Self-adaptive neural network control strategy based on preset performance

Abstract

The invention discloses a self-adaptive neural network control strategy based on preset performance, belongs to the field of intelligent control, and mainly aims at estimating and compensating unknown nonlinearity and external disturbance existing in a queue in queue control and ensuring that the whole queue realizes preset transient performance and steady-state performance. The method comprises the following steps: establishing a queue system model with unknown nonlinearity and external disturbance; establishing a tracking error; establishing a tunnel type preset performance boundary; establishing an unconstrained tracking error; establishing a queue sliding mode error; establishing a queue coupling sliding mode error; establishing an adaptive parameter estimation mechanism for external disturbance in the queue; establishing a neural network estimation mechanism for unknown nonlinearity in the queue; and establishing a queue sliding mode control strategy with preset performance based on an adaptive parameter estimation mechanism and a neural network estimation mechanism. The method is used for intelligent queue control.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent control, and mainly relates to an adaptive neural network control strategy based on preset performance. Background Art

[0002] In today's transportation system, the rapid increase in vehicle ownership has triggered a series of daunting challenges that urgently need to be addressed. Traffic congestion is worsening, energy consumption is skyrocketing, and exhaust emissions are placing a heavy burden on the ecological environment. Platoon control technology is crucial for improving traffic efficiency, enhancing traffic safety, and enhancing resource utilization. It is a key research direction and component of intelligent control systems and a key trend in the development of future intelligent transportation systems. Platoon control aims to ensure that platoons maintain a desired spacing and travel at a consistent and stable speed. This not only significantly improves traffic flow and enhances safety, but also effectively reduces energy consumption. However, current research on platoon control has certain limitations. First, there is a lack of constraints on the dynamic and transient performance of platoons, which can cause actual system performance to deviate from expectations and even fail to meet pre-determined performance indicators. Second, existing platoon disturbance observation techniques typically require prior knowledge of the disturbance's boundary values. However, in complex and changing real-world scenarios, this critical information is often difficult or even impossible to directly obtain, limiting the practical application of existing methods. In summary, to meet the higher standards and requirements for platoon control in future intelligent systems, existing platoon control methods require further innovation and improvement. Summary of the Invention

[0003] The purpose of this invention is to solve the problem that the lack of dynamic and transient performance constraints in queue control may cause the actual system performance to deviate from expectations or even fail to meet the preset performance indicators. The present invention provides an adaptive neural network control strategy based on preset performance.

[0004] An adaptive neural network control strategy based on preset performance is characterized by being able to estimate and compensate for unknown nonlinearities and external disturbances in the queue, and ensuring that the entire queue achieves preset transient and steady-state performance. The control strategy includes the following steps:

[0005] Step 1: Establish a queue system model with unknown nonlinearities and external disturbances:

[0006] Queue system model

[0007] (1)

[0008] in, is the serial number, and Representing the The position and speed of the vehicle, represents the tire radius, Indicates the transmission efficiency of the engine, Represents quality, represents the rolling resistance coefficient, is the control input, represents the aerodynamic drag coefficient, represents the acceleration due to gravity, Represents unknown nonlinearity.

[0009] Step 2: Establish tracking error:

[0010] Tracking Error

[0011] (2)

[0012] in, represents the distance between adjacent vehicles, represents the tracking error, Indicates the headway time, Indicates the resting distance.

[0013] Step 3: Establish tunnel-type preset performance boundaries:

[0014] Tunnel-type preset performance boundary

[0015] (3)

[0016] in, and It is a tunnel-type preset performance function. is a symbolic function, is the preset performance function, represents the natural constant, It's time, , , , , , , , .

[0017] Step 4: Establish unconstrained tracking error:

[0018] Unconstrained tracking error

[0019] (4)

[0020] in, is the unconstrained tracking error, , is a logarithmic function.

[0021] Step 5: Establish queue sliding mode error:

[0022] Sliding mode error

[0023] (5)

[0024] in, is the sliding mode error, is the integration variable, is the number of vehicles, .

[0025] Step 6: Establish queue coupling sliding mode error:

[0026] Coupled sliding mode error

[0027] (6)

[0028] in, is the coupled sliding mode error, .

[0029] Step 7: Establish an adaptive parameter estimation mechanism for external disturbances in the queue:

[0030] Adaptive parameter estimation mechanism

[0031] (7)

[0032] in, , , and is a constant greater than zero, , and They are and estimated value.

[0033] Step 8: Establish a neural network estimation mechanism for unknown nonlinearities in the queue:

[0034] Neural network estimation mechanism

[0035] (8)

[0036] in, , , , , and is a constant greater than zero, is the activation function, is the ideal weight, is the approximation error and satisfies , and They are and The estimated value of is the hyperbolic tangent function.

[0037] Step 9: Establish a queue sliding mode control strategy with preset performance based on the adaptive parameter estimation mechanism and the neural network estimation mechanism:

[0038] Queue Control Law

[0039] (9) in, , , .

[0040] The beneficial effect of this invention is that it effectively addresses the problem of a lack of constraints on dynamic and transient performance in queue control, which can lead to actual system performance deviating from expectations and even failing to meet preset performance indicators. It also estimates and compensates for unknown nonlinearities and external disturbances in the queue. This invention ensures that the entire queue achieves preset transient and steady-state performance, improving traffic flow, enhancing queue safety, and reducing energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Schematic diagram of the control method described in the first embodiment DETAILED DESCRIPTION

[0042] Specific implementation method 1: Combination Figure 1 This embodiment describes an adaptive neural network control strategy based on preset performance, and the control strategy includes the following steps:

[0043] Step 1: Establish a queue system model with unknown nonlinearities and external disturbances:

[0044] Queue system model

[0045] (1)

[0046] in, is the serial number, and Representing the The position and speed of the vehicle, represents the tire radius, Indicates the transmission efficiency of the engine, Represents quality, represents the rolling resistance coefficient, is the control input, represents the aerodynamic drag coefficient, represents the acceleration due to gravity, Represents unknown nonlinearity.

[0047] Step 2: Establish tracking error:

[0048] Tracking Error

[0049] (2)

[0050] in, represents the distance between adjacent vehicles, represents the tracking error, Indicates the headway time, Indicates the resting distance.

[0051] Step 3: Establish tunnel-type preset performance boundaries:

[0052] Tunnel-type preset performance boundary

[0053] (3)

[0054] in, and It is a tunnel-type preset performance function. is a symbolic function, is the preset performance function, represents the natural constant, It's time, , , , , , , , .

[0055] Step 4: Establish unconstrained tracking error:

[0056] Unconstrained tracking error

[0057] (4)

[0058] in, is the unconstrained tracking error, , is a logarithmic function.

[0059] Step 5: Establish queue sliding mode error:

[0060] Sliding mode error

[0061] (5)

[0062] in, is the sliding mode error, is the integration variable, is the number of vehicles, .

[0063] Step 6: Establish queue coupling sliding mode error:

[0064] Coupled sliding mode error

[0065] (6)

[0066] in, is the coupled sliding mode error, .

[0067] Step 7: Establish an adaptive parameter estimation mechanism for external disturbances in the queue:

[0068] Adaptive parameter estimation mechanism

[0069] (7)

[0070] in, , , and is a constant greater than zero, , and They are and estimated value.

[0071] Step 8: Establish a neural network estimation mechanism for unknown nonlinearities in the queue:

[0072] Neural network estimation mechanism

[0073] (8)

[0074] in, , , , , and is a constant greater than zero, is the activation function, is the ideal weight, is the approximation error and satisfies , and They are and The estimated value of is the hyperbolic tangent function.

[0075] Step 9: Establish a queue sliding mode control strategy with preset performance based on the adaptive parameter estimation mechanism and the neural network estimation mechanism:

[0076] Queue Control Law

[0077] (9) in, , , .

[0078] Effects of this implementation:

[0079] The proposed strategy effectively addresses the problem in queue control where a lack of constraints on dynamic and transient performance can lead to actual system performance deviating from expectations, or even failing to meet preset performance targets. It also estimates and compensates for unknown nonlinearities and external disturbances within the queue. This approach ensures that the entire queue achieves the preset transient and steady-state performance, improving traffic flow, enhancing queue safety, and reducing energy consumption.

Claims

1. An adaptive neural network control strategy based on preset performance, characterized in that: The control strategy can estimate and compensate for unknown nonlinearities and external disturbances in queue control and ensure that the entire queue achieves preset transient and steady-state performance. The control strategy includes the following steps: Step 1: Establish a queue system model with unknown nonlinearity and external disturbances; Step 2: Establish tracking error; Step 3: Establish tunnel-type preset performance boundaries; Step 4: Establish unconstrained tracking error; Step 5: Establish queue sliding mode error; Step 6: Establish queue coupling sliding mode error; Step 7: Establish an adaptive parameter estimation mechanism for external disturbances in the queue; Step 8: Establish a neural network estimation mechanism for unknown nonlinearities in the queue; Step 9: Establish a queue sliding mode control strategy with preset performance based on an adaptive parameter estimation mechanism and a neural network estimation mechanism.

2. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step 1, Queue system model (1) in, is the serial number, and Representing the The position and speed of the vehicle, represents the tire radius, Indicates the transmission efficiency of the engine, Represents quality, represents the rolling resistance coefficient, is the control input, represents the aerodynamic drag coefficient, represents the acceleration due to gravity, Represents unknown nonlinearity.

3. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step 2, Tracking Error (2) in, represents the distance between adjacent vehicles, represents the tracking error, Indicates the headway time, Indicates the resting distance.

4. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step three, Tunnel-type preset performance boundary (3) in, and It is a tunnel-type preset performance function. is a symbolic function, is the preset performance function, represents the natural constant, It's time, , , , , , , , .

5. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step 4, Unconstrained tracking error (4) in, is the unconstrained tracking error, , is a logarithmic function.

6. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step five, Sliding mode error (5) in, is the sliding mode error, is the integration variable, is the number of vehicles, .

7. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step six, Coupled sliding mode error (6) in, is the coupled sliding mode error, .

8. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step seven, Adaptive parameter estimation mechanism (7) in, , , and is a constant greater than zero, , and They are and estimated value.

9. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step eight, Neural network estimation mechanism (8) in, , , , , and is a constant greater than zero, is the activation function, is the ideal weight, is the approximation error and satisfies , and They are and The estimated value of is the hyperbolic tangent function.

10. The adaptive neural network control strategy based on preset performance according to claim 1, characterized in that: In the step nine, Queue Control Law (9) in, , , .

Images