Unmanned aerial vehicle cluster formation cooperative control method and system based on intelligent electrical control

Abstract

The invention relates to the technical field of unmanned aerial vehicle control, in particular to an unmanned aerial vehicle cluster formation cooperative control method and system based on intelligent electrical control, and the method comprises the following steps: collecting a bus voltage phase current power factor, obtaining an electrical state vector through Kalman filtering, constructing an electrical consistency constraint, and generating a formation reference state through an ant colony algorithm; and generating a control quantity sequence according to the deviation between the reference state and the real-time state, correcting and driving the formation flight in combination with the current rotating speed, and generating an unmanned aerial vehicle cluster formation cooperative control instruction according to the consistency error updating state. The formation decision is constrained by the power availability, and the current fluctuation and the rotating speed deviation are combined in optimization and feedback correction, so that the thrust and attitude dynamic matching is maintained, the individual difference amplification is inhibited, the consistency error accumulation is reduced, and the formation stability and coordination are enhanced.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and in particular to a method and system for collaborative control of UAV swarm formation based on intelligent electrical control. Background Technology

[0002] The field of UAV control technology involves the control and management of UAV flight attitude, heading, speed, altitude, and mission execution. It covers core aspects such as flight control system design, power and electrical control coordination, communication link organization, cluster operation scheduling, and multi-aircraft collaborative rule formulation. Its overall technical system revolves around the generation of flight control commands, electrical drive execution, flight status perception, and inter-aircraft information interaction to achieve orderly operation of single or multiple UAVs under complex airspace conditions.

[0003] The traditional UAV swarm formation cooperative control method refers to the unified or distributed control of the motor speed, power distribution, control surface deflection, and thrust output of multiple UAVs operating simultaneously, through pre-set formation parameters and cooperative rules. Specifically, based on the position, speed, and heading information collected by each UAV, combined with the established formation geometry, corresponding flight control commands are generated, and the power components are driven through the electrical control loop to complete attitude and position adjustments, thereby maintaining the swarm formation state and completing cooperative behavior.

[0004] Existing UAV swarm formation control relies primarily on kinematic parameters such as position, speed, and heading for decision-making. The control logic focuses on maintaining geometric relationships and rule-based scheduling. The operating status of the electric drive layer is only passively presented as an execution result. Changes in bus power supply, motor phase current imbalances, and load differences are difficult to incorporate into the collaborative control process. Under multi-aircraft heterogeneous operating conditions or external disturbances, thrust distribution bias and attitude response asynchrony are prone to occur. Formation correction relies on fixed parameter adjustments, making it difficult to reflect changes in individual energy efficiency status. This results in limited consistency maintenance capabilities and affects stability and continuous flight reliability. Summary of the Invention

[0005] To address the technical problems of existing UAV swarm formation control, which relies primarily on kinematic parameters such as position, speed, and heading for decision-making, with control logic focused on maintaining geometric relationships and rule-based scheduling, and the electric drive layer's operating state passively presented as an execution result, it is difficult to incorporate changes in bus power supply, motor phase current imbalances, and load differences into the collaborative control process. Under multi-aircraft heterogeneous operating conditions or external disturbances, thrust distribution bias and attitude response asynchrony are prone to occur. Formation correction relies on fixed parameter adjustments, making it difficult to reflect changes in individual energy efficiency, resulting in limited consistency maintenance capabilities and affecting stability and continuous flight reliability. This invention provides a UAV swarm formation collaborative control method based on intelligent electrical control.

[0006] To achieve the above objectives, this invention employs a UAV swarm formation cooperative control method based on intelligent electrical control, comprising the following steps: S1: Collect the bus voltage, motor phase current and power factor of a single drone in the drone swarm, input them into the Kalman filter algorithm for state estimation, and obtain the electrical state vector; S2: Based on the electrical state vector, construct inter-machine electrical consistency constraints, obtain the spatial position information and attitude angle information of the UAV cluster, and input the electrical consistency constraints into the ant colony algorithm for search and optimization to generate the formation reference state; S3: Construct a state deviation evaluation function based on the formation reference state and the real-time state of the UAV, and perform pheromone update and probability selection operations according to the state deviation evaluation function to generate a control quantity sequence; S4: Call the control quantity sequence to obtain the feedback current and feedback speed during the execution of the UAV electric drive, calculate the amplitude of the feedback current fluctuation and the amplitude of the feedback speed deviation, and compare them with the preset stability threshold to form a corrected control quantity sequence; S5: Drive the UAV cluster to perform formation flight based on the corrected control sequence, obtain the formation execution state and calculate the consistency error, use the consistency error as fitness feedback to update the electrical state vector, and generate UAV cluster formation cooperative control commands.

[0007] As a further aspect of the present invention, the electrical state vector includes bus voltage level, motor load state, and power factor characteristic quantity; the formation reference state includes desired relative position relationship, overall attitude reference, and heading consistency target; the control quantity sequence includes thrust distribution quantity, torque adjustment quantity, and attitude control quantity; the correction control quantity sequence includes current fluctuation suppression quantity, speed deviation compensation quantity, and stability margin index; and the UAV swarm formation cooperative control command includes formation maintenance control command, cooperative maneuver control command, and energy coordination control command.

[0008] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Collects the bus voltage, motor phase current and power factor of a single drone in the drone swarm, performs timestamp consistency check on the channel sampled data, performs interpolation correction on the time offset samples based on adjacent sampling points, performs amplitude normalization on the corrected multi-channel data, and generates a synchronous observation sequence of electrical parameters. S102: Based on the synchronous observation sequence of electrical parameters, extract the difference component of the observation vector at adjacent sampling times, perform linear mapping on the difference component and the state component at the previous time, and synchronously perform recursive update on the state covariance matrix. Perform dimensional expansion and structural rearrangement on the mapped state component to obtain the state prior vector. S103: Based on the state prior vector and the synchronous observation sequence of electrical parameters, construct the observation residual vector, calculate and update the gain matrix by combining the prior covariance and the observation noise covariance, perform weighted correction on the residual components and superimpose them onto the prior state to generate the electrical state vector.

[0009] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the electrical state vector, call the UAV state components to perform differential operation, compare the amplitude deviation and phase deviation of the differential components with the electrical consistency threshold respectively, retain the index of the state component that simultaneously satisfies the threshold constraint, and perform consistency mapping to generate the inter-machine electrical consistency constraint matrix. S202: Obtain the spatial position information and attitude angle information of the UAV cluster, perform coordinate transformation on the position information, perform normalization processing on the attitude angle information, and filter and rearrange the pose component index based on the inter-UAV electrical consistency constraint matrix to obtain the formation search state vector. S203: Based on the formation search state vector and the inter-machine electrical consistency constraint matrix, initialize the multi-path state trajectory, perform pheromone weight accumulation and path evaluation calculation based on the ant colony algorithm, jointly discriminate and select state combinations, perform state convergence and solidification, and generate formation reference state.

[0010] As a further aspect of the present invention, the electrical consistency threshold is determined by pre-acquired electrical parameter calibration data of the UAV cluster. The UAV cluster electrical parameter calibration data includes the stable voltage amplitude range and stable phase range of the UAV collected under standard power supply conditions. The electrical consistency threshold is determined by calculating the statistical dispersion of the stable voltage amplitude range and the stable phase range respectively, and setting the boundary value of the corresponding dispersion range as a fixed threshold parameter.

[0011] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Based on the formation reference state and the real-time state of the UAV, align the reference state components and the real-time state components, perform item-by-item mapping according to the state index, calculate the difference and normalize the scale of the mapped components, and generate a state deviation vector. S302: Call the state deviation vector, perform a weighted superposition operation on the deviation amplitude and the weight term according to the corresponding pheromone weight initial term, judge the superposition result and the pheromone update threshold item by item, and perform probability mapping and sequence arrangement on the component index that passes the judgment to obtain the control probability distribution sequence. The pheromone update threshold is determined by performing quantile analysis on the amplitude statistical results of the state deviation vector based on the distribution of state deviation, control convergence speed and stability index of UAV in previous formation flight missions, and selecting the corresponding quantile value. S303: Based on the control probability distribution sequence, perform random sampling and state mapping on the probability components, verify the amplitude range of the mapped state quantity and the real-time state quantity of the UAV, and perform time-series rearrangement and vectorization combination on the verification results to generate a control quantity sequence.

[0012] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Call the control quantity sequence to obtain the feedback current during the electric drive process of the UAV, perform timing alignment on the feedback current sampling sequence according to the control quantity index, perform differential processing on adjacent sampling points and extract the change amplitude to generate a set of feedback current fluctuation amplitudes. S402: Based on the feedback current fluctuation amplitude set, obtain the feedback speed during the electric drive execution process of the UAV, compare the sampled value of the feedback speed with the speed reference item according to the control quantity index, and perform difference to absolute quantity conversion on the comparison result to obtain the feedback speed deviation amplitude set. S403: Based on the feedback speed deviation amplitude set and the feedback current fluctuation amplitude set, compare them item by item with the corresponding preset stable threshold items, mark and aggregate the control quantity index that exceeds the threshold, and perform amplitude replacement and order rearrangement on the control quantity corresponding to the aggregated index to generate a corrected control quantity sequence.

[0013] As a further aspect of the present invention, the speed reference term is a reference speed sample value determined based on the target speed command corresponding to the control quantity in the control quantity sequence before the start of the UAV electric drive execution process. The reference speed sample value is obtained by averaging the feedback speed under the corresponding control quantity index of the UAV electric drive system under the no-load stable operation state through multiple samplings. The preset stability thresholds are determined by sampling and summarizing the feedback current fluctuation amplitude set and the feedback speed deviation amplitude set for multiple cycles and performing interval distribution statistics during continuous operation testing of the UAV electric drive system in the calibration phase. After removing extreme discrete values ​​of a preset proportion, the upper limit of the continuous amplitude interval with the highest proportion is selected as the current stability threshold and the speed stability threshold respectively.

[0014] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Call the corrected control sequence to drive the UAV cluster to perform formation flight, collect UAV pose state and velocity state, align the state data according to the unified time index, perform component difference calculation and vector rearrangement on the aligned pose vector, and generate a formation execution state vector set. S502: Based on the formation execution state vector set, perform structural mapping on the pose vectors according to the preset formation structure rules, construct a reference pose vector arrangement consistent with the formation number, perform vector difference operation and magnitude calculation for each UAV number, and obtain a consistency error sequence. S503: Based on the consistency error sequence, call the electrical state vector of the corresponding UAV, perform weighted update and state mapping based on the error index associated state components, aggregate and sequence the updated state vector, and generate UAV swarm formation cooperative control instructions.

[0015] A collaborative control system for drone swarms based on intelligent electrical control includes: The electrical state estimation module collects the bus voltage, motor phase current and power factor of individual drones in the drone swarm, and inputs them into the Kalman filter algorithm for state estimation to obtain the electrical state vector; The formation constraint modeling module constructs inter-machine electrical consistency constraints based on the electrical state vector, obtains the spatial position information and attitude angle information of the UAV cluster, and inputs them with the electrical consistency constraints into the ant colony algorithm for search and optimization to generate a formation reference state. The control sequence generation module constructs a state deviation evaluation function based on the formation reference state and the real-time state of the UAV, and performs pheromone update and probability selection operations according to the state deviation evaluation function to generate a control quantity sequence. The drive feedback correction module calls the control quantity sequence to obtain the feedback current and feedback speed during the execution of the UAV electric drive, calculates the amplitude of feedback current fluctuation and the amplitude of feedback speed deviation, and compares them with a preset stability threshold to form a correction control quantity sequence. The collaborative instruction update module drives the UAV cluster to perform formation flight based on the corrected control quantity sequence, obtains the formation execution state and calculates the consistency error, uses the consistency error as fitness feedback to update the electrical state vector, and generates UAV cluster formation collaborative control instructions.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by estimating the state of bus voltage, motor phase current, and power factor, quantifiable electrical operating characteristics are introduced into the coordinated control, making formation decisions subject to power availability constraints simultaneously. During the search and optimization process, inter-machine electrical consistency conditions are formed, avoiding reliance solely on spatial geometric relationships. In the generation and execution feedback correction of control sequences, current fluctuations and speed deviations are incorporated into the evaluation and update criteria, ensuring dynamic matching between thrust output and attitude adjustment. This suppresses the amplification of individual performance differences, reduces the accumulation of consistency errors, and enhances formation stability and operational coordination. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the accompanying drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention; Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0020] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0021] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0022] In this embodiment of the invention, sometimes the subscript such as W1 is written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0023] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0024] Please see Figure 1 This invention provides a method for cooperative control of unmanned aerial vehicle (UAV) swarm formation based on intelligent electrical control, comprising the following steps: S1: Collect the bus voltage, motor phase current and power factor of a single drone in the drone swarm, input them into the Kalman filter algorithm for state estimation, and obtain the electrical state vector; S2: Construct inter-machine electrical consistency constraints based on electrical state vectors, obtain spatial position information and attitude angle information of the UAV cluster, and input the electrical consistency constraints into the ant colony algorithm for search and optimization to generate formation reference state; S3: Construct a state deviation evaluation function based on the formation reference state and the real-time state variables of the UAV, and perform pheromone update and probability selection operations according to the state deviation evaluation function to generate a control variable sequence; S4: Call the control quantity sequence to obtain the feedback current and feedback speed during the execution of the UAV's electric drive, calculate the amplitude of the feedback current fluctuation and the amplitude of the feedback speed deviation, and compare them with the preset stability threshold to form a corrected control quantity sequence; S5: Drive the UAV swarm to perform formation flight based on the modified control quantity sequence, obtain the formation execution state and calculate the consistency error, use the consistency error as fitness feedback to update the electrical state vector, and generate UAV swarm formation cooperative control commands.

[0025] The electrical state vector includes bus voltage level, motor load state, and power factor characteristics. The formation reference state includes desired relative positional relationship, overall attitude reference, and heading consistency target. The control quantity sequence includes thrust distribution, torque regulation, and attitude control. The correction control quantity sequence includes current fluctuation suppression, speed deviation compensation, and stability margin index. The UAV swarm formation cooperative control commands include formation maintenance control commands, cooperative maneuver control commands, and energy coordination control commands.

[0026] Please see Figure 2 The specific steps of S1 are as follows: S101: Collects the bus voltage, motor phase current and power factor of a single drone in the drone swarm, performs timestamp consistency check on the channel sampled data, performs interpolation correction on the time offset samples based on adjacent sampling points, performs amplitude normalization on the corrected multi-channel data, and generates a synchronous observation sequence of electrical parameters. Collect bus voltage, motor phase current, and power factor of individual drones in a drone swarm, targeting drones numbered... The single unit utilizes a Hall voltage sensor and a shunt resistor connected in parallel to the onboard power management module to... Real-time capture of DC bus voltage using high-frequency sampling rate Motor phase current at the output of the three-phase inverter and the power factor calculated based on the phase difference. Transmit analog signals via The analog-to-digital converter with bit-precision conversion is used to convert the data into a digital sequence, while simultaneously reading the nanosecond-level timestamp generated by the onboard GPS / RTK module. The above electrical parameters and their corresponding timestamps are packaged and stored in a temporary buffer. The time base comparison program of the cluster central controller is started, and the time synchronization tolerance threshold is set. for The threshold was set with reference to the average latency test value of the cluster communication link. And reserved The jitter margin is calculated by individually measuring the timestamps of the sampled data packets from a single machine. With cluster reference clock deviation value Determine the time deviation at a certain sampling point Exceeding interval At that time, the sample is marked as a time-off sample, and the data of the two adjacent valid sampling points before and after the offset sample are retrieved. and ,in Representing voltage or current amplitude, perform linear interpolation operations, for example, for the offset time. The voltage correction calculation follows the formula ,like , The time interval is The offset target point is located in the middle. At this point, the correction value is calculated. After offset point correction, amplitude normalization is performed on the multi-channel data, and the voltage reference value is set. Current reference value Power factor reference value The normalized values ​​were calculated using the ratio method. If the measured voltage is The normalization result is The processed single-channel data are aligned and spliced ​​according to a unified time axis sequence to generate a synchronous observation sequence of electrical parameters.

[0027] S102: Based on the synchronous observation sequence of electrical parameters, extract the difference component of the observation vector between adjacent sampling times, perform linear mapping between the difference component and the state component of the previous time, and simultaneously perform recursive update on the state covariance matrix. Perform dimensional expansion and structural rearrangement on the mapped state components to obtain the state prior vector. Extract the current moment from the synchronous observation sequence. observation vector Compared to the previous moment observation vector ,in For inclusion Voltage, current, and power factor of the drone 1D vector, perform vector subtraction operation To obtain the differential characteristics of the system's dynamic changes, a state transition matrix is ​​set. diagonal elements are The diagonal matrix is ​​used to simulate the inertial attenuation characteristics of electrical parameters, and the input mapping matrix is ​​set. The coefficient is The gain matrix, calling the posterior state estimation vector from the previous time step. Perform linear mapping operation Assume that the voltage state component of a certain drone at the previous moment is The current observation difference is Then the prior state components are calculated as follows: Synchronous execution state covariance matrix The recursive update sets the process noise covariance matrix. The diagonal element value is A diagonal matrix, performing matrix operations. If the covariance of the previous time step Corresponding element is The updated prior covariance element value is Then, the prior state matrix obtained by mapping is expanded in dimension, transforming the original matrix structure stacked according to drone number ( Remodeled into a column vector structure The state prior vector is obtained by indexing and rearranging the sub-blocks in the order of voltage, current, and power factor.

[0028] S103: Based on the state prior vector and the synchronous observation sequence of electrical parameters, construct the observation residual vector, calculate and update the gain matrix by combining the prior covariance and the observation noise covariance, perform weighted correction on the residual components and superimpose them on the prior state to generate the electrical state vector. Define observation matrix identity matrix Read the measured normalized observation vector at the current moment. With the output state prior vector Perform subtraction operation Obtain the residual components and define the observation noise covariance matrix. Its diagonal element is set according to the sensor accuracy manual. (correspond (measurement error variance), combined with the prior covariance matrix Perform Kalman gain calculation Substituting the data from the previous example, Element is Then the scalar part of the gain is approximately calculated as follows: This gain coefficient indicates a degree of confidence in the measured value of approximately [percentage missing]. Then, weighted corrections are applied to the residual components and superimposed onto the prior state, and the operation is performed. If the observed value Corresponding components are Prior state for Then the residual The correction amount is The final state was updated to The specific electrical state parameter update process is shown in Table 1. By synchronously correcting the three dimensions of voltage, current and power factor, the optimal state estimation of all dimensions is completed, and the electrical state vector is generated.

[0029] Table 1: Example Data Table of UAV Electrical State Vector Update

[0030] Table 1 lists the drones for a specific sampling time. The update process data for the three core electrical parameters are as follows: the prior state value is derived from the state extrapolation of the previous moment; the measured observation value comes from the sensor data after normalization; and the final state value is the weighted fusion result of the two, taking into account the proportion of system noise and measurement noise (i.e., gain K). Numerical results show that the final state value... Compared to prior values To the observed values A significant shift occurred, reflecting the corrective effect of the observation data on the state estimation.

[0031] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the electrical state vector, call the UAV state components to perform differential operations, compare the amplitude deviation and phase deviation of the differential components with the electrical consistency threshold respectively, retain the index of the state component that simultaneously meets the threshold constraint, and perform consistency mapping to generate the inter-machine electrical consistency constraint matrix. First, extract the electrical state vectors of every two drones in the cluster. and At the present moment Electrical state components, including normalized bus voltage , and voltage phase , Perform differential operation to obtain voltage amplitude deviation Phase deviation Assuming The normalized voltage is Phase is , The normalized voltage is Phase is Then the calculation yields , Then, it retrieves pre-stored electrical parameter calibration data for the drone swarm, which is based on standard... In a DC regulated power supply environment Using the same type of drone The statistical set obtained from the second power-on test is shown in Table 2. This data was analyzed by examining the voltage amplitude stability range in the calibration data. With phase stable interval Perform statistical analysis and calculate the sample standard deviation. and Set confidence coefficient To cover Calculate the voltage consistency threshold within the normal fluctuation range. Phase Consistency Threshold If the normalized voltage standard deviation is obtained statistically... Phase standard deviation Then the threshold parameters are determined as follows: , The calculated deviation values ​​are compared item by item with the threshold. The judgment logic is as follows: if and If both conditions are met, then the judgment is made. and Possessing electrical isomorphism, preserving its index relationship and marking it as Conversely, it is marked as In this example and Therefore, the decision is passed, and the cluster is traversed. Each pair of combinations is constructed with dimension 1. The binary adjacency matrix, the matrix in which the first... Line 1 Column elements That is, in accordance with the above judgment result, the diagonal element is forcibly set to Generate the inter-machine electrical consistency constraint matrix.

[0032] Table 2: Examples of Electrical Consistency Threshold Calibration and Calculation for Unmanned Aerial Vehicle Swarms

[0033] Table 2 lists the statistical results of data collected under standard laboratory conditions. The "Calculation Threshold" column contains key criteria used in the algorithm steps to determine electrical consistency. For example, the threshold for normalized bus voltage is set as follows: pu means that if the voltage deviation between any two machines exceeds this value, then the electrical characteristics of the two machines are considered to be too different, and they are not suitable for close formation and coordination.

[0034] S202: Obtain the spatial position information and attitude angle information of the UAV cluster, perform coordinate transformation on the position information, perform normalization processing on the attitude angle information, filter and rearrange the pose component index based on the inter-machine electrical consistency constraint matrix, and obtain the formation search state vector. Using an airborne RTK-GNSS positioning module Frequency reading of longitude in geodetic coordinate system ,latitude and altitude Simultaneously, the roll angle of the body axis system is read through the MEMS inertial measurement unit. Pitch angle and yaw angle ,For example The collected data is , , , Perform a transformation from the geographic coordinate system to the local tangent plane Cartesian coordinate system, selecting the formation's geometric center as the origin. Calculate relative coordinates using the Mercator projection transformation formula , ,in Take the average radius of the Earth If the difference in longitude is The difference in latitude is Then, the plane coordinates are obtained. Approximately Vertical coordinates The attitude angle information is normalized, and the angle normalization reference is set as follows: radians, calculation , , If the measured yaw angle is The normalized value is Then, the generated inter-machine electrical consistency constraint matrix is ​​invoked. ,by For index rows, retrieve The Middle row value column index Filter out those that are related to Set of neighboring nodes with consistent electrical characteristics Extract only the set The normalized position vector corresponding to the UAV With attitude vector The selected pose components are rearranged according to the rule of distance from nearest to farthest, and a local augmented vector containing the state of the current user and the states of its effective neighbors is constructed. This yields the formation search state vector.

[0035] S203: Based on the formation search state vector and the inter-machine electrical consistency constraint matrix, initialize the multi-path state trajectory, perform pheromone weight accumulation and path evaluation calculation based on the ant colony algorithm, jointly discriminate and select state combinations, perform state convergence and solidification, and generate formation reference state; Spread in the formation configuration solution space Artificial ants, with an initial pheromone concentration matrix set. The elements are Define the set of formation state nodes. Based on the ant colony algorithm, iterative search is performed, and at the 1st... In this iteration, the ant starts from the current state node. Transition to the next state node The probability follows the formula: ; Setting pheromone heuristic factors Expected heuristic factor Heuristic function Defined as an element of the electrical consistency constraint matrix Reciprocal of the Euclidean distance between nodes The product of, if (Electrical inconsistency), then The probability of forced transfer is ,like And distance ,but Substituting into the calculation of the numerator of the transition probability, we get... After constructing the ant's path, calculate the comprehensive evaluation score for each path. ,in For formation error, Let be the electrical dispersion of the nodes on the path. For the weighting coefficients, select The path with the smallest value is selected as the optimal solution for this iteration, and a global update is performed on the pheromones along that path. Set the volatility coefficient Increment ,set up , ,but After the update, the pheromone concentration increased, after After the iteration converges, the node coordinate sequence and attitude sequence corresponding to the optimal path are extracted, solidified as target reference values, and the formation reference state is generated.

[0036] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the formation reference state and the real-time state of the UAV, align the reference state components with the real-time state components, perform item-by-item mapping according to the state index, calculate the difference and scale normalize the mapped components, and generate a state deviation vector. First, the current control cycle is extracted from the output formation reference state. The corresponding target status instruction, which contains the first... The three-dimensional spatial coordinates of the drone and three-axis attitude angles Simultaneously, the real-time status observation values ​​are read through the airborne sensor fusion module. Perform timestamp alignment and check the reference status timestamp. With real-time measurement timestamp The difference, if the difference is less than If a direct match is used, the reference state is advanced to the measurement time using linear extrapolation. Based on the physical meaning index of the state components (e.g., longitude corresponds to longitude, elevation corresponds to elevation), item-by-item difference calculation is performed to construct the original error vector. Assuming the target height Actual height The height deviation is If the target yaw angle The actual measurement is Then the angle deviation is Subsequently, the difference components are scaled and normalized by introducing a preset state normalization scale vector. Among them, the location scale factor Set as the maximum allowable formation position error Attitude scale factor Set to maximum attitude maneuver range Perform element-wise division. For the aforementioned height deviation, the normalized value is calculated as follows: For yaw deviation, the normalized value is calculated as follows: After completing the normalization calculations for all 6 degrees of freedom, the components are then... The sequential encapsulation generates a state deviation vector.

[0037] S302: Call the state deviation vector, perform a weighted superposition operation on the deviation amplitude and the weight term according to the corresponding pheromone weight initial term, judge the superposition result and the pheromone update threshold item by item, and perform probability mapping and sequence arrangement on the component index that passes the judgment to obtain the control probability distribution sequence. Read the pheromone concentration values ​​left on the path nodes after the ant colony algorithm converges. Use this as the initial weight term and set the deviation influence factor. With pheromone guiding factors For each component in the state deviation vector Calculate the comprehensive discrimination index Assume the normalized deviation modulus of a certain component is pheromone concentration at the corresponding path point The calculated index value is At the same time, determine the pheromone update threshold. The threshold was set based on the raw data statistics of previous UAV formation flight missions. The "UAV Raw Flight State Deviation Statistics Table" shown in Table 3 was retrieved, and the "State Deviation Statistical Quantile Value" column was analyzed. The threshold was selected from the... Using quantile values ​​as a basic reference, and considering the requirements for controlling convergence speed, adjustment coefficients are set. Calculate the threshold Based on the data in Table 3, if the "cruise mode" is selected... quantile Then the threshold is calculated as follows: Then, the superposition results With threshold Perform item-by-item judgment, the judgment logic is: if If the deviation of a component is significant or the path guidance requirement is strong, the component index is retained; otherwise, it is set to zero or removed. For the aforementioned calculated values... If the threshold is If the criteria are met, a probability mapping is performed on the component indices that pass the criteria, and the Softmax function is used to retain the index values. Convert to control response probability The components are then arranged sequentially according to their probability values ​​from largest to smallest to obtain the control probability distribution sequence.

[0038] Table 3: Statistics of Initial Flight Status Deviation of UAVs

[0039] As shown in Table 3, this table records the raw state deviation statistics of the UAV under different flight modes. The "85th percentile" column not only reflects the deviation boundary of most normal flight under this mode, but also serves as the dynamic threshold set in this embodiment. The key numerical basis, for example, in cruise mode, the threshold calculation directly references... This is derived from the statistical results.

[0040] S303: Based on the control probability distribution sequence, random sampling and state mapping are performed on the probability components. The mapped state variables and the real-time state variables of the UAV are verified by amplitude interval. The verification results are rearranged in time and combined into vectorized form to generate a control variable sequence. The roulette wheel selection algorithm is used to sample the probability distribution sequence and generate intervals. Uniformly distributed random numbers Accumulate the probability values ​​in the sequence until the sum is greater than 1. Select the corresponding state control channel index (such as "Throttle Channel" or "Pitch Channel"), perform a state mapping operation, and convert the selected abstract probability component into a specific physical control increment. Set control mapping gain If the normalized height deviation is selected Then calculate the initial throttle increment. The calculated mapped state variables are compared with the real-time state variables of the UAV using amplitude range verification, and the current PWM control signal value of the motor is read. Calculate the target control quantity Compare the allowable physical operating range of the motor ,judge If true, retain the value; otherwise, if the calculation result is... If this triggers the limiting logic, it will be forcibly corrected to... The timing of the verified channel control variables is rearranged, and the data packets are reorganized according to the channel order specified in the flight control protocol (such as Aileron, Elevator, Throttle, Rudder), and the multidimensional control variables are combined into a column vector form. Generate a control quantity sequence.

[0041] Please see Figure 5 The specific steps of S4 are as follows: S401: Call the control quantity sequence to obtain the feedback current during the execution of the UAV electric drive, perform timing alignment on the feedback current sampling sequence according to the control quantity index, perform differential processing on adjacent sampling points and extract the change amplitude, and generate a set of feedback current fluctuation amplitudes. First, parse the generated control quantity sequence. Extract the four motor throttle commands and their corresponding timestamps contained therein. The onboard power management unit's current monitoring channel is activated to read the DC bus input current of the power control branch in real time at a sampling frequency of 2kHz. The timing alignment operation is performed, and considering the inherent physical delay between the ESC response and sensor data acquisition and transmission, an alignment delay compensation time is set. The time is 20ms, and the timestamp is retrieved in the feedback current sampling buffer. If no exact matching time point exists in the buffer, then the two closest sampling points before and after that time point are selected. and Perform linear interpolation. For example, if the value at the previous sampling point is 12.50A at 100.0ms and the value at the next sampling point is 12.55A at 100.5ms, and the target time is 100.2ms, then calculate the alignment current value. A. Subsequently, differential processing is performed on the aligned current values ​​of adjacent sampling periods, and the current record value of the previous control period is read. Perform subtraction operation The absolute value of the change is taken to obtain the amplitude. Assuming the current was 12.10A at the previous moment and is 12.52A at the current moment, the current fluctuation amplitude is... A. Repeat this process for the motor channel. If an abnormally low current fluctuation amplitude is detected in a certain channel, such as 0.01A, it is necessary to check whether it is in the motor stop zone. Finally, the current fluctuation values ​​calculated by the channel are encapsulated in the order of motor number to generate a set of feedback current fluctuation amplitude values.

[0042] S402: Based on the set of feedback current fluctuation amplitude, obtain the feedback speed during the electric drive execution process of the UAV, compare the sampled value of the feedback speed with the speed reference item according to the control quantity index, perform difference and absolute quantity conversion on the comparison result, and obtain the set of feedback speed deviation amplitude. The real-time commutation frequency of the motor is directly read through the bidirectional DShot communication protocol of the electronic speed controller (ESC) and converted into a physical speed value. Simultaneously, based on the throttle percentage in the current control quantity sequence Calculation of speed reference term The pre-calibrated "throttle-speed" response characteristic curve is invoked. This curve is obtained by fitting the motor with progressively increasing throttle from 10% to 90% on a stationary test bench. The fitting function is set to a quadratic polynomial. For example, if the current control command is 60% throttle, then the base speed calculated is... RPM, the measured feedback speed is 5750 RPM, the two are compared item by item according to the control quantity index, and the difference calculation is performed. The algebraic deviation was calculated to be RPM, then performs absolute transformation on it. If the measured speed of another channel is 6100 RPM, then the absolute value of the deviation is 200 RPM. This step aims to quantify the mechanical response error of the actuator to the control command and to collect the absolute values ​​of the speed deviation of the channels to obtain the feedback speed deviation amplitude set.

[0043] S403: Based on the feedback speed deviation amplitude set and the feedback current fluctuation amplitude set, compare each item with the corresponding preset stable threshold item, mark and aggregate the control quantity index that exceeds the threshold, and perform amplitude replacement and order rearrangement on the control quantity corresponding to the aggregated index to generate a corrected control quantity sequence. First, set the current fluctuation threshold. With respect to speed deviation threshold These two thresholds are set based on durability test data of the drone's power system, among which... It is set to 10% of the rated operating current, i.e., 2.5A, to tolerate normal dynamic acceleration and deceleration current surges. The speed is set to 5% of the base speed, i.e., 295 RPM (taking 5900 RPM as an example), to determine whether the motor is stalled or has severe slip. A step-by-step comparison logic is executed. For motor channel 1, if the current fluctuation is 0.42A (less than 2.5A) and the speed deviation is 150 RPM (less than 295 RPM), the channel is considered normal, and the original control value is retained. For motor channel 2, if the current fluctuation reaches 3.0A or the speed deviation reaches 400 RPM, it is considered to have exceeded the threshold. The channel index is then marked and aggregated. For the marked channels, the control value amplitude is replaced using a damping correction strategy, and the correction amount is calculated. Set the damping coefficient Assuming the original control quantity is 60% (0.60), the speed deviation ratio is... The corrected control quantity is That is, 56.6%, in order to reduce the impact of aggressive control commands on the power system. The specific verification and correction data are shown in Table 4. After completing the traversal correction of the channels, the channel data are rearranged in order according to the original control protocol format (such as PWM pulse width sequence) to generate the corrected control quantity sequence.

[0044] Table 4: Examples of UAV Electric Drive Feedback Verification and Correction

[0045] Table 4 lists the feedback verification details of a quadcopter UAV during one control cycle, including the motor... Because the feedback current fluctuation of 3.10A exceeded the preset threshold of 2.50A, and the speed deviation of 450RPM also exceeded the allowable range of 295RPM, the protective correction mechanism was triggered. The control quantity was automatically reduced from the original 60.0% to 56.2% to suppress the risk of overcurrent and restore speed tracking stability. The other three motor channels maintained the original control command output because multiple indicators were within the safety threshold.

[0046] Please see Figure 6 The specific steps of S5 are as follows: S501: Call the corrected control sequence to drive the UAV cluster to perform formation flight, collect UAV pose and velocity state variables, align the state data according to the unified time index, perform component difference calculation and vector rearrangement on the aligned pose vector, and generate a formation execution state vector set. First, a data packet containing multiple individual drone throttle and attitude control commands, generated through a cluster communication network, is sent to drive the onboard flight control system to control motor speed and control surface deflection. Simultaneously, the onboard RTK-GPS and IMU integrated navigation module collects the three-dimensional position coordinates of multiple drones in real time at a frequency of 50Hz. Three-dimensional velocity vector and three-axis attitude angles Set a unified time synchronization index Check the timestamp of the data frame ,like Then, the state variable from the previous time step is used. State quantity at the current moment Linear interpolation alignment is performed to ensure that the UAV state data are in the same time slice. Then, component difference calculation is performed on the aligned pose vector to calculate the increment of the state at the current time step compared with the state at the previous sampling time step. To verify the continuity of data updates, if the position of a component changes by more than [a certain amount of time]... If the preset maximum physical displacement limit for a single step is not met, the data point is determined to be abnormal and replaced with a Kalman filter prediction value. Finally, the result is determined according to the UAV fuselage ID number. to The verified position, velocity, and attitude data are sequentially concatenated in the following order to construct a dimension of [missing information]. The set of execution state vectors of the formation.

[0047] S502: Based on the formation execution state vector set, perform structural mapping on the pose vector according to the preset formation structure rules, construct a reference pose vector arrangement consistent with the formation number, perform vector difference operation and magnitude calculation for each UAV number, and obtain a consistency error sequence. Retrieve the pre-stored "wedge" formation configuration definition file, which specifies the formation with the navigator as the leader. The ideal relative position offset of the wingman in the local coordinate system of the origin. ,For example The ideal offset is , The ideal offset is Extract the navigator from the execution state vector set. Real-time location coordinates Perform structure mapping operations to construct wingmen. Reference pose vector Simultaneously extract actual position coordinates Perform vector interpolation calculations on each UAV number to calculate the position error vector. And perform modulus calculation to obtain the consistency error value. Repeat the above calculation process for the nodes in the cluster, compile the calculated error magnitude values ​​into a column as shown in Table 5, and generate a consistency error sequence.

[0048] Table 5: Correlation between Unmanned Aerial Vehicle (UAV) Swarm Formation Consistency Error and Electrical Status

[0049] As shown in Table 5, the position tracking status and associated parameters of some UAVs in the cluster during the current control cycle are listed in detail. Among them, the "position error magnitude" directly reflects the accuracy of maintaining the formation geometry, while the "electrical stability factor" is derived from the electrical condition assessment and is used to adjust the weight of control corrections in subsequent steps to prevent excessive position corrections when the electrical system is unstable.

[0050] S503: Based on the consistency error sequence, call the electrical state vector of the corresponding UAV, perform weighted update and state mapping based on the error index associated state components, aggregate and sequence the updated state vector, and generate UAV swarm formation cooperative control instructions; First, read the generated electrical condition assessment results and obtain the voltage stability score for each drone as the electrical stability factor. ,For example Due to slightly large voltage fluctuations, it was assigned Set the base position feedback gain Calculate the overall corrected weights ,for Calculated The target correction amount is calculated by using this weight to perform weighted attenuation processing on the error terms in the consistency error sequence. This correction represents the additional velocity vector magnitude required to eliminate position errors. A state mapping process is then performed, decomposing and projecting this velocity correction onto the longitudinal and lateral channels of the body coordinate system to generate the corresponding pitch angle increments. With roll angle increment ,like It needs to be corrected to the northeast, then it is mapped as follows: , The attitude correction increments calculated by the UAV are aggregated and superimposed with the original navigation commands, and then sequentially converted and packaged according to the data frame format (frame header + ID + attitude data + check bit) specified by the communication protocol to generate the final UAV swarm formation cooperative control commands.

[0051] Please see Figure 7 A collaborative control system for drone swarms based on intelligent electrical control includes: The electrical state estimation module collects the bus voltage, motor phase current and power factor of individual drones in the drone swarm, and inputs them into the Kalman filter algorithm for state estimation to obtain the electrical state vector; The formation constraint modeling module constructs inter-machine electrical consistency constraints based on electrical state vectors, obtains the spatial position and attitude angle information of the UAV cluster, and inputs the electrical consistency constraints into the ant colony algorithm for search and optimization to generate the formation reference state. The control sequence generation module constructs a state deviation evaluation function based on the formation reference state and the real-time state variables of the UAV, and performs pheromone update and probability selection operations according to the state deviation evaluation function to generate a control quantity sequence. The drive feedback correction module calls the control quantity sequence to obtain the feedback current and feedback speed during the execution of the UAV's electric drive, calculates the amplitude of the feedback current fluctuation and the amplitude of the feedback speed deviation, and compares them with the preset stability threshold to form a correction control quantity sequence. The collaborative instruction update module drives the UAV swarm to perform formation flight based on the corrected control quantity sequence, obtains the formation execution status and calculates the consistency error, uses the consistency error as fitness feedback to update the electrical state vector, and generates UAV swarm formation collaborative control instructions.

[0052] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for cooperative control of UAV swarm formation based on intelligent electrical control, characterized in that, Includes the following steps: S1: Collect the bus voltage, motor phase current and power factor of a single drone in the drone swarm, input them into the Kalman filter algorithm for state estimation, and obtain the electrical state vector; S2: Based on the electrical state vector, construct inter-machine electrical consistency constraints, obtain the spatial position information and attitude angle information of the UAV cluster, and input the electrical consistency constraints into the ant colony algorithm for search and optimization to generate the formation reference state; S3: Construct a state deviation evaluation function based on the formation reference state and the real-time state of the UAV, and perform pheromone update and probability selection operations according to the state deviation evaluation function to generate a control quantity sequence; S4: Call the control quantity sequence to obtain the feedback current and feedback speed during the execution of the UAV electric drive, calculate the amplitude of the feedback current fluctuation and the amplitude of the feedback speed deviation, and compare them with the preset stability threshold to form a corrected control quantity sequence; S5: Drive the UAV cluster to perform formation flight based on the corrected control sequence, obtain the formation execution state and calculate the consistency error, use the consistency error as fitness feedback to update the electrical state vector, and generate UAV cluster formation cooperative control commands.

2. The method for cooperative control of UAV swarm formation based on intelligent electrical control according to claim 1, characterized in that, The electrical state vector includes bus voltage level, motor load state, and power factor characteristics. The formation reference state includes desired relative positional relationship, overall attitude reference, and heading consistency target. The control quantity sequence includes thrust distribution, torque regulation, and attitude control. The correction control quantity sequence includes current fluctuation suppression, speed deviation compensation, and stability margin index. The UAV swarm formation cooperative control commands include formation maintenance control commands, cooperative maneuver control commands, and energy coordination control commands.

3. The method for cooperative control of UAV swarm formation based on intelligent electrical control according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Collects the bus voltage, motor phase current and power factor of a single drone in the drone swarm, performs timestamp consistency check on the channel sampled data, performs interpolation correction on the time offset samples based on adjacent sampling points, performs amplitude normalization on the corrected multi-channel data, and generates a synchronous observation sequence of electrical parameters. S102: Based on the synchronous observation sequence of electrical parameters, extract the difference component of the observation vector at adjacent sampling times, perform linear mapping on the difference component and the state component at the previous time, and synchronously perform recursive update on the state covariance matrix. Perform dimensional expansion and structural rearrangement on the mapped state component to obtain the state prior vector. S103: Based on the state prior vector and the synchronous observation sequence of electrical parameters, construct the observation residual vector, calculate and update the gain matrix by combining the prior covariance and the observation noise covariance, perform weighted correction on the residual components and superimpose them onto the prior state to generate the electrical state vector.

4. The UAV swarm formation cooperative control method based on intelligent electrical control according to claim 3, characterized in that, The specific steps of S2 are as follows: S201: Based on the electrical state vector, call the UAV state components to perform differential operation, compare the amplitude deviation and phase deviation of the differential components with the electrical consistency threshold respectively, retain the index of the state component that simultaneously satisfies the threshold constraint, and perform consistency mapping to generate the inter-machine electrical consistency constraint matrix. S202: Obtain the spatial position information and attitude angle information of the UAV cluster, perform coordinate transformation on the position information, perform normalization processing on the attitude angle information, and filter and rearrange the pose component index based on the inter-UAV electrical consistency constraint matrix to obtain the formation search state vector. S203: Based on the formation search state vector and the inter-machine electrical consistency constraint matrix, initialize the multi-path state trajectory, perform pheromone weight accumulation and path evaluation calculation based on the ant colony algorithm, jointly discriminate and select state combinations, perform state convergence and solidification, and generate formation reference state.

5. The UAV swarm formation cooperative control method based on intelligent electrical control according to claim 4, characterized in that, The electrical consistency threshold is determined by the pre-acquired electrical parameter calibration data of the UAV cluster. The UAV cluster electrical parameter calibration data includes the stable voltage amplitude range and stable phase range of the UAV collected under standard power supply conditions. The electrical consistency threshold is determined by calculating the statistical dispersion of the stable voltage amplitude range and the stable phase range respectively, and setting the boundary value of the corresponding dispersion range as a fixed threshold parameter.

6. The UAV swarm formation cooperative control method based on intelligent electrical control according to claim 4, characterized in that, The specific steps for S3 are as follows: S301: Based on the formation reference state and the real-time state of the UAV, align the reference state components and the real-time state components, perform item-by-item mapping according to the state index, calculate the difference and normalize the scale of the mapped components, and generate a state deviation vector. S302: Call the state deviation vector, perform a weighted superposition operation on the deviation amplitude and the weight term according to the corresponding pheromone weight initial term, judge the superposition result and the pheromone update threshold item by item, and perform probability mapping and sequence arrangement on the component index that passes the judgment to obtain the control probability distribution sequence. S303: Based on the control probability distribution sequence, perform random sampling and state mapping on the probability components, verify the amplitude range of the mapped state quantity and the real-time state quantity of the UAV, and perform time-series rearrangement and vectorization combination on the verification results to generate a control quantity sequence.

7. The method for cooperative control of UAV swarm formation based on intelligent electrical control according to claim 6, characterized in that, The specific steps of S4 are as follows: S401: Call the control quantity sequence to obtain the feedback current during the electric drive process of the UAV, perform timing alignment on the feedback current sampling sequence according to the control quantity index, perform differential processing on adjacent sampling points and extract the change amplitude to generate a set of feedback current fluctuation amplitudes. S402: Based on the feedback current fluctuation amplitude set, obtain the feedback speed during the electric drive execution process of the UAV, compare the sampled value of the feedback speed with the speed reference item according to the control quantity index, and perform difference to absolute quantity conversion on the comparison result to obtain the feedback speed deviation amplitude set. S403: Based on the feedback speed deviation amplitude set and the feedback current fluctuation amplitude set, compare them item by item with the corresponding preset stable threshold items, mark and aggregate the control quantity index that exceeds the threshold, and perform amplitude replacement and order rearrangement on the control quantity corresponding to the aggregated index to generate a corrected control quantity sequence.

8. The method for cooperative control of UAV swarm formation based on intelligent electrical control according to claim 7, characterized in that, The speed reference term is a reference speed sample value determined based on the target speed command corresponding to the control quantity in the control quantity sequence before the start of the UAV electric drive execution process. The reference speed sample value is obtained by averaging the feedback speed under the corresponding control quantity index of the UAV electric drive system under the no-load stable operation state through multiple samplings. The preset stability thresholds are determined by sampling and summarizing the feedback current fluctuation amplitude set and the feedback speed deviation amplitude set for multiple cycles and performing interval distribution statistics during continuous operation testing of the UAV electric drive system in the calibration phase. After removing extreme discrete values ​​of a preset proportion, the upper limit of the continuous amplitude interval with the highest proportion is selected as the current stability threshold and the speed stability threshold respectively.

9. The method for cooperative control of UAV swarm formation based on intelligent electrical control according to claim 7, characterized in that, The specific steps of S5 are as follows: S501: Call the corrected control sequence to drive the UAV cluster to perform formation flight, collect UAV pose state and velocity state, align the state data according to the unified time index, perform component difference calculation and vector rearrangement on the aligned pose vector, and generate a formation execution state vector set. S502: Based on the formation execution state vector set, perform structural mapping on the pose vectors according to the preset formation structure rules, construct a reference pose vector arrangement consistent with the formation number, perform vector difference operation and magnitude calculation for each UAV number, and obtain a consistency error sequence. S503: Based on the consistency error sequence, call the electrical state vector of the corresponding UAV, perform weighted update and state mapping based on the error index associated state components, aggregate and sequence the updated state vector, and generate UAV swarm formation cooperative control instructions.

10. A collaborative control system for unmanned aerial vehicle (UAV) swarms based on intelligent electrical control, characterized in that: The system is used to implement the UAV swarm formation cooperative control method based on intelligent electrical control as described in any one of claims 1-9, and the system includes: The electrical state estimation module collects the bus voltage, motor phase current and power factor of individual drones in the drone swarm, and inputs them into the Kalman filter algorithm for state estimation to obtain the electrical state vector; The formation constraint modeling module constructs inter-machine electrical consistency constraints based on the electrical state vector, obtains the spatial position information and attitude angle information of the UAV cluster, and inputs them with the electrical consistency constraints into the ant colony algorithm for search and optimization to generate a formation reference state. The control sequence generation module constructs a state deviation evaluation function based on the formation reference state and the real-time state of the UAV, and performs pheromone update and probability selection operations according to the state deviation evaluation function to generate a control quantity sequence. The drive feedback correction module calls the control quantity sequence to obtain the feedback current and feedback speed during the execution of the UAV electric drive, calculates the amplitude of feedback current fluctuation and the amplitude of feedback speed deviation, and compares them with a preset stability threshold to form a correction control quantity sequence. The collaborative instruction update module drives the UAV cluster to perform formation flight based on the corrected control quantity sequence, obtains the formation execution state and calculates the consistency error, uses the consistency error as fitness feedback to update the electrical state vector, and generates UAV cluster formation collaborative control instructions.

Images