Intelligent path planning method and system for fire extinguishing aircraft

Abstract

The application relates to the technical field of path planning, and particularly discloses an intelligent path planning method and system for a fire extinguishing aircraft. The method comprises the following steps: collecting aircraft state data, extracting a load-thrust mutation feature at a load release moment, and constructing a transient dynamic reachable domain; obtaining task point coordinates, performing vector homotopy analysis on the reachable domain in combination with a target heading, calculating a cooperative gain coefficient to determine whether to assist in approaching the target; if the condition is met, a dynamic attitude preset trajectory is generated; the preset trajectory is mapped to the dynamic reachable domain for compatibility analysis under time-varying constraints, an execution saturation risk curve is calculated, and adaptive boundary contraction optimization is performed on the trajectory. The system comprises a reachable analysis module, a vector analysis module, a trajectory planning module and an intelligent optimization module, and is used for realizing the above method. The application is beneficial to improving the maneuvering efficiency of fire extinguishing operation under the premise of flight safety.

Description

Technical Field

[0001] This invention relates to the field of path planning technology, specifically to an intelligent path planning method and system for firefighting aircraft. Background Technology

[0002] When firefighting aircraft perform firefighting and rescue operations, they typically employ a pinpoint pulse release of extinguishing agents, jettisoning a large payload from the fuselage in a very short time. At the moment of payload release, the aircraft's total mass decreases abruptly. Due to inertial response delays, the electronic speed controller and propellers cannot immediately reduce their output power, causing a sudden surge in the thrust-to-weight ratio and an upward ejection tendency. Simultaneously, because the liquid tanks are usually located under the fuselage or at a non-center of gravity location, the release of the payload causes a sudden shift in the aircraft's center of gravity, generating violent and uncontrolled induced torques.

[0003] However, existing flight control and path planning technologies typically treat the aforementioned thrust-to-weight ratio abrupt changes and induced torque as harmful disturbances. Traditional control strategies tend to passively suppress these disturbances, i.e., suppressing altitude changes by rapidly reducing throttle and resisting attitude roll by counter-rudder maneuvers to maintain aircraft stability. This adversarial approach wastes the gravitational potential energy and rebound kinetic energy contained in the moment of load release, resulting in low energy efficiency. On the other hand, forcibly resisting the huge induced torque places instantaneous power demands on the actuators, which can easily trigger overload protection or even instability and crash.

[0004] Therefore, the present invention provides an intelligent path planning method and system for fire extinguishing aircraft. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent path planning method and system for firefighting aircraft to solve the aforementioned background problems.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A method for intelligent path planning of a firefighting aircraft includes the following steps:

[0008] Collect state analysis data of firefighting aircraft, extract load-thrust mutation characteristics at the moment of load release from the state analysis data; construct the transient dynamic reachability domain of the aircraft based on the mutation characteristics;

[0009] The three-dimensional coordinates of the mission point are obtained, the polarity features of the reachable power domain are extracted, and the principal direction vector of the sudden thrust is obtained; the target heading vector is obtained and combined with the principal direction vector to perform vector homogeneity analysis, obtain the cooperative gain coefficient, and determine whether it can assist the aircraft in approaching the target;

[0010] If it can assist the aircraft in approaching the target, the target guidance vector with the highest cooperative gain coefficient is extracted; based on the centroid drift characteristics caused by load release, the target guidance vector is subjected to torque balancing processing to generate a preset attitude trajectory that uses torque to assist steering.

[0011] The preset trajectory is mapped to the dynamic reachable domain, and the compatibility analysis of time-varying constraints is performed to obtain the risk curve of execution saturation; based on the risk curve, the attitude preset trajectory is optimized by adaptive boundary shrinkage.

[0012] As a further technical solution of the present invention: the method for constructing the power reachable domain is as follows:

[0013] Establish the resultant gravitational acceleration vector, synthesize the load-thrust abrupt change characteristics with the gravitational acceleration vector to obtain the omnidirectional maximum resultant acceleration vector;

[0014] The current flight speed of the aircraft is obtained and used as the initial velocity. At the same time, a maneuvering time window is established. The maximum resultant acceleration vector in all directions is used as the input boundary to calculate the terminal displacement points that the aircraft can reach in each direction in the three-dimensional prediction space within the maneuvering time window.

[0015] Connect the terminal displacement points in all directions with a convex hull to form a closed spatial geometry, which serves as the transient dynamic reachable domain of the aircraft.

[0016] As a further technical solution of the present invention: the method for obtaining the load-thrust mutation characteristics is as follows:

[0017] Obtain the real-time mass and locked thrust value of the payload to be released from the aircraft;

[0018] Obtain the aircraft's base empty mass as the remaining inertial mass value;

[0019] The ratio of the locking thrust value to the remaining inertial mass value is calculated to obtain the theoretical maximum ejection acceleration, which is used as the load-thrust mutation characteristic.

[0020] As a further technical solution of the present invention: the method for performing the vector homogeneity analysis is as follows:

[0021] Obtain the unit vector of the main direction vector of the sudden thrust and the unit vector of the target heading vector;

[0022] Performing a dot product on two unit vectors yields a scalar result that is the cosine of the angle between the two vectors, which is used as the cooperative gain coefficient.

[0023] As a further technical solution of the present invention: the method for obtaining the principal direction vector of the sudden thrust is as follows:

[0024] The aircraft's onboard navigation system reads the waypoint coordinates of the next waypoint in the preset flight route and converts the waypoint coordinates into three-dimensional coordinates in a local station-centered coordinate system with the aircraft's current position as the origin.

[0025] Construct a spatial vector from the current position of the spacecraft to the three-dimensional coordinates of the subsequent mission points, as the target heading vector;

[0026] Calculate the geometric centroid coordinates of the closed spatial geometry corresponding to the transient dynamic reachable domain;

[0027] Construct a spatial vector pointing from the current position of the spacecraft to the coordinates of the geometric centroid, and define it as the principal direction vector of the sudden thrust.

[0028] As a further technical solution of the present invention: the process of generating the preset attitude trajectory is as follows:

[0029] The positions of the center of mass before and after load release are extracted and the torque is quantified to obtain the spontaneous overturning torque caused by load release;

[0030] Establish the terminal convergence state between the aircraft's attitude angular acceleration and driving torque;

[0031] The moment of inertia is obtained, and inverse dynamics is solved based on its own overturning torque, moment of inertia, and terminal convergence state, combined with the target guidance vector, to generate a dynamic attitude preset trajectory.

[0032] As a further technical solution of the present invention: the process of performing the torque quantification is as follows:

[0033] The position of the first center of mass of the aircraft under full load before the release of the fire extinguishing agent is obtained, and the position of the second center of mass under empty load after the release of the load is obtained.

[0034] Calculate the coordinate difference between the first and second centroid positions to obtain the centroid offset vector;

[0035] Establish the thrust axis of action based on the current locked thrust value;

[0036] Calculate the vertical distance from the axis of thrust action to the position of the second center of mass, and use it as the eccentric lever arm;

[0037] By multiplying the locked thrust value with the eccentric lever arm, we obtain the self-overturning torque generated at the moment of load release because the thrust axis does not pass through the second center of mass position.

[0038] As a further technical solution of the present invention: the adaptive boundary shrinkage optimization is performed as follows:

[0039] Set a safe saturation threshold, scan the risk curve of execution saturation, and identify whether there are risk peaks that exceed the safe saturation threshold;

[0040] If no risk spike occurs, continue monitoring for risk spikes.

[0041] If a risk spike occurs, initiate adaptive contraction processing and calculate the trajectory contraction factor.

[0042] The preset tilt angle and initial angular velocity in the attitude preset trajectory are proportionally attenuated using a trajectory contraction factor.

[0043] As a further technical solution of the present invention, the risk curve is established as follows:

[0044] Construct a timeline sequence that covers the characteristic time windows of load release;

[0045] The preset trajectory is discretized and sampled on the time axis to obtain a series of discrete time-time pre-simulation state points;

[0046] Project each pre-simulated state point into the coordinate system of the dynamic reachable domain;

[0047] The Euclidean distance from the pre-simulated state point to the boundary of the transient dynamic reachable domain is calculated and defined as the state safety margin. The state safety margin is then normalized to obtain the normalized execution saturation index.

[0048] Connect the execution saturation indicators at all points in time on the timeline to form a curve, and generate a risk curve for execution saturation.

[0049] An intelligent path planning system for a firefighting aircraft includes the following modules:

[0050] Reachability Analysis Module: Used to collect state analysis data of firefighting aircraft, extract load-thrust mutation characteristics at the moment of load release from the state analysis data; and construct the transient dynamic reachability domain of the aircraft based on the mutation characteristics.

[0051] Vector Analysis Module: Used to acquire the three-dimensional coordinates of the flight mission point, extract the polarity features of the reachable power domain, and obtain the principal direction vector of the sudden thrust; acquire the target heading vector and combine it with the principal direction vector to perform vector homogeneity analysis, obtain the cooperative gain coefficient, and determine whether it can assist the aircraft in approaching the target;

[0052] Trajectory planning module: If it can assist the aircraft in approaching the target, the target guidance vector with the highest cooperative gain coefficient is extracted; based on the centroid drift characteristics caused by load release, the target guidance vector is subjected to torque balancing processing to generate a preset attitude trajectory that uses torque to assist steering.

[0053] Intelligent optimization module: used to map the preset trajectory to the dynamic reachable domain and perform time-varying constraint compatibility analysis to obtain the risk curve of execution saturation; based on the risk curve, adaptive boundary shrinkage optimization is performed on the attitude preset trajectory.

[0054] The beneficial effects of this invention are as follows:

[0055] 1. By collecting aircraft state data and extracting the load-thrust mutation characteristics at the moment of load release, a transient dynamic reachability domain for the aircraft is constructed. The ratio of locked thrust to remaining inertial mass is materialized as a reachability envelope in physical space, quantifying the spatial limit boundary that the aircraft can reach without increasing throttle opening and solely relying on fluid unloading rebound force. This transforms the originally discrete variable mass dynamics process into a visualized geometric solution space. Polarity features are extracted from the dynamic reachability domain, and vector homing analysis is performed in conjunction with the target heading to assess and make decisions regarding the utilization of mutational energy. By calculating the cooperative gain coefficient, the degree of homing between the excess thrust vector generated by load release and the mission target vector is determined, eliminating invalid maneuver paths that, although physically reachable, have low energy utilization or even generate negative work, thus improving the utilization efficiency of limited onboard electrical energy.

[0056] 2. A torque balance model is constructed based on the center-of-gravity drift characteristics induced by load release, and a dynamic attitude preset trajectory is generated to achieve active utilization and conversion of uncontrolled induced torque. To address the issue of a violent overturning torque caused by the thrust axis deviating from the center of gravity due to rapid load release, a preset angular velocity and tilt angle are introduced within a characteristic time window before release. This allows the overturning torque generated at the moment of release to become the driving force that counteracts the preset action and propels the fuselage toward the target, thus reducing the burden on actuators and response delay during attitude adjustment.

[0057] 3. Map the preset trajectory to the dynamic reachable domain for time-varying compatibility analysis and adaptive boundary contraction optimization, constructing a physical capability compatibility verification mechanism from the planning layer to the execution layer. By calculating the execution saturation risk curve, risk sections in the preset trajectory that exceed the current physical limits of the aircraft (such as motor saturation, overload exceeding limits) can be identified. Aggressive attitude commands are proportionally attenuated through an adaptive contraction factor, thereby dynamically balancing catapult maneuverability with maintaining flight envelope stability and reducing the risk of instability caused by control commands exceeding the physical bandwidth of the actuators. Attached Figure Description

[0058] The invention will now be further described with reference to the accompanying drawings.

[0059] Figure 1 This is a flowchart of an intelligent path planning method for a fire extinguishing aircraft according to the present invention;

[0060] Figure 2 This is a flowchart of the present invention that determines whether the current ejection direction can effectively assist the aircraft in approaching the target;

[0061] Figure 3This is a functional module diagram of an intelligent path planning system for a firefighting aircraft according to the present invention. Detailed Implementation

[0062] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0063] Example 1:

[0064] like Figure 1 As shown, an intelligent path planning method for a firefighting aircraft includes the following steps:

[0065] S10. Collect the state analysis data of the firefighting aircraft and extract the load-thrust mutation characteristics at the moment of load release from the state analysis data; construct the transient dynamic reachability domain of the aircraft based on the mutation characteristics.

[0066] The method for collecting state analysis data from firefighting aircraft and extracting the load-thrust mutation characteristics at the moment of load release from the state analysis data is as follows:

[0067] Preferably, the output power data of the electronic speed governor and the liquid level data of the liquid tank pressure sensor are collected synchronously by the airborne flight control computer;

[0068] Use liquid level data and output power data as status analysis data for fire extinguishing aircraft;

[0069] Read the power-thrust calibration curve table pre-stored in the control computer;

[0070] It should be noted that the power-thrust calibration curve table is a polynomial function relationship fitted by the actual tensile data corresponding to different output powers recorded by the tensile sensor during the ground test stand stage. The power-thrust calibration curve table is established based on the polynomial function relationship between power and tensile data.

[0071] Substitute the collected output power data into the power-thrust calibration curve table for interpolation index calculation to obtain the actual mechanical thrust generated by the aircraft at the current throttle opening, which is defined as the current locked thrust value.

[0072] The current liquid volume in the tank is calculated based on the liquid level height data, and the preset extinguishing agent density is obtained. The liquid volume is multiplied by the preset extinguishing agent density to obtain the real-time mass of the payload to be released by the aircraft.

[0073] Obtain the basic empty mass of the aircraft (i.e., the total mass of the aircraft excluding the liquid tank fire extinguishing agent), and define the basic empty mass as the residual inertial mass value;

[0074] It should be noted that the remaining inertial mass value represents the physical mass of the aircraft body after all the extinguishing agent has left the fuselage and passed through. Due to the core working condition of small fire extinguishing aircraft in fixed-point operation, the actual release is a concentrated pulse (completed in tens to hundreds of milliseconds), which is much shorter than the throttle response delay and is approximately an instantaneous mass change.

[0075] The ratio of the locking thrust value to the remaining inertial mass value is calculated to obtain the theoretical maximum ejection acceleration, which is used as the load-thrust mutation characteristic.

[0076] The method for constructing the transient dynamic reachability domain of an aircraft based on mutation characteristics is as follows:

[0077] Establish a three-dimensional prediction space with the aircraft's center of mass as the origin, and set a short maneuver time window;

[0078] Preferably, the maneuver time window is 2 seconds;

[0079] By combining the load-thrust abrupt change characteristics with the gravitational acceleration vector, the omnidirectional maximum resultant acceleration vector is synthesized.

[0080] For example, the synthesis is performed as follows: taking the aircraft's center of mass as the origin, the load-thrust abrupt change characteristics (i.e., the numerical value of the theoretical maximum ejection acceleration) are synthesized. Using π as the radius, construct an isotropic acceleration sphere;

[0081] It should be noted that any point on the surface of the sphere represents the pure thrust acceleration that the aircraft can generate when launched in that direction at full throttle, ignoring gravity.

[0082] For any unit direction vector in the three-dimensional prediction space The corresponding thrust acceleration vector is Obtain the local gravitational acceleration vector. (Direction vertically downwards, module length approximately) );

[0083] in, The number of the unit direction vector;

[0084] For each sampling direction in space, the thrust acceleration vector and the gravitational acceleration vector are superimposed, and the superposition method is as follows: ;

[0085] It's understandable that performing vector superposition is equivalent to vector-translating the "thrust acceleration sphere" at the origin along the direction of gravity. For example, when the aircraft accelerates vertically upwards, the resultant acceleration is... When accelerating vertically downwards, the resultant acceleration is: ;

[0086] With the current flight speed as the initial velocity and the omnidirectional maximum resultant acceleration vector as the input boundary, the closed spatial geometry formed by the displacement boundary within the dynamic time window of the kinetic integral computer is the dynamic reachable domain.

[0087] Connect the terminal displacement points in all directions with convex hulls to form a closed spatial geometry, which serves as the transient dynamic reachable domain of the aircraft.

[0088] Preferably, the displacement boundary within the computer's dynamic time window is defined as follows:

[0089] S101. Establish the initial three-dimensional prediction space and discretize it into N unit direction vectors;

[0090] Preferably, N=1024;

[0091] S102. For each unit direction vector, assume that the aircraft applies all the acceleration represented by the load-thrust mutation exponent to the direction corresponding to the unit direction vector.

[0092] S103. Combining the aircraft's current velocity vector with the air resistance model, perform a second integral for each direction to obtain the terminal displacement point of the direction corresponding to the unit direction vector at the end of the time window.

[0093] It should be noted that the transient dynamic reachable domain is geometrically represented as an asymmetric "teardrop" or "egg" shaped space starting from the current position, and the size of the dynamic reachable domain is directly proportional to the load-thrust abrupt change characteristics.

[0094] It should be noted that the physical meaning of constructing the power reachability domain is: to delineate the boundary that the aircraft absolutely cannot fly out of, relying solely on the rebound force after jettisoning water, without changing the current throttle opening (without consuming additional power), thus providing spatial constraints for finding the optimal path in the future.

[0095] The purpose of constructing the transient dynamic reachable domain is as follows:

[0096] Objective 1: To transform the complex variable mass dynamics process at the moment of load release into an intuitive three-dimensional spatial geometric envelope, thereby reducing the need for complex real-time dynamic calculations in subsequent fire and rescue planning;

[0097] Objective 2: To determine the physical limits that an aircraft can reach without increasing throttle, thereby providing a solution space constraint for subsequent path planning and reducing the number of unworkable paths that exceed physical capabilities.

[0098] S20. Obtain the three-dimensional coordinates of the flight mission point, extract the polarity features of the reachable power domain, and obtain the main direction vector of the sudden thrust; obtain the target heading vector and combine it with the main direction vector to perform vector homogeneity analysis, obtain the cooperative gain coefficient, and determine whether it can assist the aircraft in approaching the target;

[0099] The method for obtaining the three-dimensional coordinates of the flight mission point, extracting the polarity features of the reachable power domain, and obtaining the principal direction vector of the sudden thrust is as follows:

[0100] Preferably, the waypoint coordinates of the next waypoint in the preset flight route are read through the aircraft's onboard navigation system, and the waypoint coordinates are converted into three-dimensional coordinates in the local station center (ENU) coordinate system with the current position of the aircraft as the origin;

[0101] Construct a spatial vector from the current position of the spacecraft to the three-dimensional coordinates of the subsequent mission points, as the target heading vector;

[0102] Geometric features are extracted from the transient dynamic reachable domain generated in step S10 to determine the principal direction of the sudden thrust.

[0103] Preferably, the method for extracting geometric features to determine the principal direction of the abrupt thrust is as follows:

[0104] Calculate the geometric centroid coordinates of the closed spatial geometry corresponding to the transient dynamic reachable domain;

[0105] Construct a spatial vector pointing from the current position of the spacecraft to the coordinates of its geometric center of mass, and define it as the principal direction vector of the sudden thrust.

[0106] It should be noted that since the transient dynamic reachable domain is generated by the integral of the unidirectional excess thrust produced by load release, the geometry of the dynamic reachable domain exhibits the characteristic of extending along the thrust direction. Therefore, the vector pointing from the origin to the center of mass can characterize the average action direction of the sudden energy.

[0107] A vector co-directional analysis is performed on the main direction vector of the sudden thrust and the target heading vector. The cosine of the angle between the two is calculated to obtain the cooperative gain coefficient, which is used to characterize the coupling degree between the direction of the sudden thrust and the direction of the mission target.

[0108] The preferred method for performing vector homogeneity analysis is as follows:

[0109] S201. Obtain the unit vector of the main direction vector of the sudden thrust and the unit vector of the target heading vector;

[0110] S202. Perform a dot product operation on two unit vectors. The scalar result is the cosine of the angle between the two vectors, which is used as the cooperative gain coefficient.

[0111] It should be noted that the range of the cooperative gain coefficient is [-1, 1]. The closer the value is to 1, the more consistent the direction of the sudden thrust is with the direction of the target mission; a negative value indicates that the two directions are opposite.

[0112] Read the preset positive gain threshold in the control computer;

[0113] Preferably, the positive gain threshold is set to 0.5;

[0114] It should be noted that the physical meaning of setting the threshold to 0.5 is that the angle between the direction of the sudden thrust and the target direction is less than 60 degrees, ensuring that at least 50% of the energy generated by the load release is used to propel the aircraft toward the target in the positive direction, rather than for unnecessary lateral drift.

[0115] like Figure 2 As shown, the cooperative gain coefficient is compared with a preset positive gain threshold for determination;

[0116] If the cooperative gain coefficient is greater than or equal to the positive gain threshold, the condition is met, indicating that the current ejection direction can effectively assist the aircraft in approaching the target.

[0117] If the cooperative gain coefficient is less than the positive gain threshold, the condition is not met, indicating that releasing the load under the current attitude cannot bring sufficient range gain, and it is necessary to maintain normal flight or readjust the approach angle.

[0118] It is understandable that the purpose of introducing and determining the cooperative gain coefficient is as follows:

[0119] Function 1: To assess, from the perspective of energy coupling, whether the excess thrust generated by the load release is coordinated with the fire and rescue mission objective, and to identify the ejection direction that performs positive or negative work.

[0120] Secondly, if the current attitude and position prevent the ejection force from effectively targeting the target, the ejection can be abandoned directly to reduce the waste of precious onboard electrical energy caused by blind maneuvering.

[0121] Example 2:

[0122] Please see Figure 1 As shown, an intelligent path planning method for a firefighting aircraft includes the following steps:

[0123] S30. If it can assist the aircraft in approaching the target, extract the target guidance vector with the highest cooperative gain coefficient; perform torque balancing processing on the target guidance vector based on the centroid drift characteristics caused by load release, and generate a preset attitude trajectory that uses torque to assist steering.

[0124] The method for extracting the target steering vector with the highest cooperative gain coefficient is as follows:

[0125] Preferably, when the determination result in step S20 is that the cooperative gain coefficient is greater than or equal to the positive gain threshold, the flight control computer activates the transient leverage planning mode.

[0126] Extract the target heading vector corresponding to the highest cooperative gain coefficient from the calculation results of step S20, and use it as the target guidance vector for this ejection jump;

[0127] Among them, the method of generating a preset attitude trajectory that utilizes torque-assisted steering by performing torque balancing processing on the target guidance vector based on the centroid drift characteristics caused by load release is as follows:

[0128] S301. Extract the position of the center of mass before and after load release and quantify the torque to obtain the spontaneous overturning torque caused by load release;

[0129] The preferred method for torque quantification is as follows:

[0130] The position of the first center of mass of the aircraft under full load before the release of the fire extinguishing agent is obtained, and the position of the second center of mass under empty load after the release of the load is obtained.

[0131] Calculate the coordinate difference between the first and second centroid positions to obtain the centroid offset vector;

[0132] Establish the thrust axis of action based on the current locked thrust value;

[0133] Calculate the vertical distance from the axis of thrust action to the position of the second center of mass, and use it as the eccentric lever arm;

[0134] The locked thrust value is multiplied with the eccentric lever arm to obtain the self-overturning torque generated at the moment of load release because the thrust axis does not pass through the second center of mass position;

[0135] S302. Establish the terminal convergence state between the aircraft's attitude angular acceleration and driving torque;

[0136] Preferably, the method for establishing the dynamic balance relationship between the aircraft's attitude angular acceleration and driving torque is as follows:

[0137] Set the target steering vector to the terminal convergence state that the dynamic equilibrium relationship must reach at the moment of load release;

[0138] It should be noted that the terminal convergence state refers to the moment when the load release is completed, the thrust axis of the fuselage must be aligned with the target guide vector to ensure that the obtained ejection kinetic energy is directed towards the target to the greatest extent.

[0139] S303. Obtain the moment of inertia, and perform inverse dynamics solution based on its own overturning torque, moment of inertia and terminal convergence state, combined with the target guidance vector, to generate a dynamic attitude preset trajectory.

[0140] The preferred method for trajectory planning is as follows:

[0141] Read the three-axis rotational inertia tensor of the aircraft under full load. ;

[0142] It should be noted that the spontaneous overturning torque generated after the load is released The aircraft body naturally turns to the target, if the load release time is If 0, then based on the rotational dynamics equations and the rigid body rotation laws, establish the attitude evolution equations:

[0143] ;

[0144] The boundary condition for the attitude evolution equation is: within a very short time after release... (e.g., 0.5s) after which, the fuselage attitude The angle of the target guidance vector should be aligned. ;

[0145] Attitude angle at any moment The spontaneous overturning torque calculated in step S301, For rotational inertia, For time, The initial angular velocity, This is the initial preset angle;

[0146] The moment of load release ( ) is set as a boundary condition, and at the same time (i.e., the angle of the target guidance vector), substitute it into the attitude evolution equation, and solve in reverse at the beginning of the characteristic time window ( ) Required initial preset angle and initial angular velocity ;

[0147] Generate resistance to head-up moment The aircraft must begin executing a specific nose-down acceleration maneuver (i.e., the calculated acceleration) at a pre-release characteristic time (preferably 0.4s). This sequence of attitudes that changes over time is the attitude preset trajectory;

[0148] It is understandable that executing the attitude preset trajectory is similar to the pre-pressurization action before diving on a springboard. That is, within the characteristic time window before the hydraulic valve of the liquid tank is opened, the fuselage is controlled to actively perform a specific reverse pre-turn action (such as pre-nodding), so that the huge spontaneous roll torque (such as pitching torque) generated when the load is released is exactly canceled or guided by this pre-turn action, thereby transforming the fuselage's originally harmful roll tendency into a turning action pointing towards the target, and realizing ejection using induced torque to assist in turning.

[0149] The purpose of using spontaneous flipping torque to generate a preset attitude trajectory is as follows:

[0150] Objective 1: To transform the originally harmful tendency of fuselage roll into a beneficial driving force that assists the fuselage in quickly turning to the target by pre-setting attitude guiding torque;

[0151] Objective 2: To utilize the physical torque itself to complete most of the steering work, reducing the need for the motor and servo to output huge reverse torque in a very short time to maintain balance, thereby reducing the risk of overload protection or response lag in the actuator.

[0152] S40. Map the preset trajectory to the dynamic reachable domain and perform time-varying constraint compatibility analysis to obtain the risk curve of execution saturation; perform adaptive boundary contraction optimization on the attitude preset trajectory based on the risk curve;

[0153] The method for mapping the preset attitude trajectory to the transient dynamic reachability domain and performing time-varying constraint compatibility analysis to obtain the risk curve of execution saturation is as follows:

[0154] Construct a timeline sequence that covers the characteristic time windows of load release;

[0155] Preferably, the duration of the characteristic time window for the release of the coverage load is 0.1–0.5 s, and more preferably 0.3 s;

[0156] The attitude preset trajectory generated by S30 is discretized and sampled on the time axis to obtain a series of discrete time-time pre-simulation state points;

[0157] For example, the pre-simulation state points include: pre-simulation position, pre-simulation velocity value, and pre-simulation acceleration value;

[0158] Project each pre-simulated state point into the coordinate system of the transient dynamic reachable domain constructed by S10;

[0159] For each discrete moment, determine whether the predicted state point at the discrete moment is within the safe confidence interval of the transient dynamic reachability domain;

[0160] It should be noted that the transient dynamic reachable domain is not a static rigid boundary, but a dynamic envelope that gradually shrinks over time (energy dissipation). Therefore, the essence of compatibility analysis is to determine whether the aircraft has enough remaining rebound potential energy at time t to support the flight maneuver planned by the S30 preset trajectory.

[0161] The Euclidean distance from the pre-simulated state point to the boundary of the transient dynamic reachable domain is calculated and defined as the state safety margin. The state safety margin is then normalized to obtain the normalized execution saturation index.

[0162] It should be noted that if the saturation is 100%, it means that the current action requires the aircraft to exert physical limits, which poses an extremely high risk of loss of control.

[0163] Connect the execution saturation indicators at all points in time on the timeline to form a curve, and generate a risk curve for execution saturation;

[0164] The method for adaptive boundary shrinkage optimization based on the preset attitude trajectory is as follows:

[0165] Set a safe saturation threshold, scan the risk curve of execution saturation, and identify whether there are risk peaks that exceed the safe saturation threshold;

[0166] Preferably, the safe saturation level of the risk curve for row saturation is preset to 85%;

[0167] If no risk spike occurs, continue monitoring for risk spikes.

[0168] If a risk spike occurs, it indicates that the S30 plan is too aggressive, exceeding the comfort zone (power reach range) defined by S10, and an adaptive contraction process is initiated to calculate the trajectory contraction factor.

[0169] Among them, adaptive shrinkage processing is initiated, and the trajectory shrinkage factor is calculated in the following way:

[0170] The trajectory contraction factor is obtained by comparing the safety saturation threshold of S10 with the peak value of the risk spike.

[0171] The trajectory contraction factor is used to adaptively adjust the preset tilt angle and initial angular velocity in the attitude preset trajectory;

[0172] For example, the adaptive adjustment is performed as follows: assuming the preset trajectory planned by S30 requires the motor to output... The reverse torque is used to balance the overturning caused by load release, but the current physical limit of the motor is only... (That is, 120% saturation, which carries risks);

[0173] At this point, the system should not directly truncate the reverse torque command. (Otherwise it will lead to) Instead of the residual torque causing the aircraft to lose control, the following adjustments were made:

[0174] Calculate how to reduce the flipping torque to a level that the motor can withstand. The main locking thrust value needs to be reduced from 100% to 83% (i.e., );

[0175] The system generates a thrust degradation command while generating the attitude preset trajectory;

[0176] When executing preset actions, the electronic governor synchronously limits the main throttle to 83%;

[0177] It should be noted that limiting the main throttle sacrifices some launch acceleration, but this helps to ensure that the rollover torque is always within the controllable range of the actuator, thus achieving the optimal solution within the physical boundaries.

[0178] The corrected preset trajectory is defined as the final execution trajectory and sent to the aircraft's underlying controller.

[0179] Example 3:

[0180] Please see Figure 3 As shown, an intelligent path planning system for a firefighting aircraft includes the following modules:

[0181] Reachability Analysis Module: Used to collect state analysis data of firefighting aircraft, extract load-thrust mutation characteristics at the moment of load release from the state analysis data; and construct the transient dynamic reachability domain of the aircraft based on the mutation characteristics.

[0182] Vector Analysis Module: Used to acquire the three-dimensional coordinates of the flight mission point, extract the polarity features of the reachable power domain, and obtain the principal direction vector of the sudden thrust; acquire the target heading vector and combine it with the principal direction vector to perform vector homogeneity analysis, obtain the cooperative gain coefficient, and determine whether it can assist the aircraft in approaching the target;

[0183] Trajectory planning module: If it can assist the aircraft in approaching the target, the target guidance vector with the highest cooperative gain coefficient is extracted; based on the centroid drift characteristics caused by load release, the target guidance vector is subjected to torque balancing processing to generate a preset attitude trajectory that uses torque to assist steering.

[0184] Intelligent optimization module: used to map the preset trajectory to the dynamic reachable domain and perform time-varying constraint compatibility analysis to obtain the risk curve of execution saturation; based on the risk curve, adaptive boundary shrinkage optimization is performed on the attitude preset trajectory;

[0185] In one specific embodiment, an intelligent path planning system for a firefighting aircraft further includes a positioning and self-rescue module. This module comprises a backup battery unit independent of the aircraft's main power supply, a positioning unit, and a wireless communication unit. The positioning and self-rescue module monitors the aircraft's main power supply status data in real time. When a main power supply failure is detected, it automatically activates the backup battery unit to provide power, locks the current geographic coordinates using the positioning unit, and transmits the geographic coordinates to a ground control terminal using the wireless communication unit for subsequent search and recovery of the aircraft.

[0186] The foregoing detailed description of one embodiment of the present invention is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for intelligent path planning of a firefighting aircraft, characterized in that, Includes the following steps: Collect state analysis data of firefighting aircraft and extract the load-thrust mutation characteristics at the moment of load release from the state analysis data; Constructing the transient dynamic reachability domain of an aircraft based on mutation characteristics; The three-dimensional coordinates of the mission point are obtained, the polarity features of the reachable power domain are extracted, and the principal direction vector of the sudden thrust is obtained; the target heading vector is obtained and combined with the principal direction vector to perform vector homogeneity analysis, obtain the cooperative gain coefficient, and determine whether it can assist the aircraft in approaching the target; If it can help the aircraft approach the target, then extract the target guidance vector with the highest cooperative gain coefficient; Based on the centroid drift characteristics caused by load release, the target guidance vector is subjected to torque balancing processing to generate a preset attitude trajectory that uses torque to assist steering. The preset trajectory is mapped to the dynamic reachable domain, and the compatibility analysis of time-varying constraints is performed to obtain the risk curve of execution saturation. Adaptive boundary shrinkage optimization is performed on the attitude preset trajectory based on the risk curve.

2. The intelligent path planning method for a firefighting aircraft according to claim 1, characterized in that: The dynamic reachability domain is constructed as follows: Establish the resultant gravitational acceleration vector, synthesize the load-thrust abrupt change characteristics with the gravitational acceleration vector to obtain the omnidirectional maximum resultant acceleration vector; The current flight speed of the aircraft is obtained and used as the initial velocity. At the same time, a maneuvering time window is established. The maximum resultant acceleration vector in all directions is used as the input boundary to calculate the terminal displacement points that the aircraft can reach in each direction in the three-dimensional prediction space within the maneuvering time window. Connect the terminal displacement points in all directions with a convex hull to form a closed spatial geometry, which serves as the transient dynamic reachable domain of the aircraft.

3. The intelligent path planning method for a firefighting aircraft according to claim 2, characterized in that: The method for obtaining the load-thrust abrupt change characteristics is as follows: Obtain the real-time mass and locked thrust value of the payload to be released from the aircraft; Obtain the aircraft's base empty mass as the remaining inertial mass value; The ratio of the locking thrust value to the remaining inertial mass value is calculated to obtain the theoretical maximum ejection acceleration, which is used as the load-thrust mutation characteristic.

4. The intelligent path planning method for a firefighting aircraft according to claim 1, characterized in that: The method for performing the vector homogeneity analysis is as follows: Obtain the unit vector of the main direction vector of the sudden thrust and the unit vector of the target heading vector; Performing a dot product on two unit vectors yields a scalar result that is the cosine of the angle between the two vectors, which is used as the cooperative gain coefficient.

5. The intelligent path planning method for a firefighting aircraft according to claim 4, characterized in that: The method for obtaining the principal direction vector of the sudden thrust is as follows: The aircraft's onboard navigation system reads the waypoint coordinates of the next waypoint in the preset flight route and converts the waypoint coordinates into three-dimensional coordinates in a local station-centered coordinate system with the aircraft's current position as the origin. Construct a spatial vector from the current position of the spacecraft to the three-dimensional coordinates of the subsequent mission points, as the target heading vector; Calculate the geometric centroid coordinates of the closed spatial geometry corresponding to the transient dynamic reachable domain; Construct a spatial vector pointing from the current position of the spacecraft to the coordinates of the geometric centroid, and define it as the principal direction vector of the sudden thrust.

6. The intelligent path planning method for a firefighting aircraft according to claim 1, characterized in that: The process of generating the preset attitude trajectory is as follows: The positions of the center of mass before and after load release are extracted and the torque is quantified to obtain the spontaneous overturning torque caused by load release; Establish the terminal convergence state between the aircraft's attitude angular acceleration and driving torque; The moment of inertia is obtained, and inverse dynamics is solved based on its own overturning torque, moment of inertia, and terminal convergence state, combined with the target guidance vector, to generate a dynamic attitude preset trajectory.

7. The intelligent path planning method for a firefighting aircraft according to claim 6, characterized in that: The process of performing the torque quantization is as follows: The position of the first center of mass of the aircraft under full load before the release of the fire extinguishing agent is obtained, and the position of the second center of mass under empty load after the release of the load is obtained. Calculate the coordinate difference between the first and second centroid positions to obtain the centroid offset vector; Establish the thrust axis of action based on the current locked thrust value; Calculate the vertical distance from the axis of thrust action to the position of the second center of mass, and use it as the eccentric lever arm; By multiplying the locked thrust value with the eccentric lever arm, we obtain the self-overturning torque generated at the moment of load release because the thrust axis does not pass through the second center of mass position.

8. The intelligent path planning method for a firefighting aircraft according to claim 1, characterized in that: The adaptive boundary shrinkage optimization is performed as follows: Set a safe saturation threshold, scan the risk curve of execution saturation, and identify whether there are risk peaks that exceed the safe saturation threshold; If no risk spike occurs, continue monitoring for risk spikes. If a risk spike occurs, initiate adaptive contraction processing and calculate the trajectory contraction factor. The trajectory contraction factor is used to adaptively adjust the preset tilt angle and initial angular velocity in the attitude preset trajectory.

9. The intelligent path planning method for a firefighting aircraft according to claim 1, characterized in that: The risk curve is constructed as follows: Construct a timeline sequence that covers the characteristic time windows of load release; The preset trajectory is discretized and sampled on the time axis to obtain a series of discrete time-time pre-simulation state points; Project each pre-simulated state point into the coordinate system of the dynamic reachable domain; The Euclidean distance from the pre-simulated state point to the boundary of the transient dynamic reachable domain is calculated and defined as the state safety margin. The state safety margin is then normalized to obtain the normalized execution saturation index. Connect the execution saturation indicators at all points in time on the timeline to form a curve, and generate a risk curve for execution saturation.

10. An intelligent path planning system for a firefighting aircraft, used to implement the intelligent path planning method for a firefighting aircraft as described in any one of claims 1-9, characterized in that, Includes the following modules: Reachability Analysis Module: Used to collect state analysis data of firefighting aircraft and extract the load-thrust mutation characteristics at the moment of load release from the state analysis data; Constructing the transient dynamic reachability domain of an aircraft based on mutation characteristics; Vector Analysis Module: Used to acquire the three-dimensional coordinates of the flight mission point, extract the polarity features of the reachable power domain, and obtain the principal direction vector of the sudden thrust; acquire the target heading vector and combine it with the principal direction vector to perform vector homogeneity analysis, obtain the cooperative gain coefficient, and determine whether it can assist the aircraft in approaching the target; Trajectory planning module: If it can assist the aircraft in approaching the target, then extract the target guidance vector with the highest cooperative gain coefficient; Based on the centroid drift characteristics caused by load release, the target guidance vector is subjected to torque balancing processing to generate a preset attitude trajectory that uses torque to assist steering. Intelligent optimization module: used to map the preset trajectory to the dynamic reachable domain and perform time-varying constraint compatibility analysis to obtain the risk curve of execution saturation; Adaptive boundary shrinkage optimization is performed on the attitude preset trajectory based on the risk curve.

Images