An adaptive neural network course tracking control method for unmanned surface vehicle

Abstract

The application provides an unmanned surface vehicle adaptive neural network heading tracking control method, comprising: establishing a three-order response type unmanned surface vehicle heading control mathematical model according to the surrounding environment and the sea state information of surrounding other ships; introducing a linear input quantization analysis model to quantize the control input in the control system; based on a dynamic surface control algorithm, a Lyapunov function is designed to obtain the heading controller of the unmanned surface vehicle; based on the Lyapunov stability theory, it is proved that the stability of the designed unmanned surface vehicle adaptive neural network heading tracking control system with input quantization does not require prior information of fixed quantization parameters, all signals in the closed-loop system are semi-global uniformly bounded, and the tracking error can be made arbitrarily small by adjusting the controller parameters. The technical scheme of the application can improve the heading control performance of the unmanned surface vehicle and has greater universality and adaptability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular, and more particularly, to a self-adaptive neural network heading tracking control method for unmanned surface vehicles. BACKGROUND

[0002] With the rise of intelligent maritime transportation, unmanned surface vehicles (USVs) have become an important part of global maritime operations. Its strong maneuverability, advanced autonomous navigation capability, ability to carry multi-functional loads, and ability to perform monitoring, rescue and remote dangerous tasks in vast sea areas have made it a key player in maritime transportation tasks.

[0003] In complex and variable marine environments, maintaining the stability of the ship and precisely controlling its heading is one of the main challenges for unmanned surface vehicles to achieve autonomous navigation. Currently, most research is focused on designing efficient and applicable control strategies to improve the tracking stability and accuracy of unmanned surface vehicles. However, since the various components of the unmanned surface vehicle rely on communication channels to transmit information, and the bandwidth of the communication channel is limited, there is a problem of insufficient communication resources. By introducing input quantization technology before signal transmission, the working frequency of the actuator can be effectively reduced, and the communication rate within the signal transmission bandwidth can also be reduced. Therefore, in the study of heading control of unmanned surface vehicles, the application of input quantization technology is gradually receiving widespread attention.

[0004] However, simply relying on input quantization technology for heading control cannot fully meet the development needs of unmanned surface vehicles. On the one hand, since the ship's rudder system cannot achieve step steering, ignoring the characteristics of the rudder will affect the performance quality of the control system. Therefore, from a practical application point of view, in order to obtain good heading control performance, the characteristics of the rudder servo system should also be considered. On the other hand, after introducing the linear input quantization analysis model, the dynamic surface control (DSC) algorithm is used to solve the "differentiation explosion" problem existing in the traditional backstepping method. In addition, effective smoothing of the quantized signal can also be achieved. SUMMARY

[0005] In view of the above-mentioned technical problems of limited communication bandwidth of unmanned surface vehicle on the sea and collision caused by position error, an adaptive neural network heading tracking control method of unmanned surface vehicle with actuator input quantization and ship rudder power characteristics is provided, aiming at solving the model uncertainty in the control system and the external disturbance in the heading control. The linear model is mainly used to describe the quantization process, which effectively reduces the execution frequency of the actuator, thereby saving the communication resources and reducing the communication pressure. In addition, the system adaptive control law is designed, which successfully reduces the tracking error to the minimum, ensuring that the unmanned surface vehicle can accurately sail along the predetermined route. The Lyapunov stability theory is used to prove that all signals in the closed-loop system are semi-global uniformly bounded.

[0006] The technical means adopted by the present application are as follows:

[0007] An adaptive neural network heading tracking control method of unmanned surface vehicle comprises:

[0008] S1, a three-order response type unmanned surface vehicle heading control mathematical model is established according to the sea state information of the surrounding environment and other ships around;

[0009] S2, a linear input quantization analysis model is introduced to quantize the control input in the control system;

[0010] S3, based on the dynamic surface control algorithm, a Lyapunov function is designed to obtain the heading controller of the unmanned surface vehicle;

[0011] S4, based on the Lyapunov stability theory, it is proved that the stability of the adaptive neural network heading tracking control system of the unmanned surface vehicle with input quantization is designed without the prior information of fixed quantization parameters, and all signals in the closed-loop system are semi-global uniformly bounded, and the tracking error is made to be arbitrarily small by adjusting the controller parameters.

[0012] Further, step S1 specifically comprises:

[0013] S11, a ship nonlinear Nomoto model and a ship rudder power characteristic model are constructed, as follows:

[0014]

[0015] Wherein, ψ represents the heading angle of the ship, δ represents the rudder angle, K and T represent the ship maneuverability index, α and β represent the nonlinear parameters, δ E represents the command rudder angle of the rudder, K E represents the rudder control gain, T E represents the time constant of the rudder.

[0016] S12, introduce input quantization, construct a third-order response type mathematical model of unmanned surface vehicle, as follows:

[0017]

[0018] wherein, ω1, ω2 represent unknown disturbances caused by wind, flow and wave.

[0019] Further, step S2 specifically comprises:

[0020] S21, for the control input of the ship model considering input quantization, let Q(u) = q1(t)u + q2(t), take:

[0021]

[0022] S22, since the quantization process is symbol invariant, q1(t) is unknown, from the formula of step S21, q1(t) > 0; when |u(t)| < a, Q(u) is bounded, q1(t) = 1, then q2(t) is bounded, take

[0023] Further, step S3 specifically comprises:

[0024] S31, define the heading error as: S1 = x1-x ld , S2 = x2-x 2d , S3 = x3-x 3d , and derive the heading error respectively, to obtain:

[0025]

[0026] S32, define Lyapunov function, as follows:

[0027]

[0028] wherein, y i represents the error of virtual control term,

[0029] S33, define a first-order low-pass filter, as follows:

[0030]

[0031] wherein,

[0032] S34, since f(x, t) is unknown, construct RBF neural network to approximate the unknown function of the system model, there is an ideal weight vector θ * ∈R N , so that the neural network θ *Th(x) is close enough to a given function f and the absolute value of the approximation error is not greater than σ M That is,

[0033] f(x) = θ *T h(x) + ε *

[0034] S35, according to Young's inequality, derive V1, V2 in step S32, get:

[0035]

[0036] S36, define time-varying gain: Wherein, q1(t) min is the lower bound of q1(t);

[0037] S37, design Lyapunov function, as follows:

[0038]

[0039] And derive the Lyapunov function designed in step S37, get:

[0040]

[0041] S38, according to step S36 and step S37, design system control rate and adaptive rate, as follows:

[0042]

[0043] Wherein,

[0044] Further, step S4, specifically includes:

[0045] S41, from the Lyapunov function designed in step S32, get the quantization part, as follows:

[0046]

[0047] Take Wherein, l>0;

[0048] S42, put According to Young's inequality, get:

[0049]

[0050] S43, since Get Let So get According to Also get

[0051] S44. From steps S35 and S37, we obtain:

[0052]

[0053] S45. Based on Young's inequality and other inequalities get:

[0054]

[0055]

[0056] Where, λ max (·) represents the largest eigenvalue of ·;

[0057] S46, Take Then we have:

[0058]

[0059] The control parameters are selected as follows:

[0060]

[0061] Where l represents the parameter to be designed;

[0062] S47, due to Thus obtain Substituting, we get:

[0063]

[0064] S48, by It can be known that:

[0065] It has a maximum value, denoted as Q; select Therefore, we have:

[0066]

[0067] S49. Taking all factors into consideration, due to Then V is bounded, and solving the above inequality yields:

[0068]

[0069] Clearly, all signals in the closed-loop system are semi-globally bounded, when hour Then t→∞,

[0070] Compared with the prior art, the present invention has the following advantages:

[0071] 1. The application provides a kind of unmanned surface vehicle self-adapting neural network heading tracking control method, with actuator input quantization and ship rudder power characteristics, and considering the characteristics of ship rudder servo system with input quantization, solve the problem that general ship rudder system cannot realize step steering.

[0072] 2. The application provides a kind of unmanned surface vehicle self-adapting neural network heading tracking control method, introduces a kind of linear input quantization analysis model, which ensures that controller design does not depend on any prior known quantization parameter information.Compared with the fixed quantization parameter used in the controller design process of traditional quantization control method, the control method of the application has greater universality and adaptability.

[0073] 3. The application provides a kind of unmanned surface vehicle self-adapting neural network heading tracking control method, adopts dynamic surface control (DSC) algorithm, solves the "derivative explosion" problem existing in traditional backstepping control.In addition, it can realize the effective smoothing of quantization signal.

[0074] Based on the above reasons, the application can be widely popularized in the field of artificial intelligence. BRIEF DESCRIPTION OF DRAWINGS

[0075] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed to be used in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0076] Figure 1 The method flowchart of the application.

[0077] Figure 2 The simplified block diagram of the adaptive neural network control system with actuator input quantization and ship rudder power characteristics provided by the application.

[0078] Figure 3 The tracking result graph of USV heading angle, USV yaw angle velocity and USV rudder angle provided by the embodiment of the application.

[0079] Figure 4 The tracking error result graph of USV heading angle, USV yaw angle velocity and USV rudder angle provided by the embodiment of the application.

[0080] Figure 5 The control input curve comparison graph provided by the embodiment of the application. DETAILED DESCRIPTION

[0081] In the following, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application, so that those skilled in the art can better understand the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work should fall within the scope of protection of the present application.

[0082] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0083] As shown in Figure 1 The present application provides an unmanned surface vehicle adaptive neural network heading tracking control method with actuator input quantization and ship rudder power characteristics, comprising:

[0084] S1, according to the sea state information of the surrounding environment and the surrounding other ships, a third-order response type unmanned surface vehicle heading control mathematical model is established;

[0085] S2, a linear input quantization analysis model is introduced to quantize the control input in the control system;

[0086] S3, based on the dynamic surface control algorithm, a Lyapunov function is designed to obtain the heading controller of the unmanned surface vehicle;

[0087] S4, based on the Lyapunov stability theory, it is proved that the stability of the designed unmanned surface vehicle adaptive neural network heading tracking control system with input quantization is achieved without the prior information of fixed quantization parameters, and all signals in the closed-loop system are semi-global uniformly bounded, and the tracking error is made to be arbitrarily small by adjusting the controller parameters.

[0088] In specific implementation, as a preferred embodiment of the present application, step S1 specifically comprises:

[0089] S11, a ship nonlinear Nomoto model and a ship rudder power characteristic model are constructed, as follows:

[0090]

[0091] wherein, ψ represents the ship heading angle, δ is the rudder angle, K, T represent the ship motion maneuverability index, α, β represent the nonlinear parameters, δ E represents the rudder angle issued by the rudder, K E represents the rudder control gain, T E represents the rudder time constant.

[0092] S12, introduce input quantization, construct a third-order response type mathematical model of the unmanned surface vehicle, as follows:

[0093]

[0094] wherein, ω1, ω2 represent unknown disturbances caused by wind, flow and wave.

[0095] In specific implementation, as a preferred embodiment of the present application, step S2 specifically comprises:

[0096] S21, for the control input in the ship model considering input quantization, let Q(u)=q1(t)u+q2(t), take:

[0097]

[0098] S22, since the quantization process is symbol invariant, q1(t) is unknown, from the formula of step S21, q1(t)>0; when |u(t)|<a, Q(u) is bounded, q1(t)=1, then q2(t) is bounded, so take

[0099] In specific implementation, as a preferred embodiment of the present application, step S3 specifically comprises:

[0100] S31, define the heading error as: S1=x1-x ld , S2=x2-x 2d , S3=x3-x 3d , and derive the heading error respectively, to obtain:

[0101]

[0102] S32, define the Lyapunov function, as follows:

[0103]

[0104] wherein, y i represents the error of virtual control item,

[0105] S33, define a first-order low-pass filter as follows:

[0106]

[0107] wherein,

[0108] S34, since f(x, t) is unknown, construct a RBF neural network to approximate the unknown function of the system model, there is an ideal weight vector θ * ∈R N , so that the neural network θ *T h(x) is sufficient to approximate the given function f and the approximation error absolute value is not greater than σ M , that is:

[0109] f(x) = θ *T h(x) + ε *

[0110] S35, according to Young inequality, derive V1, V2 in step S32, get:

[0111]

[0112] S36, define time-varying gain: wherein, q1(t) min is the lower bound of q1(t);

[0113] S37, design Lyapunov function as follows:

[0114]

[0115] and derive the Lyapunov function designed in step S37, get:

[0116]

[0117] S38, according to step S36 and step S37, design system control rate and adaptive rate as follows:

[0118]

[0119] wherein,

[0120] In specific implementation, as a preferred embodiment of the present application, step S4 specifically comprises:

[0121] S41, from the Lyapunov function designed in step S32, get the quantization part as follows:

[0122]

[0123] take where l > 0;

[0124] S42, put into, according to Young inequality, get:

[0125]

[0126] S43, because get Let Thus get According to Also get

[0127] S44, from step S35 and step S37, get:

[0128]

[0129] S45, according to Young inequality and inequality get:

[0130]

[0131] where λ max (·) represents the maximum eigenvalue of ·;

[0132] S46, take 0 < η0≤ 1.0, then:

[0133]

[0134] Control parameter selection as follows:

[0135]

[0136] where l represents the parameter to be designed;

[0137] S47, because Thus get Substitute get:

[0138]

[0139] S48, from know:

[0140] There is a maximum value, denoted by Q; select Therefore, then:

[0141]

[0142] S49, comprehensive consideration, because ​Then V is bounded, and solving the above inequality yields:

[0143]

[0144] Clearly, all signals in the closed-loop system are semi-globally bounded, when hour Then t→∞,

[0145] Example

[0146] To verify the effectiveness of the present invention, this embodiment uses MATLAB for computer simulation research, with the parameters set as follows:

[0147] The simulation object is the "Lanxin" unmanned surface vessel from Dalian Maritime University. The parameters of this unmanned vessel are: length 7.02m, beam 2.6m, speed 35kn, full-load draft 0.32m, and full-load displacement 2.73m. 3 The square coefficient is 0.6976. The Norrbin motion model is adopted, with model parameters K = 0.71, T = 0.32, and nonlinear coefficients α = 1 and β = 0.001.

[0148] The ideal course of the unmanned vessel is set as φ d =sin(0.1*t)+cos(0.2*t), the controller designed in step S3 is used to control the unmanned surface vehicle, and the design parameters are as follows: The quantization parameter is K=8.

[0149] Experimental results are as follows Figures 3-5 As shown, Figure 3 The tracking results of the unmanned surface vehicle's heading angle, bow roll rate, and rudder angle are shown, demonstrating that the controller designed in this invention has good control performance and can quickly achieve ship heading tracking. Figure 4 The tracking errors for ship heading angle, ship bow roll rate, and ship rudder angle are shown to converge quickly to a small residual set, further verifying the effectiveness of the controller designed in this invention. Figure 5 To account for control input curves that include and do not include actuator input quantization, by Figure 5 It can be seen that the input quantization process has the effect of reducing the signal transmission burden, reducing the actuator execution frequency, and making the process closer to the control law of multi-level servo systems in nautical practice.

[0150] The simulation results verify the effectiveness of the adaptive neural network controller in the course tracking task. The control system can effectively cope with the changes and disturbances of the external environment, and ensure the stability of the course control. By introducing the input quantization technology, the communication burden of the unmanned surface vehicle is reduced, and the execution frequency of the actuator is also reduced. From the practical application, in order to obtain good course control performance, the characteristics of the ship rudder servo system are also considered. Finally, the effectiveness of the adaptive neural network course tracking control method of the unmanned surface vehicle with the input quantization of the actuator and the dynamic characteristics of the ship rudder is further verified.

[0151] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An adaptive neural network heading tracking control method for an unmanned surface vehicle, characterized in that, include: S1. Based on the surrounding environment and sea state information of other vessels, establish a mathematical model for the heading control of a third-order response unmanned surface vehicle, including: S11. Construct a nonlinear Nomoto model of the ship and a model with the ship's steering gear dynamic characteristics, as follows: wherein ψ represents a ship heading angle, δ a rudder angle, K, T both represent ship maneuverability indices, a, β both represent nonlinear parameters, δ E represents a commanded rudder angle from a rudder actuator, K E represents a rudder actuator control gain, T E represents a rudder actuator time constant; S12. Introducing input quantization, a third-order response mathematical model of the unmanned surface vehicle is constructed as follows: in, ω1 and ω2 both represent unknown disturbances caused by wind, current, and waves; S2. Introduce a linear input quantization analysis model to quantize the control inputs in the control system, including: S21. For the control input in the ship model after input quantization, let Q(u) = q1(t)u + q2(t), and take: S22. Since the sign remains unchanged during the quantization process, q1(t) is unknown. From the formula in step S21, we know that q1(t) > 0. When |u(t)| < a, Q(u) is bounded, q1(t) = 1, then q2(t) is bounded. Therefore, we take... S3. Based on the dynamic surface control algorithm, design a Lyapunov function to obtain the heading controller of the unmanned surface vehicle, including: S31. Define the heading error as: S1 = x1 - x ld S2 = x2 - x 2d S3 = x3 - x 3d And by differentiating the heading error respectively, we get: S32. Define the Lyapunov function as follows: Among them, y i This represents the error of the virtual control term. S33. Define a first-order low-pass filter as follows: in, S34. Since f(x,t) is unknown, an RBF neural network is constructed to approximate the unknown function of the system model. An ideal weight vector θ exists. * ∈R N This makes the neural network θ *T h(x) sufficiently approximates the given function f and the absolute value of the approximation error is no greater than σ. M ,Right now: f(x)=θ *T h(x)+e * S35. According to Young's inequality, taking the derivative of V1 and V2 in step S32, we get: S36. Define time-varying gain: Where, q1(t) min It is a lower bound of q1(t); S37. Design a Lyapunov function, as follows: And by differentiating the Lyapunov function designed in step S37, we obtain: S38. Based on steps S36 and S37, design the system control law and adaptive law as follows: in, S4. Based on Lyapunov stability theory, prove the stability of the designed adaptive neural network heading tracking control system for unmanned surface vehicles with input quantization when no prior information of fixed quantization parameters is required. Furthermore, all signals in the closed-loop system are semi-globally consistent and bounded, and the tracking error can be made arbitrarily small by adjusting the controller parameters.

2. The adaptive neural network heading tracking control method for an unmanned surface vehicle according to claim 1, characterized in that, Step S4 specifically includes: S41. The Lyapunov function designed in step S32 yields the quantized part, as follows: Pick Where l > 0; S42, will Substituting the values, according to Young's inequality, we get: S43, due to get make Thus obtain according to Another S44. From steps S35 and S37, we obtain: S45. Based on Young's inequality and other inequalities get: Where, λ max (·) represents the largest eigenvalue of ·; S46. Take 0 < η0 ≤ 1.0, Then we have: The control parameters are selected as follows: c1≥l+1, Where l represents the parameter to be designed; S47, due to Thus obtain Substituting, we get: S48, by It can be known that: It has a maximum value, denoted as Q; select Therefore, we have: S49. Taking all factors into consideration, due to Then V is bounded, and solving the above inequality yields: Clearly, all signals in the closed-loop system are semi-globally bounded, when hour Then t→∞,

Images