Distributed supervision full-vector motion control method for chassis of electric civil aircraft tractor

Abstract

The invention discloses a distributed supervision full-vector motion control method for a chassis of an electric civil aircraft tractor. The method comprises the following steps: establishing a vehicle six-degree-of-freedom model, and constructing a distributed supervision controller to supervise subsystem states; and when the disturbance is large or the performance is reduced, the constraint condition is updated online according to the neural network controller and transmitted to the domain controller. The weight is modified by transmitting the mark signal combination, and the distributed supervision controller adjusts the working parameters of the sub-system actuator, so that the full-vector motion control of the chassis is realized. According to the invention, intervention and quit of neural network control under the change of the vehicle state and the external environment are realized. Constraint solution is applied to a domain controller, and overall solution realizes coordination among subsystems and overall performance optimization. The innovative method is efficient and reliable, adapts to complex and changeable working environments and is expected to promote new development of related fields.

Description

Technical Field

[0001] This patent relates to the technical field of electric civil aircraft tractors, and in particular to a distributed supervisory control method for achieving full vector motion control of the chassis. The method combines neural network control and sensor information feedback to improve the handling performance, safety and adaptability of the tractor. Background Art

[0002] Traditional civil aircraft tractor chassis control methods usually adopt a centralized control strategy, that is, the status information of all subsystems is aggregated to a central controller for decision-making. However, in the face of complex and changing working environments, this method often leads to problems such as slow system response and poor adaptability, affecting the operating efficiency and safety of aircraft tractors. Summary of the invention

[0003] In order to overcome the limitations of traditional chassis control methods, the present invention provides a distributed supervisory full vector motion control method for the chassis of an electric civil aircraft tractor. The method combines distributed supervisory control and neural network control technology, and has the ability to perceive vehicle status and external environment changes in real time, thereby optimizing the chassis control strategy and improving the handling performance and adaptability of the tractor. BRIEF DESCRIPTION OF THE DRAWINGS

[0004] Figure 1 It is a principle framework diagram of the method of the present invention;

[0005] Figure 2 It is a six-degree-of-freedom model diagram of the chassis of the electric civil aircraft tractor in the present invention;

[0006] Figure 3 It is the principle diagram of the distributed supervisory controller in the present invention;

[0007] Figure 4 It is a diagram of the control method of the present invention. DETAILED DESCRIPTION

[0008] The following are the implementation steps of the method of the present invention, including specific formulas and related sensors:

[0009] 1. Establish a six-degree-of-freedom model of the vehicle: First, according to the physical structure and kinematic characteristics of the electric civil aircraft tractor, establish a six-degree-of-freedom model of the vehicle, including positional degrees of freedom (x, y, z) and attitude degrees of freedom (roll, pitch, yaw). The model can be expressed by the following formula:

[0010] Position vector: P = [x, y, z]

[0011] Pose vector: Euler_angles = [roll, pitch, yaw]

[0012] 2. Construct a distributed supervisory controller and sensor layout: Based on the six-degree-of-freedom model of the vehicle, a distributed supervisory controller is constructed to monitor the status of each subsystem of the chassis. In addition, sensors such as gyroscopes, accelerometers, position sensors, and direction sensors are arranged to collect chassis status information in real time, including: In the implementation of this patent, the steering failure emergency driving method of the four-wheel full-vector aircraft tractor can be implemented through the following specific steps:

[0013] Gyroscope: measures the rate of change of the vehicle's attitude and provides angular velocity information.

[0010] Accelerometer: Measures the acceleration of the vehicle and provides linear acceleration information.

[0014] Position sensor: measures the position change of the vehicle and provides position information.

[0015] Direction sensor: measures the vehicle's heading angle changes and provides heading information.

[0016] 3. Neural network controller updates constraints online: When the system faces large disturbances or performance degradation, the neural network controller intervenes and uses sensor information to update constraints online. Such a control strategy can adapt to complex and changing working environments. The update formula of the neural network controller is:

[0017] Constraint update: U(t+1)=NN(P(t),Euler_angles(t),sensor_data(t))

[0018] Where NN stands for deep neural network, P(t) and Euler_angles(t) are the position and attitude vectors at the current time t, and sensor_data(t) is the sensor measurement data.

[0019] 4. Transfer the marker signal combination to adjust the weight: After the constraint update output by the neural network controller is combined with the marker signal, it is transferred to the distributed supervisory controller for intelligent adjustment of the controller weight. Such an adjustment mechanism can optimize the response performance of the system while maintaining the stability of the controller. The weight adjustment formula is:

[0020] Weight adjustment: Weights = Weights + α * (U (t + 1) + Tag)

[0022] Among them, Weights is the weight of the controller, α is the learning rate, and Tag is the tag signal combination.

[0023] 5. Distributed supervisory controller adjusts actuator operating parameters: The distributed supervisory controller adjusts the actuator working parameters of each chassis subsystem in real time according to the updated weights, such as motor speed, steering gear angle, etc., to achieve full vector motion control of the chassis. The actuator working parameter adjustment formula is:

[0024] Actuator working parameter adjustment: Actuator_Params(t+1)=Actuator_Params(t)+β*Weights(t+1)

[0025] Among them, Actuator_Params(t) is the actuator operating parameter at the current time t, and β is the proportional coefficient of the adjustment parameter.

[0026] 6. Intervention and exit of neural network control: According to the vehicle status and external environment changes, the neural network controller intelligently intervenes and exits chassis control to ensure the stability and reliability of the system under different working conditions.

[0027] 7. Constraint solving is applied to the domain controller: Constraint solving is applied to the domain controller to ensure coordination between subsystems and global performance optimization, further improving the performance of the chassis control system. The constraint solving formula is:

[0028] Constraints solving: Constraints = Solve (Actuator_Params)

[0029] Among them, Constraints is the constraint condition of the subsystem actuator, and Solve is the solution function.

Claims

1. A distributed supervised full vector motion control method for an electric civil aircraft tractor chassis, characterized in that The following steps are involved: a. Establish a six-degree-of-freedom model of the electric civil aircraft tractor, including positional degrees of freedom (x, y, z) and attitude degrees of freedom (roll, pitch, yaw); b. Construct a distributed supervisory controller to monitor the status of each subsystem of the electric civil aircraft tractor chassis and collect chassis status information in real time; c. Arrange sensors such as gyroscopes, accelerometers, position sensors and direction sensors to measure the vehicle's attitude change rate, acceleration, position and heading angle changes, and provide corresponding sensor data; d. The neural network controller intervenes and updates the constraints online based on the sensor data to adapt to the complex and changing working environment; e. After combining the constraint update output by the neural network controller with the marking signal, it is transmitted to the distributed supervisory controller for adjusting the weight of the controller; f. The distributed supervisory controller adjusts the actuator working parameters of each subsystem of the chassis according to the weights to achieve full vector motion control of the chassis; g. The neural network controller intelligently intervenes and exits chassis control according to vehicle status and external environment changes to ensure system stability and reliability; h. Constraint solving is applied to the domain controller to coordinate the movement between subsystems to achieve global performance optimization.

2. The distributed supervised full vector motion control method for the electric civil aircraft tractor chassis according to claim 1, wherein: The weight adjustment formula in the distributed supervisory controller is: Weight adjustment: Weights = Weights + α * (U (t + 1) + Tag) Among them, Weights is the weight of the controller, α is the learning rate, U(t+1) is the constraint update output by the neural network controller, and Tag is the tag signal combination.

3. The distributed supervised full vector motion control method for the chassis of an electric civil aircraft tractor according to claim 1, wherein: The constraint condition update formula in the neural network controller is: Constraint update: U(t+1)=NN(P(t),Euler_angles(t),sensor_data(t)) Among them, NN is a deep neural network, P(t) and Euler_angles(t) are the position and attitude vectors at the current time t, and sensor_data(t) is the sensor measurement data.

4. The distributed supervised full vector motion control method for the electric civil aircraft tractor chassis according to claim 1, wherein: The actuator working parameter adjustment formula in the distributed supervisory controller is: Actuator working parameter adjustment: Actuator_Params(t+1)=Actuator_Params(t)+β*Weights(t+1) Among them, Actuator_Params(t) is the actuator working parameter at the current time t, and β is the proportional coefficient of the adjustment parameter.

5. The distributed supervised full vector motion control method for the electric civil aircraft tractor chassis according to claim 1, wherein: Said Constraints solving: Constraints = Solve (Actuator_Params) Among them, Constraints is the constraint condition of the subsystem actuator, and Solve is the solution function.

Images