A method and system for altitude synchronization and tension adaptive control of a UAV winch

Abstract

This invention discloses a method and system for altitude synchronization and tension adaptive control of a UAV winch, belonging to the field of tethered UAV control technology. The method includes: real-time acquisition of UAV flight altitude data and actual cable length data; obtaining the deviation by subtracting the altitude from the length; generating a winch motor speed adjustment signal for cable winding and unwinding; continuous monitoring of real-time cable tension; comparing with a safe tension threshold range to determine abnormal tension conditions; using fuzzy logic to compensate the speed adjustment signal for abnormal tension; and implementing early warning, emergency stop, or cable winding / unwinding buffer protection for sudden tension changes, cable loosening, redundant entanglement, and excessive tightness / jamming, respectively. This invention achieves coordinated control of altitude synchronization and tension adaptive control, improving the stability and safety of UAV winch operations and effectively avoiding cable failures.

Description

Technical Field

[0001] This invention belongs to the field of tethered unmanned aerial vehicle (UAV) control technology, specifically a method and system for UAV winch altitude synchronization and tension adaptive control. Background Technology

[0002] Tethered drones, with their advantages of long endurance and high operational stability, are widely used in power line inspection, environmental monitoring, and emergency communications. During operation, the control of the winch cable's winding and unwinding directly affects the drone's flight safety and operational efficiency. Currently, most existing drone winch control methods rely on manual control or simple mechanical linkage control, resulting in low control precision and lag. When the drone takes off or descends, the winch's winding and unwinding speed cannot be precisely synchronized with the drone's flight altitude, easily leading to situations where the cable is too loose and tangled or too tight and stretched. This not only affects the drone's flight attitude but can also cause cable breakage, loosening of interfaces, and ultimately, drone crashes or equipment damage. Furthermore, existing technology lacks real-time monitoring and adaptive adjustment mechanisms for cable tension, making it impossible to promptly identify risks such as abnormal cable tension, slack, or sudden load changes. This makes it difficult to cope with unexpected situations in complex operating environments, resulting in insufficient reliability and high maintenance costs for tethered drone operations. Furthermore, while some automated control schemes can automate winch cable delivery and take-up, they lack a closed-loop linkage logic between altitude acquisition, tension monitoring, and winch drive. This results in slow control response and an inability to dynamically adjust based on the real-time flight status of the UAV, making it difficult to meet the high-precision and high-safety requirements of tethered operations. Therefore, developing a control method that enables altitude synchronization between the winch and the UAV, and adaptive adjustment of cable tension, is crucial for improving the safety and reliability of tethered UAV operations. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention proposes a method and system for altitude synchronization and tension adaptive control of a UAV winch. It collects real-time UAV flight altitude data and actual cable length data, calculates the deviation by subtracting the altitude from the length, and generates a signal to adjust the winch motor's winding and unwinding speed. The system continuously monitors the real-time cable tension and compares it to a safe tension threshold range to determine abnormal tension conditions. For tension anomalies, fuzzy logic is used to compensate the winding and unwinding speed adjustment signal, and warnings, emergency stops, or winding and unwinding buffer protection are implemented for sudden tension changes, cable loosening, redundant entanglement, and excessive tightness / jamming, respectively. This invention achieves coordinated control of altitude synchronization and tension adaptive control, improving the stability and safety of UAV winch operations and effectively preventing cable failures.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A method for altitude synchronization and tension adaptive control of a UAV winch includes:

[0006] Real-time acquisition of drone flight altitude data and actual cable length data;

[0007] The difference between the flight altitude data and the actual released length data is used to generate a height-length deviation, and a speed adjustment signal for driving the winch motor to adjust the wire release and take-up speed is generated based on the height-length deviation.

[0008] During the process of the winch motor responding to the winding and unwinding speed adjustment signal to perform the winding and unwinding operation, the real-time tension data of the cable is continuously detected, and the real-time tension data is compared and analyzed with the preset safe tension threshold range. If the real-time tension data exceeds the safe tension threshold range, it is determined that the cable is in a slack state, has abnormal tension, or has a sudden change in load.

[0009] Based on the judgment result, a tension compensation signal or preset protection logic is generated; the tension compensation signal is used to correct the take-up and release speed adjustment signal; the preset protection logic includes: issuing an anti-loosening warning when a sudden tension change is detected at the cable interface, and triggering anti-tangling stop or emergency take-up and release buffer control when the cable is redundant or too tight.

[0010] Specifically, the flight altitude data is obtained by fusing a Global Navigation Satellite System receiver based on real-time dynamic differential technology with a barometric altimeter, including:

[0011] The global navigation satellite system receiver based on real-time dynamic differential technology collects pseudorange and carrier phase observations of the UAV relative to the base station, and generates the first altitude measurement value by solving the differential correction.

[0012] The static pressure value at the location of the UAV is collected using a barometric altimeter, and the static pressure value is converted into a second altitude measurement value based on a standard atmospheric model;

[0013] The first and second altitude measurements are used as observation inputs by employing a Kalman filter, and a state equation is constructed based on the vertical motion velocity and vertical acceleration of the UAV. Through the prediction and update iteration process of the Kalman filter, the flight altitude data after fusion correction is generated.

[0014] Specifically, the actual length data is obtained by combining an incremental photoelectric encoder installed on the winch shaft with a real-time diameter model of the winding hub, including:

[0015] An incremental photoelectric encoder installed on the end of the winch shaft is used to collect the number of rotation pulses and the rotation direction signal of the winch shaft. The number of rotation pulses is accumulated by a pulse counter to generate the number of shaft rotations.

[0016] Based on the preset real-time diameter model of the winding hub, the pre-calibrated cable layer-diameter mapping table is queried, and combined with the real-time record of the current cable release layer, the average winding diameter corresponding to the current working turn of the winch shaft is generated.

[0017] The actual length of the winch cable is generated by multiplying the number of rotations of the shaft by the average winding diameter.

[0018] Specifically, the difference between the flight altitude data and the actual release length data is used to generate an altitude-length deviation, including:

[0019] The data includes the drone's flight altitude data and the actual length of the winch cable. A data alignment buffer is used to synchronize the flight altitude data and the actual length data at the same timestamp, generating synchronized altitude and length values.

[0020] Subtract the synchronization length value from the synchronization height value to generate the original deviation.

[0021] An amplitude limiting filtering algorithm is used to remove outliers and smooth the original deviation to generate a height-length deviation.

[0022] Specifically, based on the height-length deviation, a signal for adjusting the take-up and release speeds of the winch motor is generated, including:

[0023] The height-length deviation is input to a proportional-integral-derivative controller, which performs proportional, integral, and derivative operations on the height-length deviation based on preset proportional, integral, and derivative coefficients to generate a basic speed control quantity.

[0024] Obtain real-time cable tension data, and based on the real-time tension data, obtain the tension influence factor corresponding to the real-time tension data through a preset tension-speed correction lookup table;

[0025] Multiply the basic speed control value by the tension influence factor to generate the comprehensive speed control value;

[0026] A pulse width modulation signal generator is used to convert the comprehensive speed control quantity into a take-up and release speed adjustment signal with a corresponding duty cycle; the take-up and release speed adjustment signal is used to drive the winch motor to perform take-up or release operations at a target speed.

[0027] Specifically, after generating the height-length deviation, the method further includes integral separation control based on the height-length deviation, specifically:

[0028] The absolute value of the height-length deviation is measured to generate the absolute value of the deviation;

[0029] The absolute value of the deviation is compared with a preset integral separation threshold. When the absolute value of the deviation is greater than the integral separation threshold, an integral separation enable signal is generated to disable the integral operation in the proportional-integral-derivative controller. When the absolute value of the deviation is less than or equal to the integral separation threshold, an integral input enable signal is generated to restart the integral operation.

[0030] Specifically, the real-time tension data is compared and analyzed with a preset safe tension threshold range. If the real-time tension data exceeds the safe tension threshold range, it is determined that the cable is in a slack state, experiencing abnormal tension, or a sudden change in load, including:

[0031] Read the pre-stored safe tension threshold range; the safe tension threshold range includes a minimum safe tension threshold and a maximum safe tension threshold.

[0032] Based on comparator logic, the real-time tension data is compared with the minimum safe tension threshold and the maximum safe tension threshold respectively; if the real-time tension data is less than the minimum safe tension threshold, it is determined that the cable is in a slack state; if the real-time tension data is greater than the maximum safe tension threshold, it is determined that the cable has abnormal tension or sudden load change; if the real-time tension data is within the safe tension threshold range, it is determined that the cable tension is normal.

[0033] Specifically, generating the tension compensation signal based on the determination result includes:

[0034] When the determination result indicates that the cable is in a slack state, the difference between the real-time tension data and the minimum safe tension threshold is calculated to generate a slack deviation, and the slack deviation is input to the first fuzzy logic controller. The input membership function of the first fuzzy logic controller adopts a triangular function, and the output membership function adopts a single-valued function. Fuzzy reasoning and centroid method defuzzification are performed through the fuzzy rule base to generate the first tension compensation coefficient.

[0035] Multiply the first tension compensation coefficient by the comprehensive speed control amount in the take-up and release speed adjustment signal to generate a corrected speed adjustment signal that increases the take-up speed or decreases the release speed.

[0036] When the determination result indicates that the cable has abnormal tension or sudden load change, the difference between the real-time tension data and the maximum safe tension threshold is calculated to generate an over-tension deviation, and the over-tension deviation is input to the second fuzzy logic controller to generate a second tension compensation coefficient; the second fuzzy logic controller is configured the same as the first fuzzy logic controller.

[0037] The second tension compensation coefficient is multiplied by the comprehensive speed control value in the take-up and release speed adjustment signal to generate a corrected speed adjustment signal that increases the release speed or decreases the take-up speed.

[0038] Specifically, the step of issuing an anti-loosening warning when a sudden tension change is detected at the cable interface includes:

[0039] Perform a first-order difference operation on the real-time tension data sequence composed of the real-time tension data to generate tension change rate data;

[0040] Based on the pre-calibrated interface pull-out force threshold at the cable interface, a tension change judgment threshold is generated;

[0041] When the real-time tension data exceeds the maximum safe tension threshold and the tension change rate data exceeds the tension change determination threshold, it is determined that a tension change has occurred at the cable interface.

[0042] Based on the judgment result, an anti-loosening warning signal is generated and sent to the ground station through the UAV's data link downlink. At the same time, the buzzer and status indicator light are triggered to issue an on-site alarm.

[0043] Specifically, the step of triggering anti-tangle stop or emergency cable rewind buffer control when cable redundancy or excessive cable tightness is determined includes:

[0044] When the determination result is that the cable is in a slack state and the duration of the slack state exceeds the preset redundancy tolerance time threshold, the cable is determined to be redundant and there is a risk of entanglement; when the determination result is that the cable has abnormal tension and the duration of the abnormal tension exceeds the preset overtightness tolerance time threshold, the cable is determined to be too tight and there is a risk of jamming; based on the determination result, an anti-entanglement stop signal is generated; the anti-entanglement stop signal is used to cut off the power input of the winch motor and simultaneously control the electromagnetic brake to grip the winch shaft, performing emergency stop protection;

[0045] When the determination result is cable redundancy, an emergency take-up buffer control signal is generated, which drives the winch motor to accelerate take-up at a preset buffer acceleration. When the determination result is cable too tight, an emergency release buffer control signal is generated, which drives the winch motor to decelerate release at a preset buffer deceleration. Until the real-time tension data recovers to the safe tension threshold range, the take-up and release speed adjustment signal is switched back for control.

[0046] A drone winch altitude synchronization and tension adaptive control system includes:

[0047] The data acquisition module is used to collect data on the drone's flight altitude and the actual length of the cable extended.

[0048] The deviation calculation module is used to synchronize and align the collected flight altitude data with the actual release length data in time, calculate the difference to generate the original deviation, and then use a limiting filter algorithm to remove outliers and smooth the data to obtain the altitude-length deviation.

[0049] The speed control generation module is used to input the height-length deviation into the PID controller to calculate the basic speed control quantity. Combined with the real-time cable tension data, the tension influence factor is obtained through the tension-speed correction lookup table to obtain the comprehensive speed control quantity and convert it into a wire take-up and release speed adjustment signal.

[0050] The anomaly detection module obtains real-time tension data by performing sliding window averaging filtering on the digital tension values ​​collected by the tension sensor, and compares it with the preset safe tension threshold range to determine whether the cable is in a slack state, has abnormal tension, or has a sudden load change.

[0051] The tension compensation module is used to generate a tension compensation coefficient through a fuzzy logic controller based on the tension anomaly judgment result. When the tension anomaly is judged to be sudden, cable redundancy, or cable too tight, it triggers anti-loosening warning, anti-tangling shutdown, or emergency cable winding buffer control.

[0052] Compared with the prior art, the beneficial effects of the present invention are:

[0053] 1. This invention proposes a method for altitude synchronization and tension adaptive control of a UAV winch. By real-time acquisition, data synchronization, and deviation calculation of UAV flight altitude and cable length, combined with PID control and tension correction, a signal for adjusting the cable winding and unwinding speed is generated. With the assistance of altitude fusion calculation, precise cable length measurement, amplitude limiting filtering, and integral separation strategies, the dynamic matching accuracy between the winch winding and unwinding and the UAV altitude is improved, significantly reducing cable tension fluctuations and length mismatch problems, and ensuring a smooth and stable operation.

[0054] 2. This invention proposes a method for altitude synchronization and adaptive tension control of UAV winches. It relies on a wheel-axle type tension sensor and sliding window filtering to achieve real-time tension monitoring. The speed control quantity is quickly corrected through safety threshold judgment and fuzzy logic compensation. It also sets up early warning, emergency stop and buffer protection logic for abnormal states such as interface loosening, redundant winding, and excessive tightness jamming. It can effectively avoid faults such as cable breakage, loosening, and winding, enhance the safety and reliability of system operation, extend the service life of equipment, and adapt to the stable operation requirements of UAV winches in complex scenarios. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of a method for altitude synchronization and tension adaptive control of a UAV winch according to the present invention;

[0056] Figure 2This is a flowchart illustrating the principle of a drone winch altitude synchronization and tension adaptive control method according to the present invention.

[0057] Figure 3 This is a diagram illustrating the architecture of a drone winch altitude synchronization and tension adaptive control system according to the present invention. Detailed Implementation

[0058] Example 1:

[0059] Please see Figure 1 and Figure 2 The present invention provides an embodiment of a method for altitude synchronization and tension adaptive control of a UAV winch, the method comprising S1 to S4, including the following steps:

[0060] S1: Real-time acquisition of drone flight altitude data and actual cable extension length data;

[0061] The flight altitude data is obtained by fusing data from a Global Navigation Satellite System receiver based on real-time dynamic differential technology and a barometric altimeter, specifically including:

[0062] S1.1: The global navigation satellite system receiver based on real-time dynamic differential technology collects pseudorange observations and carrier phase observations of the UAV relative to the base station, and generates the first altitude measurement value through differential correction calculation;

[0063] Furthermore, the specific steps of S1.1 include:

[0064] (1) Start and initialize the base station and airborne receiving equipment. The base station is deployed at a fixed location in the work area with open terrain and no signal obstruction. The receiver of the base station and the airborne receiver carried by the UAV are both configured to support the multi-constellation joint reception mode of Beidou-3, GPS and GLONASS. The satellite receiving elevation angle is set to 15 degrees to ensure stable reception of low elevation satellite signals. The data transmission communication link between the base station and the airborne receiver adopts a dedicated radio station with a communication rate of 19200bp / s and a communication delay controlled within 20ms to ensure that the differential correction data can be transmitted in real time and without delay.

[0065] (2) The airborne receiver continuously scans the navigation satellite signals in the sky according to the set receiving elevation angle, identifies and locks the valid satellite signals whose signal strength meets the working requirements, maintains the stable phase-locked state of the signal, numbers and marks each locked satellite, and continuously records the quality level of the satellite signal. During the entire operation, the number of satellites that are stably tracked is not less than 6. When the number of satellites is less than 6, it automatically switches to the single constellation priority receiving mode to ensure that the observation data is not interrupted.

[0066] (3) The airborne receiver synchronously collects the pseudorange observation value and carrier phase observation value corresponding to the UAV position under the unified Beidou time standard. The sampling rate of the pseudorange observation value is set to 1Hz and the sampling rate of the carrier phase observation value is set to 5Hz. Both observation data carry the unified Beidou time standard to ensure that the data of the base station and the airborne terminal are strictly aligned in time.

[0067] (4) Based on its own known precise three-dimensional coordinates, the base station synchronously collects the pseudorange observation value and carrier phase observation value of the corresponding satellite under the same BeiDou time scale as the airborne receiver. The base station uses its own observation data and known coordinates to calculate the pseudorange correction number and carrier phase correction number that can be used to correct the airborne observation data. In the process of calculating the correction number, the ionospheric delay error, tropospheric delay error, satellite ephemeris error and multipath effect error are eliminated. Among them, the ionospheric delay correction adopts the Klobuchar model and the parameters are set to the standard atmospheric model parameters. The tropospheric delay correction adopts the Hopfield model and the temperature parameter is set to 25℃ and the air pressure parameter is set to 1013.25Pa. The Hopfield model and the Klobuchar model are existing technologies in this field and are not the inventive solutions of this application. They will not be described in detail here.

[0068] (5) The base station encapsulates the calculated pseudorange correction and carrier phase correction according to the RTK differential dedicated RTCM3 communication protocol to form a complete differential correction data packet. The base station sends the differential correction data packet to the airborne receiver on the UAV in real time through a dedicated radio data transmission link with a communication rate of 19200bps and a link delay controlled within 20ms, to ensure that the airborne receiver can obtain effective correction data in a timely manner.

[0069] (6) The airborne receiver receives differential correction data packets sent by the reference station in real time, verifies and parses the differential correction data packets, separates the pseudorange correction and carrier phase correction, sets data check bits during parsing, and discards data packets that fail to be verified.

[0070] (7) The airborne receiver superimposes the pseudorange observation value it collects with the corresponding pseudorange correction value to obtain the corrected pseudorange observation value. At the same time, it superimposes the carrier phase observation value it collects with the corresponding carrier phase correction value to obtain the corrected carrier phase observation value. The airborne receiver inputs the corrected pseudorange observation value and the corrected carrier phase observation value into the least squares algorithm. The iteration convergence threshold of the least squares algorithm is set to 1 cm, and the upper limit of the number of iterations is set to 20. The three-dimensional coordinate result of the UAV relative to the base station is obtained through iterative calculation. The least squares algorithm is the prior art in this field and is not an inventive solution of this application. It will not be described in detail here.

[0071] (8) Extract the vertical coordinate component from the three-dimensional coordinate result obtained by the solution. This coordinate component is the first height measurement value after differential correction. This height measurement value is based on the elevation of the base station. The static measurement accuracy can reach within 1 cm, and the dynamic measurement accuracy can reach within 3 cm.

[0072] S1.2: The static pressure value at the location of the UAV is collected using a barometric altimeter, and the static pressure value is converted into a second altitude measurement value based on a standard atmospheric model;

[0073] Furthermore, the specific steps in S1.2 include:

[0074] (1) Perform hardware initialization and working parameter configuration of the barometric altimeter. The barometric altimeter used in this embodiment is a digital barometric pressure sensor with an integrated temperature compensation unit. After the sensor is powered on, it automatically completes internal calibration and enters a stable working state after calibration. The sensor's measurement range is set to 0~110kPa, the measurement resolution is set to 0.1Pa, the data output rate is set to 50Hz, the temperature compensation range is set to -40℃~85℃, and the compensation update cycle is set to 10ms to ensure that static pressure measurement data can be stably output under various environmental conditions of UAV operation.

[0075] (2) Connect the static pressure acquisition port of the barometric altimeter to the static pressure tube of the UAV fuselage. The static pressure tube is installed in an area outside the airflow disturbance range of the UAV rotor to avoid the dynamic pressure generated by the rotor from mixing into the acquisition path and affecting the measurement accuracy. After the system is started, the barometric altimeter continuously acquires the atmospheric static pressure signal of the current location of the UAV. The acquired analog static pressure signal first enters the signal amplification unit inside the sensor. The amplification factor is fixed at 50 times. The amplified signal enters the low-pass filter unit. The filter cutoff frequency is set to 10Hz to filter out high-frequency interference caused by circuit noise and mechanical vibration. After filtering, the signal enters the analog-to-digital conversion unit. The analog-to-digital conversion bit is set to 16 bits and the conversion period is set to 5ms. Finally, the digital static pressure value is output.

[0076] (3) During the ground initialization phase before the UAV takes off, the system controls the barometric altimeter to continuously collect 50 static pressure values ​​of the ground environment. The arithmetic average of the 50 collected results is calculated, and the average value is used as the local reference pressure value. The local reference pressure value represents the standard atmospheric pressure at the takeoff point.

[0077] (4) Load and configure the standard atmospheric model. In this embodiment, the international standard atmospheric model is used as the basis for the conversion between pressure and altitude. The reference atmospheric parameters set in the standard atmospheric model are: the reference air pressure value corresponds to 1013.25 Pa, the reference temperature value corresponds to 15℃, the temperature vertical lapse rate is set to decrease by 0.0065℃ for every 1 meter increase in altitude, and the applicable altitude range of the model is set to 0 to 11000 meters, which fully covers the altitude range of the UAV's regular operation.

[0078] (5) Substitute the real-time static pressure value into the standard atmospheric model to perform altitude calculation, including: reading the real-time static pressure value output by the barometric altimeter at the current moment; reading the stored local reference air pressure value and reference temperature value; determining the atmospheric layer in which the real-time static pressure value is located according to the stratification calculation rules of the standard atmospheric model. In this embodiment, the operating altitude of the UAV is within the troposphere, so the air pressure altitude calculation formula corresponding to the troposphere is directly adopted; performing the exponential and logarithmic operations required by the model in sequence; combining the calculation results with parameters such as the vertical temperature lapse rate to obtain the relative altitude value of the UAV relative to the takeoff point.

[0079] (6) Establish a sliding data window with a length of 10, store the relative height values ​​calculated at 10 consecutive times into the window, remove the earliest stored value in the window each time a new relative height value enters the window, and perform an arithmetic average on all valid values ​​in the window to obtain the smoothed height value, so as to avoid the height value from jumping due to instantaneous airflow fluctuations or electromagnetic interference.

[0080] (7) The smoothed relative height value is used as the final second height measurement value.

[0081] S1.3: Using a Kalman filter, the first altitude measurement value and the second altitude measurement value are used as observation inputs. A state equation is constructed based on the vertical motion speed and vertical acceleration of the UAV. Through the prediction and update iteration process of the Kalman filter, the flight altitude data after fusion correction is generated.

[0082] Furthermore, the specific steps in S1.3 include:

[0083] (1) Perform the initial configuration and parameter setting of the Kalman filter. In this embodiment, the state variables of the Kalman filter are selected from three physical quantities: the vertical position, vertical velocity, and vertical acceleration of the UAV. The three state variables together constitute the system state vector. The system process noise covariance matrix is ​​configured according to the mechanical characteristics of the UAV's vertical motion. The process noise variance corresponding to the vertical position is set to 0.001, the process noise variance corresponding to the vertical velocity is set to 0.005, and the process noise variance corresponding to the vertical acceleration is set to 0.01. The system observation noise covariance matrix is ​​weighted according to the actual measurement accuracy of the two altitude measurements. The observation noise variance corresponding to the first altitude measurement is set to 0.002, and the observation noise variance corresponding to the second altitude measurement is set to 0.01. The initial state covariance matrix is ​​set as a diagonal matrix, and the initial values ​​on the diagonal are 0.1, 0.05, and 0.05, respectively.

[0084] (2) Establish the system state equation and state transition model. The state equation takes the system state vector at the current moment as input and calculates the system predicted state vector at the next moment. In this embodiment, the time interval of the state transition matrix is ​​set to 20ms, which is consistent with the UAV flight control cycle. The correlation coefficient between vertical position and vertical velocity in the state transition matrix directly adopts the time interval value, and the correlation coefficient between vertical velocity and vertical acceleration also directly adopts the time interval value. The vertical acceleration is regarded as constant in a short time, which describes the dynamic transmission relationship between the UAV's position, velocity and acceleration in the vertical direction.

[0085] (3) Read and synchronize the first height measurement value and the second height measurement value, combine the first height measurement value and the second height measurement value to form an observation vector, which is used as the external input of the Kalman filter observation update stage. The dimension of the observation vector is matched with the dimension of the vertical position variable in the state vector to ensure that the observation data can effectively correct the state prediction results.

[0086] (4) Starting from the updated optimal state vector at the current moment, the prior state prediction vector at the next moment is calculated by combining the pre-established state equation and state transition matrix. This is also called the prior estimate of the state vector. At the same time, the prior state covariance matrix at the next moment is calculated based on the state covariance matrix, state transition matrix and process noise covariance matrix at the current moment. This matrix is ​​used to characterize the uncertainty of the prior state prediction result. The calculation methods of the state covariance matrix and the prior estimate of the state vector are existing technologies in this field and are not the inventive solution of this application. They will not be elaborated here.

[0087] (5) The weights are automatically assigned based on the real-time noise levels of the two height measurements, and the Kalman gain is calculated. When the accuracy of the first height measurement is stable, its weight is automatically increased. When the second height measurement is more reliable, its contribution ratio is automatically adjusted. No manual intervention is required throughout the process.

[0088] (6) Calculate the difference between the prior state prediction vector and the observation vector, multiply the difference by the calculated Kalman gain, and finally add the product to the prior state prediction vector to obtain the corrected posterior optimal state vector at the current time. Simultaneously update the posterior state covariance matrix. Use the prior state covariance matrix, Kalman gain and observation matrix to complete the iterative calculation. The updated posterior state covariance matrix will serve as the initial basis for the state prediction at the next time.

[0089] (7) Extract the fused and corrected flight altitude data. Extract the state variable value corresponding to the vertical position from the updated posterior optimal state vector. The state variable value is the flight altitude data after fusion and correction by Kalman filter, which combines the advantages of the first altitude measurement value and the second altitude measurement value.

[0090] (8) Throughout the entire operation, the UAV repeatedly performs the complete process of state prediction, Kalman gain calculation, observation update and optimal state extraction at a fixed 20ms cycle. The posterior output of each iteration is directly used as the prior input of the next iteration. The two altitude measurements are continuously fused and denoised. The iteration process is executed stably throughout without interruption or reset until the UAV completes the operation and executes the system shutdown command.

[0091] The actual length data is obtained by using an incremental photoelectric encoder installed on the winch shaft end in conjunction with a real-time diameter model of the winding hub, specifically including:

[0092] S1.4: An incremental photoelectric encoder installed on the end of the winch shaft is used to collect the number of rotation pulses and the rotation direction signal of the winch shaft. The number of rotation pulses is accumulated by a pulse counter to generate the number of shaft rotations.

[0093] Furthermore, the incremental photoelectric encoder selected in this embodiment has a resolution of 1024 lines, meaning that the encoder outputs 1024 pulse signals for each rotation of the winch shaft, and simultaneously outputs two-phase quadrature pulse signals A and B. By judging the phase difference between the two-phase pulses, the rotation direction of the winch shaft is identified; forward rotation corresponds to the unwinding operation, and reverse rotation corresponds to the rewinding operation. The encoder is tightly installed on the connecting shaft between the output shaft of the winch motor and the winch drum, without gaps or slippage, ensuring that every inch of rotation of the winch shaft can be accurately captured. The pulse counter is a 32-bit high-speed counter, capable of capturing the pulse signals output by the encoder in real time, with a counting frequency of up to 1MHz, far exceeding the output frequency of the encoder, avoiding pulse loss. During system initialization, the pulse counter is automatically reset to zero, with the count zero point being when the winch cable is fully retracted and the drum is in the initial position. When the winch motor drives the drum to rotate, the pulse counter accumulates the number of pulses in the forward direction or decrements them in the reverse direction according to the rotation direction. The system calculates the actual number of rotations of the winch shaft based on the ratio of the total number of pulses to the encoder resolution.

[0094] S1.5: Based on the preset real-time diameter model of the winding hub, query the pre-calibrated cable layer-diameter mapping table, and combine it with the real-time record of the current cable release layer to generate the average winding diameter corresponding to the current working circle of the winch shaft.

[0095] Furthermore, during the winding and unwinding process, the cable is wound layer by layer onto the drum. As the number of unwinding layers decreases, the winding diameter of the remaining cable on the drum gradually decreases. As the number of winding layers increases, the winding diameter gradually increases. If a fixed diameter is used to calculate the unwound length, a large cumulative error will occur. Therefore, this embodiment introduces a real-time diameter model of the winding hub to solve the length calculation error problem caused by variable diameter winding. Before the system leaves the factory, a large number of calibration experiments are conducted to establish a cable layer-diameter mapping table. The calibration process is based on the initial diameter of the drum, and the total diameter of the drum after each layer of cable is wound is recorded. At the same time, the cable length that can be wound in each layer and the corresponding number of rotations are recorded to form a complete cable layer-diameter mapping table and store it in the flash memory of the controller.

[0096] Furthermore, during operation, the system calculates the number of turns of the winch shaft based on the pulse counter, combines this with the number of turns corresponding to each layer of cable, and records in real time the current number of cable layers released and the remaining number of winding layers. It then calls the real-time diameter model of the winding hub, automatically queries the cable layer-diameter mapping table, and obtains the actual winding diameter of the remaining cable on the drum. Simultaneously, it considers minor factors such as the elastic deformation of the cable and the winding gap, and fine-tunes the queried diameter value to generate the average winding diameter corresponding to the current working turn. This average winding diameter is updated in real time with the cable release and take-up operations, and is recalculated after each rotation to ensure that the length calculation is always based on the real-time drum diameter, thus eliminating length errors caused by variable diameter winding at the source.

[0097] Furthermore, the specific steps in S1.5 include:

[0098] (1) Complete the initialization and basic parameter configuration of the real-time diameter model of the winding hub. The winch hub used in this embodiment is a cylindrical rigid structure. The initial bottom diameter of the hub is set to 50 mm, and the outer diameter of the cable is set to 1.2 mm. The system completes the solidification and storage of all calibration data before leaving the factory. No parameter modification is performed during operation. The maximum number of turns of a single layer of cable can be wound in the real-time diameter model of the winding hub is preset to 120 turns, and the maximum number of layers of cable allowed to be wound is 20 layers. The above parameters together constitute the basic calculation conditions of the real-time diameter model of the winding hub.

[0099] (2) Establish and load a pre-calibrated cable layer number-diameter mapping table. The cable layer number-diameter mapping table is generated by physical calibration experiments before leaving the factory. The calibration process is based on the initial bottom diameter of the hub. After each layer of cable is wound, the total winding diameter corresponding to the current layer is recorded. The number of layers increases sequentially from the 1st layer to the 20th layer. Each layer corresponds to a unique winding diameter value. The diameter after the first layer of cable is wound is 52.4 mm, and the diameter after the second layer of cable is wound is 54.8 mm. For each additional layer, the diameter value increases by 2.4 mm until the 20th layer corresponds to 98 mm. The system loads the mapping table completely into the computing memory during the initialization phase as the standard basis for real-time diameter query.

[0100] (3) The cumulative number of rotations of the winch shaft is counted in real time based on the rotation pulse data collected by the incremental photoelectric encoder at the end of the winch shaft. The cumulative number of rotations is compared with the maximum number of single-layer windings of 120. When the cumulative number of windings reaches 120, it is determined that the cable release operation of one layer is completed. The number of layers released is continuously counted. In the initial state, the cable is fully retracted and the number of released layers is recorded as 0 layers. The number of layers is increased by 1 for each layer released and decreased by 1 for each layer retracted, so as to ensure that the number of layers is completely consistent with the actual cable release and retraction status.

[0101] (4) Input the current number of cable layers into the real-time diameter model of the winding hub. The real-time diameter model of the winding hub takes the current number of cable layers as the only input variable. First, it performs a validity check on the number of layers to ensure that the number of layers is within the legal range of 0 to 20 layers. Values ​​outside the range are treated as boundary values.

[0102] (5) Based on the number of layers that have been released, locate the corresponding entry in the cable layer-diameter mapping table, and directly read the standard diameter value of the remaining cable on the hub after winding at that layer. For example, if the number of layers that have been released is 5, the diameter of the remaining cable is 86 mm. If the number of layers that have been released is 10, the diameter of the remaining cable is 74 mm. The query process directly uses the calibrated fixed value.

[0103] (6) Considering the small gaps and elastic deformations in the actual winding process of the cable, the system applies a fixed correction amount based on the queried diameter. The correction amount is set to ±0.1 mm. This correction amount is a fixed value calibrated in the experiment and does not change with the winding state.

[0104] (7) The diameter value after query and correction is used as the average winding diameter corresponding to the current working circle of the winch shaft. The average winding diameter is updated in real time as the number of cable layers is released. The query and correction process is re-executed every time a layer is changed to ensure that the average winding diameter is always consistent with the current winding state.

[0105] S1.6: Multiply the number of rotations of the shaft by the average winding diameter to generate the actual length of the winch cable released.

[0106] Furthermore, after obtaining the precise number of shaft rotations and the real-time average winding diameter, the system multiplies the number of shaft rotations by the average winding diameter and then by pi to obtain the actual cable length at the current moment. During the cable release process, the number of shaft rotations continuously increases, and the actual released length increases synchronously; during the cable retraction process, the number of shaft rotations continuously decreases, and the actual released length decreases synchronously. To further improve the stability of the length data, the system performs a moving average processing on the length data obtained from five consecutive calculations, eliminating abnormal length values ​​caused by motor vibration and pulse interference, and outputs smoothed actual released length data. The measurement error of this data is controlled within ±0.5cm, matching the accuracy of the fused and corrected flight altitude data, ensuring the accuracy of subsequent altitude-length deviation calculations.

[0107] S2: Subtract the flight altitude data from the actual released length data to generate a height-length deviation, and generate a winding and releasing speed adjustment signal for driving the winch motor based on the height-length deviation.

[0108] The difference between the flight altitude data and the actual launch length data is used to generate an altitude-length deviation, including:

[0109] S2.1: Obtain the UAV flight altitude data and the actual length of the winch cable, and use a data alignment buffer to synchronize the flight altitude data and the actual length data at the same timestamp to generate synchronized altitude and synchronized length values.

[0110] Furthermore, since the flight altitude data and the actual launch length data are acquired by different acquisition units, there are slight differences in the latency of acquisition, transmission, and processing. Even if the acquisition frequency is the same, the timestamps may not be completely consistent. Therefore, this embodiment sets up a data alignment buffer specifically responsible for achieving time synchronization of the two types of data. The data alignment buffer is a first-in-first-out (FIFO) cache unit with a cache depth of 10 sets of data, capable of storing the flight altitude data and actual launch length data of the most recent 10 moments. Each set of data carries a unique timestamp. When new flight altitude data and actual launch length data are input into the data alignment buffer, the buffer automatically matches the two sets of data with the closest timestamps. Two sets of data with a time difference of less than 1ms are determined to be data from the same moment, completing the time synchronization alignment and outputting the synchronized altitude value and synchronized length value. If the time difference is greater than 1ms, the data with the larger delay is discarded, and the valid data from the previous moment is used to supplement it, ensuring that the timestamps of the synchronized data are completely consistent, fundamentally avoiding calculation errors caused by time asynchrony.

[0111] S2.2: Subtract the synchronization length value from the synchronization height value to generate the original deviation;

[0112] Furthermore, the value and sign of the initial deviation directly reflect the mismatch between height and length. When the synchronized height value is greater than the synchronized length value, the initial deviation is positive, indicating that the drone's flight altitude is higher than the cable's extended length, the cable is too tight, and the release speed needs to be increased or the retrieval speed slowed down. When the synchronized height value is less than the synchronized length value, the initial deviation is negative, indicating that the drone's flight altitude is lower than the cable's extended length, the cable is slack, and the retrieval speed needs to be increased or the release speed slowed down. When the synchronized height value and the synchronized length value are equal, the initial deviation is zero, indicating that the height and length are perfectly matched, and the cable condition is ideal. In the operational scenario of this embodiment, the target state for the drone hovering operation is that the initial deviation approaches zero. The system continuously calculates the initial deviation to monitor the cable tension in real time.

[0113] S2.3: The original deviation is processed by an amplitude limiting filter algorithm to remove outliers and smooth the data, generating a height-length deviation. The height-length deviation is used to characterize the degree of dynamic mismatch between the actual hovering height of the UAV and the theoretical cable extension length.

[0114] During actual data collection, flight altitude data and actual launch length data may be affected by factors such as electromagnetic interference, motor vibration, and sudden changes in airflow, resulting in instantaneous abnormal jumps and significant fluctuations in the original deviation. If the original deviation is used directly to control the motor, it will cause frequent sudden changes in motor speed, affecting system stability. Given that the flight altitude measurement accuracy is ±0.02m and the actual cable length measurement accuracy is ±0.005m, the normal maximum fluctuation range of these two types of data is the sum of their measurement accuracies, i.e., ±0.025m. Therefore, this embodiment employs an amplitude limiting filtering algorithm, setting the maximum allowable variation range of the deviation to ±0.025m. When the difference between the original deviation values ​​at two consecutive moments exceeds the maximum allowable variation range, the current original deviation value is determined to be an outlier and is discarded, replaced by the filtered deviation value from the previous moment. If the difference is within the allowable range, the current original deviation value is retained and weighted averaged with the previous five valid deviation values ​​to generate the altitude-length deviation value. This altitude-length deviation value has no abnormal jumps, is smooth and stable, and can accurately reflect the dynamic mismatch between the actual hovering altitude of the UAV and the theoretical cable length.

[0115] Based on the height-length deviation, a signal is generated to adjust the winding and unwinding speed of the winch motor, including:

[0116] S2.4: The height-length deviation is input to the proportional-integral-derivative controller, which performs proportional, integral, and derivative operations on the height-length deviation based on preset proportional, integral, and derivative coefficients to generate a basic speed control quantity;

[0117] Furthermore, the specific steps in S2.4 include:

[0118] (1) Complete the initialization and parameter tuning configuration of the proportional-integral-derivative controller. In this embodiment, an incremental proportional-integral-derivative controller is used. The proportional-integral-derivative controller takes the height-length deviation after amplitude limiting filtering and smoothing as the only input and the basic speed adjustment of the winch motor as the output. The operation cycle of the incremental proportional-integral-derivative controller is consistent with the control cycle of the UAV system and is set to 20ms. The proportional coefficient, integral coefficient and derivative coefficient are obtained through the closed-loop tuning of the UAV winch system on site. The proportional coefficient is set to 25, the integral coefficient is set to 0.8 and the derivative coefficient is set to 1.2. All coefficients are calibrated and stored before the system leaves the factory and are not dynamically modified during operation.

[0119] (2) The proportional-integral-derivative controller acquires the height-length deviation that has been eliminated and smoothed after the outlier removal process. It multiplies the current height-length deviation by the preset proportional coefficient to obtain the proportional operation output. The proportional component is used to provide an instant response capability based on the real-time magnitude of the deviation. The larger the deviation, the higher the proportional operation output, which can drive the winch motor to quickly approach the target state and eliminate the mismatch between height and length.

[0120] (3) The height-length deviation of the current time and all previous valid times are continuously accumulated to obtain the cumulative total of the deviation. The cumulative total is then multiplied by the preset integral coefficient to obtain the integral operation output. The integral step is used to eliminate the static deviation of the system and avoid the situation where the height and length have a small mismatch that cannot be corrected for a long time. The integral accumulation operation is only performed when the integral separation logic is allowed to be put into operation to ensure that there is no integral saturation phenomenon.

[0121] (4) Calculate the difference between the current height-length deviation and the previous height-length deviation to obtain the rate of change of the deviation. Then multiply the rate of change by the preset differential coefficient to obtain the differential operation output. The differential link is used to suppress the sudden change trend of the deviation in advance, offset the instantaneous fluctuations caused by factors such as airflow disturbance and motor vibration, and improve the stability of the controller output.

[0122] (5) Add the proportional operation output, integral operation output and differential operation output to obtain the original control quantity without tension correction. This original control quantity is the basic speed control quantity. Its value corresponds to the target speed adjustment range of the winch motor, and the positive or negative value corresponds to the rotation direction of the motor for winding or unwinding.

[0123] (6) Set upper and lower limit thresholds for the generated basic speed control quantity to obtain the basic speed control quantity after amplitude limiting. The upper limit is set to 250 and the lower limit is set to -250. The control quantity exceeding the amplitude limit range is truncated according to the threshold boundary value to ensure that the output signal is safe and reliable.

[0124] S2.5: Obtain real-time cable tension data, and based on the real-time tension data, obtain the tension influence factor corresponding to the real-time tension data through a preset tension-speed correction lookup table;

[0125] Furthermore, the specific steps of S2.5 include:

[0126] (1) Initialize the construction and parameter configuration of the tension-speed correction lookup table. The tension-speed correction lookup table used in this embodiment is established based on the cable safety tension threshold range. The safety tension threshold range is from the minimum safety tension threshold of 5N to the maximum safety tension threshold of 40N. This safety tension threshold range serves as the sole basis for the construction of the tension-speed correction lookup table and has a direct reference basis. The tension-speed correction lookup table divides the entire tension range into 5 consecutive intervals. Each interval corresponds to a fixed tension influence factor. All interval boundaries and factor values ​​are calibrated and stored in the read-only storage area of ​​the control unit before leaving the factory and remain unchanged during operation. Specifically, the division method is as follows: tension greater than or equal to 5N and less than or equal to 15N is the first interval; tension greater than 15N and less than or equal to 25N is the second interval; tension greater than 25N and less than or equal to 35N is the third interval; tension greater than 35N and less than or equal to 40N is the fourth interval. Intervals less than 5N or greater than 40N do not participate in the regular correction and directly enter the tension compensation logic.

[0127] (2) Set the tension influence factor corresponding to the first interval to 1.1, the tension influence factor corresponding to the second interval to 1.0, the tension influence factor corresponding to the third interval to 0.9, and the tension influence factor corresponding to the fourth interval to 0.8.

[0128] (3) Obtain the real-time tension data after sliding window averaging filtering. The reading period is consistent with the system control period, which is 20ms. After reading, the real-time tension data is validated. The validation range is set to 0N to 50N. Tension data exceeding this range is considered abnormal and the valid tension value of the previous period is directly used as the replacement.

[0129] (4) The real-time tension data that has passed the verification is compared with the interval boundary of the tension-speed correction lookup table one by one. Specifically, starting from the lowest interval, the interval range to which the real-time tension data belongs is judged in turn. First, it is judged whether it belongs to the first interval. If it does not belong, it is judged whether it belongs to the second interval. The process is carried out in turn until the corresponding target interval is matched. The entire judgment process is executed in the interval order.

[0130] (5) After determining the interval to which the real-time tension data belongs, retrieve the tension influence factor value stored in advance for that interval from the tension-speed correction lookup table;

[0131] (6) Confirm whether the extracted tension influence factor is within the legal range of 0.8 to 1.1, and output the confirmed tension influence factor. The legal range is the safety correction range set by the system. Factors exceeding the range will be restricted to boundary values ​​to avoid over-correction of the basic speed control quantity.

[0132] S2.6: Multiply the basic speed control quantity by the tension influence factor to generate the comprehensive speed control quantity;

[0133] Furthermore, the integrated speed control quantity is the final control parameter that integrates the requirements for high synchronization and tension safety. By multiplying the basic speed control quantity and the tension influence factor, the two control requirements are organically combined. When the basic speed control quantity needs to increase the motor speed, if the tension is too high, the tension influence factor will reduce the integrated speed control quantity to prevent the tension from increasing further due to excessive motor speed. If the tension is too low, the tension influence factor will increase the integrated speed control quantity, allowing the motor to respond quickly and eliminate deviations. This integrated control method ensures rapid synchronization between the drone altitude and the winch cable length, while always keeping the cable tension within a safe range, achieving dual control objectives.

[0134] S2.7: The comprehensive speed control quantity is converted into a take-up and release speed adjustment signal with a corresponding duty cycle using a pulse width modulation signal generator; the take-up and release speed adjustment signal is used to drive the winch motor to perform take-up or release operations at the target speed.

[0135] Furthermore, the winch motor is a brushless DC motor, which requires a pulse width modulation (PWM) signal for driving. The duty cycle of the PWM signal directly determines the motor speed; a larger duty cycle results in a faster motor speed, and a smaller duty cycle results in a slower motor speed. The signal level direction determines the motor's rotation direction: a high level corresponds to wire release, and a low level corresponds to wire take-up. In this embodiment, the PWM signal generator receives the comprehensive speed control quantity and converts it into a pulse signal with the corresponding duty cycle based on the magnitude of the comprehensive speed control quantity. The direction of the signal is determined based on the positive or negative sign of the control quantity, generating a stable and interference-free wire release / take-up speed adjustment signal. This signal is directly input to the motor driver, which amplifies the signal and drives the winch motor to rotate at the target speed, achieving precise wire release / take-up speed control. The entire signal conversion process takes less than 1ms, ensuring real-time motor response.

[0136] Furthermore, the pulse width modulation signal generator is integrated into the UAV winch drive control unit. After power-on, it first performs register clearing and clock source calibration operations. The working clock source frequency of the pulse width modulation signal generator is set to 16MHz, the signal output period is set to 20ms, and the maximum count value of the pulse width modulation signal is set to 10000. This value is used to determine the full-scale reference for duty cycle calculation. All parameters are configured at once during the initialization phase and remain fixed during operation.

[0137] Furthermore, the comprehensive speed control quantity is obtained by multiplying the basic speed control quantity by the tension influence factor, and its value range is limited to -250~250. When the system preset value is positive, it represents the wire feeding operation; when the value is negative, it represents the wire reeling operation; and when the value is zero, it represents the motor stopping. This polarity judgment rule corresponds completely to the rotation direction control logic of the winch motor.

[0138] Furthermore, the specific steps in S2.7 include:

[0139] (1) The pulse width modulation signal generator reads the comprehensive speed control quantity at the current moment, first extracts the symbol information of the comprehensive speed control quantity and stores it temporarily, then calculates the absolute value of the comprehensive speed control quantity, and converts the input range into a positive value range of 0 to 250. Then, the value is mapped to the count value range of 0 to 10000 according to the linear proportional relationship. The mapping ratio coefficient is set to 40. The pulse count value corresponding to the amplitude of the comprehensive speed control quantity is obtained by conversion.

[0140] (2) The pulse width modulation signal generator compares the obtained pulse count value with the maximum count value of 10000, and uses the ratio of the pulse count value to the maximum count value as the duty cycle parameter of the final output. The duty cycle value is limited to 0%~100%. When the comprehensive speed control value is 0, the duty cycle is 0% and when the comprehensive speed control value reaches ±250, the duty cycle is 100%. The intermediate values ​​are output according to the linear relationship.

[0141] (3) The pulse width modulation signal generator calls the symbol information temporarily stored in (1). When the comprehensive speed control quantity is positive, it outputs a high-level direction indicator, and the corresponding motor performs the wire feeding action. When the comprehensive speed control quantity is negative, it outputs a low-level direction indicator, and the corresponding motor performs the wire taking-up action. When the comprehensive speed control quantity is zero, it does not output a direction indicator, and the motor remains in a stopped state.

[0142] (4) The pulse width modulation signal generator takes 20ms as a complete output cycle. In each cycle, it outputs a pulse waveform with alternating high and low levels according to the calculated duty cycle. It outputs a high level at the beginning of the cycle and switches to a low level when the internal counter reaches the pulse count value, and maintains it until the end of the current cycle. The waveform of each cycle is generated independently and strictly aligned with the system clock to ensure that the output signal is free from jitter and distortion.

[0143] (5) The pulse width modulation signal generator synchronously binds the generated pulse width modulation waveform with the direction indicator signal so that the two are perfectly aligned in time. The combined signal is the winding and unwinding speed adjustment signal. The duty cycle of the pulse width modulation waveform determines the target speed of the winch motor. The larger the duty cycle, the faster the speed. The direction indicator determines the rotation direction of the winch motor, thereby controlling the execution of winding or unwinding actions.

[0144] (6) The take-up and release speed adjustment signal is transmitted to the motor driver through the dedicated output interface of the control unit. The signal transmission adopts a single-ended level output mode, with the high level voltage set to 3.3V and the low level voltage set to 0V, which conforms to the level recognition standard of the motor driver. After receiving the take-up and release speed adjustment signal, the motor driver drives the motor to run according to the duty cycle and direction command, so that the motor performs take-up or release operation at the target speed.

[0145] After generating the height-length deviation, the method further includes integral separation control based on the height-length deviation, specifically:

[0146] A1: Take the absolute value of the height-length deviation to generate the absolute value of the deviation;

[0147] A2: Compare the absolute value of the deviation with a preset integral separation threshold; when the absolute value of the deviation is greater than the integral separation threshold, generate an integral separation enable signal to disable integral operation in the proportional-integral-derivative controller; when the absolute value of the deviation is less than or equal to the integral separation threshold, generate an integral input enable signal to restart integral operation.

[0148] Furthermore, integral separation control is used to address the integral saturation problem of PID controllers under large deviations. When the drone suddenly ascends or descends, or experiences a sudden change in load, the altitude-length deviation increases instantaneously. If the integral term continues to operate at this time, it will lead to excessive integral accumulation, causing motor overshoot, oscillation, or even damage to cables or motors. In this embodiment, the preset integral separation threshold is 5cm. When the absolute value of the deviation is greater than 5cm, it indicates a large deviation, requiring a rapid response. The integral term is then shut down, and the motor is controlled only through proportional and derivative components to avoid integral saturation. When the absolute value of the deviation decreases to within 5cm, it indicates a small deviation, requiring the elimination of static error. The integral term is then reactivated, and the absolute value of the deviation is gradually brought closer to zero through integral calculation.

[0149] Furthermore, when the integral separation enable signal is received, the deviation accumulation operation is immediately stopped, the current integral accumulation value remains unchanged, and no new deviation is accumulated to avoid integral saturation. When the integral input enable signal is received, the deviation accumulation operation is immediately resumed, and the integral operation continues to be performed according to the preset integral coefficient to eliminate the static deviation of the system.

[0150] S3: During the process of the winch motor responding to the winding and unwinding speed adjustment signal to perform the winding and unwinding operation, the real-time tension data of the cable is continuously detected, and the real-time tension data is compared and analyzed with the preset safe tension threshold range. If the real-time tension data exceeds the safe tension threshold range, it is determined that the cable is in a slack state, has abnormal tension, or has a sudden change in load.

[0151] Furthermore, during the process of the winch motor responding to the winding and unwinding speed adjustment signal to perform the winding and unwinding operation, the real-time tension data of the cable is continuously detected, including: acquiring the positive pressure signal of the cable on the sensor axle based on a wheel-axle tension sensor installed in series at the cable lead-out end, wherein the wheel-axle tension sensor integrates a strain gauge full-bridge circuit; based on the signal conditioning circuit, the millivolt-level differential voltage signal output by the strain gauge full-bridge circuit is sequentially amplified by an instrumentation amplifier, filtered by a low-pass filter, and quantized by an analog-to-digital converter to generate a digital tension value; and based on a sliding window averaging filter, the digital tension value at multiple consecutive sampling times is averaged to generate real-time tension data; the sampling frequency of the real-time tension data is not less than twice the control cycle of the winch motor.

[0152] The real-time tension data is compared and analyzed with a preset safe tension threshold range. If the real-time tension data exceeds the safe tension threshold range, it is determined that the cable is in a slack state, experiencing abnormal tension, or a sudden change in load, including:

[0153] S3.1: Read the pre-stored safe tension threshold range; the safe tension threshold range includes a minimum safe tension threshold and a maximum safe tension threshold;

[0154] Furthermore, the safe tension threshold range is determined based on the material, diameter, rated tensile force, and operating load of the cable used in this embodiment. The cable is a high-strength aramid fiber braided cable with a rated tensile force of 50N. The minimum safe tension threshold is set to 5N. If it is lower than the minimum safe tension threshold, it indicates that the cable is too loose and there is a risk of tangling or knotting. The maximum safe tension threshold is set to 40N, which is lower than the rated tensile force of the cable to leave a safety margin. If it is higher than the maximum safe tension threshold, it indicates that the cable is under too much force and there is a risk of breakage or loosening of the interface. The safe threshold range is stored in the controller's non-volatile memory and can be fine-tuned during operation according to the actual load and cable condition to ensure the rationality and safety of the threshold.

[0155] S3.2: Based on comparator logic, the real-time tension data is compared with the minimum safe tension threshold and the maximum safe tension threshold respectively; if the real-time tension data is less than the minimum safe tension threshold, it is determined that the cable is in a slack state; if the real-time tension data is greater than the maximum safe tension threshold, it is determined that the cable has abnormal tension or sudden load change; if the real-time tension data is within the safe tension threshold range, it is determined that the cable tension is normal.

[0156] In this embodiment, the comparator logic is a hardware real-time comparison logic with a response time of less than 1ms. It can compare the real-time tension data with the threshold in real time. When the real-time tension data is between 5N and 40N, the tension is determined to be normal, and the winch motor continues to operate according to the cable reeling speed adjustment signal. When the real-time tension data is less than 5N, the cable is determined to be slack. At this time, the cable is drooping and swaying, which may easily cause it to get tangled with the drone's fuselage or rotor, leading to a safety accident. When the real-time tension data is greater than 40N, the cable is determined to have abnormal tension or sudden load change. This may be due to the drone suddenly rising, the load below getting stuck, or the cable getting caught on an obstacle. If not handled in time, it may lead to consequences such as the cable breaking, the load falling, and the drone going out of control. The comparator logic outputs the judgment result in real time.

[0157] Furthermore, during the process of the winch motor responding to the winding and unwinding speed adjustment signal to perform the winding and unwinding operation, the winding uniformity control is also included based on the unwound cable length and the number of cable layers. Specifically, this includes: firstly, generating the remaining cable length data by subtraction based on the actual unwound length data and the pre-stored total cable length data; simultaneously, generating the theoretical lateral displacement of the cable guide based on the number of shaft rotations recorded by the incremental photoelectric encoder and the cable lead; then, generating the target position of the cable guide based on the remaining cable length data and the number of cable layers by querying a preset cable guide position mapping table; and finally, using the difference between the theoretical lateral displacement of the cable guide and the target position of the cable guide as input, outputting the position adjustment signal of the cable guide drive motor based on the closed-loop proportional-integral-derivative controller, so that the cable guide guides the cable to be neatly arranged on the winding hub. In this embodiment, the effective winding width of the winding hub is set to 120 mm, the outer diameter of the cable is set to 1.2 mm, the number of winding turns per layer is set to 120, the maximum number of winding layers is set to 20, and the total cable length is set to 100 meters. The cable guide position mapping table uses the number of cable layers as the first index and the remaining cable length as the second index, dividing the effective winding width of the hub into 120 position levels evenly in the horizontal direction. Each turn of cable corresponds to an independent horizontal target position of the cable guide. The starting point, ending point, and step displacement of the cable guide corresponding to each layer are marked according to the rule of changing direction layer by layer. The first layer is arranged from left to right, the second layer is arranged from right to left, and the subsequent layers alternate in direction to ensure that each layer of cable is tightly arranged without overlapping or depression. Each item in the cable guide position mapping table contains three sets of associated data: the number of layers, the remaining cable length, and the corresponding target position of the cable guide. All position parameters are quantified and calibrated in millimeters and stored in the non-volatile storage area of ​​the controller, which is directly queried and called during operation.

[0158] Furthermore, the cable guide position mapping table is constructed through physical winding calibration before the equipment leaves the factory: based on the structure size of the winding hub, the cable diameter, and the number of turns per layer, the cable is manually wound layer by layer for calibration. The first layer of cable is wound from the left end of the hub to the right, and the second layer is folded back from the right end to the left, alternating in a cycle. Each layer of 120 turns is divided into 120 horizontal positions. The remaining cable length is calculated by combining the cable consumption length per turn. The horizontal coordinates of the cable guide corresponding to each layer and remaining length are matched one by one and entered into the data table line by line.

[0159] S4: Generate a tension compensation signal or preset protection logic based on the judgment result; the tension compensation signal is used to correct the take-up and release speed adjustment signal; the preset protection logic includes: issuing an anti-loosening warning when a sudden tension change is detected at the cable interface, and triggering anti-tangling stop or emergency take-up and release buffer control when the cable is redundant or too tight.

[0160] The step of generating a tension compensation signal based on the determination result includes:

[0161] B1: When the determination result is that the cable is in a slack state, calculate the difference between the real-time tension data and the minimum safe tension threshold to generate a slack deviation, and input the slack deviation to the first fuzzy logic controller; the input membership function of the first fuzzy logic controller adopts a triangular function, the output membership function adopts a single-valued function, and fuzzy reasoning and centroid method defuzzification are performed through a pre-set fuzzy rule library to generate the first tension compensation coefficient;

[0162] Furthermore, when the cable is in a slack state, the release speed needs to be slowed down or the take-up speed needs to be accelerated through the tension compensation signal, so that the cable can be quickly tensioned to the safe tension threshold range. The slack deviation reflects the severity of the cable slack; the larger the deviation, the more severe the slack, and the greater the compensation force required. The first fuzzy logic controller is specifically designed for cable slack scenarios. The input membership function adopts a triangular function, which can accurately classify the slack deviation into three levels: slight, moderate, and severe. The output membership function adopts a single-valued function, which outputs a stable compensation coefficient without oscillation. The fuzzy rule base is obtained based on multiple sets of whole machine prototype test calibrations. Combined with the hierarchical logic of three-level input and three-level output, three core control rules are condensed, such as "small slack deviation, small compensation coefficient" and "large slack deviation, large compensation coefficient". The slack deviation is converted into the corresponding compensation coefficient through fuzzy inference, and then the centroid method is used to defuzzify and output a precise first tension compensation coefficient. The first tension compensation coefficient is greater than 1, which is used to increase the take-up speed of the motor or reduce the release speed.

[0163] Furthermore, the specific steps for B1 include:

[0164] (1) Read the judgment result of the current cycle;

[0165] (2) Obtain the real-time tension data after the sliding window averaging process, and perform validity verification on the real-time tension data after reading. Only retain the valid data in the range of 0N~5N, and replace the data outside the range with the valid value of the previous period.

[0166] (3) Subtract the current effective real-time tension data from the minimum safe tension threshold of 5N to obtain the slack deviation. The slack deviation is a non-negative value. The larger the value, the more serious the cable slack. This slack deviation is used as the only input variable of the first fuzzy logic controller.

[0167] (4) Configure the fuzzy subsets and triangular membership function parameters of the input variables of the first fuzzy logic controller. The relaxation bias is divided into three fuzzy subsets: slight relaxation, moderate relaxation, and severe relaxation. The universe of discourse of the input variables is set to 0N~5N. The vertex of the triangular membership function corresponding to slight relaxation is 0, the left endpoint is -1N, and the right endpoint is 2.5N. The vertex of the triangular membership function corresponding to moderate relaxation is 2.5N, the left endpoint is 0N, and the right endpoint is 5N. The vertex of the triangular membership function corresponding to severe relaxation is 5N, the left endpoint is 2.5N, and the right endpoint is 6N. All membership function parameters are set during the initialization stage and are used to fuzzify the relaxation bias.

[0168] (5) Substitute the actual input relaxation bias into the three configured triangular membership functions, and calculate the membership degree of the relaxation bias to the three fuzzy subsets of slight relaxation, moderate relaxation and severe relaxation respectively. Each fuzzy subset outputs a membership value between 0 and 1, which represents the degree to which the current relaxation bias belongs to the fuzzy state. The larger the membership value, the higher the matching degree. The formula for calculating the triangular membership function is the prior art in this field and is not an inventive solution of this application. It will not be elaborated here.

[0169] (6) Configure the fuzzy subset and single-value membership function parameters of the output variable. The output variable of the first fuzzy logic controller is the first tension compensation coefficient. The output domain is set to 1.0~1.3. The output variable is divided into three fuzzy subsets, namely small compensation, medium compensation and large compensation. The single-value function output corresponding to small compensation is 1.1, the single-value function output corresponding to medium compensation is 1.2, and the single-value function output corresponding to large compensation is 1.3. The single-value function directly outputs a fixed value during operation without interval change.

[0170] (7) Load and enable the pre-set fuzzy rule library, which contains three fixed rules. The first rule is that if the input is slightly relaxed, the output will be small compensation. The second rule is that if the input is moderately relaxed, the output will be medium compensation. The third rule is that if the input is severely relaxed, the output will be large compensation. The rule library adopts a fixed logical relationship, which is consistent with the actual physical characteristics of the cable relaxation degree and compensation force. It is calibrated and stored before leaving the factory and is not modified during operation.

[0171] (8) Match the three fuzzy subset membership values ​​obtained in sequence according to the fuzzy rule library, calculate the excitation intensity of each rule by using the maximum-minimum synthesis inference method, obtain the activation degree of the output fuzzy subset corresponding to each rule, and retain the maximum membership degree of each rule as the inference result;

[0172] Furthermore, the specific steps for calculating the incentive intensity of each rule using the max-min composite inference method include:

[0173] 1) Load the relaxed bias membership data that has been fuzzified, load the configuration information of all fuzzy rules in the first fuzzy logic controller, and then load the fixed parameters of the single-value membership function of the output variable;

[0174] 2) Read the input membership degree values ​​of the first fuzzy rule sequentially. These membership degree values ​​represent the degree of membership of the relaxation bias to the slightly relaxed fuzzy subset, and their range is between 0 and 1. This is the sole input basis for calculating the excitation intensity of this rule. Then, perform a minimum value operation on the first fuzzy rule. It should be explained that the minimum value operation in the max-min composite inference method refers to directly taking the value of the current rule's input membership degree as the initial excitation intensity of this rule. Since the first fuzzy rule contains only a single input variable, this input membership degree value is directly used as the excitation intensity of this rule. Finally, store the excitation intensity obtained after the minimum value operation into a dedicated register as the weight basis for this rule to participate in the centroid method defuzzification operation.

[0175] 3) Next, read the membership value of the input fuzzy subset corresponding to the second fuzzy rule. This membership value is the degree of membership of the relaxation bias to the moderately relaxed fuzzy subset. Perform the minimum value operation on the second fuzzy rule and store the excitation intensity calculated by this rule into an independent storage unit, which is stored separately from the excitation intensity of the first rule to avoid data interference between different rules.

[0176] 4) Continue reading the membership degree value of the input fuzzy subset corresponding to the third fuzzy rule. This membership degree value is the degree of membership of the relaxation bias to the severely relaxed fuzzy subset. Perform the minimum value operation on the third fuzzy rule and store the excitation intensity obtained by this rule into the corresponding storage location. At this point, the excitation intensity of all three fuzzy rules has been calculated and safely stored.

[0177] 5) Check in turn whether the incentive intensity values ​​of the three rules are all between 0 and 1. This range is the legal range of fuzzy inference incentive intensity. If a value is found that is outside the range, it will be forcibly limited to the boundary value.

[0178] 6) The excitation intensity of the three rules that have been verified is synchronously transmitted to the centroid method for defuzzification calculation as the weight parameters for weighted calculation.

[0179] (9) Using the values ​​of each single-valued function as the weighting center and the excitation intensity of the corresponding rule as the weighting coefficient, the cumulative denominator term is obtained by weighted summation; the cumulative summation of the excitation intensity values ​​of each rule is calculated to obtain the cumulative numerator term; then the cumulative denominator term is divided by the cumulative numerator term to obtain the first tension compensation coefficient.

[0180] (10) Perform amplitude constraint and validity verification on the first tension compensation coefficient, and output the first tension compensation coefficient after verification. The system limits the effective range of the first tension compensation coefficient to 1.0~1.3. The calculation results exceeding the effective range are truncated according to the boundary value to ensure that the compensation force is within a safe and controllable range and will not cause excessive adjustment to the motor speed.

[0181] B2: Multiply the first tension compensation coefficient by the comprehensive speed control amount in the take-up and release speed adjustment signal to generate a corrected speed adjustment signal that increases the take-up speed or decreases the release speed;

[0182] Furthermore, the corrected speed adjustment signal is amplified based on the original take-up and release speed adjustment signal through the first tension compensation coefficient, allowing the winch motor to take up the line at a faster speed, quickly reduce the length of the released cable, eliminate the slack state, and make the tension quickly rise back to the safe threshold range. During the correction process, the system will limit the maximum compensation amplitude to avoid excessive compensation that would cause the tension to momentarily exceed the safe upper limit, ensuring that the correction process is smooth and safe.

[0183] B3: When the determination result is that the cable has abnormal tension or sudden load change, calculate the difference between the real-time tension data and the maximum safe tension threshold, generate the over-tension deviation, and input the over-tension deviation to the second fuzzy logic controller to generate the second tension compensation coefficient; the second fuzzy logic controller is configured the same as the first fuzzy logic controller;

[0184] Furthermore, when there is abnormal tension or sudden load change in the cable, it is necessary to slow down the take-up speed or speed up the release speed through the tension compensation signal to release the cable tension and prevent the tension from continuing to rise. The over-tension deviation reflects the severity of the tension exceeding the limit. The larger the deviation, the greater the tension and the greater the compensation force. The second fuzzy logic controller has the same structure and parameters as the first fuzzy logic controller. It only adjusts the fuzzy rule base for the over-tension scenario. The rule logic is "small over-tension deviation, small compensation coefficient" and "large over-tension deviation, large compensation coefficient". The output second tension compensation coefficient is less than 1, which is used to suppress over-speed release or reduce the take-up speed.

[0185] Furthermore, the second fuzzy logic controller is specifically designed for tension compensation calculations in scenarios involving abnormal tension or sudden load changes in cables. Its overall structure and operating parameters are completely consistent with those of the first fuzzy logic controller. During the system power-on initialization phase, the second fuzzy logic controller completes the loading and fixing of all parameters. The input variable is the over-tension deviation, and the output variable is the second tension compensation coefficient. The input membership function uniformly adopts a triangular function, and the output membership function uniformly adopts a single-valued function. The fuzzy inference process uses the max-min synthesis inference method, and the defuzzification method uses the centroid method. The controller's operation cycle is consistent with the overall system control cycle, set to 20ms. All structural and algorithm parameters are calibrated before leaving the factory and are not modified during operation.

[0186] Furthermore, the specific steps for B3 include:

[0187] (1) Obtain the cable safety tension threshold range, and obtain and confirm the valid judgment results of abnormal cable tension or load change;

[0188] (2) Obtain the real-time tension data after stabilization. After reading, verify the validity of the data and retain only the valid data within the range of 40N or greater and 50N or less. Data outside this range is replaced by the valid value of the previous cycle.

[0189] (3) Subtract the maximum safe tension threshold of 40N from the current effective real-time tension data to obtain the over-tension deviation. The over-tension deviation is a non-negative value. The larger the value, the more serious the tension on the cable exceeds the safe limit, and the more obvious the load change. This over-tension deviation is the only input variable of the second fuzzy logic controller.

[0190] (4) Configure the fuzzy subsets and triangular membership function parameters of the input variables of the second fuzzy logic controller. Specifically, the over-tension deviation is divided into three fuzzy subsets: slight over-tension, moderate over-tension, and severe over-tension. The universe of discourse of the input variables is set to 0N~10N. The vertex of the triangular membership function corresponding to slight over-tension is 0N, the left endpoint is -1N, and the right endpoint is 5N. The vertex of the triangular membership function corresponding to moderate over-tension is 5N, the left endpoint is 0N, and the right endpoint is 10N. The vertex of the triangular membership function corresponding to severe over-tension is 10N, the left endpoint is 5N, and the right endpoint is 11N. All membership function parameter setting rules are exactly the same as the input membership function parameter setting rules of the first fuzzy logic controller.

[0191] (5) Substitute the actual input over-tension deviation into the three configured triangular membership functions, and calculate the membership of the over-tension deviation to the three fuzzy subsets of slight over-tension, moderate over-tension, and severe over-tension respectively. Each fuzzy subset outputs a membership value between 0 and 1, which represents the degree to which the current over-tension deviation belongs to the fuzzy state. The larger the membership value, the higher the matching degree.

[0192] (6) Configure the fuzzy subset and single-value membership function parameters of the output variable. The output variable of the second fuzzy logic controller is the second tension compensation coefficient. The output domain is set to 0.7~1.0. The output variable is divided into three fuzzy subsets, namely small compensation, medium compensation and large compensation. The single-value function output corresponding to small compensation is 0.9, the single-value function output corresponding to medium compensation is 0.8, and the single-value function output corresponding to large compensation is 0.7. The single-value function directly outputs a fixed value during operation, and the setting logic of the output compensation coefficient maintains a symmetrical correspondence with the first fuzzy logic controller.

[0193] (7) Load and enable the pre-set fuzzy rule library, which contains three fixed rules. The first rule is to output small compensation when the input is slightly over-tension, the second rule is to output medium compensation when the input is moderate over-tension, and the third rule is to output large compensation when the input is severely over-tension. The rule library adopts fixed logic relationship, which is consistent with the actual physical characteristics of the cable over-tension degree and compensation force. It is calibrated and stored before leaving the factory. The number and structure of the rules are the same as those of the first fuzzy logic controller.

[0194] (8) The three fuzzy subset membership values ​​obtained in (5) are matched sequentially according to the fuzzy rule library. The excitation intensity of each rule is calculated by using the maximum-minimum synthesis reasoning method. The activation degree of the output fuzzy subset corresponding to each rule is obtained. The maximum membership degree of each rule is retained as the reasoning result. The reasoning operation process is completely consistent with the first fuzzy logic controller.

[0195] (9) The centroid method is used to perform defuzzification operation to generate accurate second tension compensation coefficient, and the defuzzification operation is exactly the same as that of the first fuzzy logic controller;

[0196] (10) The amplitude constraint and validity verification of the second tension compensation coefficient are performed to obtain the second tension compensation coefficient after verification. The effective range of the second tension compensation coefficient is limited to 0.7~1.0. The calculation results exceeding this range are truncated according to the boundary value to ensure that the compensation force is within a safe and controllable range and will not cause excessive adjustment to the motor speed.

[0197] B4: Multiply the second tension compensation coefficient by the comprehensive speed control amount in the take-up and release speed adjustment signal to generate a corrected speed adjustment signal that increases the release speed or decreases the take-up speed.

[0198] Furthermore, the corrected speed adjustment signal reduces the amplitude of the original speed control quantity through the second tension compensation coefficient, slowing down the take-up and constraining the release of the wire too quickly, reducing the tension of the released cable, and allowing the tension to quickly fall back to the safe threshold range; similarly, the system will limit the minimum compensation amplitude to avoid excessive compensation causing the tension to drop below the safe lower limit momentarily, ensuring that the tension recovers smoothly.

[0199] The method of issuing an anti-loosening warning when a sudden change in tension is detected at the cable interface includes:

[0200] C1: Perform a first-order difference operation on the real-time tension data sequence composed of the real-time tension data to generate tension change rate data. The first-order difference operation is a prior art in this field and is not an inventive solution of this application. It will not be described in detail here.

[0201] Furthermore, tension mutations are not only manifested as tension values ​​exceeding the maximum safety threshold, but also as rapid changes in tension within a short period. Relying solely on tension values ​​cannot accurately identify the risk of interface loosening; therefore, tension change rate data is introduced. The system performs a first-order difference operation on a sequence of 10 consecutive real-time tension data points, calculating the ratio of the change in tension to the change time at adjacent moments, generating tension change rate data. This data accurately reflects the rate of tension change; a higher change rate indicates a more severe tension mutation and a higher risk of interface loosening.

[0202] C2: Generate a tension change judgment threshold based on the pre-calibrated interface pull-out force threshold at the cable interface;

[0203] Furthermore, the interface between the cable and the drone / load is a weak point under stress. The interface pull-out force threshold is pre-calibrated through experiments. This threshold represents the maximum rate of tension change the interface can withstand; exceeding this value may cause the interface screws, clips, and other connectors to loosen or slip. Based on the interface pull-out force threshold and a safety factor, the system generates a tension mutation judgment threshold, which serves as the standard for determining whether the interface is at risk of loosening.

[0204] In this embodiment, the cable interface includes two key parts: the connection end between the cable and the UAV body and the connection end between the cable and the load. Before leaving the factory, the system conducts a pull-out force calibration test on the two types of interfaces. By applying tension step by step and recording the critical tension value when the interface tends to loosen, the interface pull-out force threshold is finally determined to be uniformly calibrated to 45N. This value is the safe critical value for the interface to withstand sudden tension. After calibration, it is stored in the read-only storage area of ​​the controller and remains fixed during operation without dynamic modification.

[0205] Furthermore, the tension mutation judgment threshold is used to determine whether the tension increase in a short period of time reaches the risk level of interface loosening. It is generated based on the interface pull-out force threshold, combined with the safety factor and actual working conditions. In this embodiment, the safety factor is set to 0.8 to retain sufficient safety margin and avoid frequent false alarms caused by setting the tension mutation judgment threshold too tightly.

[0206] In this embodiment, the pre-calibrated interface pull-out force threshold of 45N is multiplied by a safety factor of 0.8 to obtain the maximum allowable tension rise value that the interface can withstand within a single calculation cycle. The calculation result is 36N, which means that the interface is in a safe state when the tension rise does not exceed 36N within 20ms. If the tension rise exceeds the threshold, there is a risk of loosening. This tension rise is the tension change judgment threshold.

[0207] C3: When the real-time tension data exceeds the maximum safe tension threshold and the tension change rate data exceeds the tension change determination threshold, it is determined that a tension change has occurred at the cable interface;

[0208] It should be noted that only when both the tension value and the tension change rate exceed the limit are met simultaneously will it be determined as a sudden tension change at the interface. This avoids misjudgment due to a single abnormal parameter. This dual judgment logic improves the accuracy of the warning and reduces false and missed warnings.

[0209] C4: Generate an anti-loosening warning signal based on the judgment result, and send the anti-loosening warning signal to the ground station through the UAV's data link downlink, while triggering the buzzer and status indicator light to issue an on-site alarm.

[0210] Furthermore, the anti-loosening warning signal includes information such as warning time, drone location, real-time tension value, tension change rate, and interface status. This information is transmitted to the ground station in real time via the drone's data transmission radio. The ground station's monitoring interface immediately displays a red warning window and plays a warning voice message to remind operators to handle the situation promptly. At the same time, the drone's buzzer emits a high-frequency alarm sound, and the red status indicator light flashes rapidly, achieving on-site audible and visual alarms. This allows on-site personnel to detect abnormalities immediately and take timely emergency measures to prevent interface loosening and load drop.

[0211] The mechanism for triggering anti-tangle shutdown or emergency cable rewind / unwinding buffer control when cable redundancy or excessive cable tightness is determined includes:

[0212] D1: When the determination result is that the cable is in a slack state and the duration of the slack state exceeds the preset redundancy tolerance time threshold, it is determined that the cable is redundant and there is a risk of entanglement; when the determination result is that the cable has abnormal tension and the duration of the abnormal tension exceeds the preset overtightness tolerance time threshold, it is determined that the cable is too tight and there is a risk of jamming; based on the determination result, an anti-entanglement stop signal is generated; the anti-entanglement stop signal is used to immediately cut off the power input of the winch motor and simultaneously control the electromagnetic brake to grip the winch shaft, and perform emergency stop protection;

[0213] Furthermore, if the cable slack condition persists for more than 3 seconds beyond the redundancy tolerance time threshold, it indicates that the tension compensation cannot eliminate the slack, the cable has a large amount of redundancy, and is very prone to tangling and knotting. In this case, it is determined to be cable redundancy. If the abnormal tension condition of the cable persists for more than 2 seconds beyond the overtightness tolerance time threshold, it indicates that the tension compensation cannot release the tension, the cable may have caught on an obstacle or the load is stuck, and there is a risk of breakage. In this case, it is determined to be cable overtightness. When either of the above two serious abnormalities occurs, the system immediately generates an anti-tangling stop signal. The anti-tangling stop signal is the highest priority instruction, which directly cuts off the power input of the motor driver and triggers the electromagnetic brake. The electromagnetic brake clamps the winch shaft within 100ms, preventing the drum from rotating, achieving an emergency stop, and preventing safety accidents such as cable tangling, breakage, and load drop, thus protecting the safety of the drone, winch, cable, and load.

[0214] D2: When the determination result is cable redundancy, an emergency take-up buffer control signal is generated. The emergency take-up buffer control signal drives the winch motor to accelerate take-up at a preset buffer acceleration until the real-time tension data is restored to the safe tension threshold range. Then, the take-up and release speed adjustment signal is switched back for control.

[0215] Furthermore, if the cable redundancy is relatively low and the shutdown threshold has not been reached, the system initiates emergency take-up buffer control. The preset buffer acceleration is 0.5 m / s² to prevent excessively fast take-up speed from causing sudden tension changes. The motor smoothly accelerates the take-up with buffer acceleration, gradually retrieving the redundant cable. At the same time, the tension data is monitored in real time. When the tension recovers to the safe tension threshold range of 5N~40N, the system automatically exits the buffer control and switches back to the regular take-up and take-up speed adjustment signal control, restoring height synchronization and adaptive tension adjustment, achieving automatic recovery without shutdown or impact.

[0216] D3: When the determination result is that the cable is too tight, an emergency cable release buffer control signal is generated. The emergency cable release buffer control signal drives the winch motor to decelerate and release the cable at a preset buffer deceleration until the real-time tension data is restored to the safe tension threshold range. Then, the control is switched back to the cable winding and releasing speed adjustment signal.

[0217] Furthermore, if the cable tension is only slightly too high and has not reached the shutdown threshold, the system will activate emergency cable release buffer control. The preset buffer deceleration is 0.5 m / s² to prevent the cable release speed from dropping too quickly and causing a sudden drop in tension. The motor will decelerate smoothly with the buffer deceleration to gradually release the cable tension. At the same time, the tension data will be monitored in real time. When the tension returns to the safe tension threshold range, the system will automatically exit the buffer control and switch back to the normal control mode to ensure continuous operation without manual intervention.

[0218] Furthermore, the method also includes torque feedforward compensation of the winch motor based on the rate of change of tension and the rate of change of height-length deviation, including: performing time difference on real-time tension data to generate a real-time tension rate of change, and simultaneously performing time difference on the height-length deviation to generate a deviation rate of change, then multiplying the real-time tension rate of change by a preset first feedforward gain coefficient to generate a first torque compensation amount, multiplying the deviation rate of change by a preset second feedforward gain coefficient to generate a second torque compensation amount, and finally summing the first torque compensation amount and the second torque compensation amount to generate a total torque feedforward compensation amount, and superimposing the total torque feedforward compensation amount onto the torque loop setpoint of the winch motor. The first feedforward gain coefficient is used to convert the tension change rate into the corresponding torque compensation amount. It is determined based on the rated tension range of the cable, the rated torque of the winch motor, and the 20ms control cycle of the system. In this embodiment, it is set to 0.012. The second feedforward gain coefficient is used to convert the height-length deviation change rate into the corresponding torque compensation amount. It is determined based on the maximum allowable change range of the height-length deviation, the measurement accuracy of the winch cable length, and the fusion accuracy of the UAV flight altitude. In this embodiment, it is set to 0.095. Both the first and second feedforward gain coefficients are calibrated through bench testing and on-site debugging before the system leaves the factory and remain fixed during operation. They can output torque compensation in advance when the tension or deviation shows a sudden change trend, effectively suppressing system overshoot and oscillation, and improving the speed and stability of the winch motor response.

[0219] Furthermore, the method also includes fault self-diagnosis based on the real-time tension data and the height-length deviation: First, a fault feature vector is constructed, which includes the real-time tension data, the height-length deviation, the speed command value in the winding and unwinding speed adjustment signal, and the real-time current feedback value of the winch motor; then, the fault feature vector is input to a classifier pre-trained based on a support vector machine, which uses a radial basis function kernel function and outputs fault type labels through a decision function; the fault type labels include: tension sensor fault, encoder fault, motor drive fault, and cable jamming fault; based on the fault type labels, corresponding fault alarm signals and degradation control are generated. The radial basis function kernel function is existing technology in this field and is not an inventive solution of this application, and will not be elaborated upon here.

[0220] Example 2:

[0221] Please see Figure 3 Another embodiment of the present invention provides: a drone winch altitude synchronization and tension adaptive control system, comprising:

[0222] The data acquisition module 10 is used to collect UAV flight altitude data and actual cable length data in real time. The flight altitude data is obtained by the fusion of RTK-GNSS receiver and barometric altimeter, and the actual length data is obtained by the incremental photoelectric encoder at the winch shaft end combined with the real-time diameter model of the winding hub. At the same time, the digital tension value is collected by the wheel axle tension sensor.

[0223] The deviation calculation module 20 is used to synchronize and align the collected flight altitude data with the actual release length data in time, calculate the difference to generate the original deviation, remove outliers and smooth the data using a limiting filter algorithm to obtain the altitude-length deviation.

[0224] The speed control generation module 30 is used to input the height-length deviation into the PID controller to calculate the basic speed control quantity. Combined with the real-time cable tension data, the tension influence factor is obtained through the tension-speed correction lookup table. The two are multiplied to obtain the comprehensive speed control quantity, which is then converted into a winding and unwinding speed adjustment signal by the pulse width modulation signal generator to drive the winch motor to wind and unwind the cable at the target speed.

[0225] The anomaly detection module 40 obtains real-time tension data by performing sliding window averaging filtering on the digital tension values ​​collected by the tension sensor, and compares them with the preset safe tension threshold range to determine whether the cable is in a slack state, has abnormal tension, or experiences a sudden load change. At the same time, it calculates the tension change rate to determine the tension change state at the interface.

[0226] The tension compensation module 50 is used to generate a tension compensation coefficient through a fuzzy logic controller to correct the winding and unwinding speed adjustment signal based on the tension anomaly judgment result. When a sudden tension change, cable redundancy, or cable being too tight is judged, it triggers anti-loosening warning, anti-tangling stop, or emergency winding and unwinding buffer control to ensure the safety of system operation.

[0227] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the present invention. All of these variations are within the protection scope of the present invention.

Claims

1. A method for unmanned aerial vehicle winch height synchronization and tension self-adaptive control, characterized in that, include: Real-time acquisition of drone flight altitude data and actual cable length data; The difference between the flight altitude data and the actual released length data is used to generate a height-length deviation, and a speed adjustment signal for driving the winch motor to adjust the wire release and take-up speed is generated based on the height-length deviation. During the process of the winch motor responding to the winding and unwinding speed adjustment signal to perform the winding and unwinding operation, the real-time tension data of the cable is continuously detected, and the real-time tension data is compared and analyzed with the preset safe tension threshold range. If the real-time tension data exceeds the safe tension threshold range, it is determined that the cable is in a slack state, has abnormal tension, or has a sudden change in load. Based on the judgment result, generate a tension compensation signal or preset protection logic; The tension compensation signal is used to correct the take-up and release speed adjustment signal; The preset protection logic includes: issuing an anti-loosening warning when a sudden tension change is detected at the cable interface; and triggering anti-tangling shutdown or emergency cable rewinding buffer control when cable redundancy or excessive tightness is detected. Based on the height-length deviation, a signal is generated to adjust the winding and unwinding speed of the winch motor, including: The height-length deviation is input to a proportional-integral-derivative controller, which performs proportional, integral, and derivative operations on the height-length deviation based on preset proportional, integral, and derivative coefficients to generate a basic speed control quantity. Obtain real-time cable tension data, and based on the real-time tension data, obtain the tension influence factor corresponding to the real-time tension data through a preset tension-speed correction lookup table; Multiply the basic speed control value by the tension influence factor to generate the comprehensive speed control value; A pulse width modulation signal generator is used to convert the comprehensive speed control quantity into a take-up and release speed adjustment signal with a corresponding duty cycle; the take-up and release speed adjustment signal is used to drive the winch motor to perform take-up or release operations at a target speed.

2. The unmanned aerial vehicle winch height synchronization and tension self-adaptive control method of claim 1, wherein, The flight altitude data is obtained by fusing data from a Global Navigation Satellite System receiver based on real-time dynamic differential technology and a barometric altimeter, specifically including: The global navigation satellite system receiver based on real-time dynamic differential technology collects pseudorange and carrier phase observations of the UAV relative to the base station, and generates the first altitude measurement value by solving the differential correction. The static pressure value at the location of the UAV is collected using a barometric altimeter, and the static pressure value is converted into a second altitude measurement value based on a standard atmospheric model; The first and second altitude measurements are used as observation inputs by employing a Kalman filter, and a state equation is constructed based on the vertical motion velocity and vertical acceleration of the UAV. Through the prediction and update iteration process of the Kalman filter, the flight altitude data after fusion correction is generated.

3. The unmanned aerial vehicle winch height synchronization and tension self-adaptive control method of claim 2, wherein, The actual length data is obtained by using an incremental photoelectric encoder installed on the winch shaft end in conjunction with a real-time diameter model of the winding hub, specifically including: An incremental photoelectric encoder installed on the end of the winch shaft is used to collect the number of rotation pulses and the rotation direction signal of the winch shaft. The number of rotation pulses is accumulated by a pulse counter to generate the number of shaft rotations. Based on the preset real-time diameter model of the winding hub, the pre-calibrated cable layer-diameter mapping table is queried, and combined with the real-time record of the current cable release layer, the average winding diameter corresponding to the current working turn of the winch shaft is generated. The actual length of the winch cable is generated by multiplying the number of rotations of the shaft by the average winding diameter.

4. The unmanned aerial vehicle winch height synchronization and tension self-adaptive control method of claim 3, wherein, The difference between the flight altitude data and the actual launch length data is used to generate an altitude-length deviation, including: The data includes the drone's flight altitude data and the actual length of the winch cable. A data alignment buffer is used to synchronize the flight altitude data and the actual length data at the same timestamp, generating synchronized altitude and length values. Subtract the synchronization length value from the synchronization height value to generate the original deviation. An amplitude limiting filtering algorithm is used to remove outliers and smooth the original deviation to generate a height-length deviation.

5. The unmanned aerial vehicle winch height synchronization and tension self-adaptive control method of claim 4, wherein, After generating the height-length deviation, the method further includes integral separation control based on the height-length deviation, specifically: The absolute value of the height-length deviation is measured to generate the absolute value of the deviation; The absolute value of the deviation is compared with a preset integral separation threshold. When the absolute value of the deviation is greater than the integral separation threshold, an integral separation enable signal is generated to disable the integral operation in the proportional-integral-derivative controller. When the absolute value of the deviation is less than or equal to the integral separation threshold, an integral input enable signal is generated to restart the integral operation.

6. The unmanned aerial vehicle winch height synchronization and tension self-adaptive control method of claim 5, wherein, The real-time tension data is compared and analyzed with a preset safe tension threshold range. If the real-time tension data exceeds the safe tension threshold range, it is determined that the cable is in a slack state, experiencing abnormal tension, or a sudden change in load, including: Read the pre-stored safe tension threshold range; the safe tension threshold range includes a minimum safe tension threshold and a maximum safe tension threshold. Based on comparator logic, the real-time tension data is compared with the minimum safe tension threshold and the maximum safe tension threshold respectively; if the real-time tension data is less than the minimum safe tension threshold, it is determined that the cable is in a slack state; if the real-time tension data is greater than the maximum safe tension threshold, it is determined that the cable has abnormal tension or sudden load change; if the real-time tension data is within the safe tension threshold range, it is determined that the cable tension is normal.

7. The unmanned aerial vehicle winch height synchronization and tension self-adaptive control method of claim 6, wherein, The step of generating a tension compensation signal based on the determination result includes: When the determination result indicates that the cable is in a slack state, the difference between the real-time tension data and the minimum safe tension threshold is calculated to generate a slack deviation, and the slack deviation is input to the first fuzzy logic controller. The input membership function of the first fuzzy logic controller adopts a triangular function, and the output membership function adopts a single-valued function. Fuzzy reasoning and centroid method defuzzification are performed through the fuzzy rule base to generate the first tension compensation coefficient. Multiply the first tension compensation coefficient by the comprehensive speed control amount in the take-up and release speed adjustment signal to generate a corrected speed adjustment signal that increases the take-up speed or decreases the release speed. When the determination result indicates that the cable has abnormal tension or sudden load change, the difference between the real-time tension data and the maximum safe tension threshold is calculated to generate an over-tension deviation, and the over-tension deviation is input to the second fuzzy logic controller to generate a second tension compensation coefficient; the second fuzzy logic controller is configured the same as the first fuzzy logic controller. The second tension compensation coefficient is multiplied by the comprehensive speed control value in the take-up and release speed adjustment signal to generate a corrected speed adjustment signal that increases the release speed or decreases the take-up speed.

8. The method for altitude synchronization and tension adaptive control of a UAV winch as described in claim 7, characterized in that, The method of issuing an anti-loosening warning when a sudden change in tension is detected at the cable interface includes: Perform a first-order difference operation on the real-time tension data sequence composed of the real-time tension data to generate tension change rate data; Based on the pre-calibrated interface pull-out force threshold at the cable interface, a tension change judgment threshold is generated; When the real-time tension data exceeds the maximum safe tension threshold and the tension change rate data exceeds the tension change determination threshold, it is determined that a tension change has occurred at the cable interface. Based on the judgment result, an anti-loosening warning signal is generated and sent to the ground station through the UAV's data link downlink. At the same time, the buzzer and status indicator light are triggered to issue an on-site alarm.

9. The method for altitude synchronization and tension adaptive control of a UAV winch as described in claim 8, characterized in that, The mechanism for triggering anti-tangle shutdown or emergency cable rewind / unwinding buffer control when cable redundancy or excessive cable tightness is determined includes: When the determination result is that the cable is in a slack state and the duration of the slack state exceeds the preset redundancy tolerance time threshold, the cable is determined to be redundant and there is a risk of entanglement; when the determination result is that the cable has abnormal tension and the duration of the abnormal tension exceeds the preset overtightness tolerance time threshold, the cable is determined to be too tight and there is a risk of jamming; based on the determination result, an anti-entanglement stop signal is generated; the anti-entanglement stop signal is used to cut off the power input of the winch motor and simultaneously control the electromagnetic brake to grip the winch shaft, performing emergency stop protection; When the determination result is cable redundancy, an emergency take-up buffer control signal is generated, which drives the winch motor to accelerate take-up at a preset buffer acceleration. When the determination result is cable too tight, an emergency release buffer control signal is generated, which drives the winch motor to decelerate release at a preset buffer deceleration. Until the real-time tension data recovers to the safe tension threshold range, the take-up and release speed adjustment signal is switched back for control.

10. A drone winch altitude synchronization and tension adaptive control system, used to implement the drone winch altitude synchronization and tension adaptive control method according to any one of claims 1-9, characterized in that, include: The data acquisition module is used to collect data on the drone's flight altitude and the actual length of the cable extended. The deviation calculation module is used to synchronize and align the collected flight altitude data with the actual release length data in time, calculate the difference to generate the original deviation, and then use a limiting filter algorithm to remove outliers and smooth the data to obtain the altitude-length deviation. The speed control generation module is used to input the height-length deviation into the PID controller to calculate the basic speed control quantity. Combined with the real-time cable tension data, the tension influence factor is obtained through the tension-speed correction lookup table to obtain the comprehensive speed control quantity and convert it into a wire take-up and release speed adjustment signal. The anomaly detection module obtains real-time tension data by performing sliding window averaging filtering on the digital tension values ​​collected by the tension sensor, and compares it with the preset safe tension threshold range to determine whether the cable is in a slack state, has abnormal tension, or has a sudden load change. The tension compensation module is used to generate a tension compensation coefficient through a fuzzy logic controller based on the tension anomaly judgment result. When the tension anomaly is judged to be sudden, cable redundancy, or cable too tight, it triggers anti-loosening warning, anti-tangling shutdown, or emergency cable winding buffer control.

Images