A safety management system for tethered unmanned aerial vehicles

Abstract

The present invention relates to the field of unmanned aerial vehicle technology, and in particular to a tethered unmanned aerial vehicle safety management system, including multi-source data acquisition, data fusion and analysis, intelligent risk assessment, control instruction generation and execution mechanism modules. The multi-source data acquisition module uses a variety of sensors to collect the flight and tethered cable data of the unmanned aerial vehicle, and transmits them to the data fusion and analysis module through a time division multiple access protocol with low packet loss rate and low delay; the intelligent risk assessment module determines the risk level based on the information, combined with a risk assessment model based on a support vector machine and a Bayesian network algorithm, and generates a prompt message if it exceeds a preset range; the control instruction generation module generates instructions according to the preset strategy based on the risk prompt and sends them to the execution mechanism module after sorting; the flight control system in the execution mechanism module cooperates with the electric retracting and releasing device of the tethered cable to realize real-time adjustment and control of the flight attitude and cable status of the unmanned aerial vehicle, ensuring safe and stable execution of tasks in complex environments.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to a tethered unmanned aerial vehicle safety management system. Background Art

[0002] Tethered drones are widely used in many fields due to their unique advantages. For example, in military reconnaissance, they can hover over a specific area for a long time and transmit intelligence information in real time. In power inspection, they can conduct close-range and long-term monitoring of transmission lines and promptly discover hidden dangers of line faults. In emergency rescue scenarios, they can take off quickly, provide an aerial perspective for the rescue site, and assist in command decision-making.

[0003] However, with the continuous expansion of applications, tethered drones also face many challenges. When performing tasks in complex environments, their safety issues become increasingly prominent. Complex meteorological conditions, such as strong winds, heavy rains, low temperatures, etc., will have a serious impact on the flight stability of drones, and may cause the flight attitude to be out of control; electromagnetic interference environment, such as the presence of a large number of electronic devices in cities or near communication base stations, will interfere with the flight control system and communication links of drones, causing signal transmission interruptions or erroneous command execution; and geographical environmental factors, such as complex mountainous terrain and urban areas with numerous buildings, increase the risk of drones colliding with obstacles. At the same time, as a key component of tethered drones, the status of tethered cables is also directly related to flight safety. The cables may be over-stretched, entangled, and other problems, which will affect the normal flight of the drone and even cause serious accidents such as crashes;

[0004] Some existing safety management measures on the market often have limitations. Some systems can only monitor a single factor, such as only focusing on the flight attitude or only monitoring the tension of the tethered cable, and lack comprehensive consideration of multiple factors. In terms of risk assessment, the accuracy and real-time performance are insufficient, and it is difficult to accurately judge the risk level of drones under complex working conditions, resulting in the inability to issue effective risk warnings in a timely manner. In addition, the control measures are not intelligent enough, and targeted control instructions cannot be automatically generated and prioritized according to the degree of risk and emergency situations, making it difficult to achieve efficient adjustment and control of the drone's flight attitude and tethered cable status. These problems restrict the safe and stable operation of tethered drones in complex environments. Therefore, a tethered drone safety management system is proposed to address the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide a tethered drone safety management system to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A tethered drone safety management system, comprising:

[0008] Multi-source data acquisition module: integrated sensors are deployed on the drone body and tethered cables to collect multi-source raw data of the drone flight and tethered cables, and then the collected multi-source raw data is transmitted to the data fusion and analysis module in real time;

[0009] Data fusion and analysis module: Receives multi-source raw data from the multi-source data acquisition module, performs fusion processing on the multi-source raw data, removes data errors and redundancy, mines the inherent correlation of the data, forms effective data information that can accurately characterize the actual operating status of the drone and the environmental conditions in which it is located, and outputs the effective data information to the intelligent risk assessment module;

[0010] Intelligent risk assessment module: Based on the valid data information output by the received data fusion and analysis module, combined with the preset risk assessment model, the risk assessment model comprehensively considers multiple factors including the flight attitude of the drone, environmental factors and the status of the tethered cable, and makes a real-time judgment on the risk level currently faced by the drone. When the risk level is determined to exceed the preset normal range, the corresponding risk warning information is generated and passed to the control instruction generation module;

[0011] Control instruction generation module: Based on the risk warning information transmitted by the intelligent risk assessment module, it generates targeted control instructions according to the preset control strategy, and prioritizes the generated control instructions according to the severity and urgency of the risk, and then sends the prioritized control instructions to the actuator module;

[0012] Actuator module: Receives control instructions from the control instruction generation module, and various components including the UAV flight control system, the electric retractable wire device of the tethered cable and other related actuators work together to perform corresponding actions, so as to achieve real-time adjustment and control of the UAV flight attitude and the status of the tethered cable, and ensure that the tethered UAV can perform tasks safely and stably in complex environments.

[0013] As a preferred solution, the multi-source raw data collected by the multi-source data acquisition module is transmitted to the data fusion and analysis module through the time division multiple access TDMA communication protocol. The data transmission packet loss rate is less than 1%, the transmission delay time is within 10 milliseconds, and the data transmission follows the following formula:

[0014] ;

[0015] ;

[0016] in, Indicates the packet loss rate, is the number of packets lost during transmission, is the total number of packets sent, Indicates the transmission delay time, is the time it takes for the data packet to arrive at the receiving end, The timestamp of the data packet sent from the sender.

[0017] As a preferred solution, the data fusion and analysis module uses a data fusion algorithm that combines the Kalman filter algorithm with the weighted average algorithm to fuse the multi-source raw data. The state update equation of the Kalman filter algorithm is:

[0018] ;

[0019] ;

[0020] in, is the prior state estimate, is the posterior state estimate at the previous moment, is the state transfer matrix, is the control input matrix, is the control input vector, is the prior estimation error covariance matrix, is the posterior estimation error covariance matrix of the previous moment, is the process noise covariance matrix;

[0021] The measurement update equation is:

[0022] ;

[0023] ;

[0024] ;

[0025] in, is the Kalman gain, is the measurement matrix, is the measurement value vector, is the posterior state estimate, is the posterior estimation error covariance matrix, is the measurement noise covariance matrix, is the identity matrix;

[0026] In the weighted average algorithm, for a certain data dimension , the fused data The calculation formula is:

[0027] ;

[0028] in, is the number of data sources involved in the fusion, Indicates Data sources in the data dimension The data value on is the weight of the corresponding data source.

[0029] As a preferred solution, the risk assessment model adopted by the intelligent risk assessment module is based on the combination of support vector machine algorithm and Bayesian network algorithm for continuous training and optimization. For the support vector machine algorithm, in the binary classification problem, given the training data set ,in, is the feature vector, is the category label, and its goal is to find a hyperplane To divide the data into two categories, the optimization objective function is:

[0030] ;

[0031] ; ; ;

[0032] in, is the normal vector of the hyperplane, is the intercept, is the penalty parameter, is the slack variable;

[0033] Bayesian network constructs a directed acyclic graph To express the probability dependence between variables, is a collection of nodes, is a set of edges, nodes The conditional probability distribution of Indicates that in its parent node set In the known case The probability distribution of the entire Bayesian network is:

[0034] ;

[0035] By combining the support vector machine algorithm with the Bayesian network algorithm, the support vector machine is used to classify and identify sample data, and the Bayesian network is used to model the complex dependencies between variables. Based on the learning of the accumulated flight data, the weight distribution of each dimensional factor in the risk assessment is adaptively adjusted, and the risk level faced by drones under different working conditions is accurately judged. The generated risk warning information clearly includes the risk category, severity, and urgency.

[0036] As a preferred solution, the control strategy preset in the control instruction generation module formulates control instruction generation rules for different risk categories, different severity levels and different urgency levels. For the adjustment of the UAV flight attitude, a proportional-integral-differential controller is used, and the mathematical expression of its control law is: ,in, is the controller output, is the deviation between the set value and the actual measured value, is the proportionality coefficient, is the integration coefficient, Indicates that from the initial time To current time Deviation The integral of

[0037] For the tension adjustment of moored cables, based on the fuzzy logic control algorithm, the cable tension deviation and its deviation change rate are used as input fuzzy quantities. The cable tension adjustment amount is determined by setting fuzzy sets, membership functions, fuzzy inference rules and defuzzification methods. The adjustment range is set according to the cable tension situation and environmental conditions to achieve reasonable control of the cable tension.

[0038] For emergency operations, when the risk level reaches the preset threshold, the corresponding emergency operation process is triggered. The parameters in the process are determined based on pre-set mathematical relationships to ensure response to various risk situations and ensure the safety of the drone.

[0039] As a preferred solution, in the actuator module, the flight control system adjusts the relevant control instructions according to the flight attitude to accurately control the speed of each motor and the angle of the rudder surface, so as to realize the real-time correction and stable maintenance of the UAV's flight attitude; the electric retractable wire device of the tethered cable has an automatic tension adjustment function, and the built-in tension sensor feeds back the cable tension signal in real time, adjusts the retractable wire speed and strength according to the relevant control instructions, maintains the tethered cable tension within a reasonable range, and prevents the cable from abnormal conditions including over-stretching and entanglement, and the response time of the automatic tension adjustment function is controlled within 0.5 seconds, and its tension adjustment speed control follows the following formula: ,in, is the speed of cable tension adjustment, is the proportionality factor related to the device characteristics, The cable tension adjustment amount determined according to the control command.

[0040] It can be seen from the technical solution provided by the present invention that the tethered drone safety management system provided by the present invention has the following beneficial effects:

[0041] 1. Efficiency and accuracy of data collection and transmission:

[0042] Adopting multi-source data acquisition module: integrating multiple sensors on the drone body and tethered cables, it can fully acquire multi-source raw data related to the drone flight and tethered cables, covering key information such as flight attitude, environmental factors and cable status, providing a rich data foundation for subsequent precise analysis and control;

[0043] The time division multiple access (TDMA) communication protocol ensures the quality of data transmission: ensuring that the data transmission packet loss rate is less than 1% and the transmission delay time is within 10 milliseconds; low packet loss rate ensures data integrity and avoids analysis deviations caused by data loss; low latency enables the system to obtain the latest data in a timely manner, realize real-time monitoring and rapid response to the status of the drone, which is particularly important in complex flight environments. For example, when responding to sudden weather changes or emergency maneuvers of drones, the system can respond quickly to ensure flight safety;

[0044] 2. Accurate data fusion and risk assessment:

[0045] Data fusion combining Kalman filtering and weighted average algorithm: effectively removes data errors and redundancy, and mines the inherent correlation of data; Kalman filtering performs well in dynamic system state estimation, and can continuously update the estimation of the drone state based on the state at the previous moment and the current measurement value, such as accurately estimating the drone's position, speed, and attitude; the weighted average algorithm assigns corresponding weights according to the reliability of different data sources, integrates information from multiple data sources, obtains more accurate fusion data, and improves the accuracy of characterizing the actual operating status and environmental conditions of the drone;

[0046] Risk assessment model combining support vector machine and Bayesian network: Based on a large amount of historical and simulated data training and optimization, support vector machine is good at handling classification problems. It distinguishes data of different risk levels by finding the optimal hyperplane. Combined with the modeling ability of Bayesian network for complex dependencies between variables, such as considering the mutual influence between flight attitude, environmental factors and cable status, it can adaptively adjust the weight distribution of factors in each dimension in risk assessment. This enables the system to accurately judge the risk level faced by drones under different working conditions, generate detailed risk warning information including risk category, severity and urgency, provide accurate basis for subsequent control command generation, help prevent potential risks in advance, and reduce the probability of accidents.

[0047] 3. Intelligent control and stable execution:

[0048] The targetedness and priority strategy of the control instruction generation module: preset control instruction generation rules for different risk categories, severity and urgency; the proportional-integral-differential (PID) controller is used for flight attitude adjustment, which can accurately adjust the motor speed and rudder angle according to the deviation between the set value and the actual measured value to achieve fast and stable attitude correction; the tethered cable tension adjustment is based on the fuzzy logic control algorithm, which fully considers the cable tension deviation and its rate of change, reasonably controls the tension, and avoids excessive stretching or entanglement of the cable; the emergency operation can trigger the corresponding process in time when the risk level reaches the threshold, and the process parameters are determined according to mathematical relationships, ensuring that the system can take effective measures to ensure the safety of the drone in an emergency, and at the same time, execute the control instructions in an orderly manner according to the risk priority to avoid resource conflicts and chaotic operations;

[0049] Collaborative execution and reliable performance of actuator modules: The flight control system and the electric retractable cable device for the tethered cable work together. The flight control system accurately controls the motor speed and rudder angle, corrects the flight attitude in real time, and ensures flight stability. The automatic tension adjustment function of the electric retractable cable device responds quickly (within 0.5 seconds), and can quickly adjust the cable tension according to the control command to maintain it within a reasonable range to prevent cable failure. The reliable operation and efficient collaboration of each actuator ensure the system's real-time and effective control of the UAV's flight attitude and cable status, enabling tethered UAVs to perform tasks safely and stably in complex environments, improving the reliability and adaptability of UAV operations and expanding their scope of application, such as playing a greater role in long-term monitoring, emergency rescue and other scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The present invention is a schematic diagram of the overall structure of a tethered drone safety management system. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0053] like Figure 1 As shown, an embodiment of the present invention provides a tethered drone safety management system, including a multi-source data acquisition module, a data fusion and analysis module, an intelligent risk assessment module, a control instruction generation module and an actuator module.

[0054] In this embodiment, the multi-source data acquisition module integrates sensors and is deployed on the drone body and tethered cables to collect multi-source raw data of the drone flight and tethered cables, and then transmits the collected multi-source raw data to the data fusion and analysis module in real time;

[0055] Further, the specific operation steps of the multi-source data acquisition module are as follows:

[0056] Sensor Integration and Deployment:

[0057] Determine the type and quantity of sensors:

[0058] According to the flight parameters of the drone and the status of the tethered cable that need to be monitored, select the appropriate sensor type, such as accelerometer, gyroscope, barometer, GPS (Global Positioning System) sensor, geomagnetic sensor, wind speed and direction sensor, temperature sensor, tension sensor, etc.;

[0059] Determine the number and location of each sensor installed on the drone body and tethered cables to ensure that relevant data can be fully and accurately obtained; for example, install multiple accelerometers and gyroscopes at key locations on the drone body to accurately measure flight attitude; install tension sensors at different locations on the tethered cables to monitor cable tension changes in real time;

[0060] Sensor installation and debugging:

[0061] Install the selected sensors on the drone body and tethered cables. For sensors on the drone body, ensure that they are firmly installed and do not affect the flight performance of the drone, and that the sensor's measurement direction is correct. For sensors on tethered cables, take protective measures to prevent them from being damaged during flight.

[0062] After completing the sensor installation, conduct debugging work; check whether each sensor is working properly, calibrate the sensor's measurement parameters to ensure the accuracy of the data; for example, calibrate the accelerometer and gyroscope so that their measured values ​​are consistent with the actual acceleration and angular velocity; test the positioning accuracy of the GPS sensor to ensure that it can accurately obtain the drone's location information;

[0063] Data collection:

[0064] To set the data collection frequency:

[0065] Determine the appropriate data collection frequency based on the flight characteristics and safety management requirements of the drone. For key parameters such as flight attitude, such as acceleration, angular velocity, attitude angle, etc., a higher collection frequency (e.g., dozens or even hundreds of times per second) may be required to capture rapid changes in the drone's flight status. For environmental factors such as temperature, wind speed and direction, the collection frequency can be relatively low (e.g., several times per second), but it must also ensure that changes in the environment can be reflected in a timely manner. For the tension of the tethered cable, set an appropriate collection frequency based on the cable characteristics and the possible tension change speed to ensure that tension anomalies can be detected in a timely manner.

[0066] Start data collection:

[0067] The multi-source data acquisition module is turned on, and each sensor starts to collect data according to the set acquisition frequency; the sensor converts the physical quantity measured in real time into an electrical signal, and performs preliminary signal processing, such as filtering and amplification, to improve the data quality; for example, the accelerometer converts the measured acceleration signal into a digital signal, removes the noise interference, and then temporarily stores the processed data in the cache of the acquisition module;

[0068] Data transmission:

[0069] To establish a TDMA communication connection:

[0070] Establish a time division multiple access (TDMA) communication connection between the multi-source data acquisition module and the data fusion and analysis module; ensure that the communication parameters of the two modules are set consistently, including communication frequency, time slot allocation, etc.; for example, set the communication frequency to 2.4GHz, and reasonably allocate the time slot for each sensor data transmission according to the data volume and real-time requirements to ensure the orderliness and efficiency of data transmission;

[0071] Data packaging and transmission:

[0072] The collected multi-source raw data is packaged in a predetermined format, and information such as data identification and timestamp is added so that the data can be accurately identified and processed at the receiving end; for example, the data from the accelerometer is marked as "ACC" and the acquisition timestamp is attached, and then the packaged data is placed in the time slot specified by the TDMA communication protocol for transmission;

[0073] The multi-source data acquisition module sends the data of each sensor to the data fusion and analysis module in turn according to the time slot arrangement of the TDMA communication protocol; during the transmission process, the data transmission status is monitored in real time to ensure the integrity and accuracy of the data; if a transmission error or lost data packet is found, retransmission or error correction measures are taken in a timely manner;

[0074] Data reception confirmation:

[0075] Data reception and verification:

[0076] The data fusion and analysis module receives data from the multi-source data acquisition module in a specified time slot; performs integrity check on the received data to check whether the data is complete and the format is correct; for example, through methods such as checksum or cyclic redundancy check (CRC), it verifies whether errors occur during data transmission; if the check finds data errors, it sends a retransmission request to the multi-source data acquisition module;

[0077] Receive confirmation feedback:

[0078] When the data fusion and analysis module successfully receives complete and correct data, it sends a reception confirmation signal to the multi-source data acquisition module; after receiving the confirmation signal, the multi-source data acquisition module continues with the next round of data acquisition and transmission; if no confirmation signal is received, the data is resent after a certain period of time until a confirmation signal is received or the upper limit of the retransmission number is reached; this ensures that the data can be reliably transmitted from the acquisition module to the analysis module, providing an accurate data basis for subsequent operations such as data fusion and risk assessment.

[0079] In this embodiment, the data fusion and analysis module receives the multi-source raw data from the multi-source data acquisition module, performs fusion processing on the multi-source raw data, removes data errors and redundancy, mines the internal association of the data, forms effective data information that can accurately characterize the actual operating status of the drone and the environmental conditions in which it is located, and outputs the effective data information to the intelligent risk assessment module;

[0080] Further, the specific operation steps of the data fusion and analysis module are as follows:

[0081] Data reception:

[0082] Establish communication interface and buffer:

[0083] The data fusion and analysis module first establishes a communication interface with the multi-source data acquisition module to ensure that data can be stably received; at the same time, a data buffer is set up to temporarily store the received multi-source raw data; for example, a certain size of memory space is opened as a buffer, and when data is received, the data is stored in the buffer in order and waits for subsequent processing;

[0084] Data reception and preliminary inspection:

[0085] According to the preset communication protocol, receive the multi-source raw data from the multi-source data acquisition module; during the receiving process, perform a preliminary check on the data, including whether the data format, data length, etc. meet expectations; if the data format is wrong or the data length is abnormal, record the error information and try to request to resend the data; for example, check whether the header and tail of each data packet are correct, whether the number of bits in the data field meets the regulations, etc.;

[0086] Data fusion processing:

[0087] Data synchronization and time alignment:

[0088] Since the data acquisition frequency and transmission delay of different sensors may be different, it is necessary to perform time synchronization and alignment operations on multi-source raw data; using a reference clock (such as the drone system clock or GPS time signal) as a reference, adjust the data from each sensor to the same time scale; for example, for accelerometer and gyroscope data with slightly different acquisition times, interpolation or data resampling methods can be used to make them correspond in time, so as to accurately analyze the instantaneous state of the drone later;

[0089] Data cleaning and denoising:

[0090] Use filtering algorithms to denoise the raw data and remove the noise interference introduced during the measurement process; common filtering methods include Kalman filtering and mean filtering. For accelerometer data, Kalman filtering can update the optimal estimate of acceleration in real time based on the previous state estimate and the current measurement value, effectively reducing the impact of noise on the data and making the data smoother and more accurate.

[0091] Identify and remove outliers; detect and remove obviously erroneous data points by setting reasonable data thresholds or using statistical methods (such as calculating the mean and standard deviation of the data, and treating data that deviates too much from the mean as outliers); for example, if the wind speed measurement at a certain moment exceeds the normal wind speed range (based on historical data and environmental conditions), mark the data point as an outlier and correct or delete it;

[0092] Data fusion algorithm application (combination of Kalman filter and weighted average algorithm):

[0093] Kalman filter status update:

[0094] According to the state update equation of Kalman filter algorithm , using the posterior state estimate of the previous moment , state transfer matrix , control input matrix and the control input vector , calculate the prior state estimate at the current moment ; For example, for the position and velocity state estimation of a drone, if the position and velocity estimates at the previous moment, as well as the drone's motion model (represented by the state transfer matrix) and possible control inputs (such as motor speed changes, etc.) are known, the current position and velocity can be preliminarily predicted;

[0095] Update equation based on prior estimation error covariance matrix , using the posterior estimation error covariance matrix of the previous moment , state transfer matrix and the process noise covariance matrix , calculate the prior estimation error covariance matrix at the current moment ; Process noise covariance matrix It reflects the influence of disturbance factors not considered in the system model on the state estimation;

[0096] Kalman filter measurement update:

[0097] Calculate Kalman gain , where the measurement matrix describes the measurement values The relationship between the state variables, the measurement noise covariance matrix Represents the noise level in the measurement process; for example, for GPS measurement position information, the measurement matrix The latitude and longitude information measured by GPS is associated with the position state variables of the UAV, and the noise covariance matrix is ​​measured. Consider the influence of GPS measurement error to calculate the appropriate Kalman gain;

[0098] The Kalman gain calculated from the measurement vector and , update the posterior state estimate ; Compare the measured value with the predicted value based on the prior estimate, and use the Kalman gain to adjust the prior estimate to obtain a more accurate posterior state estimate; for example, the position measured by GPS is fused with the position predicted by the motion model to obtain a more accurate estimate of the drone's position;

[0099] Update the posterior estimation error covariance matrix , used for the next iterative calculation;

[0100] Weighted average algorithm fusion (taking a certain data dimension as an example):

[0101] Determine the number of data sources involved in the fusion And the data value of each data source in this data dimension and the corresponding weights ; The weight can be determined based on the accuracy and reliability of the sensor. For example, for the gyroscope and accelerometer data for measuring the attitude angle of the drone, if the gyroscope has high accuracy but has drift problems, and the accelerometer has slightly lower accuracy but good stability, the gyroscope can be given a higher weight (such as 0.7) and the accelerometer a lower weight (such as 0.3);

[0102] According to the weighted average algorithm formula , calculate the fused data value; for the attitude angle data, the attitude angle data measured by the gyroscope and accelerometer at the same time are weighted and summed according to their respective weights to obtain the fused attitude angle estimation value, which combines the advantages of both and more accurately reflects the actual attitude of the drone;

[0103] Data Association and Mining:

[0104] Feature extraction and parameter calculation:

[0105] Extract characteristic parameters related to the UAV's operating status and environmental conditions from the fused data; for example, calculate the UAV's flight speed, acceleration change rate, attitude angle change rate, and tethered cable tension change trend; these characteristic parameters can more intuitively reflect the UAV's dynamic characteristics and cable status changes;

[0106] Analyze the correlation between different data; determine the degree of correlation between flight attitude data and environmental factors (such as wind speed and direction), as well as the relationship between the tether cable tension and the UAV flight status (such as altitude and speed changes) through statistical methods or machine learning algorithms (such as correlation analysis algorithms); for example, it is found that when the wind speed increases, the roll angle and pitch angle of the UAV will change accordingly, and the tether cable tension will also be affected. By establishing these relationship models, it is helpful to have a deeper understanding of the behavior of the UAV system;

[0107] Build a data association model:

[0108] According to the extracted features and the data association relationship obtained by analysis, a data association model is constructed; the model can be in the form of mathematical equations, logical rules or neural network models; for example, a mathematical model describing the relationship between the flight attitude of the UAV and environmental factors and cable tension is established, through which the cable tension change can be predicted based on the known environmental data and part of the flight attitude data, or vice versa, possible flight attitude problems or environmental interference factors can be judged based on abnormal cable tension;

[0109] Generation and output of effective data information:

[0110] Effective data integration and packaging:

[0111] Integrate the characteristic parameters, correlation model results and other relevant data information obtained after fusion processing to form effective data information that can comprehensively and accurately characterize the actual operating status of the drone and the environmental conditions in which it is located; for example, organize the drone's current position, speed, attitude angle, cable tension, ambient temperature, wind speed and direction and other information according to a certain structure to facilitate the use of subsequent modules;

[0112] Encapsulate the integrated valid data information and add necessary identification and description information, such as data source, collection time, data type, etc. This helps the intelligent risk assessment module to correctly identify and parse the data;

[0113] Data output and transmission:

[0114] Establish a communication connection with the intelligent risk assessment module, and send the encapsulated valid data information to the intelligent risk assessment module according to the predetermined communication protocol; ensure the reliability and real-time performance of data transmission, and adopt appropriate communication technology (such as wireless communication or wired communication, determined according to the system architecture) and data transmission format (such as JSON or XML format) to enable the intelligent risk assessment module to smoothly receive and process data; for example, send data in JSON format to the intelligent risk assessment module through the wireless data transmission module, and after receiving the data, the module can easily parse and extract key information for risk assessment.

[0115] In this embodiment, the intelligent risk assessment module performs analysis based on the valid data information output by the received data fusion and analysis module in combination with a preset risk assessment model. The risk assessment model comprehensively considers multiple factors including the flight attitude of the drone, environmental factors, and the status of the tethered cable, and makes a real-time judgment on the risk level currently faced by the drone. When the risk level is determined to exceed the preset normal range, the corresponding risk warning information is generated and transmitted to the control instruction generation module;

[0116] Further, the specific operation steps of the intelligent risk assessment module are as follows:

[0117] Data reception and preprocessing:

[0118] Establish communication interface and data reception:

[0119] The intelligent risk assessment module establishes a stable communication interface with the data fusion and analysis module to ensure timely and accurate reception of valid data information; when the data fusion and analysis module sends data, the intelligent risk assessment module starts the receiving program and temporarily stores the received data in the buffer;

[0120] Perform format check and integrity verification on the received data; check whether the data conforms to the predetermined format specifications, such as whether the data fields are complete and whether the data types are correct; if the data format is wrong or the data is incomplete, promptly notify the data fusion and analysis module to resend the data;

[0121] Data parsing and feature extraction:

[0122] Parse the received and verified data to extract key characteristic parameters related to the UAV's flight attitude, environmental factors, and tethered cable status; for example, obtain the UAV's current attitude angle (roll angle, pitch angle, yaw angle), flight speed, acceleration, location information, environmental temperature, humidity, wind speed and direction, and tethered cable tension, length change and other parameters from the data;

[0123] Normalize the extracted characteristic parameters to make them in the same order of magnitude or specific numerical range, so as to facilitate unified calculation and comparison in the subsequent risk assessment model; for example, normalize the flight speed to between 0 and 1, convert the attitude angle to radians and perform appropriate scaling, etc.;

[0124] Risk assessment model calculation:

[0125] Support vector machine algorithm classification and discrimination (based on training data set):

[0126] Load a pre-trained support vector machine (SVM) model, which is trained based on a large amount of historical flight data and simulation data. The training data set contains feature vectors under various working conditions (such as the vector composed of the parameters such as the flight attitude of the drone extracted above, environmental factors and cable status) and corresponding category labels (indicating different risk levels or status categories, such as normal, low risk, medium risk, high risk, etc.);

[0127] The feature vector composed of the currently extracted normalized feature parameters is input into the SVM model, and the model uses the learned hyperplane (composed of the normal vector and the intercept Confirm, satisfy ) classify and discriminate the data, and calculate the probability or score of the feature vector belonging to different risk level categories; for example, by calculating the distance from the feature vector to the hyperplane, and based on the mapping relationship between the distance and the risk level, preliminarily determine the approximate range of the current risk of the drone;

[0128] During the SVM calculation process, if the data points are located near the hyperplane or the classification is unclear, consider the relaxation variable and penalty parameters The role of the slack variable is to allow a certain degree of data points to violate the constraints (i.e. ), penalty parameter Control the degree of penalty for data points that violate the constraints, thereby making a trade-off between model complexity (to avoid overfitting) and classification accuracy;

[0129] Bayesian network algorithm modeling and probability calculation:

[0130] Construct a Bayesian network model, which is a directed acyclic graph Represents the probabilistic dependency between variables, where the node set Including drone flight attitude variables (such as attitude angle change rate, speed change rate, etc.), environmental factor variables (wind speed and direction changes, temperature changes, etc.) and tethered cable state variables (tension change rate, cable length change, etc.), edge set Represents the causal or dependency relationship between variables; for example, changes in wind speed and direction may affect the attitude angle of the drone, and changes in the attitude angle of the drone may cause changes in the tension of the tethered cable. These relationships are reflected in the Bayesian network through directed edges;

[0131] Conditional probability distribution according to Bayesian network , combined with the currently received valid data information, calculate the conditional probability of each node; for example, given the current wind speed and direction data, calculate the conditional probability of the change in the attitude angle of the drone; then calculate the conditional probability of the change in the tension of the tethered cable based on the change in the attitude angle; calculate the conditional probability of the whole system in different state combinations (i.e., joint probability distribution) through the chain rule ) to further refine the assessment of drone risk status;

[0132] The Bayesian network can update the probability distribution of nodes in real time based on new data, thereby dynamically adjusting the judgment of the drone risk level; for example, when new environmental data or drone flight status data is received, the Bayesian network will recalculate the probability of each node to reflect the impact of changes in system status on risk assessment;

[0133] Comprehensive determination of risk level:

[0134] Comprehensively analyze the classification and discrimination results of the support vector machine algorithm and the probability calculation results of the Bayesian network algorithm; for example, different weights can be set for the results of the two algorithms (the weight values ​​are determined according to their accuracy and reliability in model training and verification), and then the weighted sum is used to obtain the final risk assessment score;

[0135] According to the preset risk level classification threshold, the comprehensive risk assessment score is converted into the corresponding risk level; for example, 0-0.3 is set as low risk, 0.3-0.6 is set as medium risk, and 0.6-1 is set as high risk. When the calculated risk assessment score is 0.4, it is determined that the drone is currently facing medium risk;

[0136] Generation and delivery of risk warning information:

[0137] Risk warning information generation:

[0138] When the risk level is determined to be beyond the preset normal range (such as the upper limit of low risk), detailed risk warning information is generated based on the risk assessment results; the risk warning information includes risk category (such as abnormal flight attitude risk, environmental interference risk, tethered cable failure risk, etc.), severity (corresponding to risk level, such as medium risk or high risk) and urgency (determined according to the speed and degree of impact of the risk on the safety of the drone, for example, rapidly changing risk parameters may correspond to a high degree of urgency);

[0139] The risk warning information may also include relevant suggested measures or reference information. For example, for the risk of abnormal flight attitude, possible causes (such as motor failure, sensor deviation, etc.) and preliminary troubleshooting methods may be indicated so that operators can quickly understand the situation and take corresponding measures.

[0140] The information is passed to the control instruction generation module:

[0141] Establish a communication connection with the control instruction generation module, and send the generated risk warning information to the control instruction generation module according to the predetermined communication protocol; ensure the timeliness and accuracy of information transmission, and adopt reliable communication methods (such as reliable wireless transmission protocols or wired communication links) to prevent information loss or erroneous transmission; for example, send the risk warning information to the control instruction generation module through an encrypted wireless communication channel, and the module will parse and process the information after receiving it, and generate corresponding control instructions according to the risk warning.

[0142] In this embodiment, the control instruction generation module generates targeted control instructions according to the preset control strategy based on the risk warning information transmitted by the intelligent risk assessment module, and prioritizes the generated control instructions according to the severity and urgency of the risk, and then sends the prioritized control instructions to the execution mechanism module;

[0143] Further, the specific operation steps of the control instruction generation module are as follows:

[0144] Receiving and analyzing risk warning information:

[0145] Establish communication interface and data reception:

[0146] The control instruction generation module establishes a communication connection with the intelligent risk assessment module and monitors the transmission of risk warning information in real time; when receiving risk warning information, the receiving program is started to receive and store the data completely in the buffer;

[0147] Perform format verification and integrity check on the received data to ensure the accuracy and completeness of the risk warning information; if the data format is wrong or the data is missing, send feedback to the intelligent risk assessment module in a timely manner and request to resend the correct information;

[0148] Information parsing and parameter extraction:

[0149] Analyze the risk warning information and extract key parameters, including risk category (such as abnormal flight attitude, excessive tether cable tension, sudden change in ambient wind speed, etc.), severity (such as low risk, medium risk, high risk, etc.) and urgency (such as urgent, relatively urgent, general, etc.), as well as other relevant data that may be included (such as abnormal attitude angle value, specific cable tension value, ambient wind speed change, etc.); these parameters will serve as an important basis for generating control instructions;

[0150] Control command generation:

[0151] Formulate control instruction rules for different risk categories:

[0152] According to the risk category, the corresponding control instruction generation rules are searched in the preset control strategy library; for example:

[0153] If the risk is an abnormal flight attitude, possible control rules include adjusting the speed of each motor of the drone to correct the attitude angle; specifically, when the drone is detected to be tilted to the left (abnormal roll angle), the speed of the right motor is increased, while the speed of the left motor is appropriately reduced to restore the drone to a horizontal attitude; for abnormal pitch angles, the speed difference between the front and rear motors is adjusted accordingly;

[0154] For the risk of excessive tension in the moored cable, the control rule may be to start the electric reeling and releasing device of the moored cable to adjust the tension; if the tension exceeds a certain range of the upper limit, control the reeling device to reel in the cable at an appropriate speed to reduce the cable length and thus reduce the tension; if the tension is lower than the lower limit, control the releasing device to release the cable to restore the cable tension to a reasonable range;

[0155] When facing the risk of sudden changes in wind speed, adjust the flight direction and speed of the drone according to the wind speed and direction. For example, when flying against the wind, increase the thrust of the drone appropriately to maintain the speed and position stability. When flying with the wind and the wind speed is too high, reduce the flight speed to prevent loss of control due to the wind speed.

[0156] Combine severity and urgency to optimize control instructions:

[0157] According to the severity and urgency of the risk, adjust and optimize the parameters of the initially generated control instructions; for example:

[0158] In more serious situations (such as high risk), increase control strength; for flight attitude adjustment, change the motor speed more significantly; for tether cable tension adjustment, speed up the reeling and releasing speed; for environmental interference, adjust flight parameters more quickly;

[0159] When the urgency is high (such as in an emergency), key control instructions are executed first to ensure the safety of the drone. For example, when there are both abnormal flight attitude and tethered cable tension problems, but the abnormal flight attitude poses a more urgent threat to the safety of the drone, resources are concentrated on adjusting the flight attitude first, and the cable tension problem is dealt with after the attitude stabilizes. At the same time, in an emergency, some special control strategies may be adopted, such as immediately starting the backup power supply or switching to safe flight mode.

[0160] Control instruction priority sorting:

[0161] Prioritization Evaluation Criteria:

[0162] Establish a priority assessment system that comprehensively considers factors such as the severity and urgency of the risk and the degree of impact on drone flight safety; for example, set weights for severity and urgency, and obtain the priority value corresponding to each risk warning through weighted calculation; high-risk and urgent situations are given the highest priority, and low-risk and moderately urgent situations are given a lower priority;

[0163] Sort the control instructions:

[0164] Sort the generated control instructions according to the priority values; a sorting algorithm (such as bubble sort, quick sort, etc.) can be used to sort the control instruction list to ensure that the control instructions with high priority are ranked first; for example, the control instructions for correcting serious flight attitude abnormalities are ranked first, followed by the instructions for handling emergency tether cable tension problems, and then the instructions for dealing with general environmental interference, etc.;

[0165] Control command sending:

[0166] Establish a communication connection with the actuator module:

[0167] Establish a reliable communication link with the actuator module to ensure that the control instructions can be transmitted accurately. The communication method can be selected according to the system architecture, such as wireless communication (Wi-Fi, Bluetooth, 4G / 5G, etc.) or wired communication (Ethernet, serial communication, etc.). During the connection establishment process, configure and test the communication parameters to ensure the stability of communication.

[0168] Control command packaging and sending:

[0169] The sorted control instructions are packaged in a format that can be recognized by the actuator module, and necessary instruction headers, instruction numbers, parameter information, etc. are added; for example, for the motor speed control instructions of the flight control system, the target motor number, speed value and other information are packaged into a data frame according to a specific protocol format;

[0170] The packaged control instructions are sent to the actuator module in sequence. During the sending process, the sending status is monitored in real time to ensure that the instructions are sent completely. If the sending fails, resend the instructions according to the preset resending mechanism until the actuator module successfully receives the control instructions. At the same time, the sending log of the control instructions is recorded, including the sending time, instruction content, reception confirmation information, etc., for subsequent query and troubleshooting.

[0171] In this embodiment, the actuator module receives the control command sent by the control command generation module, and various components including the UAV flight control system, the electric retractable wire device of the tethered cable and other related actuator components cooperate to perform corresponding actions, so as to achieve real-time adjustment and control of the UAV flight attitude and the state of the tethered cable, and ensure that the tethered UAV can perform tasks safely and stably in complex environments;

[0172] Further, the specific operation steps of the actuator module are as follows:

[0173] Control command reception and analysis:

[0174] Communication interface establishment and command reception:

[0175] Each component in the actuator module (UAV flight control system, tethered cable electric retracting and releasing device, etc.) establishes a communication interface with the control instruction generation module to ensure that the control instruction can be received in time; when the control instruction generation module sends an instruction, each component starts the receiving program and stores the received instruction data in the corresponding buffer;

[0176] Perform integrity check and format verification on the received control instructions; check whether the instruction data is complete and whether there is any data loss or error; at the same time, verify whether the instruction format conforms to the pre-defined protocol specifications to ensure that the instruction content can be correctly parsed; if the instruction is found to be wrong or incomplete, promptly feedback to the control instruction generation module and request resending;

[0177] Instruction parsing and task allocation:

[0178] Parse the verified control instructions and extract the operation information and parameters contained therein; for example, for the instructions of the UAV flight control system, parse out the target flight attitude (such as the desired attitude angle, speed, etc.), motor control signals (such as the speed setting value of each motor), and other information; for the instructions of the electric retractable wire device for the tethered cable, parse out the parameters such as the target value for adjusting the cable tension, the direction and speed of the retractable wire;

[0179] According to the command analysis results, the tasks are assigned to the corresponding execution components; the flight attitude control related instructions are sent to the UAV flight control system, and the tethered cable tension adjustment instructions are sent to the tethered cable electric retracting and releasing device to ensure that each execution component is clear about the specific actions it needs to perform;

[0180] Drone flight attitude adjustment:

[0181] Flight control system response:

[0182] The UAV flight control system calculates the motor control signals required to achieve the target attitude based on the received flight attitude control instructions; for example, through the flight control algorithm (such as the PID control algorithm), the speed value that needs to be adjusted for each motor is calculated based on the deviation between the current actual attitude and the target attitude; if the instruction requires the UAV to roll to the right at a certain angle, the flight control system will calculate the speed that the right motor should reduce and the speed that the left motor should increase based on the difference between the current attitude angle and the expected attitude angle;

[0183] Send control signals to the motor drive module to drive the motor to rotate; the motor drive module accurately controls the current and voltage of each motor according to the instructions of the flight control system, so that the motor runs at the set speed; at the same time, the flight control system monitors the actual speed feedback of the motor in real time, compares it with the target speed, and continuously adjusts the motor drive signal through closed-loop control to ensure that the actual speed of the motor is consistent with the target speed, thereby achieving accurate adjustment of the UAV's flight attitude;

[0184] Rudder angle adjustment (if applicable):

[0185] For drones equipped with rudders (such as fixed-wing drones), the flight control system will also adjust the angle of the rudder according to the flight attitude control command; for example, when the heading of the drone needs to be changed, the flight control system will control the rudder to rotate to a certain angle, generate a corresponding yaw torque, and turn the drone; the servo drives the rudder to rotate to the specified angle and maintain stability according to the signal of the flight control system; the flight control system will also monitor the rudder angle feedback in real time to ensure that the rudder angle is accurate to achieve the expected flight attitude change;

[0186] Tether cable tension adjustment:

[0187] Electric pay-off and take-up device response:

[0188] After receiving the tension adjustment command, the electric retractable and retractable device of the moored cable first compares the current cable tension value fed back by the built-in tension sensor with the target tension value in the command, and calculates the tension difference that needs to be adjusted;

[0189] According to the tension difference and the requirements of the reeling and unreeling speed in the instruction, the motor is controlled to drive the reeling and unreeling mechanism to work; if the current cable tension is too large and the cable needs to be unreeled to reduce the tension, the electric reeling and unreeling device will start the unreeling motor and unreel a certain length of cable at the set speed; conversely, if the cable tension is too small and the cable needs to be reeled in to increase the tension, the reeling motor is started to reel in the cable; during the reeling and unreeling process, the tension sensor monitors the cable tension changes in real time and transmits the feedback signal to the control unit of the electric reeling and unreeling device to achieve closed-loop control of the tension;

[0190] Automatic tension adjustment function realizes:

[0191] The electric retractable wire device has an automatic tension adjustment function. Its control unit uses the cable tension deviation and its deviation change rate as input fuzzy quantities according to a preset tension adjustment algorithm (such as a fuzzy logic control algorithm), and determines the cable tension adjustment amount through pre-set fuzzy sets, membership functions, fuzzy inference rules and defuzzification methods; for example, when the tension deviation is large and the deviation change rate is fast, the retractable wire speed is increased; when the tension is close to the target value and the deviation change rate is small, the retractable wire speed is slowly adjusted to avoid tension over-adjustment;

[0192] The response time of the automatic tension adjustment function of the electric take-up and pay-off device is controlled within 0.5 seconds, and its tension adjustment speed follows the formula ,in, is the speed of cable tension adjustment, is the proportionality factor related to the device characteristics, The cable tension adjustment amount is determined according to the control command; in this way, the tethered cable tension is adjusted quickly and accurately to keep it within a reasonable range at all times, preventing the cable from being overstretched, entangled, or other abnormal conditions;

[0193] Collaborative execution and status monitoring:

[0194] Each execution component works together:

[0195] The UAV flight control system and the electric retractable cable device for the tethered cable and other executive components coordinate with each other while performing their respective tasks; for example, when the UAV is performing rapid maneuvering flight, the change in flight attitude will cause the tension of the tethered cable to change. At this time, the flight control system and the electric retractable cable device need to work together; when the flight control system adjusts the flight attitude, it transmits the attitude change information to the electric retractable cable device in real time. The electric retractable cable device adjusts the retractable cable action in advance according to the trend of cable tension changes, ensuring that the cable tension remains stable during the change of flight attitude, avoiding the impact of sudden tension changes on the flight safety of the UAV;

[0196] Condition monitoring and feedback:

[0197] During the execution of the action, each execution component monitors its own status and execution effect in real time, and feeds back the relevant information to the control unit of the actuator module; for example, the UAV flight control system monitors the status parameters such as motor temperature, current, battery power, and the actual changes in flight attitude; the tethered cable electric retractable device monitors the tension sensor status, motor working status, retractable cable length and other information; the control unit integrates and analyzes this feedback information to determine whether the execution action has achieved the expected effect; if an abnormal situation is found (such as motor overheating, abnormal fluctuations in cable tension, etc.), take corresponding measures in a timely manner, such as adjusting control instructions, issuing alarm information or starting backup equipment, etc., to ensure that the tethered UAV can perform tasks safely and stably in complex environments.

[0198] In this embodiment, the multi-source raw data collected by the multi-source data acquisition module is transmitted to the data fusion and analysis module through the time division multiple access TDMA communication protocol, and its data transmission packet loss rate is less than 1%, and the transmission delay time is within 10 milliseconds. The data transmission follows the following formula:

[0199] ;

[0200] ;

[0201] in, Indicates the packet loss rate, is the number of packets lost during transmission, is the total number of packets sent, Indicates the transmission delay time, is the time it takes for the data packet to arrive at the receiving end, The timestamp of the data packet sent from the sender.

[0202] In this embodiment, the data fusion and analysis module uses a data fusion algorithm that combines the Kalman filter algorithm with the weighted average algorithm to fuse the multi-source raw data. The state update equation of the Kalman filter algorithm is:

[0203] ;

[0204] ;

[0205] in, is the prior state estimate, is the posterior state estimate at the previous moment, is the state transfer matrix, is the control input matrix, is the control input vector, is the prior estimation error covariance matrix, is the posterior estimation error covariance matrix of the previous moment, is the process noise covariance matrix;

[0206] The measurement update equation is:

[0207] ;

[0208] ;

[0209] ;

[0210] in, is the Kalman gain, is the measurement matrix, is the measurement value vector, is the posterior state estimate, is the posterior estimation error covariance matrix, is the measurement noise covariance matrix, is the identity matrix;

[0211] In the weighted average algorithm, for a certain data dimension , the fused data The calculation formula is:

[0212] ;

[0213] in, is the number of data sources involved in the fusion, Indicates Data sources in the data dimension The data value on is the weight of the corresponding data source.

[0214] In this embodiment, the risk assessment model adopted by the intelligent risk assessment module is based on the support vector machine algorithm combined with the Bayesian network algorithm for continuous training and optimization. For the support vector machine algorithm, in the binary classification problem, given a training data set ,in, is the feature vector, is the category label, and its goal is to find a hyperplane To divide the data into two categories, the optimization objective function is:

[0215] ;

[0216] ; ; ;

[0217] in, is the normal vector of the hyperplane, is the intercept, is the penalty parameter, is the slack variable;

[0218] Bayesian network constructs a directed acyclic graph To express the probability dependence between variables, is a collection of nodes, is a set of edges, nodes The conditional probability distribution of Indicates that in its parent node set In the known case The probability distribution of the entire Bayesian network is: ;

[0219] By combining the support vector machine algorithm with the Bayesian network algorithm, the support vector machine is used to classify and identify sample data, and the Bayesian network is used to model the complex dependencies between variables. Based on the learning of the accumulated flight data, the weight distribution of each dimensional factor in the risk assessment is adaptively adjusted, and the risk level faced by drones under different working conditions is accurately judged. The generated risk warning information clearly includes the risk category, severity, and urgency.

[0220] In this embodiment, the control strategy preset by the control instruction generation module formulates control instruction generation rules for different risk categories, different severity levels and different urgency levels. For the adjustment of the UAV flight attitude, a proportional-integral-differential controller is adopted, and the mathematical expression of its control law is: ,in, is the controller output, is the deviation between the set value and the actual measured value, is the proportionality coefficient, is the integration coefficient, Indicates that from the initial time To current time Deviation The integral of

[0221] For the tension adjustment of moored cables, based on the fuzzy logic control algorithm, the cable tension deviation and its deviation change rate are used as input fuzzy quantities. The cable tension adjustment amount is determined by setting fuzzy sets, membership functions, fuzzy inference rules and defuzzification methods. The adjustment range is set according to the cable tension situation and environmental conditions to achieve reasonable control of the cable tension.

[0222] For emergency operations, when the risk level reaches the preset threshold, the corresponding emergency operation process is triggered. The parameters in the process are determined based on pre-set mathematical relationships to ensure response to various risk situations and ensure the safety of the drone.

[0223] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A tethered drone safety management system, characterized by: include: Multi-source data acquisition module: integrated sensors are deployed on the drone body and tethered cables to collect multi-source raw data of the drone flight and tethered cables, and then the collected multi-source raw data is transmitted to the data fusion and analysis module in real time; Data fusion and analysis module: Receives multi-source raw data from the multi-source data acquisition module, performs fusion processing on the multi-source raw data, removes data errors and redundancy, mines the inherent correlation of the data, forms effective data information that can accurately characterize the actual operating status of the drone and the environmental conditions in which it is located, and outputs the effective data information to the intelligent risk assessment module; Intelligent risk assessment module: Based on the valid data information output by the received data fusion and analysis module, combined with the preset risk assessment model, the risk assessment model comprehensively considers multiple factors including the flight attitude of the drone, environmental factors and the status of the tethered cable, and makes a real-time judgment on the risk level currently faced by the drone. When the risk level is determined to exceed the preset normal range, the corresponding risk warning information is generated and passed to the control instruction generation module; Control instruction generation module: Based on the risk warning information transmitted by the intelligent risk assessment module, it generates targeted control instructions according to the preset control strategy, and prioritizes the generated control instructions according to the severity and urgency of the risk, and then sends the prioritized control instructions to the actuator module; Actuator module: Receives control instructions from the control instruction generation module, and various components including the UAV flight control system, the electric retractable wire device for the tethered cable and other related actuators work together to perform corresponding actions, so as to achieve real-time adjustment and control of the UAV flight attitude and the state of the tethered cable, and ensure that the tethered UAV can perform tasks safely and stably in complex environments; The risk assessment model adopted by the intelligent risk assessment module is based on the continuous training optimization of the support vector machine algorithm combined with the Bayesian network algorithm. For the support vector machine algorithm, in the binary classification problem, given a training data set ,in, is the feature vector, is the category label, and its goal is to find a hyperplane To divide the data into two categories, the optimization objective function is: ； ； ； ； in, is the normal vector of the hyperplane, is the intercept, is the penalty parameter, is the slack variable; Bayesian network constructs a directed acyclic graph To express the probability dependence between variables, is a collection of nodes, is a set of edges, nodes The conditional probability distribution of Indicates that in its parent node set In the known case The probability distribution of the entire Bayesian network is: ; By combining the support vector machine algorithm with the Bayesian network algorithm, the support vector machine is used to classify and identify sample data, and the Bayesian network is used to model the complex dependencies between variables. Based on the learning of the accumulated flight data, the weight distribution of each dimensional factor in the risk assessment is adaptively adjusted, and the risk level faced by drones under different working conditions is accurately judged. The generated risk warning information clearly includes the risk category, severity, and urgency.

2. A tethered drone safety management system according to claim 1, characterized in that: The multi-source raw data collected by the multi-source data acquisition module is transmitted to the data fusion and analysis module through the time division multiple access TDMA communication protocol. The data transmission packet loss rate is less than 1%, the transmission delay time is within 10 milliseconds, and the data transmission follows the following formula: ； ； in, Indicates the packet loss rate, is the number of packets lost during transmission, is the total number of packets sent, Indicates the transmission delay time, is the time it takes for the data packet to arrive at the receiving end, The timestamp of the data packet sent from the sender.

3. A tethered drone safety management system according to claim 1, characterized in that: The data fusion and analysis module uses a data fusion algorithm that combines the Kalman filter algorithm with the weighted average algorithm to fuse the multi-source raw data. The state update equation of the Kalman filter algorithm is: ； ； in, is the prior state estimate, is the posterior state estimate at the previous moment, is the state transfer matrix, is the control input matrix, is the control input vector, is the prior estimation error covariance matrix, is the posterior estimation error covariance matrix of the previous moment, is the process noise covariance matrix; The measurement update equation is: ； ； ； in, is the Kalman gain, is the measurement matrix, is the measurement value vector, is the posterior state estimate, is the posterior estimation error covariance matrix, is the measurement noise covariance matrix, is the identity matrix; In the weighted average algorithm, for a certain data dimension , the fused data The calculation formula is: ； in, is the number of data sources involved in the fusion, Indicates Data sources in the data dimension The data value on is the weight of the corresponding data source.

4. A tethered drone safety management system according to claim 1, characterized in that: The control strategy preset in the control instruction generation module formulates control instruction generation rules for different risk categories, different severity levels and different urgency levels. For the adjustment of the UAV flight attitude, a proportional-integral-differential controller is used, and the mathematical expression of its control law is: ,in, is the controller output, is the deviation between the set value and the actual measured value, is the proportionality coefficient, is the integration coefficient, Indicates that from the initial time To current time Deviation The integral of For the tension adjustment of moored cables, based on the fuzzy logic control algorithm, the cable tension deviation and its deviation change rate are used as input fuzzy quantities. The cable tension adjustment amount is determined by setting fuzzy sets, membership functions, fuzzy inference rules and defuzzification methods. The adjustment range is set according to the cable tension situation and environmental conditions to achieve reasonable control of the cable tension. For emergency operations, when the risk level reaches the preset threshold, the corresponding emergency operation process is triggered. The parameters in the process are determined based on pre-set mathematical relationships to ensure response to various risk situations and ensure the safety of the drone.

5. The tethered drone safety management system according to claim 1, characterized in that: In the actuator module, the flight control system adjusts the relevant control instructions according to the flight attitude to accurately control the speed of each motor and the angle of the rudder surface, so as to realize the real-time correction and stable maintenance of the UAV's flight attitude; the electric retractable wire device of the tethered cable has an automatic tension adjustment function, and the built-in tension sensor feeds back the cable tension signal in real time, adjusts the retractable wire speed and strength according to the relevant control instructions, maintains the tethered cable tension within a reasonable range, and prevents the cable from abnormal conditions including over-stretching and winding, and the response time of the automatic tension adjustment function is controlled within 0.5 seconds, and its tension adjustment speed control follows the following formula: ,in, is the speed of cable tension adjustment, is the proportionality factor related to the device characteristics, The cable tension adjustment amount determined according to the control command.

Images