CN-121978934-A - Cross-domain bidirectional coupling optimization method and device based on hydrogen energy unmanned aerial vehicle

CN121978934ACN 121978934 ACN121978934 ACN 121978934ACN-121978934-A

Abstract

The invention relates to a cross-domain bidirectional coupling optimization method and a cross-domain bidirectional coupling optimization device based on a hydrogen unmanned aerial vehicle, which relate to the technical field of autonomous control of the hydrogen unmanned aerial vehicle and comprise the following steps of firstly, establishing a motion system and an energy system model of the hydrogen unmanned aerial vehicle; the method comprises the steps of constructing a safe flight reference track based on a Bezier curve method, tracking by utilizing a differential flat controller, acquiring real-time efficiency feedback of a fuel cell from an energy domain in the motion process, introducing a time reconstruction factor to construct on-line track re-optimization aiming at minimizing fuel consumption and tracking error, and finally tracking the reference track by utilizing self-adaptive equivalent hydrogen consumption minimum strategy by utilizing SOC global reference information from the motion domain, and completing on-line power distribution. The method can realize cross-domain bidirectional coupling collaborative optimization of the motion domain and the energy domain of the hydrogen unmanned aerial vehicle, improve the overall energy efficiency and the motion performance, and is suitable for the hydrogen unmanned aerial vehicle to execute multi-scene autonomous flight tasks.

Inventors

GUO XIAOYU
LIU GUOWEI
SONG XIAOWEI
DONG ZHEN
WANG CHENLIANG

Assignees

北京航空航天大学

Dates

Publication Date: 20260505
Application Date: 20260129

Claims (10)

1. The cross-domain bidirectional coupling optimization method based on the hydrogen energy unmanned aerial vehicle is characterized by comprising the following steps of: firstly, establishing a hydrogen energy unmanned aerial vehicle motion system and an energy system model, wherein the motion system model comprises unmanned aerial vehicle motion dynamics and required power models, and the energy system model comprises a hydrogen fuel cell and a lithium battery model; Second, based on the known map information and task requirements, A is used The algorithm and the Bezier curve parameterization method generate a safe and feasible reference flight track meeting dynamics constraint, and a differential flat controller is further adopted to complete tracking control of the reference flight track; Thirdly, in the autonomous movement process, performing online movement re-optimization of energy domain information coupling, namely, by acquiring real-time efficiency feedback of a fuel cell from an energy domain, introducing a time reconstruction factor as an optimization variable, and constructing online trajectory re-optimization aiming at minimizing fuel consumption and tracking error on the basis of a reference flight trajectory and a differential flat controller in the second step; And fourthly, performing on-line energy management of the motion domain information coupling to meet the flight energy requirement of the unmanned aerial vehicle, namely acquiring a globally optimal SOC reference curve based on the motion track of the third step of re-optimization by using a dynamic programming method, further tracking the SOC reference curve by adopting a self-adaptive equivalent hydrogen consumption minimum strategy, performing on-line energy management, and completing real-time power distribution of a hydrogen-electricity hybrid system in the hydrogen energy unmanned aerial vehicle.
2. The cross-domain bidirectional coupling optimization method based on the hydrogen energy unmanned aerial vehicle according to claim 1, wherein in the first step, a hydrogen energy unmanned aerial vehicle motion system and an energy system model are established, wherein the motion system model comprises unmanned aerial vehicle motion dynamics and required power models, the energy system model comprises a hydrogen fuel cell and a lithium battery model, and the method comprises the following specific implementation steps: Taking a four-rotor hydrogen energy unmanned aerial vehicle as an object, establishing a motion dynamics model: , Wherein, the 、 And Respectively representing a position vector, a speed vector and a gesture quaternion vector of the unmanned plane, Is the quality of the unmanned aerial vehicle, Is the gravitational acceleration vector in the world coordinate system, The thrust of the unmanned aerial vehicle is represented, Is the body coordinate system of the unmanned aerial vehicle in the current state The unit direction vector of the axis, Represents the triaxial moment vector of the unmanned plane, Represents the angular velocity vector of the unmanned aerial vehicle in the body coordinate system, For unmanned aerial vehicle inertia matrix, operator Representing an antisymmetric matrix determined by angular velocity, namely: , The required power model is established as follows: , Wherein, the Represent the first The power of the electric motor is calculated, Is an approximation of the total propulsion power of the unmanned aerial vehicle, Is the first The moment of the electric motor is set up in the form of a torque, Is a moment coefficient; The output model of the hydrogen fuel cell is: , Wherein, the And Is the output current and voltage of the hydrogen fuel cell stack, Indicating the open circuit voltage (open circuit voltage), Is the activation loss voltage, which is the voltage at which the activation is lost, Is the ohmic loss voltage and is used to control the voltage, Is the number of galvanic pile, A is the Tafil slope, Indicating the maximum allowable current value of the current, Representing the film resistance; The output model of the lithium battery is as follows: , Wherein, the Is the output voltage of the lithium battery, Is the output current of the lithium battery, Indicating the output power of the lithium battery, And The open circuit voltage and equivalent internal resistance of the lithium battery are determined by the characteristics of the lithium battery, Is the charge state of lithium battery, its change rate Is in a functional relation with the output current of the lithium battery, Is the charge capacity of lithium battery, simultaneously And Will be calculated by the formula of (2) Represented as And thus (2) Represented as Is a function of (2) 。
3. The cross-domain bi-directional coupling optimization method based on the hydrogen energy unmanned aerial vehicle according to claim 2, wherein in the second step, according to the known map information and task requirements, A is used The algorithm and the Bezier curve parameterization method generate a safe and feasible reference flight track meeting dynamics constraint, and further adopt a differential flat controller to complete tracking control of the reference flight track, and the specific implementation steps are as follows: based on the known environment map information and task requirements, A is adopted Generating an initial shortest safety path by an algorithm, and smoothing the initial safety path by utilizing a Bezier curve to generate a reference track meeting the motion requirement of the unmanned aerial vehicle; The reference track is expressed in the form of a segmented Bezier curve: , Wherein, the As a function of the time variable, For the number of track segments, Is the order of the polynomial trajectory, Is the first Segment No The vector of the individual control points is used, For the initially allocated time interval, i.e. time To the point of The reference track of the unmanned aerial vehicle in the period is the first The track of the segments is provided with a track, Is the Bernstein polynomial base, Is an independent variable, is taken from the above formula , Is a parameterized formula of the reference track; In view of the track smoothness requirement, track acceleration is selected as a smoothness index based on polynomial track microminiaturization and space decoupling characteristics, and a track smooth cost function is established: , Wherein, the Is on track of The Bezier curve expression in the direction shows that the cost function is expressed in quadratic form because each derivative of the Bezier curve is linearly expressed by the control point , wherein, For the sequence of control points, For a symmetric positive definite matrix, constraints related to track differential characteristics including track point position, speed and acceleration constraints expressed as linear equality or inequality constraints, and the optimization problem of track smoothing is converted into a quadratic programming problem: , Wherein, the Is a matrix of coefficients constrained by an equation, Is the right-hand vector of the equality constraint, Is a coefficient matrix constrained by an inequality, Is the right-hand vector of the inequality constraint; after solving and obtaining the reference track by using the method, tracking the track by adopting a differential flat controller: , Wherein, the A thrust vector is expected for the drone in world coordinate system, The current position, the speed and the acceleration state vectors of the unmanned aerial vehicle are respectively, Respectively the expected state vectors corresponding to the reference trajectories, And The expected attitude calculation formulas are as follows: , Wherein, the The three-axis unit direction vectors of the machine body coordinate system of the unmanned aerial vehicle under the expected gesture are respectively, In order to be a yaw angle, Is a world coordinate system when the world coordinate system is transformed to a desired machine body coordinate system Rotation of the shaft Obtained after that Unit direction vector of axis, symbol Defining a rotation matrix from a machine body coordinate system to a world coordinate system under the expected attitude of the unmanned aerial vehicle as The unmanned aerial vehicle expects angular velocity Sum angular acceleration The method comprises the following steps: , Wherein, the Is under the world coordinate system The axis unit direction vector is set, Is the desired yaw angle of the drone, The three-axis unit direction vector is the coordinate system of the current state machine body; Then, the drone control input is defined as: , Wherein, the The amount of thrust is desired for the unmanned aerial vehicle, Is a three-axis moment vector which is a three-axis moment vector, Is a rotation matrix from a body coordinate system to a world coordinate system in the current state of the unmanned aerial vehicle, Is the three-axis unit direction vector of the current state machine body coordinate system, Gain matrices for attitude error and angular velocity error respectively, Is a rotation matrix from the body coordinate system to the world coordinate system in the expected state of the unmanned aerial vehicle.
4. The method for optimizing cross-domain bi-directional coupling of a hydrogen energy unmanned aerial vehicle according to claim 3, wherein in the third step, in the autonomous movement process, on-line movement re-optimization of energy domain information coupling is performed, by acquiring real-time efficiency feedback of a fuel cell from an energy domain, introducing a time reconstruction factor as an optimization variable, and constructing on-line trajectory re-optimization targeting fuel consumption and tracking error minimization on the basis of a reference trajectory and a differential flat controller in the second step, wherein the method comprises the following specific implementation steps: Discretizing a motion dynamics equation and a reference track of the four-rotor hydrogen unmanned aerial vehicle to obtain It is discretized for the sampling time, and the discrete equation is expressed as: , Wherein, the Is the first The state vector of the step(s), Is the first The position vector of the unmanned aerial vehicle is obtained, First, the The speed vector of the step unmanned aerial vehicle, Is the first The quaternion vector of the unmanned aerial vehicle, Is the first Angular velocity vector of unmanned plane and system dynamics pass through fourth-order Dragon lattice tower numerical integration algorithm Update, the algorithm is based on the current state Control input Fixed sampling time Calculating the next state , Is the first The control input vector of the step is used, Is the first The step thrust is of a magnitude that is equal to the step thrust, Is the first The step moment vector is used for generating a step moment vector, Is the first obtained after discretizing the reference track Reference state vectors of steps, control input vectors being entirely determined by current state and reference state, i.e. control law ; By introducing time factors Scaling the sampling time, i.e. Reconstructing the speed, acceleration and other time attributes of the reference track, regarding the output power of the fuel cell as the required power, and based on the power, obtaining the real-time efficiency feedback of the fuel cell from the energy domain, wherein the approximate fuel consumption cost is expressed as follows: , Wherein, the Is the first The power required for the steps, i.e. the fuel cell power, Is the first The efficiency of the fuel cell is obtained by real-time feedback of the energy domain, Representing time scaled first Discrete time length of steps; thus, a multi-objective optimization problem is established that aims at minimizing tracking errors and fuel consumption: , Wherein, the Is a range of discrete times in which, Is the motion state vector of the kth step, Is the first The desired motion state vector of the step, Is the first The desired power of the step(s), Is the first The fuel cell efficiency of the step is that, Is the control input vector of the kth step of the unmanned aerial vehicle, A discretized form of the unmanned aerial vehicle dynamic equation after time scaling is represented; indicating that the tracking error is to be taken, The weight coefficient of the tracking error is the greater the weight is, the higher the tracking precision of the unmanned aerial vehicle on the reference track is; the energy weight coefficient is that the larger the weight is, the better the energy efficiency of the hydrogen energy unmanned aerial vehicle is; For punishing Is used in the present invention, Is a corresponding weight coefficient, the larger the weight is, the more conservative the algorithm is for time scaling; And Is the minimum and maximum value of the control input, reflects the saturation performance of the unmanned aerial vehicle actuator, And Representing boundary constraint of time factors, and ensuring that time is scaled within a reasonable range; and using IPOPT and other open source solvers to solve the nonlinear optimization problem and performing time reconstruction on the flight trajectory.
5. The cross-domain bidirectional coupling optimization method based on the hydrogen energy unmanned aerial vehicle according to claim 4, wherein in the fourth step, in order to meet the flight energy requirement of the unmanned aerial vehicle, the online energy management of the motion domain information coupling is performed, based on the motion track of the third step of re-optimization, a global optimal SOC reference curve is obtained by using a dynamic programming method, the SOC reference curve is further tracked by adopting a self-adaptive equivalent hydrogen consumption minimum strategy, the online energy management is performed, and the real-time power distribution of a hydrogen-electricity hybrid system in the hydrogen energy unmanned aerial vehicle is completed, and the method comprises the following specific implementation steps: according to the pre-re-optimization reference track, a global optimal SOC reference curve is obtained by using a dynamic programming method, and online energy management is performed by adopting an equivalent hydrogen consumption minimum strategy: , Wherein, the Is the total equivalent hydrogen mass consumption rate, including the hydrogen consumption rate of the hydrogen fuel cell and the equivalent hydrogen consumption rate of the lithium cell, Indicating the rate of hydrogen fuel mass consumption, Is the molar mass of the hydrogen and, Is the number of the electric pile pieces, Is the faraday constant of the device, Is the output current of the hydrogen fuel cell stack, Is the equivalent hydrogen consumption of the lithium battery, Is an equivalent factor to the number of the elements, Is the low heating value of the hydrogen gas, Representing the output model of a lithium battery, Is the power of the fuel cell and, Is the power of the lithium battery and is provided with a power supply, 、、 And The power constraints imposed on the fuel cell and the lithium cell, respectively, to ensure safe operation of the system, And Is the charge state of the lithium battery To maintain normal operation of the lithium battery, Representing the efficiency of the DC/DC converter, Is an estimate of the required power of the drone.
6. The cross-domain bi-directional coupling optimization method based on the hydrogen energy unmanned aerial vehicle of claim 5, wherein, Equivalent factor The on-line adjustment is realized by the following formula: , Wherein, the Is the first The equivalent factor of the steps is that, And Respectively the first Step by step A state and a reference state, the state, And updating an equation for the scale factor, thereby realizing tracking of the SOC reference curve.
7. Cross-domain bidirectional coupling optimizing device based on hydrogen energy unmanned aerial vehicle, characterized by comprising: The system comprises a model building module, a hydrogen energy unmanned aerial vehicle motion system and an energy system model, wherein the motion system model comprises unmanned aerial vehicle motion dynamics and required power models, and the energy system model comprises a hydrogen fuel cell and a lithium battery model; the track generation module is used for using A according to the known map information and task requirements The algorithm and the Bezier curve parameterization method generate a safe and feasible reference flight track meeting dynamics constraint, and a differential flat controller is further adopted to complete tracking control of the reference flight track; The re-optimizing module is used for executing online motion re-optimization of energy domain information coupling in the autonomous motion process, namely, by acquiring the real-time efficiency feedback of the fuel cell from the energy domain, introducing a time reconstruction factor as an optimization variable, and constructing online track re-optimization aiming at minimizing fuel consumption and tracking error on the basis of the reference flight track and the differential flat controller in the second step; And the energy management module is used for performing online energy management of the motion domain information coupling for meeting the flight energy requirement of the unmanned aerial vehicle, acquiring a globally optimal SOC reference curve based on a motion track of third step re-optimization by using a dynamic programming method, further tracking the SOC reference curve by adopting a self-adaptive equivalent hydrogen consumption minimum strategy, performing online energy management, and completing real-time power distribution of a hydrogen-electricity hybrid system in the hydrogen energy unmanned aerial vehicle.
8. A computing device comprising at least one processor and a memory storing program instructions that, when read and executed by the processor, cause the computing device to perform the hydrogen energy unmanned cross-domain bi-directional coupling optimization method of any of claims 1-6.
9. A readable storage medium storing program instructions, which when read and executed by a computing device, cause the computing device to perform the hydrogen energy unmanned aerial vehicle-based cross-domain bi-directional coupling optimization method of any of claims 1-6.
10. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements a cross-domain bi-directional coupling optimization method based on a hydrogen energy drone according to any one of claims 1-6.

Description

Cross-domain bidirectional coupling optimization method and device based on hydrogen energy unmanned aerial vehicle Technical Field The invention relates to the technical field of autonomous control (G05D 1/00) of a hydrogen unmanned aerial vehicle, in particular to a cross-domain bidirectional coupling optimization method and device based on the hydrogen unmanned aerial vehicle, which solve the problems of low energy efficiency and weak synergy of the hydrogen unmanned aerial vehicle in an autonomous flight process. Background Unmanned aerial vehicles show wide application prospects in the fields of industry, commerce and exploration. However, the limited energy density of the lithium battery carried by the traditional unmanned aerial vehicle restricts further development and application of the unmanned aerial vehicle in multi-scene and long-endurance autonomous tasks. Hydrogen fuel cells are considered to be ideal choices for new generation unmanned energy systems due to their high energy density and environmentally friendly nature. Recent technological advances in hydrogen unmanned aerial vehicles have fully validated their outstanding performance in terms of cruising ability, however, hydrogen unmanned aerial vehicles still face many challenges and difficulties, especially in terms of cooperative control of motion dynamics and hydrogen hybrid energy systems. It is worth noting that compared with the traditional unmanned aerial vehicle, the energy system structure of the hydrogen energy unmanned aerial vehicle is more complex, and the high dynamic characteristic of the flight scene puts severe demands on stable power supply of the power supply system. Therefore, it is needed to realize deep integration and cross-domain cooperation of a motion system and a power system at a system level so as to cope with complex dynamics and external environment and realize efficient, quick and durable autonomous flight of a hydrogen-powered unmanned aerial vehicle. Furthermore, based on the existing motion planning and energy management method, the design of the hydrogen energy unmanned aerial vehicle planning and management integrated method is of great importance, and is a key technology for realizing cross-domain coordination of a motion domain and an energy domain. Currently, research on motion planning and energy management of a hydrogen energy unmanned aerial vehicle mainly focuses on a layered optimization strategy of a motion layer and an energy layer, information interaction of the hydrogen energy unmanned aerial vehicle is more unidirectional transmission from a motion domain to an energy domain, and research on an integrated integration method is relatively deficient. The patent application with the application number of CN202411254543.6 provides a two-stage gear optimization control method for speed planning and energy management aiming at a net-linked hybrid vehicle, wherein the upper layer design considers the speed track optimization of the oil consumption of an engine, and transmits an optimization target to the lower layer for energy distribution and gear optimization. The patent application CN202411579882.1 proposes an integrated method for high-energy-efficiency planning and management based on a hydrogen unmanned aerial vehicle, by inputting track planning information into an energy layer to obtain an optimal battery state of charge (SOC) track, the energy layer allocates power between a fuel cell and a lithium cell accordingly. The patent application with the application number of CN202411198543.9 provides a track planning and energy management coupling optimization method based on a solar energy distribution map aiming at a solar energy hydrogen energy hybrid power flying vehicle, but because of high complexity of a coupling problem, a sequential quadratic programming algorithm is adopted for solving, the risk of converging on a local optimal solution exists, and the global optimality is difficult to ensure. Literature "Flight trajectory and energy management coupled optimization for hybrid electric UAVs with adaptive sequential convex programming method"（ discloses a hybrid unmanned aerial vehicle flight trajectory and energy management coupling optimization based on a self-adaptive sequential convex planning method), which is used for simultaneously optimizing the hybrid unmanned aerial vehicle flight trajectory and energy management. In summary, the existing method has the remarkable limitations that the existing method is limited by an offline solving mode and lacks practical application feasibility, or only considers unidirectional information transfer from a motion layer to an energy layer, ignores reverse constraint of an energy domain on motion performance, so that a motion planning result cannot match with an optimal state of the energy domain, and meanwhile, energy management is easy to fall into local optimization due to insufficient utilization of motion domain information and is diffic