High-precision automatic docking control method and system for airport scene boarding bridge

Abstract

The invention discloses a high-precision automatic docking control method and system for an airport scene boarding bridge, and relates to the field of automatic control of airport ground equipment. Aiming at the interference problems of flexible vibration and the like caused by variable structure inertia characteristics, nonlinear dynamic coupling and a composite environment long cantilever structure in the docking process of the boarding bridge, the method comprises the following steps: firstly, establishing a three-degree-of-freedom Euler-Lagrange dynamic model considering a telescopic coupling effect; secondly, constructing a data guide learning control framework, designing a smooth soft projection operator to process physical parameter constraints, and updating a physical parameter estimation value, a repetitive interference estimation value and a non-repetitive interference boundary on line by utilizing iterative domain process data; and finally, designing a control law in combination with an instruction filtering backstepping method, and introducing a robust damping term to suppress random interference.

Description

Technical Field

[0001] This invention belongs to the field of automated control technology for airport ground service equipment, specifically relating to a high-precision automatic docking control method and system for airport surface boarding bridges. Background Technology

[0002] Passenger boarding bridges (PBBs) are crucial hubs connecting terminals and aircraft, and their docking efficiency directly determines aircraft turnaround time. Currently, PBB docking primarily relies on manual operation, resulting in low efficiency and high risk. Achieving fully automated docking is a core requirement for smart airport construction. However, achieving high-precision automated docking faces the following key challenges:

[0003] 1. Variable Structure Dynamics: The boarding bridge is a massive cantilever structure. As the tunnel extends or retracts, its moment of inertia relative to the center of rotation (especially the yaw axis inertia) undergoes dramatic nonlinear changes. For example, the moment of inertia during extension may be several times greater than that during retraction.

[0004] 2. Strong nonlinear coupling: Due to the RPR (rotation-translation-rotation) configuration, the telescoping motion generates a huge Coriolis force, which strongly affects the stability of yaw and pitch motion.

[0005] 3. Actuator saturation risk: Existing optimization control methods (such as control barrier function CBF) often require extremely large control torque when the system approaches the safety boundary, which can easily lead to motor saturation and failure.

[0006] 4. Complex environmental interference: Airport aprons are subject to steady winds (repetitive interference) and gusts (random non-repetitive interference). Traditional PID control is difficult to balance low-speed stability and high-speed disturbance rejection. As a long cantilever structure, the boarding bridge is prone to low-frequency (0.1-0.5Hz) flexible modal vibration during movement, which affects the accuracy of the end effector.

[0007] In existing control schemes, traditional feedback control cannot adapt to large-scale parameter changes, while simple adaptive control ignores the highly repetitive nature of airport operations (same aircraft type, same parking position). Therefore, this invention proposes a "data-guided learning control" method. Summary of the Invention

[0008] The purpose of this invention is to solve the problems of insufficient accuracy and actuator saturation caused by time-varying parameters and interference in the automatic docking of boarding bridges, and to provide a control method and system that combines prior physical model data with iterative data.

[0009] The technical solution of the present invention is as follows:

[0010] A high-precision automatic docking control method for airport surface boarding bridges, based on the Eulerian-Lagrange modeling and command filtering backstepping framework, is fundamentally based on the introduction of a "data-guided" mechanism: that is, constructing a smooth projection operator using prior physical constraints and learning parameters using iterative data, including the following steps:

[0011] Step S1: Establish a dynamic model of the variable structure; including:

[0012] Establish a three-degree-of-freedom dynamic model of the boarding bridge system and define the generalized coordinate vector. ,in For the yaw angle of the rotating column, For the channel's telescopic length, The angle of the arrival port. Indicates the first Sub-dock iterations; the model includes variations in length as the length increases. Nonlinearly changing inertia matrix and Coriolis matrix The dynamic equations;

[0013] Step S2: Design a command filtering kinematic controller, including:

[0014] Define position tracking error Design of virtual control laws for kinematics A first-order command filter is introduced to obtain the filtered velocity reference signal. and its derivative Simultaneously construct the auxiliary compensation system status To compensate for filtering errors;

[0015] Step S3: Construct a linear parameterized regression matrix, including: defining the velocity tracking error based on the dynamic model. and the compensated speed error Transforming the system dynamics equations into linear parameterized form And calculate the regression matrix. Specific elements, among which The gravity vector Given a vector of physical parameters, and perform analytical calculations. It contains specific elements that exhibit the stretch-rotation coupling effect;

[0016] Step S4: Design the data-driven parameter update law, including: defining the smooth soft projection operator. By utilizing the error information on the iteration axis, we construct the physical parameter adaptive law, the repetitive interference learning law, and the non-repetitive interference boundary learning law, respectively, and update the physical parameter estimates in real time. Repetitive interference estimates Upper bound of non-repetitive interference ;

[0017] Step S5: Calculate and apply the control torque, including: based on the estimated value updated in step S4, combined with the compensated position error. and speed error Calculate the control torque including the robust damping term. This drives the movement of the boarding bridge joints.

[0018] A high-precision automatic docking control system for boarding bridges includes:

[0019] Sensing module: Includes high-precision rotary encoders installed at each joint of the boarding bridge and laser rangefinders or wire-type displacement sensors installed inside the passageway, used to collect yaw angles in real time. , telescopic length l and pitch angle Status data;

[0020] Storage module: Used to store preset reference trajectories And historical data generated during each iteration, including the generated physical parameter estimates. Repetitive interference estimates and boundary estimates ;

[0021] Calculation module: Configured to execute the control method described above, reads data from the sensing module and historical data from the storage module, and uses the data-guided learning law to calculate the control torque at the current moment. And write the updated parameters back to the storage module;

[0022] Drive module: includes servo motor driver, used to receive torque commands output by the calculation module and control the motor action of the boarding bridge rotation mechanism, telescopic mechanism and lifting mechanism.

[0023] The present invention has the following beneficial effects:

[0024] 1. Explicit variable inertia modeling: Explicitly including variable inertia in the controller design. Changing inertia matrix terms This compensates for the difference in gain requirements caused by changes in inertia from a physical perspective.

[0025] 2. Smooth soft projection adaptive law: designed The continuous smooth soft projection operator restricts the physical parameter estimates to a reasonable engineering range (e.g., the mass must be positive and have an upper bound), avoiding parameter drift in the adaptive process and thus preventing actuator saturation caused by control divergence.

[0026] 3. Complex Interference Suppression: Decompose the interference into a repetitive part (eliminated in the iteration domain by the learning law) and a non-repetitive part (eliminated by...). Robustness term in time-domain suppression). Parameters of the robust term. As the iteration count decreases, a control strategy of "seeking stability in the early stage and refining in the later stage" is achieved. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of an automated boarding bridge at an airport.

[0028] Figure 2 This is a kinematic equivalent diagram of an automated boarding bridge at an airport.

[0029] Figure 3 The figure shows the simulation results of the position tracking error converging with the number of iterations.

[0030] Figure 4 The waveform of the control torque input during the final iteration.

[0031] Figure 5 This is a complete tracking graph of the ideal trajectory during the final iteration. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other. To achieve the above objectives, this invention adopts the following technical solution.

[0033] This invention provides a high-precision automatic docking control method for airport surface boarding bridges, comprising the following steps: (See schematic diagram of airport surface boarding bridge as shown below) Figure 1 As shown in the diagram, the kinematic equivalent of the automated boarding bridge at the airport is as follows: Figure 2 As shown; Step S1. Dynamic modeling of the boarding bridge, including:

[0034] Define the generalized coordinates of the boarding bridge as follows: ,in For the yaw angle of the rotating column, For the channel's telescopic length, The angle of the arrival port. Indicates the first Sub-dock iterations; the model includes variations in length as the length increases. Nonlinearly changing inertia matrix and Coriolis matrix The dynamic equations;

[0035] The system dynamic equations are:

[0036] ;

[0037] in, It is a symmetric positive definite inertia matrix, whose diagonal elements contain elements that vary with the scaling length. Items that change ; For the Coriolis and centrifugal force matrices; It is the gravity vector; To control the input torque; The total disturbance includes the periodic flexible vibration component caused by the cantilever structure of the boarding bridge, and is decomposed into repetitive disturbances. Non-repetitive interference .

[0038] The explicit expression of the key matrix is ​​as follows:

[0039] Inertia matrix :

[0040] ;

[0041] The equivalent moment of inertia of the yaw axis; It is the equivalent moment of inertia about the yaw axis (Z-axis). Components: The moment of inertia of the fixed end connector (Rotunda) of the boarding bridge itself. The inherent rotational inertia component of the tunnel structure about the Z-axis. Moment of inertia of the pitch axis; Physical meaning: It is the moment of inertia of the tunnel / cabin assembly about the pitch axis (Y-axis).

[0042] Here This is the distance (radius of rotation) from the center of mass of the channel to the center of rotation at the current moment. Based on length, This represents the total mass of the telescopic aisle and docking bay assembly (telescopic end mass). This matrix indicates the yaw axis inertia ( ) The square and change.

[0043] Coriolis matrix :

[0044] ;

[0045] (Current iteration) (cosine value of pitch angle at time)

[0046] (Current iteration) (Sine value of pitch angle at time)

[0047] (Current iteration) (twice the sine of the pitch angle)

[0048] State Derivatives

[0049] These are the derivatives of the generalized coordinates with respect to time (i.e., the velocity terms), which play a crucial role in the Cleo force:

[0050] : Yaw rate.

[0051] Tunnel extension rate. This is a key term for generating the Coriolis force, because the mass distribution varies with... change.

[0052] : Cabin pitch rate.

[0053] in, This matrix accurately describes the coupling interference between the stretching motion and the rotational motion.

[0054] Step S2: Design a command filtering kinematic controller

[0055] Define position tracking error Design of virtual control laws for kinematics A first-order command filter is introduced to obtain the filtered velocity reference signal. and its derivative Simultaneously construct the auxiliary compensation system status To compensate for filtering errors;

[0056] Kinematic Loops and Command Filtering: Designing Virtual Control ;

[0057] It is a positive definite diagonal gain matrix. : Expected trajectory vector.

[0058] Introducing an instruction filter: ;

[0059] in, This is the filter bandwidth. This step avoids directly affecting... The problem of noise amplification caused by differentiation.

[0060] Step S3, Regression Matrix Construction

[0061] This is the core of adaptive control. The dynamic equations are written as... In the form of.

[0062] Based on the dynamic model of this embodiment, the regression matrix The third column (corresponding to unknown parameters) The specific derivation results are as follows:

[0063] ;

[0064] ;

[0065] ;

[0066] The instantaneous distance from the channel's centroid to the center of rotation. Defined as... ,in It is the base length. This is the current stretch length.

[0067] Actual tunnel extension length. , , Actual system velocities. These are feedback values ​​obtained through sensor measurements, corresponding to telescopic speed, yaw rate, and pitch rate, respectively. Note: In the regression matrix design for adaptive control, the Coriolis force term must use the actual velocities. The nonlinear coupling is calculated using a reference velocity () and the velocity vector multiplied by it is used. ).

[0068] (Current iteration) (cosine value of pitch angle at time)

[0069] (Current iteration) (Sine value of pitch angle at time)

[0070] (Current iteration) (twice the sine of the pitch angle)

[0071] : (The square of the current pitch angle cosine value). , , Reference velocity vector The three components; . : Yaw axis ( The reference angular velocity. Telescopic shaft ( The reference linear velocity. Pitch axis ( The reference angular velocity.

[0072] Reference acceleration vector The three components. The reference accelerations corresponding to the three axes mentioned above.

[0073] Gravitational acceleration

[0074] in, and These are the reference accelerations. and reference speed The amount.

[0075] Step S4, Data-Driven Update Law

[0076] Using smooth soft projection operator The operator at the physical boundary An external smoothing buffer was added to ensure the continuity of the derivative.

[0077] Update Law Design:

[0078] Physical parameters: Using the regression vector and the current velocity error Revise the previous round of estimates.

[0079] Repetitive interference: The error is directly integrated and learned to offset structural friction and steady wind load.

[0080] Non-repetitive interference boundary: The amplitude envelope of random disturbances is estimated in a monotonically increasing manner.

[0081] Step S5: Control Torque Synthesis

[0082] Final output torque It consists of five parts:

[0083] Model compensation terms: ;

[0084] Repeated interference cancellation term: ;

[0085] Feedback calming items: ;

[0086] Robust inhibition term: .in As the number of iterations increases Increase, The function becomes steeper, approximating the sign function, thereby improving the ability to suppress minute disturbances and achieving sub-centimeter level accuracy.

[0087] In step S3, based on the dynamic model, the velocity tracking error is defined. and the compensated speed error Transforming the system dynamics equations into linear parameterized form And calculate the regression matrix. Specific elements, among which The gravity vector Given a vector of physical parameters, and perform analytical calculations. It contains specific elements that exhibit the stretch-rotation coupling effect;

[0088] Regression Matrix The formula for calculating the elements is as follows: in, Reference acceleration vector The amount; Reference velocity vector The amount;

[0089] The coupling term elements are specifically:

[0090] ;

[0091] ;

[0092] ;

[0093] in, , , , This is the acceleration due to gravity.

[0094] In step S4, design a data-driven parameter update law.

[0095] Define the smooth soft projection operator By utilizing the error information on the iteration axis, we construct the physical parameter adaptive law, the repetitive interference learning law, and the non-repetitive interference boundary learning law, respectively, and update the physical parameter estimates in real time. Repetitive interference estimates Upper bound of non-repetitive interference ;

[0096] In step S5, the control torque is calculated and applied.

[0097] Based on the estimated value updated in step S4, combined with the compensated position error and speed error Calculate the control torque including the robust damping term. This drives the movement of the boarding bridge joints.

[0098] The core of this invention lies in the "data guidance" mechanism:

[0099] Prior physical data: Using engineering parameters such as the mass range of the boarding bridge and the upper limit of motor torque, a smoothing soft projection operator is designed. This operator ensures that the parameter estimates are always within the physically feasible range, fundamentally preventing control gain divergence and actuator saturation.

[0100] Iterative process data: Using historical iterative data, the learning law gradually "memorizes" and counteracts repetitive flexible vibrations and friction.

[0101] Simulation Examples

[0102] The control system is deployed based on a distributed architecture:

[0103] Sensing module: Uses an absolute encoder to collect data. Data is collected using laser ranging or wire sensors. .

[0104] Calculation module: can be an industrial PC (IPC) or an embedded controller (such as ARM / FPGA), running a data-guided learning control algorithm, with a calculation cycle set to 10ms.

[0105] Drive module: frequency converter or servo drive, operating in torque control mode.

[0106] To verify the effectiveness of this invention, a boarding bridge model with the following parameters was constructed in a simulation environment:

[0107] Physical parameters: Total mass kg; yaw inertia kg·m²; pitch inertia kg·m²; base length m.

[0108] Controller gain: , , where diag represents a diagonal matrix.

[0109] Filter bandwidth: rad / s.

[0110] Interference settings:

[0111] Repetitive disturbances: simulating periodic structural vibrations. Nm. And the periodic flexible vibration caused by the cantilever structure. In the simulation, this flexible disturbance is modeled as The structural modes were simulated at a frequency of 0.2 Hz.

[0112] Non-repetitive disturbance: Simulated random gusts, amplitude range N / Nm.

[0113] Soft projection boundary: set on the true parameters Within the range.

[0114] Simulation results show that:

[0115] Convergence: After 60 iterations, the maximum norm of the position tracking error converged from the initial 0.25m to 0.005m (i.e., 5 millimeters). Figure 3 As shown, it meets the sub-centimeter precision requirements for boarding bridge docking.

[0116] Disturbance resistance: The system remained stable even under a strong random wind load of 20 kN, proving its stability. Robust terms and boundary estimates The effectiveness.

[0117] Safety: Smooth control torque output, such as... Figure 4 As shown, it is always within the normal operating range of the actuator (such as yaw torque). (kNm), without the high-frequency jitter or numerical explosion common in traditional sliding mode control, verifying the suppressive effect of the soft projection operator on parameter drift, such as Figure 5 As shown.

Claims

1. A high-precision automatic docking control method for airport surface boarding bridges, characterized in that, Includes the following steps: Step S1: Establish a dynamic model of the variable structure; including: Establish a three-degree-of-freedom dynamic model of the boarding bridge system and define the generalized coordinate vector. ,in For the yaw angle of the rotating column, For the channel extension length, The angle of the arrival port. Indicates the first Sub-dock iterations; the model includes variations in length as the length increases. Nonlinearly changing inertia matrix and Coriolis matrix The dynamic equations; Step S2: Design a command filtering kinematic controller, including: Define position tracking error Design of virtual control laws for kinematics A first-order command filter is introduced to obtain the filtered velocity reference signal. and its derivative Simultaneously construct the auxiliary compensation system status To compensate for filtering errors; Step S3: Construct a linear parameterized regression matrix, including: defining the velocity tracking error based on the dynamic model. and the compensated speed error Transforming the system dynamics equations into linear parameterized form And calculate the regression matrix. Specific elements, among which The gravity vector Given a vector of physical parameters, and perform analytical calculations. It contains specific elements that exhibit the stretch-rotation coupling effect; Step S4: Design the data-driven parameter update law, including: defining the smooth soft projection operator. By utilizing the error information on the iteration axis, we construct the physical parameter adaptive law, the repetitive interference learning law, and the non-repetitive interference boundary learning law, respectively, and update the physical parameter estimates in real time. Repetitive interference estimates Upper bound of non-repetitive interference ; Step S5: Calculate and apply the control torque, including: based on the estimated value updated in step S4, combined with the compensated position error. and speed error Calculate the control torque including the robust damping term. This drives the movement of the boarding bridge joints.

2. The method according to claim 1, characterized in that, The dynamic model described in step S1 is as follows: ； in, It is a symmetric positive definite inertia matrix, whose diagonal elements contain elements that vary with the scaling length. Items that change ; For the Coriolis and centrifugal force matrices; It is the gravity vector; To control the input torque; The total disturbance includes the periodic flexible vibration component caused by the cantilever structure of the boarding bridge, and is decomposed into repetitive disturbances. Non-repetitive interference .

3. The method according to claim 1, characterized in that, The regression matrix mentioned in step S3 The formula for calculating the elements is as follows: ,in, Reference acceleration vector The amount; Reference velocity vector The amount; The coupling term elements are specifically: ； ； ； in, , , , This is the acceleration due to gravity.

4. The method according to claim 1, characterized in that, The smoothing soft projection operator described in step S4 Defined as: ； in, The high-confidence physical constraint interval is defined by parameters; the operator has a continuous first derivative at the boundary.

5. The method according to claim 4, characterized in that, The update law in step S4 is as follows: Physical parameter update law: ; Repetitive Disturbance Update Law: ; Non-repetitive disturbance boundary update law: ; in, Positive learning gain This is the compensated speed tracking error.

6. The method according to claim 1, characterized in that, The control torque described in step S5 The calculation formula is: ； in, For the feedback gain matrix, To compensate for the position error, It represents the Hadamah accumulation. These are robust control parameters that decay with the number of iterations.

7. A high-precision automatic docking control system for boarding bridges, characterized in that, include: Sensing module: Includes high-precision rotary encoders installed at each joint of the boarding bridge and laser rangefinders or wire-type displacement sensors installed inside the passageway, used to collect yaw angles in real time. , telescopic length l and pitch angle Status data; Storage module: Used to store preset reference trajectories And historical data generated during each iteration, including the generated physical parameter estimates. Repetitive interference estimates and boundary estimates ; Calculation module: configured to execute the control method as described in any one of claims 1 to 6, read data from the sensing module and historical data from the storage module, and calculate the control torque at the current moment using a data-guided learning law. And write the updated parameters back to the storage module; Drive module: includes servo motor driver, used to receive torque commands output by the calculation module and control the motor action of the boarding bridge rotation mechanism, telescopic mechanism and lifting mechanism.

8. An electronic device, characterized in that, include: One or more processors; A memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed by a processor, cause the processor to perform the method described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the method described in any one of claims 1 to 6.

Images