An observer-based data-driven heading control method for unmanned surface vehicle
By using an observer-based data-driven approach, a model-free adaptive controller was designed using heading angle and rudder angle data. This solved the problems of model dependence and noise influence in the heading control of unmanned surface vessels, and achieved efficient heading control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- QINGDAO UNIV
- Filing Date
- 2026-05-08
- Publication Date
- 2026-07-31
AI Technical Summary
Existing unmanned surface vessel (USV) heading control strategies rely on difficult-to-obtain model information of the heading system. Standard model-free adaptive control strategies cannot be directly applied, and data-driven controllers are poorly robust to measurement noise.
A data-driven approach based on observers is adopted. An observer is constructed using heading angle and control rudder angle data. A model-free adaptive heading controller is designed. A data model is established through tight-format dynamic linearization technology. Combined with a pseudo-partial derivative estimation module and a rudder angle change rate constraint, a data-driven heading controller is designed.
It achieves unmanned surface vessel bow heading control independent of model information, solves the "quasi-linearity" problem, improves system robustness, and can achieve good control performance under measurement noise conditions.
Smart Images
Figure CN122488733A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned surface vessel (USV) motion control technology, and more specifically, to a data-driven bow control method for USVs based on observers. Background Technology
[0002] The ocean is the largest body of water on Earth, containing abundant natural resources and serving as a crucial strategic space for the sustainable development of human society. Unmanned surface vessels (USVs), as remotely controlled and autonomous marine transport platforms, possess advantages such as small size, flexibility, high automation, remote control capability, and strong maneuverability. Developing USVs plays a vital role in promoting marine development, protecting marine resources, and maintaining maritime security.
[0003] Heading control of unmanned surface vessels (USVs) is a fundamental problem in motion control and one of the earliest research areas in which control theory has been applied. Scholars both domestically and internationally have conducted extensive research on heading control of USVs, proposing numerous control methods, such as backstepping, adaptive control, and robust control. The core idea of the backstepping-based heading control strategy for USVs is to decompose the heading system into subsystems, then recursively design virtual control laws for the subsystems based on Lyapunov functions, and finally derive the actual control laws. However, its control effect is highly dependent on the accuracy of the mathematical model; if the model is severely mismatched, the performance of the controller designed based on the backstepping method will significantly decrease. The core of the adaptive control-based heading control strategy for USVs is to adjust the controller parameters online in real time to cope with dynamic changes caused by model uncertainties and environmental disturbances. To accurately estimate the parameters, the adaptive control method requires continuous excitation of the system. If the USV is in a constant-speed, straight-ahead state for a long time, the parameter estimation may not converge, and the control effect may worsen. The core of the unmanned surface vessel (USV) heading control strategy based on robust control methods is to treat parameter perturbations, unmodeled dynamics, and external disturbances in the USV heading system model as uncertainties, describe them mathematically, and consider these uncertainties in the heading controller design to ensure heading control effectiveness under harsh conditions. However, its heading control effect is conservative, the dynamic response is not fast enough, energy consumption may be relatively high, and the control performance may significantly degrade when the USV's operating conditions change drastically beyond the preset uncertainty range. The heading control effect of the above USV heading control strategies heavily depends on the model information and modeling accuracy of the heading system, which brings great inconvenience to its real-world engineering applications. Model-free adaptive control, as an important data-driven control strategy, has been widely used in various industrial production fields, but it suffers from the "quasi-linearity" problem in its application to USV heading control systems and cannot be directly applied to USV heading systems. In summary, existing heading control technologies have the following shortcomings:
[0004] (1) Existing unmanned surface vessel heading control strategies require heading system model information. However, obtaining an accurate heading model is very difficult, which is detrimental to the implementation of the project.
[0005] (2) Existing standard model-free adaptive control strategies cannot be directly applied to the bow control of unmanned surface vessels (USVs). The bow control system of USVs does not satisfy the "quasi-linear" assumption in the standard model-free adaptive control strategy.
[0006] (3) Existing data-driven heading controllers directly use heading data with measurement noise, which will have an adverse effect on control performance and reduce the robustness of the controller. Summary of the Invention
[0007] To address the aforementioned technical problems, a data-driven heading control method for unmanned surface vessels based on an observer is proposed. This invention constructs an observer using measured heading angle and control rudder angle data, and designs a model-free adaptive heading controller using the observed heading and control rudder angles.
[0008] The technical means employed in this invention are as follows:
[0009] A data-driven bow control method for unmanned surface vessels based on an observer includes: a data-driven bow controller, an observer, and a pseudo-partial derivative estimation module; wherein:
[0010] The input of the data-driven heading controller is connected to the given desired heading signal, the observer, and the estimation module of pseudo-partial derivatives, and the output of the data-driven heading controller is connected to the unmanned surface vessel.
[0011] The input of the observer is connected to the bow sensor of the unmanned surface vessel, and the output is connected to the pseudo-partial derivative estimation module and the data-driven bow controller, respectively.
[0012] The input of the pseudo-partial derivative estimation module is connected to the data-driven heading controller and the observer, and the output is connected to the data-driven heading controller.
[0013] Furthermore, the design process of the observer-based unmanned surface vessel data-driven bow controller is as follows:
[0014] The equivalent data model of the heading system is established using the compact scheme dynamic linearization technique as follows:
[0015] (1)
[0016] in, , These are the heading angle and control input rudder angle at the sampling time, respectively. It is a pseudo-partial derivative. , .
[0017] Furthermore, the observer for the unmanned surface vessel bow-heading system data model is designed as follows:
[0018] (2)
[0019] in, and These are the observed values of the heading angle and the estimated values of the pseudo-partial derivatives, respectively. This represents the observation error of the heading angle; The observer gain and satisfying .
[0020] Furthermore, the pseudo-partial derivative estimation algorithm is designed as follows:
[0021] (3)
[0022] (4)
[0023] in, It is a constant. To improve the estimation performance of the estimation algorithm for time-varying pseudo-partial derivatives, the following pseudo-partial derivative reset algorithm is introduced:
[0024] (5)
[0025] in, It is a constant.
[0026] Furthermore, a data-driven bow controller is designed using the observed values and observation errors of the bow angle, as well as the pseudo-partial derivative estimates, as follows:
[0027] (6)
[0028] in, Step size factor; , For the weighting factors of the controller; This represents the actual rudder angle acting on the heading system at past moments. Further considering the rate of change limit of the rudder angle, the heading control algorithm is improved as follows:
[0029] when hour:
[0030] (7)
[0031] otherwise:
[0032] (8)
[0033] in, This represents the maximum permissible rate of change of the rudder angle.
[0034] Compared with existing unmanned surface vessel heading control methods, the present invention has the following advantages:
[0035] The observer-based data-driven heading control strategy provided by this invention, compared with the existing model-based heading control strategy, only requires the input and output data of the unmanned surface vessel heading system and does not depend on the system's model information, effectively solving the problem of difficulty in obtaining system models.
[0036] Compared with the standard model-free adaptive control strategy, the observer-based data-driven heading control strategy provided by this invention does not rely on the assumption of pseudo-partial derivative sign invariance, thus effectively solving the "quasi-linearity" problem in the heading control application of model-free adaptive control for unmanned surface vessels.
[0037] The observer-based data-driven heading control strategy provided by this invention, compared with existing data-driven unmanned surface vessel heading control strategies, uses heading observations in the controller design instead of directly using heading data with measurement noise, which can improve the robustness of the system to a certain extent.
[0038] Based on the above advantages, this invention can be widely promoted in the field of unmanned surface vessel control technology. Attached Figure Description
[0039] Figure 1 This is a schematic diagram of the control method of the present invention.
[0040] Figure 2 This is a diagram illustrating the heading angle tracking effect of the unmanned surface vessel of this invention.
[0041] Figure 3 This is a diagram showing the estimation effect of the heading angle observer constructed in this invention.
[0042] Figure 4 This is the control input rudder angle diagram of the present invention.
[0043] Figure 5 This is a diagram showing the pseudo-partial derivative estimation effect of the present invention. Detailed Implementation
[0044] This invention provides a data-driven bow control method for unmanned surface vessels based on an observer. The specific control structure block diagram is shown below. Figure 1 As shown, it mainly includes: a data-driven bow controller, a bow observer, and a pseudo-partial derivative estimation module; wherein:
[0045] A data-driven bow control method for unmanned surface vessels based on an observer includes: a data-driven bow controller, an observer, and a pseudo-partial derivative estimation module; wherein:
[0046] The input of the data-driven heading controller is connected to the given desired heading signal, the observer, and the estimation module of the pseudo-partial derivative, and the output of the data-driven heading controller is connected to the unmanned surface vessel.
[0047] The input of the observer is connected to the bow sensor of the unmanned surface vessel, and the output is connected to the pseudo-partial derivative estimation module and the data-driven bow controller, respectively.
[0048] The input of the pseudo-partial derivative estimation module is connected to the data-driven heading controller and the observer, and the output is connected to the data-driven heading controller.
[0049] In this embodiment, a periodic, smooth square wave signal passed through a low-pass filter is input to the heading controller. External disturbances and control signals generated by the data-driven heading controller work together to affect the heading system of the unmanned surface vessel (USV). The USV heading observer uses the signal measured by the heading sensor to observe the heading angle of the USV. The pseudo-partial derivative estimation module uses the heading angle observation error and the rudder angle change to calculate and update the pseudo-partial derivative. Finally, the data-driven controller uses the updated heading angle observation value, observation error, and pseudo-partial derivative estimate to update the controller, thereby achieving the heading control objective.
[0050] In a specific implementation, as a preferred embodiment of the present invention, the observer-based unmanned surface vessel data-driven bow control method is designed as follows:
[0051] The equivalent data model of the heading system is established using the compact scheme dynamic linearization technique as follows:
[0052] (1)
[0053] in, , These are the heading angle and control input rudder angle at the sampling time, respectively. It is a pseudo-partial derivative. , .
[0054] In practical implementation, the observer designed for the unmanned surface vessel bow-oriented system data model is as follows:
[0055] (2)
[0056] in, and These are the observed values of the heading angle and the estimated values of the pseudo-partial derivatives, respectively. This represents the observation error of the heading angle; The observer gain and satisfying .
[0057] In practice, the pseudo-partial derivative estimation algorithm is designed as follows:
[0058] (3)
[0059] (4)
[0060] in, It is a constant. Furthermore, to improve the estimation performance of the estimation algorithm for time-varying pseudo-partial derivatives, the following pseudo-partial derivative reset algorithm is introduced:
[0061] (5)
[0062] in, It is a constant.
[0063] In practical implementation, the data-driven bow controller is designed using the observed values and observation errors of the bow angle, as well as the pseudo-partial derivative estimates, as follows:
[0064] (6)
[0065] in, Step size factor; , For the weighting factors of the controller; This represents the actual rudder angle acting on the heading system at past moments. Further considering the rate of change limit of the rudder angle, the heading control algorithm is improved as follows:
[0066] when hour:
[0067] (7)
[0068] otherwise:
[0069] (8)
[0070] in, This represents the maximum permissible rate of change of the rudder angle.
[0071] Example
[0072] To verify the control effect of the proposed observer-based unmanned surface vessel (USV) data-driven bow control method, simulation verification was performed using the parameters of an USV. The USV is 25.7 meters long and 6.0 meters wide. The model of the USV bow control system is shown below:
[0073] (9)
[0074] in, , and The heading angle, angular velocity, and rudder angle at the sampling time are given. and These are the control coefficients related to the unmanned surface vessel, which are as follows in this embodiment: , . It is the sampling period, and its value is... . It refers to external ocean disturbances acting on the unmanned surface vessel. It includes the nonlinear component of the unmanned surface vessel's bow system, specifically in the form of: , It is a constant.
[0075] In this embodiment, the heading control objective is to track a periodic, smooth square wave signal output by a low-pass filter as follows:
[0076] (10)
[0077] in, and These are the desired forward signal and intermediate variables, respectively; , . The input signal of the system is in the following form:
[0078] (11)
[0079] During simulation verification, the controller parameters were selected as follows: , , , , .
[0080] The simulation results of this embodiment are as follows: Figure 2-5 As shown. Among them. Figure 2 As shown in the tracking effect diagram, even under conditions of ocean disturbance and measurement noise, the unmanned surface vessel (USV) data-driven heading control method based on the observer provided by this invention can achieve the heading control target of the USV. Figure 3 The image shows the estimation results of the heading observer. As can be seen from the estimation results, the observer has a good observation effect on the heading direction, and the observation error has a fast convergence speed. Figure 4 To control the change in input rudder angle, by Figure 4 It can be seen that a large control input rudder angle is only generated when the desired heading signal changes abruptly. When the desired heading signal is a slow time-varying signal, the rudder angle approaches zero, which is consistent with the theoretical analysis. Figure 5 The estimation effect of the pseudo-partial derivative is given by Figure 5 It can be seen that the estimated value of the pseudo-partial derivative is bounded.
[0081] The simulation results demonstrate that the observer-based unmanned surface vessel (USV) data-driven heading control method proposed in this invention can achieve good heading control using only the system's input and output data. Furthermore, this invention maintains good heading control performance even in the presence of measurement noise.
[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A data-driven bow control method for unmanned surface vessels based on observers, characterized in that... include: Data-driven bow controller, bow observer, and pseudo-partial derivative estimation module; wherein: The input of the data-driven heading controller is connected to the given desired heading signal, the observer, and the estimation module of pseudo-partial derivatives, and the output of the data-driven heading controller is connected to the unmanned surface vessel. The input terminal of the observer is connected to the bow sensor of the unmanned surface vessel, and the output terminal of the observer is connected to the pseudo-partial derivative estimation module and the data-driven bow controller, respectively. The input of the pseudo-partial derivative estimation module is connected to the data-driven heading controller and the observer, and the output of the pseudo-partial derivative estimation is connected to the data-driven heading controller.
2. The design process of the observer-based unmanned surface vessel data-driven bow controller according to claim 1 is as follows: The equivalent data model of the heading system is established using the compact scheme dynamic linearization technique as follows: (1) in, , These are the heading angle and control input rudder angle at the sampling time, respectively. It is a pseudo-partial derivative. , .
3. According to claim 1, the observer for the unmanned surface vessel bow-heading system data model is designed as follows: (2) in, and These are the observed values of the heading angle and the estimated values of the pseudo-partial derivatives, respectively. This represents the observation error of the heading angle; The observer gain and satisfying .
4. According to claim 1, the pseudo-partial derivative estimation algorithm is designed as follows: (3) (4) in, It is a constant. To improve the estimation performance of the estimation algorithm for time-varying pseudo-partial derivatives, the following pseudo-partial derivative reset algorithm is introduced: (5) in, It is a constant.
5. According to claim 1, the data-driven bow controller is designed using the observed values and observation errors of the bow angle, as well as the pseudo-partial derivative estimates, as follows: (6) in, Step size factor; , For the weighting factors of the controller; This represents the actual rudder angle acting on the heading system at past moments. Further considering the rate of change limit of the rudder angle, the heading control algorithm is improved as follows: when hour: (7) otherwise: (8) in, This represents the maximum permissible rate of change of the rudder angle.