Intelligent Planning Method and System for Aircraft Transportation Missions Based on Real-Time Battery Prediction

Abstract

This invention discloses an intelligent planning method and system for aircraft transportation missions based on real-time power prediction. The method includes: constructing a full-physics power model that couples the dynamic characteristics of the inertial measurement unit, the environmental wind field vector, and battery characteristics, and completing offline calibration; during mission execution, using an extended Kalman filter to jointly estimate the load mass and wind field vector in real time, and dynamically solving for the optimal cruise speed that minimizes energy consumption per unit distance; predicting the aircraft's remaining support distance based on the real-time calculated state and "energy bucket" model; and automatically triggering a hierarchical closed-loop strategy—prioritizing waypoint elimination, replanning the route, or immediate return—when energy is insufficient by continuously monitoring and comparing with mission requirements. This invention achieves accurate endurance prediction and autonomous mission adjustment under load changes and wind field disturbances, significantly improving the safety, adaptability, and completion efficiency of aircraft in multi-point transportation missions such as express delivery.

Description

Technical Field

[0001] This invention belongs to the field of aircraft mission planning and energy management technology, and particularly relates to an intelligent planning method and system for aircraft transportation missions based on real-time power prediction. Background Technology

[0002] Aircraft are playing an increasingly important role in modern urban logistics and express delivery, with typical missions involving multi-point cargo delivery, dynamic route adjustment, and adaptation to complex outdoor environments. The reliable completion of these missions heavily relies on accurate prediction of the aircraft's real-time energy consumption and dynamic assessment of its remaining range.

[0003] Inaccurate remaining battery power predictions could cause an aircraft to run out of power mid-mission, preventing it from completing deliveries to subsequent drop points and resulting in cargo loss, delays, or safety incidents. Conversely, overly conservative range estimates can limit the payload and coverage area of ​​a single mission, reducing overall delivery efficiency. Especially in outdoor flights, sudden changes in load mass (reduced after cargo drop), real-time changes in environmental wind fields, nonlinear battery discharge characteristics, and anisotropy of energy consumption under different flight paths can significantly amplify prediction errors and increase the risk of forced landings. Therefore, accurately and dynamically predicting the aircraft's remaining support distance under varying operating conditions and enabling intelligent mission replanning based on this is a core requirement for improving the success rate and safety of delivery missions.

[0004] Currently, research on aircraft mission planning mainly focuses on static path optimization, multi-aircraft collaborative scheduling, and basic battery status monitoring. However, there is a lack of systematic solutions based on physical mechanisms for real-time energy consumption modeling and range prediction. Existing technologies mostly use linear extrapolation of battery SOC percentage or simplified empirical formulas for estimation. This method is difficult to effectively cope with the coupled effects of multiple variables such as load changes, wind field interference, heat loss, and optimal speed adjustment, resulting in insufficient prediction accuracy.

[0005] In reality, aircraft range prediction is essentially a highly nonlinear, multivariate process. It is not only constrained by the current battery status, but also requires real-time decoupling of implicit states (such as load mass and wind field vector), optimization of cruise strategies, and reservation of safety margins.

[0006] Therefore, there is an urgent need for a physical model-driven real-time prediction and closed-loop planning method to overcome the limitations of existing technologies in dynamic transportation scenarios and provide more reliable and efficient core technology support for applications such as aircraft express delivery. Summary of the Invention

[0007] The purpose of this invention is to solve the problems in existing aircraft delivery mission planning, such as the lack of accurate models driven by physical mechanisms for range prediction, reliance on linear extrapolation of battery SOC or simple empirical estimation leading to large prediction deviations, insufficient power during missions, or low efficiency. Therefore, this invention proposes an intelligent planning method and system for aircraft transportation missions based on real-time power prediction.

[0008] To achieve one of the above-mentioned objectives, embodiments of the present invention provide an intelligent planning method for aircraft transportation missions based on real-time power prediction, the method comprising:

[0009] Step 1: Construct a full-physics power model of the aircraft that couples the dynamic field, wind field, and battery characteristics:

[0010] ,

[0011] in, For real-time power, Based on maintaining power, Indicates the length of the modulus. It is a three-axis specific force vector. For the overall efficiency of the flight propulsion system, For load quality, For wind speed vectors, Ground velocity vector, This represents the real-time current of the battery. The basic internal resistance of the battery, This is the internal resistance compensation coefficient. This represents the remaining percentage of battery charge. For induced power coefficient, The parasitic power coefficient;

[0012] Step 2: Substitute the current load mass and wind speed vector into the full physics power model of the aircraft to solve for the optimal cruising ground speed that minimizes energy consumption per unit ground distance.

[0013] Step 3: Calculate the remaining support distance of the aircraft under the current planned path. If the remaining support distance is insufficient to cover the remaining tasks, the task points are eliminated in order of preset priority from low to high, and the path is replanned. Otherwise, the transportation task is completed according to the current planned path.

[0014] As a further improvement to the embodiment of the present invention, step 1 further includes a method combining full envelope flight excitation and gradient descent parameter identification to evaluate the coefficients of the full physics power model of the aircraft. Calibration is performed, including,

[0015] Control the aircraft to perform a combination of flight maneuvers, including variable load hovering, multi-speed cruise, and high-G maneuvering acceleration and deceleration, and collect data. Using synchronized measured samples, a loss function is constructed to measure the mean square error between the real-time power prediction and the real-time measured power of the full physics power model of the aircraft. The gradient of the loss function with respect to each coefficient is then calculated, and the function is iteratively updated according to a set learning rate until convergence. The iterative update formula for each coefficient is as follows:

[0016] ,

[0017] ,

[0018] ,

[0019] in, The first Second and third The next iteration , The first Second and third The next iteration , The first Second and third The next iteration , For learning rate, For the first Real-time power prediction values ​​in the measured sample set. For the first Real-time measured power values ​​in the set of measured samples. For the first The preset load mass in the group of measured samples, For the first The specific force in the group of measured samples For the first The composite airspeed vector in the set of measured samples.

[0020] As a further improvement to one embodiment of the present invention, the overall efficiency of the flight propulsion system is... The expression is:

[0021] ,

[0022] in, For ideal induced power, The total thrust generated by the rotor. air density, For rotor sweep area, This is the operating voltage.

[0023] As a further improvement to one embodiment of the present invention, step 2 utilizes an extended Kalman filter to jointly estimate the current load mass and wind field vector in real time, including:

[0024] Define the state vector: , Let be the state vector at time t. These are the x, y, and z wind speed vectors at time t in the navigation coordinate system.

[0025] Constructing a state prediction model: ,in, Let be the state vector at time t-1. The process noise at time t-1;

[0026] Establish the nonlinear observation equation: ,in, Let be the observation vector at time t. , The measured real-time power value at time t. These represent the x, y, and z-axis ratios at time t in the navigation coordinate system. Let be the observation function at time t. , Let be the load mass at time t. Let be the triaxial specific force vector at time t. Let be the ground velocity vector at time t. Let be the wind speed vector at time t. Let be the current of the battery at time t. Let be the thrust vector of the aircraft at time t. Let be the air resistance vector at time t;

[0027] Calculate the Jacobian matrix of the observation function with respect to the state vector: ,in, Let be the Jacobian matrix at time t;

[0028] Calculate the Kalman gain and correct the state vector in real time:

[0029] ,

[0030] in, for The posterior state estimation vector, Based on The one-step predicted state vector at time t. Kalman gain;

[0031] Based on the real-time corrected state vector, the current load mass and wind speed vector are obtained.

[0032] As a further improvement to one embodiment of the present invention, the calculation of the optimal cruising ground speed in step 2 includes,

[0033] ,

[0034] in, This is a scalar value representing the optimal cruising ground speed.

[0035] As a further improvement to one embodiment of the present invention, the calculation of the remaining support distance of the aircraft under the current planned path in step 3 includes,

[0036] Obtain the effective remaining energy of the spacecraft under the current planned path. :

[0037] ,

[0038] in, This represents the current remaining percentage of battery power. This is a warning line for the non-linear voltage drop of the battery. The rated capacity of the battery. The average discharge voltage of the battery. This represents the battery's discharge efficiency coefficient. The dimension conversion coefficient is... Reserved energy for return and landing;

[0039] Calculate the remaining support distance of the aircraft on the current planned path:

[0040] ,

[0041] in, This is a power reference obtained by smoothing real-time power measurements over a set time range using a sliding window filter. The target cruise speed scalar.

[0042] As a further improvement to one embodiment of the present invention, step 3 includes,

[0043] like If the duration exceeds the set time threshold, the transportation task is determined to be unsustainable, and tasks are removed in ascending order of preset priority until the conditions are met. And replan the route; among them, This represents the remaining support distance for the aircraft along its current planned path. The remaining total flight distance for the mission. The set safety distance margin is determined; otherwise, it is determined that the transport mission can continue to be carried out, and the aircraft continues to complete the transport mission according to the current planned path at the target cruising speed.

[0044] As a further improvement to one embodiment of the present invention, the expression for the target cruising speed scalar is:

[0045] ,

[0046] in, To achieve the required velocity scalar for piecewise time windows, , Indexes for unfinished task points; This is the set of unfinished task points. From the current location to the unfinished task point The cumulative flight distance, To reach the unfinished task point The latest time constraint, For the current time, To allow for a time margin, This is a scalar value representing the optimal cruising ground speed.

[0047] As a further improvement to one embodiment of the present invention, if task points are eliminated sequentially, it still cannot satisfy the requirement. If this happens, an immediate return-to-base command will be triggered.

[0048] To achieve one of the above-mentioned objectives, another embodiment of the present invention provides an intelligent planning system for aircraft transportation missions based on real-time power prediction, the system comprising:

[0049] The power model building module is used to build a full physics power model of an aircraft that couples dynamic fields, wind fields, and battery characteristics.

[0050] The optimal cruising ground speed calculation module substitutes the current load mass and wind speed vector into the full physics power model of the aircraft to solve for the optimal cruising ground speed that minimizes energy consumption per unit ground distance.

[0051] The remaining support distance calculation module is used to calculate the remaining support distance of the aircraft under the current planned path;

[0052] The remaining task planning module is used to determine whether the remaining support distance is sufficient to cover the remaining tasks. If not, task points are removed in order of priority from low to high, and the route is replanned. If so, the transportation task is completed according to the currently planned route.

[0053] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0054] 1. This invention constructs a full-physics power model that couples dynamics, wind field, and battery characteristics, and utilizes extended Kalman filtering for online joint state estimation, achieving real-time and dynamic perception of aircraft load, wind field, and energy state. This method fundamentally changes the traditional static mission planning mode that relies on preset fixed speeds and conservative battery margins. It can calculate the optimal cruise speed (sweet spot speed) that dynamically changes with heading and wind speed in real time, and accurately predict the remaining support distance. Thus, it maintains an energy-optimal flight strategy in complex environments, significantly improving the aircraft's mission range and economy, and providing core intelligence for reliable and efficient transportation in dynamic environments.

[0055] 2. This invention uses real-time perceived aircraft energy state, load, and environmental disturbances as direct criteria for mission feasibility, achieving a closed-loop automatic connection from "state perception" to "mission decision-making." Through hysteresis decision logic and a hierarchical replanning strategy, this method can automatically adjust mission objectives or trigger a safe return when insufficient energy is predicted, significantly enhancing the aircraft's mission autonomy and overall reliability under uncertainties such as load changes, wind disturbances, and battery degradation. This provides key technical support for building unmanned systems with self-situational awareness and intelligent survivability. Attached Figure Description

[0056] Figure 1 This is a flowchart of the method of the present invention.

[0057] Figure 2 This is a system structure diagram of the present invention. Detailed Implementation

[0058] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0059] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0060] The intelligent planning method for aircraft transportation missions based on real-time power prediction proposed in this invention, such as... Figure 1 As shown, it includes the following steps:

[0061] Step 1, Construction of the full-state sensing physical power model: Construct a full-physical power model of the aircraft that couples the dynamic characteristics of the inertial measurement unit (IMU), the environmental wind field vector, and the electrochemical characteristics of the battery.

[0062] Specifically, the expression for the full physics power model of the aircraft is as follows:

[0063] ,

[0064] in: Real-time power (W); The baseline sustaining power (W) includes the power consumption of avionics equipment and the power consumption of the rotor at zero lift drag. Indicates the length of the modulus. Triaxial specific force vector measured by IMU Modulus length ( ), representing the total thrust acceleration requirement to overcome gravity and maneuver overload ( ); The ground velocity vector measured by the sensor ( ); All of these are real-time state variables to be solved. Load mass (kg) Wind speed vector ( ); These are the model coefficients to be calibrated, unaffected by changes in wind field and load; Induced power coefficient ), Parasitic power coefficient ( ); The overall efficiency of the flight propulsion system is dimensionless and can be calculated using bench tension.

[0065] Therefore, 0.5 is usually taken as an empirical constant; For ideal induced power, The total thrust generated by the rotor. air density, For rotor sweep area, This refers to the system operating voltage. The real-time current (A) is collected by the airborne power management module. This represents the percentage of remaining battery power, dimensionless. The basic internal resistance of the battery (Ω); It is a low-electricity internal resistance compensation coefficient, used to characterize the heat loss characteristics that increase linearly as the electricity is depleted.

[0066] Specifically, the solution process for the wind field coupling parasitic term includes the following steps and formulas:

[0067] (1) Reconstruct the thrust vector of the airframe: based on the real-time speed of each motor j fed back by the electronic speed controller. Combined with propeller thrust coefficient (Given parameters of the flight control system) Calculate the total thrust modulus along the vertical axis (Z-axis) of the aircraft. And construct the thrust vector in the body coordinate system. (The negative sign indicates that the thrust direction is usually defined as the negative direction of the Z-axis of the aircraft):

[0068]

[0069] (2) Inverse dynamics solution to obtain the air resistance vector: Based on Newton's second law, the triaxial specific force vector measured by IMU is used. (i.e., accelerometer readings, including the acceleration due to motion and the component that counteracts gravity) and the real-time estimated load mass. Establish the force balance equation of the body and deduce the aerodynamic drag:

[0070]

[0071] This step utilizes the physical residuals of sensor data to separate the resultant external forces acting on the aircraft that are neither thrust nor gravity, namely air resistance.

[0072] (3) Analyze the composite airspeed vector: based on the aerodynamic drag formula To establish the mapping relationship between the drag vector and the airspeed vector, assuming the airframe is isotropic and simplifying through equivalent projected area, the magnitude of the airspeed vector is... With direction They are respectively:

[0073]

[0074] This leads to the airspeed vector in the body coordinate system:

[0075]

[0076] in The lumped aerodynamic drag coefficient, The air density is given; the equivalent drag coefficient of the machine body is given. The offline calibration method is as follows: control the aircraft to perform multiple sets of different speeds in a windless environment. Horizontal constant speed cruise, using IMU to collect the body balance tilt angle at each speed range. According to the horizontal force equilibrium equation , build about The linear regression model was used to fit and obtain the lumped aerodynamic drag coefficient of this aircraft model. .

[0077] (4) Coordinate system transformation and wind speed decoupling: using attitude rotation matrix (Convert the body coordinate system to the navigation coordinate system), project the airspeed vector onto the navigation coordinate system, and compare it with the ground speed vector measured by GPS. Perform vector subtraction to calculate the ambient wind speed vector:

[0078]

[0079] (5) Substitute the calculated wind speed vector into the parasitic power model to obtain the final correction term that includes the influence of the wind field:

[0080]

[0081] This is to compensate for the actual aerodynamic energy consumption under headwind, tailwind and crosswind conditions.

[0082] Step 2, offline calibration and solidification of model coefficients: Under known load, the aerodynamic parameters and basic power consumption parameters in the power model are identified by combining full envelope flight excitation and gradient descent parameter identification. After the accuracy is verified by an independent dataset, the calibration coefficients and battery internal resistance model parameters are stored in the flight controller.

[0083] Specifically, the aircraft is controlled to perform a combination of flight maneuvers, including variable load hovering, multi-speed cruise, and high-maneuver acceleration and deceleration, and data is collected. Using synchronized measured power and sensor state data, a mean squared error loss function is constructed to measure the deviation between the model's predicted power and the measured power. Subsequently, the gradient of the loss function with respect to each coefficient is calculated, and then adjusted according to the learning rate. The coefficients are updated iteratively until convergence. Non-negativity constraints are applied to the coefficients during the iteration process, and the globally optimal parameter set obtained after convergence is finally fixed into the flight controller.

[0084] Among them, the induced power coefficient The iterative update formula is:

[0085]

[0086] Parasitic power coefficient The iterative update formula is:

[0087]

[0088] base power The iterative update formula is:

[0089]

[0090] in, The first Second and third The next iteration , The first Second and third The next iteration , The first Second and third The next iteration , For learning rate, For the first Real-time power prediction values ​​in the measured sample set. For the first Real-time measured power values ​​in the set of measured samples. For the first The preset load mass in the group of measured samples, For the first The specific force in the group of measured samples For the first The composite airspeed vector in the set of measured samples.

[0091] Step 3, Joint State Estimation During Flight: During the transport mission, an extended Kalman filter is established, and an observer is constructed using the inverse kinematics principle. By combining IMU measurements with ground speed information, the current load mass and environmental wind field vector are decoupled and estimated in real time during dynamic flight.

[0092] Specifically, the extended Kalman filter (EKF) is defined as follows for real-time decoupling and solution:

[0093] (a) Define the system state vector: Couple the load mass and the three-axis wind speed vector in the navigation coordinate system into a unified state vector. , Let be the state vector at time t. These are the x, y, and z wind speed vectors at time t in the navigation coordinate system.

[0094] (b) Constructing a state prediction model: Assuming that in an extremely short sampling period Inside, the load mass and wind speed satisfy a random walk model, that is... ,in, Let be the state vector at time t-1. The process noise at time t-1;

[0095] (c) Establishing nonlinear observation equations :

[0096] Construct observation vectors using sensor measurements: , Let be the observation vector at time t. The measured real-time power value at time t. These represent the x, y, and z-axis ratios at time t in the navigation coordinate system; their observation functions are... Driven by a combination of power conservation and dynamic equilibrium:

[0097] , Let be the load mass at time t. Let be the triaxial specific force vector at time t. Let be the ground velocity vector at time t. Let be the wind speed vector at time t. Let be the current of the battery at time t. Let be the thrust vector of the aircraft at time t. Let be the air resistance vector at time t;

[0098] (d) Perform recursive correction: Calculate the Jacobian matrix of the observation function with respect to the state vector:

[0099] ,in, Let be the Jacobian matrix at time t;

[0100] Calculate Kalman gain The state vector is corrected in real time using measured power and acceleration residuals:

[0101] ,

[0102] in, for The posterior state estimation vector, Based on The one-step predicted state vector at time t. This is the Kalman gain.

[0103] (e) Based on the real-time corrected state vector, obtain the current load mass and wind speed vector.

[0104] It is worth noting that this process of decoupling the two variables in real time using the EKF method only exists in the first replanning cycle after the mission begins. Since the payload mass remains constant during independent flight, apart from determining the payload mass of the new mission cycle in the first replanning cycle, the real-time power can be accurately calculated throughout the entire mission cycle by simply calculating the wind field changes in real time using conventional methods in the remaining replanning cycles.

[0105] Step 4, Real-time solution of anisotropic sweet spot velocity: Based on the real-time wind speed vector and load mass calculated in Step 3, substitute them into the power model to solve for the optimal cruising ground speed that minimizes energy consumption per unit ground distance. This speed is dynamically adjusted with the angle between the heading and the wind direction.

[0106] Specifically, the optimal ground speed that minimizes energy consumption per unit ground distance (Wh / km) is numerically determined in the flight controller:

[0107]

[0108] in This is a scalar value for the optimal cruising ground speed; the formula implicitly includes the influence of wind: in headwinds, the algorithm automatically increases... To reduce loiter time; when tailwind, reduce It utilizes wind power for propulsion.

[0109] Step 5, Dynamic Remaining Support Distance Prediction and Closed-Loop Planning: Based on the real-time calculated load, wind field, and battery internal resistance status, calculate the remaining support distance of the aircraft under the current planned path; if the predicted distance is insufficient to cover the remaining tasks and meet the safety margin, adjust the transportation target according to the preset priority and replan the path.

[0110] Specifically, the calculation of the remaining support distance is based on the "energy bucket" model, and the specific formula is as follows:

[0111]

[0112] in, To use a sliding window filter for past The power benchmark, which is a smoothed version of the real-time power within a second, is used to filter out instantaneous fluctuations. The target cruise speed scalar.

[0113] Specifically, the expression for the target cruise speed scalar is:

[0114] ,

[0115] in, To achieve the required velocity scalar for piecewise time windows, This is a scalar value representing the optimal cruising ground speed.

[0116] Furthermore, based on the remaining path length and the deadline, the required velocity scalar for each segmented time window is calculated:

[0117] , Indexes for unfinished task points; This is the set of unfinished task points. From the current location to the unfinished task point The cumulative flight distance, To reach the unfinished task point The latest time constraint, For the current time, This is to allow for a time margin.

[0118] Specifically, the effective remaining energy The specific formula is:

[0119]

[0120] in, This represents the current remaining percentage of battery power. This is a warning line for the non-linear voltage drop of the battery. The rated capacity of the battery. The average discharge voltage of the battery. This represents the battery's discharge efficiency coefficient. The dimension conversion coefficient is... Reserved energy for return and landing.

[0121] Furthermore, the effective remaining energy The parameters that accurately reflect the aircraft's current true range are obtained as follows:

[0122] Battery rated capacity Initially, the nominal watt-hours (Wh) of the battery are used, and subsequent aging reduction corrections are made based on the cumulative charge-discharge cycle count and SOH (State of Health) curve recorded by the flight controller.

[0123] Discharge efficiency coefficient The calculation formula is: ,in The estimated current corresponding to the target cruising speed. This is the battery internal resistance under current operating conditions, which is used to deduct the Joule heat loss during high-current discharge.

[0124] Nonlinear warning threshold Based on the battery discharge characteristic curve measured offline, the voltage drop slope was selected. The inflection point value when the voltage exceeds a preset threshold (typically 15%) is used to eliminate the inefficient region under low voltage.

[0125] Forced energy reservation Defined as ,in The hovering power is based on real-time load estimation. A preset safe landing operation time (e.g., 120 seconds) is set to ensure that the aircraft has a sufficient vertical landing window before the battery runs out.

[0126] Based on the above definition, the "energy bucket" model constructed in this invention maps the traditional percentage of electricity to physically available joules of work, thereby improving the safety of remaining mileage prediction.

[0127] Specifically, if the predicted distance is insufficient to cover the remaining tasks and meet the safety margin, the transportation targets are adjusted according to preset priorities and the routes are replanned, including:

[0128] like And the duration exceeds the set time threshold. If the transportation task is deemed unsustainable, task points will be eliminated sequentially from low to high priority until the condition is met. And replan the route; among them, This represents the remaining support distance for the aircraft along its current planned path. The remaining total flight distance for the mission. The set safety distance margin is determined; otherwise, it is determined that the transport mission can continue to be carried out, and the aircraft continues to complete the transport mission according to the current planned path at the target cruising speed.

[0129] Furthermore, if task points are eliminated one by one, it still cannot satisfy the requirements. If this happens, an immediate return-to-base command will be triggered.

[0130] Step 6, Closed-loop control execution: Repeat steps 4 to 6 periodically to achieve closed-loop control for dynamic load estimation after deployment or load changes, supporting distance prediction and task replanning.

[0131] Specifically, the closed-loop control parameters are set as follows: EKF state estimation frequency: 10-50Hz, used to capture load and wind field dynamics in real time; energy bucket update frequency: 1-5Hz, used for integral calculation. Power smoothing window Confirmation time with replanning All times are 10-15 seconds, ensuring the robustness and stability of the decision-making system.

[0132] like Figure 2 As shown, this invention also proposes an intelligent planning system for aircraft transportation missions based on real-time power prediction. The system includes...

[0133] The power model building module is used to build a full physics power model of an aircraft that couples dynamic fields, wind fields, and battery characteristics.

[0134] The optimal cruising ground speed calculation module substitutes the current load mass and wind speed vector into the full physics power model of the aircraft to solve for the optimal cruising ground speed that minimizes energy consumption per unit ground distance.

[0135] The remaining support distance calculation module is used to calculate the remaining support distance of the aircraft under the current planned path;

[0136] The remaining task planning module is used to determine whether the remaining support distance is sufficient to cover the remaining tasks. If not, task points are removed in order of priority from low to high, and the route is replanned. If so, the transportation task is completed according to the currently planned route.

[0137] The technical solution of this invention enables sustainable mission planning:

[0138] ① Path tracking and speed control: The aircraft follows the currently planned trajectory, with... Flying at speed, among which The minimum speed scalar required to meet time constraints.

[0139] ② Layered task adjustment strategy: Tasks run in parallel, with task sequence priorities fine-tuned based on real-time status. Specifically, the layered task adjustment strategy includes:

[0140] a) Eliminate low-priority tasks: Eliminate tasks in order of their preset priority from low to high.

[0141] b) Replan the path: Based on the removed task point subset, recalculate and generate feasible paths.

[0142] c) Return immediately: If the minimum feasible subset of tasks still cannot be satisfied. If this occurs, an immediate return-to-base command will be triggered to ensure flight safety.

[0143] ③ Continuous status monitoring: Periodically return to step 1 to update power, status and distance predictions, forming a closed-loop monitoring system.

[0144] Step S4, Unsustainable task handling path: Initiate the hierarchical replanning strategy as described in claim 9:

[0145] In summary, this invention achieves core state perception and accurate prediction of aircraft load, environmental wind field, optimal speed, and remaining endurance. Utilizing these predictions, it enables dynamic assessment and autonomous decision-making regarding mission feasibility. These two aspects form a closed loop through continuous state monitoring, allowing the aircraft to adjust its strategy in real time in complex dynamic environments, intelligently degrade the mission, or safely return to base, significantly improving the autonomy, adaptability, and reliability of mission execution.

[0146] The technical solution of the present invention will be further described below with reference to examples.

[0147] A multi-rotor UAV is required to perform an emergency resupply and transportation mission, planning to deliver medical supplies to three mission points (A, B, and C). The total planned flight distance is 15 kilometers, and the information and time windows for each segment are shown in Table 1. Stable crosswinds exist in the flight area. The system is required to dynamically assess mission sustainability based on real-time battery power prediction and execute intelligent replanning.

[0148] Table 1. Flight Segment Information and Time Windows

[0149]

[0150] The "Intelligent Planning Method for Aircraft Transportation Mission Based on Real-time Power Prediction" of this invention is now used for mission execution simulation. The core parameters are as follows:

[0151] 1. Initial state and model parameters:

[0152] Battery rated capacity ,Voltage ,initial .

[0153] Safety threshold Forced energy reservation for corresponding time .

[0154] Offline calibrated power model coefficients: , , System efficiency .

[0155] Current load (including machine body) estimate .

[0156] Real-time wind speed sensing (3 m / s eastward, 5 m / s northward).

[0157] 2. Dynamic calculation and decision-making process:

[0158] 2.1 Solving for the sweet spot speed: Based on the current load and wind speed, substitute into the power model to solve for the speed with the lowest energy consumption per unit distance.

[0159]

[0160] 2.2 Energy Prediction: Calculate the available energy and remaining support distance.

[0161]

[0162] At this time, by Convert to Dimensional conversion coefficient The value is 3.6; assuming the average power demand under the current planned path is... ,but:

[0163]

[0164] 2.3 Task Continuity Assessment:

[0165] Total remaining distance for the mission Safety margin .

[0166] because The system determines that the task is sustainable.

[0167] aircraft cruise.

[0168] 3. Contingency Planning and Closed-Loop Replanning:

[0169] When the aircraft completes the mission point A (load reduced to...) On the way to mission point B, they encountered a persistent headwind, and the wind speed changed to The system recalculates:

[0170] ① New dessert speed (To counter headwinds).

[0171] ② New average power (Due to increased power consumption caused by headwind).

[0172] ③ Re-predict the remaining support distance .

[0173] ④ At this point, the remaining mission distance (B→C→base) is: .

[0174] Judgment conditions: The condition is not met (although it is still satisfied), but it is close to the threshold. The system startup status is continuously monitored, and vigilance is maintained.

[0175] If the headwind continues to intensify or the task is delayed, when the system detects... below 10 seconds At that time, hierarchical replanning is triggered.

[0176] a) Remove low-priority task point C.

[0177] b) The system replans the return route directly from mission point B, and the total distance of the new route is shortened to... .

[0178] c) After recalculation Confirmed. The system executes the new "Base → A → B → Base" path and sends a decision log to the control station stating "Due to energy constraints, mission point C delivery is canceled."

[0179] In this example, the "Intelligent Planning Method for Aircraft Transportation Missions Based on Real-Time Battery Prediction" is used for mission execution, which enables:

[0180] 1. In the initial stage of the mission, based on an accurate physical model and real-time state estimation, a sufficient support distance is predicted to support the original mission.

[0181] 2. When encountering sudden changes in the wind field during mission execution, energy risks can be detected in advance through real-time dynamic prediction.

[0182] 3. When the prediction reaches the safety threshold, the mission will be automatically downgraded (low-priority waypoints will be removed) and a feasible path will be replanned according to the preset priority, so as to maximize the mission completion rate while ensuring flight safety.

[0183] 4. Throughout the process, autonomous and intelligent operation was achieved from "state perception" → "energy prediction" → "decision judgment" → "closed-loop replanning", which significantly improved the robustness and intelligence of aircraft transportation missions in complex environments.

[0184] Finally, it should be noted that the above examples are only for clearly illustrating the technical solutions and effects of the present invention and should not be construed as limiting the present invention. Under the framework of the present invention, any technical solutions obtained by adaptively adjusting or combining parameters and strategies for different mission scenarios, flight platforms, and constraints should be considered to fall within the protection scope of the present invention.

[0185] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the modules described above can be referred to the corresponding process in the aforementioned method implementation, and will not be repeated here.

[0186] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0187] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in a combination of hardware and software functional modules.

[0188] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer system (which may be a personal computer, server, or network system, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0189] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for intelligent planning of aircraft transportation tasks based on real-time power prediction, characterized in that: include, Step 1: Construct a full-physics power model of the aircraft that couples the dynamic field, wind field, and battery characteristics: ， in, For real-time power, Based on maintaining power, Indicates the length of the modulus. It is a three-axis specific force vector. For the overall efficiency of the flight propulsion system, For load quality, For wind speed vectors, Ground velocity vector, This represents the real-time current of the battery. The basic internal resistance of the battery, This is the internal resistance compensation coefficient. This represents the remaining percentage of battery charge. For induced power coefficient, The parasitic power coefficient; Step 2: Substitute the current load mass and wind speed vector into the full physics power model of the aircraft to solve for the optimal cruising ground speed that minimizes energy consumption per unit ground distance. Step 3: Calculate the remaining support distance of the aircraft under the current planned path. If the remaining support distance is insufficient to cover the remaining tasks, the task points are eliminated in order of preset priority from low to high, and the path is replanned. Otherwise, the transportation task is completed according to the current planned path.

2. The method according to claim 1, characterized in that: Step 1 also includes a method combining full-envelope flight excitation and gradient descent parameter identification to evaluate the coefficients of the full-physics power model of the aircraft. Calibration is performed, including, Control the aircraft to perform a combination of flight maneuvers, including variable load hovering, multi-speed cruise, and high-G maneuvering acceleration and deceleration, and collect data. Using synchronized measured samples, a loss function is constructed to measure the mean square error between the real-time power prediction and the real-time measured power of the full physics power model of the aircraft. The gradient of the loss function with respect to each coefficient is then calculated, and the function is iteratively updated according to a set learning rate until convergence. The iterative update formula for each coefficient is as follows: ， ， ， in, The first Second and third The next iteration , The first Second and third The next iteration , The first Second and third The next iteration , For learning rate, For the first Real-time power prediction values ​​in the measured sample set. For the first Real-time measured power values ​​in the set of measured samples. For the first The preset load mass in the group of measured samples, For the first The specific force in the group of measured samples For the first The composite airspeed vector in the set of measured samples.

3. The method according to claim 1, characterized in that: The overall efficiency of the flight propulsion system The expression is: ， in, For ideal induced power, The total thrust generated by the rotor. air density, For rotor sweep area, This is the operating voltage.

4. The method according to claim 2, characterized in that: Step 2 utilizes an extended Kalman filter to jointly estimate the current load mass and wind field vector in real time, including: Define the state vector: , Let be the state vector at time t. These are the x, y, and z wind speed vectors at time t in the navigation coordinate system. Constructing a state prediction model: ,in, Let be the state vector at time t-1. The process noise at time t-1; Establish the nonlinear observation equation: ,in, Let be the observation vector at time t. , The measured real-time power value at time t. These represent the x, y, and z-axis ratios at time t in the navigation coordinate system. Let be the observation function at time t. , Let be the load mass at time t. Let be the triaxial specific force vector at time t. Let be the ground velocity vector at time t. Let be the wind speed vector at time t. Let be the current of the battery at time t. Let be the thrust vector of the aircraft at time t. Let be the air resistance vector at time t; Calculate the Jacobian matrix of the observation function with respect to the state vector: ,in, Let be the Jacobian matrix at time t; Calculate the Kalman gain and correct the state vector in real time: ， in, for The posterior state estimation vector, For based on The one-step predicted state vector at time t. Kalman gain; Based on the real-time corrected state vector, the current load mass and wind speed vector are obtained.

5. The method according to claim 1, characterized in that: The calculation of the optimal cruising ground speed in step 2 includes, ， in, This is a scalar value representing the optimal cruising ground speed.

6. The method according to claim 1, characterized in that: The calculation of the remaining support distance of the aircraft under the current planned path in step 3 includes, Obtain the effective remaining energy of the spacecraft under the current planned path. : ， in, This represents the current remaining percentage of battery power. This is a non-linear voltage drop warning line for the battery. The rated capacity of the battery. The average discharge voltage of the battery. This represents the battery's discharge efficiency coefficient. The dimension conversion coefficient is... Reserved energy for return and landing; Calculate the remaining support distance of the aircraft on the current planned path: ， in, This is a power reference obtained by smoothing real-time power measurements over a set time range using a sliding window filter. The target cruise speed scalar.

7. The method according to claim 1, characterized in that: Step 3 includes, like If the duration exceeds the set time threshold, the transportation task is determined to be unsustainable, and tasks are removed in ascending order of preset priority until the conditions are met. And replan the route; among them, This represents the remaining support distance for the aircraft along its current planned path. The remaining total flight distance for the mission. The set safety distance margin is determined; otherwise, it is determined that the transport mission can continue to be carried out, and the aircraft continues to complete the transport mission according to the current planned path at the target cruising speed.

8. The method according to claim 6 or 7, characterized in that: The expression for the target cruise speed scalar is: ， in, To achieve the required velocity scalar for piecewise time windows, , Indexes for unfinished task points; This is the set of unfinished task points. From the current location to the unfinished task point The cumulative flight distance, To reach the unfinished task point The latest time constraint, For the current time, To allow for a time margin, This is a scalar value representing the optimal cruising ground speed.

9. The method according to claim 7, characterized in that: If the task points are eliminated one by one and still cannot be satisfied... If this happens, an immediate return-to-base command will be triggered.

10. An intelligent planning system for aircraft transportation missions using the method described in any one of claims 1 to 9, characterized in that: include, The power model building module is used to build a full physics power model of an aircraft that couples dynamic fields, wind fields, and battery characteristics. The optimal cruising ground speed calculation module substitutes the current load mass and wind speed vector into the full physics power model of the aircraft to solve for the optimal cruising ground speed that minimizes energy consumption per unit ground distance. The remaining support distance calculation module is used to calculate the remaining support distance of the aircraft under the current planned path; The remaining task planning module is used to determine whether the remaining support distance is sufficient to cover the remaining tasks. If not, task points are removed in order of priority from low to high, and the route is replanned. If so, the transportation task is completed according to the currently planned route.

Images