multiple flight patterns, the system can adapt to the flight characteristics of different types of unmanned aerial vehicles, enhancing the universality and practicality of the method.

[0019] The process of dynamically extracting the control parameter group from the unmanned aerial vehicle navigation system realizes the acquisition of the core control parameters of the unmanned aerial vehicle. This dynamic extraction mechanism can adapt to the differences in navigation systems of different models of unmanned aerial vehicles, ensuring the integrity and accuracy of parameter acquisition. The acquisition of the control parameter group provides a key data basis for the construction of subsequent interference signals, enabling the decoy operation to directly act on the control system of the unmanned aerial vehicle.

[0020] The method of constructing an interference activation signal sequence using the control parameter group can generate interference signals that match the control system of the unmanned aerial vehicle. This signal construction method based on actual control parameters ensures the effectiveness and pertinence of the interference signals. Through timing analysis of the interference activation signal sequence, the system can accurately grasp the timing and rules of parameter changes, creating conditions for subsequent parameter change event positioning.

[0021] The accurate positioning of the parameter change event point provides an important basis for the synthesis of the flight path deviation parameter. The classification method of dividing the parameter change event point into intra-mode event points and inter-mode event points can better distinguish the characteristics of different types of events. This detailed classification method enables the system to adopt differentiated processing strategies according to the different types of event points, thereby improving the fineness of the decoy operation.

[0022] The process of generating decoy control parameters based on flight path deviation parameters realizes accurate interference with the flight path of the unmanned aerial vehicle. The decoy control parameters generated based on actual flight parameters can effectively guide the unmanned aerial vehicle to deviate from the predetermined flight path. The execution process of the decoy control parameters can realize flight interference without damaging the hardware of the unmanned aerial vehicle, thereby embodying the non-destructive characteristics of the method. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 The working principle diagram of the decoy method for preventing the autonomous flight of the unmanned aerial vehicle; Figure 2 The method flowchart of dynamically extracting the control parameter group; Figure 3 The method flowchart of constructing the interference activation signal sequence; Figure 4 The matching analysis diagram of the interference activation signal sequence and the reference time sequence; Figure 5 The flight path deviation parameter analysis diagram. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0025] Please refer to Figure 1The present application provides a system and method for preventing autonomous flight of a UAV, the method comprising: continuously monitoring flight state data stream of the UAV, identifying current flight mode type by analyzing pattern features in the data stream such as heading angle rate of change, speed fluctuation or altitude deviation, for example hover, cruise or evasion mode; dynamically extracting control parameter set from UAV navigation system according to the identified flight mode type, this process involves reading internal parameters of the navigation system such as PID controller set value or path planning point in real time, and filtering relevant parameters according to mode characteristics; constructing interference activation signal sequence using the control parameter set, the sequence consists of a series of time-stamped parameter change pulses, and performing time series analysis on the sequence to locate parameter change event points, the time series analysis uses sliding window algorithm to compare signal differences of adjacent time periods; dividing the parameter change event points into intra-mode event points and inter-mode event points, intra-mode event points refer to parameter adjustment caused by environmental disturbance within the same flight mode, while inter-mode event points involve parameter reset during mode switching, calculating parameter change amplitude based on these event points and synthesizing flight path deviation parameters; finally generating decoy control parameters based on flight path deviation parameters, for example adjusting navigation instructions by interpolation or extrapolation method, and executing these parameters to inject misleading signals, making the UAV deviate from the predetermined path.

[0026] Embodiment 1: refer to Figure 2 In specific implementation, the process of dynamically extracting control parameter set starts with capturing control reference parameter set and its duration window according to state change curve of flight mode type, the state change curve of flight mode type is constructed by continuously monitoring flight state data stream of the UAV, for example real-time sampling of acceleration sensor output or gyroscope data to generate curve trajectory, so as to identify mode features such as linear segment of uniform flight phase or nonlinear fluctuation of maneuvering flight phase; according to these curves, the control reference parameter set is extracted, the parameters include target waypoint coordinates, speed set value or attitude stability threshold, and the duration window of each parameter is recorded, that is, the time interval from the parameter taking effect to invalidation, the window boundary is marked by time stamp to ensure time sequence accuracy. In some embodiments, the generation of state change curve uses moving average filter to smooth the original data, so as to reduce noise interference, and uses derivative analysis to detect curve inflection point, so as to accurately divide the state phase; when capturing the control reference parameter set, the parameter selection is based on the inherent characteristics of the mode type, for example, the hover mode focuses on vertical speed parameter, while the cruise mode focuses on horizontal heading parameter, to ensure that the parameter set is highly related to the mode. It can be understood that the determination of the duration window depends on the persistence verification of the flight state data, the window length is confirmed by comparing the consistency of parameters at adjacent time points, to avoid misjudgment caused by instantaneous fluctuation.

[0027] The control update parameter set and its valid time interval are captured according to the transition trigger point of the flight mode type, which is detected by analyzing the mutation events in the flight state data stream, such as when the UAV switches from hover mode to cruise mode, the speed reading changes in steps or the heading angle rate exceeds the threshold, i.e. identified as a transition trigger point; the control update parameter set includes controller gain coefficients, path planning update intervals or sensor calibration parameters, which are dynamically adjusted when the mode is switched, and the valid time interval is calculated from the transition trigger point until the next stable state is established. In specific implementation, the detection of the transition trigger point adopts a multi-condition fusion strategy, combining time series analysis and pattern recognition algorithms such as Hidden Markov Model or Decision Tree to improve detection robustness; when capturing the control update parameter set, the parameter values are read from the real-time data bus of the UAV navigation system, and are associated with the valid time interval, the interval length is based on historical mode transition data statistics, or is dynamically adjusted through real-time feedback. Optionally, the setting of the valid time interval can consider environmental factor compensation, such as extending the interval to accommodate longer adjustment period under strong wind conditions, but the core is still based on the transition trigger point.

[0028] The control update parameter set is processed by the parameter standardization program to output the standardized update parameter set, and the core of the parameter standardization program is the deviation evaluation algorithm, which calculates the parameter difference measure of each state stage; the deviation evaluation algorithm uses multiple input elements: the state influence factor weighted parameter change contributes to the flight stability, the state influence factor assigns weights according to the role of the parameter in the control loop, such as the navigation parameter weight is higher than the auxiliary parameter; the time decay coefficient adjusts the contribution of historical parameter data, the time decay coefficient reduces the influence of old data in an exponential form, emphasizing recent changes; the time offset of the state transition trigger point and the parameter update time measures the synchronization of parameter adjustment, a small time offset indicates timely response, and a large time offset indicates lag; and the reference parameter setting time is used as the reference time for calculating the relative difference. In specific implementation, the calculation of the parameter difference measure is realized by weighted summation, which is expressed as difference measure equal to state influence factor multiplied by parameter change, plus offset compensation adjusted by time decay coefficient, and divided by time offset normalization factor, but the specific numerical processing avoids using mathematical formula, but is described as scalar operation; the output of the deviation evaluation algorithm is a scalar value, indicating the abnormality of the parameter in the current state. It can be understood that the parameter difference measure serves the subsequent normalization, and its calculation needs to ensure numerical stability, such as preventing overflow by limiting amplitude processing.

[0029] The parameter normalization algorithm adjusts the parameter values based on the proportion of each parameter difference measure in the total difference, obtaining a normalized update parameter set. The parameter normalization algorithm first aggregates the parameter difference measures of all state stages, calculates the total difference as the normalization benchmark, and the total difference is the cumulative sum or vector length of the difference measures of each stage. Then, based on the proportion of each parameter difference measure in the total difference, the original parameter values are adjusted, for example, linearly mapping the parameter values to the 0-1 interval, or using a sigmoid function for nonlinear scaling, to ensure the consistency of the parameter scale. In some embodiments, the parameter normalization algorithm introduces an adaptive adjustment mechanism, using a minimum threshold to avoid division by zero error when the total difference is too small, and using truncation processing to prevent distortion when the total difference is too large; the output of the normalized update parameter set makes the parameter values comparable, eliminating the influence of dimensional differences and inter-mode fluctuations. Optionally, the parameter normalization algorithm can combine machine learning methods, such as cluster analysis to identify parameter distribution patterns, but the basic implementation is still based on proportional adjustment.

[0030] The control parameter group is formed by fusing the control reference parameter set and the normalized update parameter set; the fusion process uses a weighted average strategy, in which the weight of the control reference parameter set is allocated based on the length of the time window, the longer the window, the higher the weight, indicating strong parameter stability, while the weight of the normalized update parameter set is allocated based on the inverse of the parameter difference measure, the smaller the difference, the greater the weight, emphasizing reliability; the parameter group after fusion contains integrated parameter values and their time attributes, ensuring a comprehensive reflection of the current flight mode type control requirements. In specific implementation, the fusion operation is realized by parameter-by-parameter merging, and for overlapping parameters, the value of the normalized update parameter set is preferred to reflect real-time performance while retaining the background information of the control reference parameter set; the formation of the control parameter group is a dynamic process that updates with the flight state data stream to maintain the adaptability of the deception method. It can be understood that the design of the fusion strategy needs to balance historical reference and real-time update to avoid excessive bias towards either side leading to control deviation.

[0031] Embodiment 2: see Figure 3In specific implementations, the initial step of constructing the interference activation signal sequence is to filter an update parameter subset from the control parameter group, the update parameter subset contains parameter members in the control parameter group that have higher dynamic change frequency, such as real-time updated heading angle setpoint, thrust output percentage or position feedback increment; the filtering process is based on the historical change rate threshold of the parameter, the change rate is obtained by calculating the absolute value of the difference between the parameter values in adjacent sampling periods, when the change rate continuously exceeds the pre-set threshold, the parameter is included in the update parameter subset. When generating the interference activation signal based on the update parameter subset, the value of each member of the update parameter subset is converted into a discrete time sequence pulse, the amplitude of the pulse is proportional to the degree of deviation of the parameter from its reference value, and the width of the pulse is associated with the update time interval of the parameter, thereby forming a group of signal segments reflecting the real-time fluctuations of the parameters. In some embodiments, the filtering of the update parameter subset uses a sliding window mechanism to monitor the parameter change rate, the window size is adjusted according to the dynamic characteristics of the flight mode type, for example, a smaller window is used in agile maneuver mode to capture rapid changes; when generating the interference activation signal based on the update parameter subset, the generation time of the pulse is synchronized with the control cycle of the unmanned aerial vehicle navigation system to ensure the time alignment of the signal with the underlying control logic. It can be understood that the interference activation signal based on the update parameter subset is essentially a time encoding of parameter change events, and its integrity depends on the accurate analysis of the update mechanism of the control parameter group.

[0032] The normalized update parameter set is extracted from the control parameter group, which is a parameter set processed by the parameter normalization program, with uniform numerical scale and time characteristics; when generating the interference activation signal based on the normalized update parameter set, each parameter value of the normalized update parameter set is mapped to a standardized pulse, the amplitude of the standardized pulse is directly determined by the normalized value of the parameter, and the time position of the pulse corresponds to the effective time of the parameter. In specific implementations, extracting the normalized update parameter set is a direct read operation, because the normalized update parameter set has been stored in a specific buffer area after the parameter normalization program is output; the generation of the interference activation signal based on the normalized update parameter set emphasizes the relative change relationship between the parameters, and the form of the pulse sequence can reveal the overall pattern of parameter change rather than individual fluctuations. Optionally, the interference activation signal generation process based on the normalized update parameter set can introduce smoothing filtering to suppress high-frequency noise that may be introduced in the normalization process, but the core is to maintain the coordination relationship between the parameters.

[0033] Comparing the interference activation signal based on the updated parameter subset with the interference activation signal based on the normalized parameter set is a key step to obtain the activation signal pattern. The comparison is performed in the time dimension, and the two signal sequences are aligned and compared within the same state phase time window. The comparison method includes calculating the cross-correlation coefficient between the signal segments to quantify the waveform similarity, or calculating the integral of the absolute difference of the pulse amplitude at the corresponding time points to quantify the difference. Through systematic comparison, the activation signal patterns of the same flight mode type and different flight mode types in the state phase can be obtained. In the same flight mode type, the interference activation signal based on the updated parameter subset and the interference activation signal based on the normalized parameter set usually show high similarity, and the envelope shape and main peak position of the pulse sequence tend to be consistent. However, in the state phase of the switch between different flight mode types, the two signals may show significant differences. For example, the interference activation signal based on the updated parameter subset appears a sharp jitter pulse, while the interference activation signal based on the normalized parameter set is relatively flat. Such differences constitute the recognition features of the inter-mode jump signal. In some embodiments, the extraction of the activation signal pattern uses a pattern recognition algorithm to perform cluster analysis on the comparison results, and classifies the signals with a similarity higher than a threshold as a stable pattern, and classifies the signals with a similarity lower than the threshold as a jump pattern. It can be understood that the activation signal pattern is a quantitative description of the consistency of the unmanned aerial vehicle control behavior, and is an important basis for judging the nature of the parameter change.

[0034] According to the activation signal pattern, the optimization of the generation time point of the interference activation signal sequence is a feedback regulation process; the optimization decision is based on the comparison result of the activation signal pattern value and the preset pattern threshold value, and the preset pattern threshold value is an empirical value determined through historical data analysis or simulation verification, which is used to distinguish between high consistency patterns and low consistency patterns. If the activation signal pattern is greater than the preset pattern threshold value, it indicates that the parameter change under the current flight state has high predictability and repeatability, for example, the parameter adjustment in the stable cruise stage follows a fixed pattern, at this time, the generation of the interference activation signal sequence is triggered in advance, and the amount of advance triggering is dynamically calculated according to the amplitude of the pattern value exceeding the threshold value, for example, if the pattern value exceeds the threshold value by ten percent, the triggering is advanced by one control period. The purpose is to inject interference signals before the expected parameter change of the unmanned aerial vehicle navigation system, and to preempt the control timing. If the activation signal pattern is less than the preset pattern threshold value, it indicates that the parameter change behavior is irregular or in a transition state, for example, in the initial stage of mode conversion, the parameter adjustment has uncertainty, at this time, the generation of the interference activation signal sequence is delayed, and the delay time is determined according to the degree of the pattern being below the threshold value and the average duration of historical mode switching. The purpose is to wait for the flight mode to be stable and the parameter change rule to be clear before applying interference, so as to avoid the failure of deception or the triggering of abnormal behavior of the unmanned aerial vehicle due to improper timing. In specific implementation, the optimization process is realized by a programmable time sequence trigger, which receives the activation signal pattern as input and outputs the adjusted signal generation instruction. Optionally, the optimization strategy can use a proportional-integral regulator to smooth the adjustment amplitude of the generation time point, to prevent excessive advance or delay from causing the signal to be out of step with the control event. It can be understood that the optimization of the generation time point of the interference activation signal sequence is essentially to keep the deception action in synchronization or slightly ahead of the rhythm of the autonomous flight control of the unmanned aerial vehicle, thereby enhancing the concealment and effectiveness of the deception.

[0035] In some embodiments, the construction of the reference time series employs a method of aligning and averaging a large number of historical event signals to eliminate random noise and retain common features. The interference activation signal sequence needs to be preprocessed by time window truncation and amplitude normalization before matching, in order to eliminate the influence of scale difference on the matching result. The calculated matching degree index is a value within a certain range, and the higher the value, the stronger the similarity. If the matching degree index exceeds the preset matching threshold, which is a threshold value determined by statistical learning, the time corresponding to the current interference activation signal sequence is identified as a consistent type parameter change event point. A consistent type parameter change event point means that the current detected event is highly consistent with the parameter fine-tuning event features that occur within the same flight mode in historical records, and therefore it is classified as an in-mode parameter change event point. An in-mode parameter change event point usually corresponds to the regular parameter adjustment made by the flight control system to respond to slight environmental disturbances (such as gusts).

[0036] The specific calculation of the matching degree index employs an algorithm based on covariance and standard deviation. This algorithm first needs to calculate the covariance value of the time data of the interference activation signal sequence and the reference time series. The covariance value reflects the consistency of the fluctuations of the two sequences around their respective means. Its calculation involves multiplying the deviations of the values at corresponding time points of the two sequences from their respective means, and then taking the average of these products. Let the interference activation signal sequence be with mean , and the reference time series be with mean . The calculation of the covariance value can be represented as: where: represents the value of the interference activation signal sequence at the th time point, represents the value of the reference time series at the th time point, represents the total number of time points involved in the calculation, is the arithmetic mean of all values of the interference activation signal sequence, is the arithmetic mean of all values of the reference time series. It can be understood that the sign and magnitude of the covariance value indicates the co-directionality or anti-directionality of the two sequences' changes and the strength of the linear correlation.

[0037] The algorithm needs to calculate the standard deviation of the time data of the interference activation signal sequence and the standard deviation of the reference time series. The standard deviation is a statistical quantity that measures the degree to which the data points in a sequence deviate from their mean value. The standard deviation of the interference activation signal sequence The calculation formula is: The standard deviation of the reference time series The calculation formula is: The calculation of the standard deviation completes the evaluation of the dispersion degree of each sequence. Finally, the covariance value is divided by the product of the standard deviation of the interference activation signal sequence and the standard deviation of the reference time series to obtain the matching degree index : This matching degree index is mathematically equivalent to the Pearson correlation coefficient, whose value range is between negative one and positive one. The closer the value is to positive one, the stronger the positive linear correlation between the two sequences, and the more similar the shapes. In some embodiments, in order to simplify the subsequent judgment logic, the matching degree index may be linearly transformed to map it to the interval of zero to one hundred, but the core comparison logic remains unchanged. It can be understood that this calculation method can effectively capture the shape similarity between sequences and is not sensitive to the absolute numerical amplitude, and is suitable for comparing signals that may have different baselines or gains.

[0038] The matching degree index of the interference activation signal sequence and the reference time sequence corresponding to the historical parameter reset event point is calculated, and the reference time sequence corresponding to the historical parameter reset event point represents a standard parameter reset signal waveform recorded when the flight mode is switched. The calculation method of the matching degree index is exactly the same as described above, that is, the covariance value between the interference activation signal sequence and the reference time sequence corresponding to the historical parameter reset event point is also calculated, and then divided by the product of the respective standard deviations. If the matching degree index obtained this time is lower than the preset matching threshold, it indicates that the form of the current interference activation signal sequence is quite different from the parameter reset event characteristics when the typical mode is switched, and has no similarity, so the current time is marked as a conflict type parameter change event point. The conflict type parameter change event point means that the detected event does not conform to the known intra-mode adjustment mode, but is identified as a possible event triggered by the flight mode type conversion due to its low correlation with the mode switching event, so it is classified as an inter-mode parameter reset event point. In some embodiments, there may be a situation where the matching degree index is both higher than the threshold of the intra-mode event and lower than the threshold of the inter-mode event, at which time a more complex classifier or multi-feature fusion decision can be introduced, but in the basic implementation, it is usually classified according to the preset priority or nearest neighbor principle. Optionally, the preset matching threshold for judging the intra-mode event and the inter-mode event can be the same value, or different values set according to the false alarm rate and the false alarm rate requirement. The core of the whole time sequence analysis process is to interpret the continuous interference activation signal sequence into discrete parameter change event points with type labels through signal matching technology.

[0039] Referring to Figure 4 , the matching analysis results of the interference activation signal sequence and the reference time sequence are shown. The figure contains a reference time sequence and multiple interference activation signal sequences, and the Pearson correlation coefficient is calculated as a matching degree index to identify parameter change event points. The blue solid line in the chart represents the reference time sequence of the intra-mode parameter change, which reflects the characteristics of the regular parameter adjustment of the unmanned aerial vehicle in the stable flight mode to respond to environmental disturbances. The other colored lines represent different interference activation signal sequences, the solid lines represent signals identified as consistent type parameter change event points, and the dashed lines represent signals identified as conflict type parameter change event points. The circular point area marked in the figure represents the time when the matching degree exceeds the preset threshold, and these times are marked as parameter change event points. Through this time sequence analysis method, the system can accurately distinguish between regular parameter adjustment in the mode and parameter reset events between modes, providing basic data for subsequent flight path deviation parameter synthesis. The chart clearly shows the differences in the form of signal sequences, and embodies the key role of matching degree calculation in event classification.

[0040] In a specific implementation, the process of dividing parameter change event points into intra-mode event points and inter-mode event points starts with extracting parameter change event points within the same flight mode type triggered by environmental disturbances, which are time points located via temporal analysis of the disturbance activation signal sequence; the extraction operation is based on the attribute label of the event points, when the label of a parameter change event point is marked as consistent-type parameter change event point, it indicates that the event originates from state adjustment within the flight mode, and then it is labeled as intra-mode parameter change event point, which is usually associated with control parameter updates triggered by the UAV to compensate for wind disturbances, maintain heading, or make minor path corrections. In a specific implementation, the extraction process is completed by querying the event point database, each event point record in the database contains a timestamp, an event type identifier, and associated flight mode context information, and the selection logic is to select records with event type identifier corresponding to "consistent" and flight mode context remaining unchanged before and after the event. It can be understood that accurate labeling of intra-mode parameter change event points depends on the accuracy of the matching degree index calculation in the previous temporal analysis and the rationality of the preset matching threshold setting.

[0041] Simultaneously, parameter reset event points triggered during the switching of different flight mode types are extracted, which are a subset of conflict-type parameter change event points identified in temporal analysis; the extraction operation focuses on event points near the flight mode type transition trigger points, when the time difference between the occurrence time of a parameter change event point and a known flight mode type transition trigger point is within the preset tolerance, and its label is conflict-type parameter change event point, then it is labeled as inter-mode parameter change event point, which represents the fundamental reset of the UAV navigation system's control law or path planning parameters in response to high-level instructions. In a specific implementation, the labeling of inter-mode parameter change event points needs to be cross-verified with the mode recognition results in the flight state data stream to ensure that the event points indeed occur within the transition period of mode switching rather than isolated abnormal fluctuations. Optionally, a confidence score can be set for inter-mode parameter change event points, which is calculated based on the closeness of the time to the transition trigger point and the prominence of the conflict-type feature.

[0042] Based on the intra-mode parameter change event points, the parameter variation amplitude within the same flight mode type is calculated, which aims to quantify the cumulative amount of control parameter change during a single flight mode duration; the calculation process is performed for each control parameter associated with an intra-mode parameter change event point, first determining the numerical difference of the parameter before and after the event point, i.e. the instantaneous change amount, then aggregating the instantaneous change amounts caused by all intra-mode parameter change event points within a time window, the aggregation method can adopt absolute value summation, square root of square sum or maximum absolute value, etc., and the final output value is taken as the internal path deviation parameter. The internal path deviation parameter reflects the potential trend of path deviation due to continuous external interference or internal adjustment demand in stable flight mode. In specific implementation, the calculation of parameter variation amplitude needs to distinguish parameter types, for example, position-related parameters and attitude-related parameters may adopt different aggregation weights to accurately reflect their comprehensive influence on the actual flight path. Referring to Table 1, a simplified internal path deviation parameter calculation is shown, in which the aggregation method adopts absolute value summation.

[0043] Table 1: Internal path deviation parameter calculation (for cruise mode) Based on the inter-mode parameter change event points, the parameter variation amplitude during flight mode type conversion is calculated, which focuses on the jump of parameter value at the moment of mode switching or short transition period; for each inter-mode parameter change event point, identify the control parameter directly associated with it, and calculate the numerical difference of the parameter in the two stable states before and after mode switching, this difference is usually much larger than the fine-tuning amount within the mode, for example, from the target speed of 10 m / s in cruise mode to the target speed of 0 m / s in hover mode, the calculated difference is the contribution of this event point to the switching path deviation parameter; the switching path deviation parameter can be synthesized from these jump values in various ways, for example, taking the maximum value of the absolute value of the change of all associated parameters in the mode switching event as the representative, or calculating the weighted average value of these change amounts, the weight is determined by the importance of the parameter to the flight path. The switching path deviation parameter describes the degree of path dispersion introduced by the fundamental change of flight task. In some embodiments, for complex mode switching with multiple parameters resetting at the same time, the vector synthesis of parameter change is considered when calculating the switching path deviation parameter to more comprehensively evaluate the path deviation.

[0044] The combined result of the internal path deviation parameter and the switching path deviation parameter is taken as the flight path deviation parameter, and the combination operation is not simply numerical addition, but data fusion according to the needs of the decoy strategy; a typical combination method is to assign different weight coefficients to the internal path deviation parameter and the switching path deviation parameter, and the weight coefficients reflect the importance of different types of deviation in the overall path prediction, for example, if the decoy strategy pays more attention to the cumulative drift within the mode, the weight of the internal path deviation parameter is higher, and if more attention is paid to the sudden change when the mode is switched, the weight of the switching path deviation parameter is higher, and the result of weighted summation is the flight path deviation parameter. The flight path deviation parameter is a comprehensive index for quantifying the overall degree of deviation of the UAV from the expected path under the current and near-term control behavior. In specific implementation, the determination of the weight coefficient can be based on historical flight data analysis or obtained by simulation optimization. Optionally, the combination process can also introduce a nonlinear function, for example, different fusion rules are selected according to the relative size of the internal path deviation parameter and the switching path deviation parameter. It can be understood that the accurate synthesis of the flight path deviation parameter is the basis for generating effective decoy control parameters, which needs to take into account the continuity and discrete jump characteristics of the flight behavior.

[0045] Referring to Figure 5 , the analysis result of the flight path deviation parameter is shown, and the feature difference of different types of parameter change events is presented in the form of a combination of column chart and scatter chart. The column chart part shows the average deviation amplitude and standard deviation of the parameter change event within each flight mode, reflecting the parameter change range caused by environmental disturbance in the stable flight state. The column bodies of different colors correspond to different flight modes, and the column height represents the average level of parameter change in this mode, and the error line shows the fluctuation range of the change. The red dots in the scatter chart represent the inter-mode parameter reset events, which occur when the flight mode is switched, and the parameter change amplitude is usually significantly greater than the regular adjustment within the mode. The distribution position of these scatter points shows the relevance of parameter reset events to each flight mode and the strength of parameter jump. By comparing the data of the column chart and the scatter chart, the essential difference between the intra-mode event and the inter-mode event in the parameter change amplitude can be clearly seen. This difference analysis provides an important basis for synthesizing the flight path deviation parameter, so that the system can more accurately predict the path deviation trend of the UAV and generate effective decoy control parameters.

[0046] In specific implementations, the process of generating decoy control parameters based on flight path deviation parameters is developed with flight state phases as time reference, which are time intervals divided according to flight mode types and their internal sub-states (e.g. acceleration, cruise, deceleration). Analyzing common patterns in flight path deviation parameters aims to discover those deviation characteristics that repeat across different flight mode types, with commonality, for example, the small periodic oscillation around the preset heading that can be observed in multiple modes such as hovering, cruising, and orbiting, which is caused by sensor noise or the characteristics of the underlying controller. Extracting parameter groups that are common across flight mode types is an operation based on common patterns, through clustering analysis or frequent pattern mining algorithms, from historical flight path deviation parameter data, identify those parameter combinations that are statistically significant in a multi-mode context, for example, a group of parameters including the underlying heading control gain, position loop integral coefficient, and vertical velocity limit, this group of parameters is marked as shared parameters. Shared parameters constitute the control basis that has an impact on different flight modes. In some embodiments, the analysis of common patterns uses time series pattern matching techniques to align and compare flight path deviation parameter sequences in different modes to identify similar segments and their corresponding control parameters.

[0047] Simultaneously, specific patterns in flight path deviation parameters are analyzed, which refer to those deviation characteristics that are closely associated with a specific flight mode type, not significant or morphologically different in other modes, for example, the vertical height fine-tuning caused by rotor downwash disturbance in hovering mode, or the anti-crosswind roll moment compensation in high-speed cruising mode to maintain straight flight; extracting parameter groups that only correspond to a specific flight mode type is an operation for specific patterns, through comparative analysis, those parameters that are only active in the flight path deviation parameters of a specific flight mode type and are strongly related to the characteristics of the mode are separated out, for example, the height lock threshold parameter in hovering mode, the forward speed planning curve parameter in cruising mode, these parameters are marked as exclusive parameters. Exclusive parameters reflect the unique control requirements of a specific flight task. In specific implementations, the analysis of specific patterns needs to combine the semantic information of flight mode types, for example, when the mode identifier outputs "precise hovering", then focus on the deviation components related to position holding accuracy and their control parameters.

[0048] Applying the shared parameters as a global decoy reference to all flight mode types of decoy process means that no matter the UAV is in hover, cruise or any other mode, the decoy system will first load and use the basic misleading strategy defined by the shared parameters, for example, if the shared parameters contain a basic heading offset, then this offset will be superimposed as a constant or slowly varying component to the decoy command in all modes; the adaptive adjustment of decoy parameters for specific flight mode types according to the dedicated parameters is a fine-tuning adjustment on top of the global reference, when the system detects that the UAV enters a certain mode (such as from cruise to hover), it will dynamically call the dedicated parameter set corresponding to this mode, and based on the current values of these parameters and the deviation indicated by the flight path deviation parameter, calculate the specific decoy adjustment amount for this mode, for example, in hover mode, according to the height lock threshold in the dedicated parameters, dynamically fine-tune the height instruction disturbance amplitude in the decoy signal. In specific implementation, the application of shared parameters is usually implemented through a global parameter lookup table or configuration file, while the adaptive adjustment of dedicated parameters is completed through a callback function or rule engine bound to the flight mode state machine. Optionally, there may be overlap or conflict between shared parameters and dedicated parameters, the processing strategy can be to set a priority lower limit for shared parameters, or to solve the conflict in a weighted fusion manner.

[0049] The execution of the decoy control parameter to complete the decoy action adopts a shared parameter to establish a basic decoy framework, and the basic decoy framework defines a basic structure and constraint of the decoy behavior, for example, the framework stipulates that the injection point of the decoy instruction is located after the navigation solution module and before the controller execution, and the framework also defines the maximum amplitude limit and the change rate limit of the decoy signal, and the reference values of the limits are derived from the shared parameter; the integration of the specific values of the decoy parameter is adjusted by the exclusive parameter, which is a dynamic assignment process within the basic decoy framework, the system reads the current effective exclusive parameter set, and according to the real-time calculated decoy demand, the amplitude, frequency, duration and other attributes of the decoy signal are specifically determined within the boundary set by the shared parameter, to realize dynamic decoy control. For example, the shared parameter may set the maximum range of the heading decoy amount as ±30 degrees, which is a safety boundary; when the UAV is in the cruise mode, the exclusive parameter may indicate that the change rate of the decoy amount can be faster in this mode, and the system will generate a heading misleading signal that fluctuates at a faster rate within the ±30 degree range; and when the UAV switches to the hovering mode, the corresponding exclusive parameter may indicate that the decoy should focus on the slow drift of the position rather than the rapid change of the heading, and the system will generate a decoy signal that slowly increases in the position dimension and changes little in the heading dimension. In specific implementation, the integration operation is usually completed in a real-time scheduling loop, and the decoy instruction is recalculated according to the latest flight state, shared parameter and exclusive parameter in each control period. Optionally, the integration process can introduce a feedback mechanism to fine-tune the scaling factor of the exclusive parameter according to the actual response of the UAV after the execution of the decoy action, so as to optimize the decoy effect. It can be understood that this layered decoy strategy based on shared parameters and exclusive parameters not only ensures the consistency of the decoy behavior between different modes, but also has the adaptability to different flight scenes, so as to realize precise and concealed dynamic decoy control. In some embodiments, the adjustment of the exclusive parameter may not only depend on the current mode, but also consider the historical sequence of mode conversion or environmental context information, so that the decoy strategy is more forward-looking. Optionally, the basic decoy framework itself may also contain some preset complex decoy scripts that can be triggered by the exclusive parameter, further enhancing the realism and effectiveness of the decoy.

[0050] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A decoy method for preventing autonomous flight of a drone, characterized in that, The method comprises: continuously monitoring the flight state data stream of the unmanned aerial vehicle to identify the current flight mode type; dynamically extracting a control parameter group from the unmanned aerial vehicle navigation system according to the identified flight mode type; constructing an interference activation signal sequence using the control parameter group, and performing timing analysis on the interference activation signal sequence to locate a parameter change event point; dividing the parameter change event point into an intra-mode event point and an inter-mode event point, thereby synthesizing a flight path deviation parameter; generating a decoy control parameter based on the flight path deviation parameter, and executing the decoy control parameter to complete a decoy action.

2. The decoy method of claim 1, wherein, The dynamically extracting a control parameter group from the unmanned aerial vehicle navigation system according to the identified flight mode type comprises: capturing a control reference parameter set and its duration window according to the state change curve of the flight mode type; capturing a control update parameter set and its effective time interval according to the transition trigger point of the flight mode type; applying a parameter standardization program to process the control update parameter set to output a standardized update parameter set; and fusing the control reference parameter set and the standardized update parameter set to form the control parameter group.

3. The decoy method of claim 2, wherein, The applying a parameter standardization program to process the control update parameter set to output a standardized update parameter set comprises: calculating a parameter difference measure for each state stage by a deviation evaluation algorithm, wherein the deviation evaluation algorithm uses a state influence factor, a time decay coefficient, a time offset between a state transition trigger point and a parameter update time, and a reference parameter setting time; and adjusting the parameter value based on the proportion of the parameter difference measure in the total difference by a parameter normalization algorithm to obtain the standardized update parameter set.

4. The decoy method of claim 1, wherein, The constructing an interference activation signal sequence using the control parameter group comprises: selecting an update parameter subset from the control parameter group to generate an interference activation signal based on the update parameter subset; extracting a standardized update parameter set from the control parameter group to generate an interference activation signal based on the standardized update parameter set; comparing the interference activation signal based on the update parameter subset with the interference activation signal based on the standardized update parameter set to obtain activation signal patterns of the same flight mode type and different flight mode types in the state stage; and optimizing the generation time point of the interference activation signal sequence according to the activation signal pattern.

5. The decoy method of claim 1, wherein, The performing timing analysis on the interference activation signal sequence to locate a parameter change event point comprises: calculating a matching degree index of the interference activation signal sequence and a reference time sequence corresponding to a historical parameter change event point, and if the matching degree index exceeds a preset matching threshold, identifying it as a consistent type parameter change event point as an intra-mode parameter change event point; and calculating a matching degree index of the interference activation signal sequence and a reference time sequence corresponding to a historical parameter reset event point, and if the matching degree index is lower than a preset matching threshold, identifying it as a conflict type parameter change event point as an inter-mode parameter reset event point.

6. The decoy method of claim 5, wherein, The matching degree index of the interference activation signal sequence and the reference time sequence corresponding to the historical parameter reset event point comprises: calculating the covariance value of the time data of the interference activation signal sequence and the reference time sequence, and calculating the standard deviation of the time data of the interference activation signal sequence and the standard deviation of the reference time sequence, and obtaining the matching degree index by dividing the product of the covariance value and the standard deviation.

7. The decoy method of deterring autonomous flight of drones of claim 4, wherein, The generation time point of the interference activation signal sequence is optimized according to the activation signal mode, which comprises: if the activation signal mode is greater than a preset mode threshold, triggering the generation of the interference activation signal sequence in advance; if the activation signal mode is less than the preset mode threshold, delaying the generation of the interference activation signal sequence.

8. The decoy method of deterring autonomous flight of drones of claim 1, wherein, The parameter change event points are divided into mode-in event points and mode-out event points, and a flight path deviation parameter is synthesized, which comprises: extracting parameter change event points caused by environmental disturbance in the same flight mode type, and marking them as mode-in parameter change event points; extracting parameter reset event points triggered when the flight mode type is switched, and marking them as mode-out parameter change event points; calculating the parameter change amplitude in the same flight mode type based on the mode-in parameter change event points as an internal path deviation parameter; calculating the parameter change amplitude when the flight mode type is switched based on the mode-out parameter change event points as a switching path deviation parameter; and combining the internal path deviation parameter and the switching path deviation parameter as a flight path deviation parameter.

9. The decoy method of deterring autonomous flight of drones of claim 1, wherein, The flight path deviation parameter is used to generate a decoy control parameter, which comprises: taking the flight state stage as a time reference, analyzing the common mode in the flight path deviation parameter, extracting a parameter group common to all flight mode types, and marking it as a shared parameter; analyzing the specific mode in the flight path deviation parameter, extracting a parameter group corresponding only to a specific flight mode type, and marking it as an exclusive parameter; applying the shared parameter as a global decoy reference to the decoy process of all flight mode types; and adaptively adjusting the decoy parameters of the specific flight mode type according to the exclusive parameters. The decoy control parameter is executed to complete the decoy action, which comprises: establishing a basic decoy framework using the shared parameter, and integrating the specific values of the decoy parameters adjusted by the exclusive parameters to realize dynamic decoy control.

10. A system for preventing autonomous flight of a drone, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein, The processor, when executing the computer program, realizes the steps of the decoy method for preventing the unmanned aerial vehicle from autonomous flight according to any one of claims 1 to 9.

Images