Parameter self-tuning unmanned ship expansion state observer

Abstract

The invention discloses a parameter self-tuning unmanned ship expansion state observer. Belongs to the technical field of unmanned ship automatic control. Comprising a resultant velocity self-tuning expansion state observer module and a yawing angular velocity self-tuning expansion state observer module. The resultant velocity self-tuning expansion state observer module comprises a resultant velocity sensor module, a resultant velocity expansion state observer module and a resultant velocity observer adaptive law module; the yawing angular velocity self-tuning expansion state observer module comprises a yawing angular velocity sensor module, a yawing angular velocity expansion state observer module and a yawing angular velocity observer adaptive law module. The parameter self-tuning unmanned ship expansion state observer does not depend on simulation or experimental feedback, a cost function is designed through a variational method, observation errors are optimized theoretically, and then an observer gain self-adaptive law is constructed.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned surface vehicle control, and particularly to an extended state observer for an unmanned surface vehicle with parameter self-tuning. Background Art

[0002] In recent years, with the growth of ocean operation requirements, unmanned surface vehicles (ASVs) have attracted much attention due to their advantages such as intelligence and economy. To ensure the safe and reliable operation of unmanned surface vehicles in complex ocean environments, efficient control technologies are crucial. Traditional PID control is difficult to cope with external disturbances such as wind, waves, and currents, which limits the performance of unmanned surface vehicles. Therefore, the active disturbance rejection control technology (ADRC) has gradually become a research hotspot. Its core lies in using an extended state observer (ESO) to dynamically estimate and compensate unknown disturbances in the system, significantly enhancing the anti-disturbance ability and control stability of unmanned surface vehicles in complex dynamic environments.

[0003] Regarding the estimation of unknown lumped disturbances of unmanned surface vehicles using extended state observers, domestic and foreign research scholars have made a series of research progress. L. Liu et al. proposed a multi-modal extended state observer, where each extended state observer is associated with different nominal input values, and combined with a monitoring mechanism, it can be applied to various navigation conditions with different dynamic characteristics. G. Tang et al. proposed a parallel extended state observer, which improves the estimation accuracy of the observer by paralleling multiple extended state observers. J. Yue et al. proposed a data-driven extended state observer, which can simultaneously estimate unknown input gains and unmeasured velocities based on a parallel learning method. M. Naghdi et al. proposed a fuzzy logic-based extended state observer, which designs five fuzzy rules to dynamically adjust its bandwidth, significantly enhancing the intelligence and adaptability of the extended state observer.

[0004] Regarding the problem of using extended state observers to estimate unknown disturbances, there have been some feasible technical solutions, but the existing patents still have the following problems:

[0005] First, in the existing disturbance estimation methods based on extended state observers, most observer gains rely heavily on simulation or experimental feedback and generally cannot be adaptively adjusted. Therefore, constructing an observer gain adaptive law that can be effectively applied to disturbance estimation in real time is a challenging task.

[0006] Second, in the existing unmanned surface vehicle disturbance estimation methods based on extended state observers, most of them ignore the case where the velocity sensor has measurement noise. This simplified assumption ignores the interference of noise on measurement accuracy, so it may be difficult to fully meet the control requirements of actual unmanned surface vehicles in a sensor noise environment. Summary of the Invention

[0007] In view of the deficiencies of the prior art, the present invention provides an unmanned boat extended state observer with parameter self-tuning. The present invention utilizes the idea of variational method, designs a continuous and variable observer gain adaptive law by optimizing the objective function considering the influence of measurement noise and lumped disturbance simultaneously, constructs an unmanned boat extended state observer with parameter self-tuning, and realizes the accurate estimation of the disturbance in the presence of measurement noise.

[0008] The technical means adopted by the present invention are as follows:

[0009] An unmanned boat extended state observer with parameter self-tuning, comprising an extended state observer module for self-tuning of the combined velocity and an extended state observer module for self-tuning of the yaw angular velocity;

[0010] The extended state observer module for self-tuning of the combined velocity includes a combined velocity sensor module, a combined velocity extended state observer module, and a combined velocity observer adaptive law module. The combined velocity observer adaptive law module includes a combined velocity cost function design module, a combined velocity Lagrange multiplier method module, and a combined velocity Gaussian random process module. The input end of the combined velocity sensor module is respectively connected to the output end of the unmanned boat module and the measurement noise signal, and the output end of the combined velocity sensor module is connected to the input end of the combined velocity extended state observer module. The input end of the combined velocity extended state observer module is respectively connected to the longitudinal control input signal, the output end of the combined velocity sensor module, and the output end of the combined velocity Gaussian random process module, and the output end of the combined velocity extended state observer module is connected to the input end of the combined velocity cost function design module. The input end of the combined velocity cost function design module is respectively connected to the output end of the unmanned boat module, the output end of the combined velocity extended state observer module, the measurement noise signal, and the lumped disturbance in the direction of the combined velocity of the unmanned boat, and the output end of the combined velocity cost function design module is connected to the input end of the combined velocity Lagrange multiplier method module. The input end of the combined velocity Lagrange multiplier method module is respectively connected to the output end of the combined velocity cost function design module and the Lagrange multiplier, and the output end of the combined velocity Lagrange multiplier method module is connected to the input end of the combined velocity Gaussian random process module. The input end of the combined velocity Gaussian random process module is connected to the output end of the combined velocity Lagrange multiplier method module, and the output end of the combined velocity Gaussian random process module is connected to the input end of the combined velocity extended state observer module. The input end of the unmanned boat module is connected to the longitudinal control input signal;

[0011] The yaw angular velocity self-tuning extended state observer module includes a yaw angular velocity sensor module, a yaw angular velocity extended state observer module, and a yaw angular velocity observer adaptation law module. The yaw angular velocity observer adaptation law module includes a yaw angular velocity cost function design module, a yaw angular velocity Lagrange multiplier method module, and a yaw angular velocity Gaussian random process module. The input end of the yaw angular velocity sensor module is respectively connected to the output end of the unmanned boat module and the measurement noise signal, and the output end of the yaw angular velocity sensor module is connected to the input end of the yaw angular velocity extended state observer module; the input end of the yaw angular velocity extended state observer module is respectively connected to the yaw direction control input signal, the output end of the yaw angular velocity sensor module, and the output end of the yaw angular velocity Gaussian random process module, and the output end of the yaw angular velocity extended state observer module is connected to the input end of the yaw angular velocity cost function design module; the input end of the yaw angular velocity cost function design module is respectively connected to the output end of the unmanned boat module, the output end of the yaw angular velocity extended state observer module, the measurement noise signal, and the concentrated disturbance of the yaw angular velocity direction of the unmanned boat, and the output end of the yaw angular velocity cost function design module is connected to the input end of the yaw angular velocity Lagrange multiplier method module; the input end of the yaw angular velocity Lagrange multiplier method module is respectively connected to the output end of the yaw angular velocity cost function design module and the Lagrange multiplier, and the output end of the yaw angular velocity Lagrange multiplier method module is connected to the input end of the yaw angular velocity Gaussian random process module; the input end of the yaw angular velocity Gaussian random process module is connected to the output end of the yaw angular velocity Lagrange multiplier method module, and the output end of the yaw angular velocity Gaussian random process module is connected to the input end of the yaw angular velocity extended state observer module; the input end of the unmanned boat module is connected to the yaw direction control input signal.

[0012] Further, the resultant velocity extended state observer module is expressed as:

[0013]

[0014] In the formula, y U = U + w U is the actual resultant velocity signal of the unmanned boat obtained by the resultant velocity sensor module, w U is the sensor noise signal, G U is the observer gain matrix, τ u represents the surge velocity control vector of the unmanned boat, m u is the resultant velocity control coefficient, is the estimated value of the resultant velocity, is the derivative with respect to time; is the estimated value of the concentrated disturbance of the resultant velocity direction, is Derivative with respect to time.

[0015] Furthermore, the combined velocity cost function design module is expressed as:

[0016]

[0017] X1 = [U, σ U T , Y1 = y U

[0018] where J U represents the signal at the output end of the combined velocity cost function design module, D1 = (1 0), V1 is a positive constant, W1 is a weight matrix, P U is the estimation error covariance matrix, and both W1 and P U are 2×2 positive definite symmetric matrices, w U is the sensor noise signal, represents the time integration variable, and t represents time; is the estimated value of X1, is the initial value of, X1(0) is the initial value of X1; P U (0) is the initial value of P U the initial value of.

[0019] Furthermore, the combined velocity Lagrange multiplier method module is expressed as:

[0020]

[0021] Respectively, take the partial derivatives of the Lagrangian function J U (t, λ U ) with respect to the variable B1, the Lagrange multiplier λ U and X1 to obtain the signal at the output end of the combined velocity Lagrange multiplier method module

[0022]

[0023] Furthermore, the combined velocity Gaussian random process module is expressed as:

[0024]

[0025] where P U is the estimation error covariance matrix, and G U is the adaptive law of the combined velocity extended state observer parameter.

[0026] Furthermore, the yaw angular velocity extended state observer module is expressed as: ​

[0027]

[0028] where y r = r + w r is the actual yaw angular velocity signal of the unmanned boat obtained by the yaw angular velocity sensor module, w r is the sensor noise signal, r is the yaw angular velocity vector, G r is the observer gain matrix, which is a 2×1 matrix, τ r is the yaw angular velocity control vector of the unmanned boat, m r is the yaw angular velocity control coefficient; is the estimated value of the yaw angular velocity, is the derivative with respect to time; is the estimated value of the concentrated disturbance in the yaw direction, is the derivative with respect to time.

[0029] Furthermore, the yaw angular velocity cost function design module is expressed as:

[0030]

[0031] where D2 = (1 0), V2 is a positive constant, W2 is the weight matrix, P r is the estimated error covariance matrix, W2 and P r are both 2×2 positive definite symmetric matrices, represents the time integration variable, t represents time; is the estimated value of X2 = [r, σ r T , is the initial value of, X2(0) is the initial value of X2; P r (0) is the initial value of P r .

[0032] Furthermore, the yaw angular velocity Lagrange multiplier method module is expressed as:

[0033]

[0034] Taking the partial derivatives of the Lagrangian function J r (t, λ r ) with respect to the variables B2, the Lagrange multiplier λ r and X2 respectively, the output signal of the yaw angular velocity Lagrange multiplier method module is obtained

[0035] ​

[0036] Furthermore, the yaw angular velocity Gaussian random process module is expressed as:

[0037]

[0038] where P r is the estimated error covariance matrix, and G r is the adaptive law of the yaw angular velocity extended state observer parameter.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] First, compared with the existing unmanned surface vehicle extended state observer method, the parameter self-tuning unmanned surface vehicle extended state observer method proposed by the present invention does not rely on simulation or experimental feedback. By designing a cost function using the variational method, the observation error is theoretically optimized, and then the observer gain adaptive law is constructed.

[0041] Second, compared with the existing unmanned surface vehicle extended state observer method, the parameter self-tuning unmanned surface vehicle extended state observer method proposed by the present invention optimizes the influence of measurement noise on the system while minimizing the observation error, which is beneficial to the practical application of the unmanned surface vehicle in a noisy environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0043] Figure 1 is a schematic structural diagram of the parameter self-tuning unmanned surface vehicle extended state observer system in the embodiment of the present invention.

[0044] Figure 2 is the estimated error of the combined velocity of the unmanned surface vehicle under measurement noise in the embodiment of the present invention simulation comparison diagram.

[0045] Figure 3 is the estimated error of the yaw angular velocity of the unmanned surface vehicle under measurement noise in the embodiment of the present invention simulation comparison diagram.

[0046] Figure 4 is the simulation comparison diagram of the disturbance estimation of the combined velocity direction of the unmanned surface vehicle under measurement noise in the embodiment of the present invention.

[0047] Figure 5This is the simulation comparison diagram of the disturbance estimation of the yaw direction of the unmanned boat under measurement noise in the embodiment of the present invention.

[0048] Figure 6 This is the parameter change curve diagram of the combined velocity adaptive extended state observer in the embodiment of the present invention, where (a) is the curve diagram of the combined velocity observation gain, and (b) is the curve diagram of the concentrated disturbance observation gain of the combined velocity direction.

[0049] Figure 7 This is the parameter change curve diagram of the yaw angular velocity adaptive extended state observer in the embodiment of the present invention, where (a) is the curve diagram of the yaw angular velocity observation gain, and (b) is the curve diagram of the concentrated disturbance observation gain of the yaw direction. Detailed implementation manners

[0050] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0051] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily need to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0052] The dynamic model equation of the unmanned boat applied in the embodiment of the present invention is described as:

[0053]

[0054] In the formula, respectively represent the surge velocity, sway velocity and yaw angular velocity vector of the unmanned boat; represents the control input force and moment vector of the unmanned boat; represents the disturbance vector caused by wind, wave and current; is the hull inertia mass matrix; F(ν) = -C(ν)ν - D(ν)ν + g(ν), is the centripetal force and Coriolis force coefficient matrix; is the damping matrix; g(ν) includes unmodeled fluid dynamics and modeling errors; M is the hull inertia coefficient.

[0055] Let F(ν) = [f u (u, v, r), f v (u, v, r), f r (u, v, r)] T , for the underactuated unmanned surface vehicle, rewrite Equation (1) as follows:

[0056]

[0057] Furthermore, since u = U cos(β), v = U sin(β), the dynamic equations of U and r can be written as

[0058]

[0059] where, σ U and σ r are the concentrated disturbances acting on the unmanned surface vehicle;

[0060]

[0061] β = arctan(v / u) is the sideslip angle of the unmanned surface vehicle.

[0062] The input end of the unmanned surface vehicle dynamics system is connected to the unmanned surface vehicle control input signals τ u 、τ r , and the output end of the unmanned surface vehicle dynamics system is connected to the unmanned surface vehicle speed state signals U, r.

[0063] As Figure 1 shown, a parameter self-tuning structure of an unmanned surface vehicle extended state observer includes: an extended state observer module for self-tuning of the combined velocity and an extended state observer module for self-tuning of the yaw angular velocity. Among them, the extended state observer module for self-tuning of the combined velocity includes a combined velocity sensor module, a combined velocity extended state observer module, a combined velocity observer adaptive law module (combined velocity cost function design module, combined velocity Lagrange multiplier method module, combined velocity Gaussian random process module) and an unmanned surface vehicle module; the extended state observer module for self-tuning of the yaw angular velocity includes a yaw angular velocity sensor module, a yaw angular velocity extended state observer module, a yaw angular velocity observer adaptive law module (yaw angular velocity cost function design module, yaw angular velocity Lagrange multiplier method module, yaw angular velocity Gaussian random process module) and an unmanned surface vehicle module.

[0064] The input end of the combined velocity sensor module is respectively connected to the output end of the unmanned surface vehicle module and the measurement noise signal wU are connected; the output end of the combined velocity sensor module is connected to the input end of the combined velocity extended state observer module. The input end of the combined velocity extended state observer module is respectively connected to the longitudinal control input signal τ u , the output end of the combined velocity sensor module, and the output end of the combined velocity Gaussian random process module; the output end of the combined velocity extended state observer module is connected to the input end of the combined velocity cost function design module. The input end of the combined velocity cost function design module is respectively connected to the output end of the unmanned boat module, the output end of the combined velocity extended state observer module, the measurement noise signal w U , and the concentrated disturbance σ in the direction of the combined velocity of the unmanned boat U are connected; the output end of the combined velocity cost function design module is connected to the input end of the combined velocity Lagrange multiplier method module. The input end of the combined velocity Lagrange multiplier method module is respectively connected to the output end of the combined velocity cost function design module and the Lagrange multiplier λ U are connected; the output end of the combined velocity Lagrange multiplier method module is connected to the input end of the combined velocity Gaussian random process module. The input end of the combined velocity Gaussian random process module is connected to the output end of the combined velocity Lagrange multiplier method module; the output end of the combined velocity Gaussian random process module is connected to the input end of the combined velocity extended state observer module. The input end of the unmanned boat module is connected to the longitudinal control input signal τ u are connected; the output end of the unmanned boat module is respectively connected to the input end of the combined velocity sensor module and the input end of the combined velocity cost function design module.

[0065] The input end of the yaw angular velocity sensor module is respectively connected to the output end of the unmanned boat module and the measurement noise signal w r are connected; the output end of the yaw angular velocity sensor module is connected to the input end of the yaw angular velocity extended state observer module. The input end of the yaw angular velocity extended state observer module is respectively connected to the yaw direction control input signal τ r , the output end of the yaw angular velocity sensor module, and the output end of the yaw angular velocity Gaussian random process module; the output end of the yaw angular velocity extended state observer module is connected to the input end of the yaw angular velocity cost function design module. The input end of the yaw angular velocity cost function design module is respectively connected to the output end of the unmanned boat module, the output end of the yaw angular velocity extended state observer module, the measurement noise signal w r , and the concentrated disturbance σ in the direction of the yaw angular velocity of the unmanned boat r are connected; the output end of the yaw angular velocity cost function design module is connected to the input end of the yaw angular velocity Lagrange multiplier method module. The input end of the yaw angular velocity Lagrange multiplier method module is respectively connected to the output end of the yaw angular velocity cost function design module and the Lagrange multiplier λ rare connected; the output end of the yaw angular velocity Lagrange multiplier method module is connected to the input end of the yaw angular velocity Gaussian random process module. The input end of the yaw angular velocity Gaussian random process module is connected to the output end of the yaw angular velocity Lagrange multiplier method module; the output end of the yaw angular velocity Gaussian random process module is connected to the input end of the yaw angular velocity extended state observer module. The input end of the unmanned boat module is connected to the yaw direction control input signal τ r are connected; the output end of the unmanned boat module is respectively connected to the input end of the yaw angular velocity sensor module and the input end of the yaw angular velocity cost function design module.

[0066] As a preferred embodiment of the present invention, the combined velocity extended state observer module is established according to the following method.

[0067] The input end of the combined velocity extended state observer module is respectively connected to the longitudinal control input signal τ u , the output signal y U of the combined velocity sensor module, and the output signal G U of the combined velocity Gaussian random process module. After the following transformation, the output signal of the combined velocity extended state observer module is obtained

[0068]

[0069] In the formula, y U = U + w U is the actual combined velocity signal of the unmanned boat obtained by the combined velocity sensor module, w U is the sensor noise signal, τ u represents the surge velocity control vector of the unmanned boat, m u is the combined velocity control coefficient, is the estimated value of the combined velocity, is the derivative with respect to time; is the estimated value of the concentrated disturbance in the combined velocity direction, is the derivative with respect to time. G U is the observer gain matrix, which is a 2×1 matrix.

[0070] As a preferred embodiment of the present invention, the combined velocity cost function design module is established according to the following method.

[0071] Expand the concentrated disturbance into a new state, let X1 = [U, σ U T , Y1 = y U , and the following transformation is performed on Equation (3) to obtain the new combined velocity state equation as follows:

[0072] ​

[0073] wherein D1 = (1 0).

[0074] The input end of the resultant velocity cost function design module is respectively connected to the output signal U of the unmanned boat module, the output signal of the resultant velocity extended state observer module, the sensor noise signal w U and the resultant velocity direction disturbance σ U and, through the following transformation, obtains the output end signal J of the resultant velocity cost function design module U :

[0075]

[0076] wherein V1 is a positive constant, W1 is a weight matrix, P U is the estimated error covariance matrix, both W1 and P U are 2×2 positive definite symmetric matrices, w U is the sensor noise signal, represents the time integration variable, t represents time; is the estimated value of X1, is the initial value of, X1(0) is the initial value of X1; P U (0) is the initial value of P U of. X1(0) = [0,0] T ; P U (0) = W1,. W1 is a designed positive definite symmetric matrix representing the weight, W1 = diag{0.4, 4000}; V1 = 0.08.

[0077] As a preferred embodiment of the present invention, the resultant velocity Lagrange multiplier method module is established according to the following method.

[0078] The input end of the resultant velocity Lagrange multiplier method module is respectively connected to the output signal J of the resultant velocity cost function design module U and the Lagrange multiplier λ U and, through the following transformation, obtains the Lagrangian function J U (t, λ U ):

[0079]

[0080] Respectively take partial derivatives of the Lagrangian function J U (t, λ U ) with respect to the variables B1, λ U and X1 to obtain the output end signal of the resultant velocity Lagrange multiplier method module

[0081]

[0082] As a preferred embodiment of the present invention, the combined velocity Gaussian random process module is established according to the following method.

[0083] Introduce the idea of Gaussian random process to obtain:

[0084]

[0085] In the formula, P U is the estimated error covariance matrix.

[0086] Combined with equations (8) and (9), the input end of the combined velocity Gaussian random process module is connected to the output signal of the combined velocity Lagrange multiplier method module and obtained through the following transformation:

[0087]

[0088] Eliminate and the following can be obtained:

[0089]

[0090] Furthermore, we separate the random term containing λ U and the deterministic term not containing λ U and obtain the output signal G of the combined velocity Gaussian random process module through the following transformation U :

[0091]

[0092] In the formula, G U is the adaptive law of the parameters of the combined velocity extended state observer.

[0093] As a preferred embodiment of the present invention, the yaw angular velocity extended state observer module is established according to the following method.

[0094] The input end of the yaw angular velocity extended state observer module is respectively connected to the yaw direction control input signal τ r , the output signal y of the yaw angular velocity sensor module r and the output signal G of the yaw angular velocity Gaussian random process module r and obtained the output signal of the yaw angular velocity extended state observer module through the following transformation

[0095]

[0096] In the formula, yr = r + w r is the actual yaw angular velocity signal of the unmanned boat obtained by the yaw angular velocity sensor module, w r is the sensor noise signal, r is the yaw angular velocity vector, G r is the observer gain matrix, which is a 2×1 matrix, τ r is the yaw angular velocity control vector of the unmanned boat, m r is the yaw angular velocity control coefficient; is the estimated value of the yaw angular velocity, is the derivative with respect to time; is the estimated value of the concentrated disturbance in the yaw direction, is the derivative with respect to time.

[0097] As a preferred embodiment of the present invention, the yaw angular velocity cost function design module is established according to the following method

[0098] Expand the concentrated disturbance into a new state, let X2 = [r, σ r T , Y2 = y r , and the new yaw angular velocity state equation is obtained through the following transformation from Equation (3) as follows:

[0099]

[0100] In the formula, D2 = (1 0).

[0101] The input end of the yaw angular velocity cost function design module is respectively connected to the output signal r of the unmanned boat module, the output signal of the yaw angular velocity extended state observer module the sensor noise signal w r and the yaw angular velocity direction disturbance σ r , and the output end signal J of the yaw angular velocity cost function design module is obtained through the following transformation r :

[0102]

[0103] In the formula, V2 is a positive constant, W2 is a weight matrix, P r is the estimated error covariance matrix, W2, P r are both 2×2 positive definite symmetric matrices. represents the time integration variable, t represents time; is the estimated value of X2 = [r, σ r T , is​​ The initial value, X2(0) is the initial value of X2; P r (0) is the initial value of P r . X2(0) = [0,0] T ; P r (0) = W2,. W2 is a designed positive definite symmetric matrix, representing the weight, W2 = diag{0.4, 4000}; V2 = 0.1.

[0104] As a preferred embodiment of the present invention, the yaw angular velocity Lagrange multiplier method module is established according to the following method.

[0105] The input end of the yaw angular velocity Lagrange multiplier method module is respectively connected to the output signal J of the yaw angular velocity cost function design module r and the Lagrange multiplier λ r , and through the following transformation, the Lagrangian function J r (t, λ r ) is obtained:[[]]END]]

[0106]

[0107] Respectively take the partial derivatives of the Lagrangian function J r (t, λ r ) with respect to the variables B2, λ r and X2, and obtain the output end signal of the yaw angular velocity Lagrange multiplier method module

[0108]

[0109] As a preferred embodiment of the present invention, the yaw angular velocity Gaussian random process module is established according to the following method.

[0110] Introduce the idea of Gaussian random process to obtain:[[]]END]]

[0111]

[0112] In the formula, P r is the estimated error covariance matrix.

[0113] Combined with equations (17) and (18), the input end of the yaw angular velocity Gaussian random process module is connected to the output signal of the yaw angular velocity Lagrange multiplier method module , and through the following transformation, the following is obtained:[[]]END]]

[0114]

[0115] Eliminate to obtain:[[]]END]]

[0116]

[0117] Furthermore, we separate the random term containing λ r from the deterministic term not containing λ r and obtain the output signal G of the yaw angular velocity Gaussian random process module through the following transformation r :

[0118]

[0119] where G r is the adaptive law of the yaw angular velocity extended state observer parameters.

[0120] The following further illustrates the solution and effect of the present invention through specific application examples.

[0121] The simulation test of this embodiment is as follows.

[0122] The parameters of the unmanned surface vehicle model are as follows:

[0123]

[0124] g(ν) = [0, 0, 0] T ; τ w (t) = [-0.2cos(t)cos(1.5t), 0.01sin(0.1t), -0.3sin(2t)cos(2.3t)] T .

[0125] The external measurement noise signal of the system is random Gaussian noise obeying the normal distribution N(0, 0.015 2 ).

[0126] The parameters of the fixed-gain extended state observer: G U = [5, 500] T , G r = [4, 400] T .

[0127] The initial parameters of the parameter self-tuning extended state observer: G U = [5, 500] T , G r = [4, 400] T .

[0128] The simulation results are as shown in Figures 2 to 7 Tables 1 to 2.

[0129] Figure 2 and Figure 3They are respectively the comparison diagram of the combined velocity estimation error of the unmanned boat and the comparison diagram of the yaw angular velocity estimation error of the unmanned boat. It can be seen from the figure that in the case of measurement noise, the estimation performance of the parameter self-tuning extended state observer for the combined velocity and the yaw angular velocity is more excellent than that of the fixed-gain extended state observer. Figure 4 and Figure 5 They are respectively the comparison diagram of the combined velocity direction disturbance estimation of the unmanned boat and the comparison diagram of the yaw direction disturbance estimation of the unmanned boat. It can be seen from the figure that the extended state observer with parameter self-tuning can still ensure the accurate estimation of the concentrated disturbance under measurement noise. Figure 6 and Figure 7 They are respectively the curve diagram of the parameter change of the combined velocity adaptive extended state observer and the curve diagram of the parameter change of the yaw angular velocity adaptive extended state observer. It can be seen from the figure that the gain of the observer changes with time and finally tends to a stable value to adapt to the adverse effects of disturbances and noise. Table 1 is the numerical analysis table of the estimation errors of the fixed-gain extended state observer and the parameter self-tuning extended state observer. It quantitatively shows the superiority of the parameter self-tuning extended state observer compared with the fixed-gain extended state observer in state recovery and disturbance estimation under measurement noise. Table 2 is the numerical comparison table of the error cost functions of the fixed-gain extended state observer and the parameter self-tuning extended state observer. It can be seen from the table that the proposed observer parameter adaptive law can optimize the error cost function value, which quantitatively illustrates the effectiveness of the proposed method.

[0130] Table 1 Numerical analysis table of the estimation errors of the fixed-gain extended state observer (ESO) and the parameter self-tuning extended state observer (AESO) under measurement noise

[0131]

[0132] Table 2 Numerical comparison table of the error cost functions of the fixed-gain extended state observer (ESO) and the parameter self-tuning extended state observer (AESO)

[0133]

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some 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 invention.

Claims

1. An unmanned boat extended state observer with parameter self-tuning, characterized in that It includes an extended state observer module with combined velocity self-tuning and an extended state observer module with yaw angular velocity self-tuning; The extended state observer module with combined velocity self-tuning includes a combined velocity sensor module, an extended state observer module for combined velocity, and an adaptive law module for the combined velocity observer. The adaptive law module for the combined velocity observer includes a combined velocity cost function design module, a combined velocity Lagrange multiplier method module, and a combined velocity Gaussian random process module. The input end of the combined velocity sensor module is respectively connected to the output end of the unmanned boat module and the measurement noise signal, and the output end of the combined velocity sensor module is connected to the input end of the extended state observer module for combined velocity. The input end of the extended state observer module for combined velocity is respectively connected to the longitudinal control input signal, the output end of the combined velocity sensor module, and the output end of the combined velocity Gaussian random process module, and the output end of the extended state observer module for combined velocity is connected to the input end of the combined velocity cost function design module. The input end of the combined velocity cost function design module is respectively connected to the output end of the unmanned boat module, the output end of the extended state observer module for combined velocity, the measurement noise signal, and the concentrated disturbance in the direction of the combined velocity of the unmanned boat, and the output end of the combined velocity cost function design module is connected to the input end of the combined velocity Lagrange multiplier method module. The input end of the combined velocity Lagrange multiplier method module is respectively connected to the output end of the combined velocity cost function design module and the Lagrange multiplier, and the output end of the combined velocity Lagrange multiplier method module is connected to the input end of the combined velocity Gaussian random process module. The input end of the combined velocity Gaussian random process module is connected to the output end of the combined velocity Lagrange multiplier method module, and the output end of the combined velocity Gaussian random process module is connected to the input end of the extended state observer module for combined velocity. The input end of the unmanned boat module is connected to the longitudinal control input signal; The yaw angular velocity self-tuning extended state observer module includes a yaw angular velocity sensor module, a yaw angular velocity extended state observer module, and a yaw angular velocity observer adaptive law module. The yaw angular velocity observer adaptive law module includes a yaw angular velocity cost function design module, a yaw angular velocity Lagrange multiplier method module, and a yaw angular velocity Gaussian random process module. The input end of the yaw angular velocity sensor module is respectively connected to the output end of the unmanned boat module and the measurement noise signal, and the output end of the yaw angular velocity sensor module is connected to the input end of the yaw angular velocity extended state observer module; the input end of the yaw angular velocity extended state observer module is respectively connected to the yaw direction control input signal, the output end of the yaw angular velocity sensor module, and the output end of the yaw angular velocity Gaussian random process module, and the output end of the yaw angular velocity extended state observer module is connected to the input end of the yaw angular velocity cost function design module; the input end of the yaw angular velocity cost function design module is respectively connected to the output end of the unmanned boat module, the output end of the yaw angular velocity extended state observer module, the measurement noise signal, and the concentrated disturbance of the unmanned boat yaw angular velocity direction, and the output end of the yaw angular velocity cost function design module is connected to the input end of the yaw angular velocity Lagrange multiplier method module; the input end of the yaw angular velocity Lagrange multiplier method module is respectively connected to the output end of the yaw angular velocity cost function design module and the Lagrange multiplier, and the output end of the yaw angular velocity Lagrange multiplier method module is connected to the input end of the yaw angular velocity Gaussian random process module; the input end of the yaw angular velocity Gaussian random process module is connected to the output end of the yaw angular velocity Lagrange multiplier method module, and the output end of the yaw angular velocity Gaussian random process module is connected to the input end of the yaw angular velocity extended state observer module; the input end of the unmanned boat module is connected to the yaw direction control input signal.

2. A parameter self-tuning unmanned boat extended state observer according to claim 1, characterized in that, The combined velocity extended state observer module is expressed as: where y U = U + w U is the actual combined velocity signal of the unmanned boat obtained by the combined velocity sensor module, and w U is the sensor noise signal, G U is the observer gain matrix, τ u represents the surge velocity control vector of the unmanned boat, m u is the combined velocity control coefficient, is the estimated value of the combined velocity, is the derivative with respect to time; is the estimated value of the concentrated disturbance in the combined velocity direction, is the derivative with respect to time.

3. A parameter self-tuning unmanned boat extended state observer according to claim 2, characterized in that, The combined velocity cost function design module is expressed as: X1 = [U, σ U T , Y1 = y U ​ Where, J U represents the output signal of the combined velocity cost function design module, D1 = (1 0), V1 is a positive constant, W1 is a weight matrix, P U is the estimated error covariance matrix, both W1 and P U are 2×2 positive definite symmetric matrices, w U is the sensor noise signal, represents the time integration variable, and t represents time; is the estimated value of X1, is the initial value of, X1(0) is the initial value of X1; P U (0) is the initial value of P U the initial value of.

4. A parameter self-tuning unmanned boat extended state observer according to claim 3, characterized in that, The combined velocity Lagrange multiplier method module is expressed as: Derive the partial derivatives of the Lagrangian function \(J\) U (t,\(\lambda\) U ) with respect to the variable \(B1\), the Lagrange multiplier \(\lambda\) U and \(X1\) respectively to obtain the output signal at the output end of the combined velocity Lagrange multiplier method module 5. A parameter self-tuning unmanned boat extended state observer according to claim 4, characterized in that, The combined velocity Gaussian random process module is expressed as: where, P U is the estimated error covariance matrix, and G U is the adaptive law of the combined velocity expansion state observer parameter.

6. The extended state observer of an unmanned boat with parameter self-tuning according to claim 5, characterized in that, The yaw angular velocity extended state observer module is expressed as: where y r = r + w r is the actual yaw angular velocity signal of the unmanned boat obtained by the yaw angular velocity sensor module, w r is the sensor noise signal, r is the yaw angular velocity vector, G r is the observer gain matrix, which is a 2×1 matrix, τ r is the yaw angular velocity control vector of the unmanned boat, m r is the yaw angular velocity control coefficient; is the estimated value of the yaw angular velocity, is the derivative with respect to time; is the estimated value of the concentrated disturbance in the yaw direction, is the derivative with respect to time.

7. A parameter self-tuning unmanned boat extended state observer according to claim 6, characterized in that, The yaw angular velocity cost function design module is expressed as: wherein, V2 is a positive constant, W2 is a weight matrix, and P r is the estimated error covariance matrix. Both W2 and P r are 2×2 positive definite symmetric matrices. represents the time integration variable, and t represents time; is the estimated value of X2 = [r, σ r T , is the initial value of, X2(0) is the initial value of X2; P r (0) is the initial value of P r .​ 8. A parameter self-tuning unmanned boat extended state observer according to claim 7, characterized in that, The yaw angular velocity Lagrange multiplier method module is expressed as: Derive partial derivatives of the Lagrangian function \(J\) r with respect to the variable \(B2\), the Lagrange multiplier \(\lambda\) r , and \(X2\) respectively, to obtain the output signal at the output end of the yaw angular velocity Lagrange multiplier method module r 9. A parameter self-tuning unmanned boat extended state observer according to claim 8, characterized in that, The yaw angular velocity Gaussian random process module is expressed as: where P r is the estimated error covariance matrix, and G r is the adaptive law of the yaw angular velocity extended state observer parameter.

Images