An autonomous control method for high-speed and low-energy consumption of an underwater vehicle based on a prediction model optimization

Abstract

The application discloses a kind of high-speed low-energy consumption autonomous control method of underwater vehicle based on prediction model optimization, belongs to underwater vehicle autonomous control technical field.The application establishes the longitudinal prediction model of vehicle, in each control cycle, with current navigation state as initial condition, based on the prediction model is carried out rolling optimization.Optimization function not only includes depth tracking error, more critically, introduce the navigation resistance estimation term directly related with angle of attack and rudder angle, to minimize total resistance as target, solve the optimal angle of attack sequence in future time domain and corresponding rudder angle, thrust control instruction.The application embeds hydrodynamic optimization in real-time closed-loop control, so that the vehicle can actively maintain the optimal drag-reducing attitude navigation, rather than passive maintenance level, thereby significantly reducing the total resistance and energy consumption of underwater vehicle high-speed navigation, effectively suppresses the phenomenon of buried bow, improves navigation stability.

Description

Technical Field

[0001] This invention relates to the field of autonomous control technology for underwater vehicles, specifically to a high-speed, low-energy-consumption autonomous control method for underwater vehicles based on predictive model optimization. Background Technology

[0002] When underwater vehicles perform tasks such as long-range cruising and rapid approach, their hydrodynamic characteristics change significantly, posing a severe challenge to traditional attitude and depth control methods. Currently, PID control strategies and their improved algorithms are the mainstream methods for depth and attitude regulation of underwater vehicles. These methods belong to "error-driven" feedback control, and their core design lies in directly minimizing the deviation between the navigation state and the set value. However, at high speeds, the hydrodynamic pressure on the bow of the vehicle increases sharply, generating a bow-buoyancy moment. To maintain the preset depth, the controller outputs a command rudder angle to deflect the rudder at a large angle to generate a compensating moment. This not only causes significant induced drag but also causes the vehicle to navigate in a non-streamlined attitude, further increasing drag. While traditional methods can achieve basic stability control, they completely ignore the intrinsic relationship between navigation attitude and hydrodynamic resistance, resulting in low energy efficiency and significant energy waste at high speeds. Although advanced methods such as model predictive control have been introduced into the field of vehicle control to improve control performance due to their ability to handle constraints and perform multi-objective optimization, their predictive models typically do not consider the characteristics of the vehicle changing with speed and depth, and the optimization objective functions mostly still focus on trajectory tracking accuracy and maneuverability stability, failing to take "drag," a key physical quantity that determines energy consumption, as an optimization objective. Therefore, existing technologies suffer from a fundamental problem of disconnect between control objectives and hydrodynamic performance optimization, and there is an urgent need for an autonomous control method that can actively seek the optimal drag-reducing attitude during high-speed navigation. Summary of the Invention

[0003] To address the problems of traditional methods, this invention provides a high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: This invention provides a high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization, comprising the following steps: S1: Establish a longitudinal prediction model for underwater vehicles; S2: In each control cycle, acquire the current state information of the underwater vehicle; using the current state information as the initial condition, based on the longitudinal prediction model, and with minimizing the preset objective function as the optimization objective, continuously solve for the optimal control sequence for a future period; wherein, the objective function includes at least the depth tracking error term, the navigation resistance estimation term, the rudder angle change rate limit term, and the angle of attack tracking error term; S3: Output the first control quantity in the optimal control sequence to the actuator of the underwater vehicle to control the navigation attitude and thrust of the underwater vehicle.

[0005] Furthermore, in step S1, the set of state variables involved in the longitudinal prediction model Adopt the following form: In the formula, For depth, The pitch angle is the angle of inclination. For forward speed, Vertical velocity, It is the pitch angular velocity; The set of control variables involved in the longitudinal prediction model Adopt the following form: In the formula, For the stern horizontal rudder angle, For thruster thrust; The longitudinal prediction model consists of a set of differential equations, including kinematic equations and longitudinal dynamic equations: The kinematic equations are: The longitudinal dynamic equation is: In the formula, For the additional quality coefficient, The coefficient of viscosity resistance. For the mass of the aircraft, For pitch rotation inertia, For the angle of attack.

[0006] Moreover, in step S1, the angle of attack The following formula is used to obtain: .

[0007] Furthermore, in step S1, the CFD method is used to calculate in advance the vehicle's speed and depth at different times. A corresponding continuous function is formed by fitting the model, and the longitudinal prediction model is updated in real time according to the vehicle's state during model operation. The expression is as follows: in, These are hydrodynamic parameters.

[0008] Furthermore, in step S1, the longitudinal prediction model is discretized under the sampling period to obtain a discrete state-space model, as follows: In the formula, For discrete time steps, A function to describe the changes in the model; Parameters and speed and depth This indicates that the hydrodynamic parameters are updated online.

[0009] Furthermore, in step S2, the objective function expression is as follows: In the formula, the symbol Indicates in Always The predicted value of the state at time step. To predict the length of the time domain, To control the length of the time domain, Depth tracking error weights, Target depth value, Resistance term weight, Total resistance during navigation Weight of rudder angle change rate The rudder angle is for the stern horizontal rudder. Total resistance Calculate according to the following formula: In the formula, The induced drag coefficient, This refers to the rudder angle drag coefficient; Furthermore, during the rolling solution process in step S2, constraints are applied to the real-time angle of attack of the underwater vehicle. The range of these constraints is determined based on the current speed and depth. The constraint conditions are set as follows: .

[0010] Furthermore, the optimal control sequence obtained from step S2 Extract its first element, which corresponds to the current time. The optimal control command, used for aircraft control, is expressed as follows: In the formula, for Optimal control command for stern rudder angle at all times. for Optimal thrust control command for the constant thruster.

[0011] The advantages and positive effects of this invention are as follows: 1. The prediction model construction process of this invention fully considers the impact of angle of attack on the hydrodynamics of the vehicle. During the control process, the hydrodynamic parameters related to the angle of attack are updated in real time according to the real-time speed and depth, which improves the accuracy of the prediction model.

[0012] 2. The objective function of this invention not only includes the traditional depth tracking error, but also creatively introduces a sailing resistance estimation term related to the real-time angle of attack and rudder angle. This can actively plan the optimal sailing attitude with the lowest energy consumption while ensuring control accuracy, and at the same time suppress the high-speed bow-down phenomenon and improve sailing stability.

[0013] 3. This invention embeds hydrodynamic optimization into real-time closed-loop control, enabling the vehicle to actively maintain the optimal drag-reducing attitude rather than passively maintaining a horizontal position. This significantly reduces the total drag and energy consumption of the underwater vehicle during high-speed navigation, effectively suppresses bow-burying phenomena, and improves navigation stability.

[0014] In summary, this invention can solve the problems of high resistance, high energy consumption, and poor stability caused by neglecting hydrodynamic characteristics in existing underwater vehicle control methods at high speeds. Attached Figure Description

[0015] Figure 1 A flowchart of a high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization provided by the present invention; Figure 2 This is a comparison chart of the angle-of-attack control results provided in this embodiment of the invention with those of conventional solutions; Figure 3 This is a comparison chart of the depth control results provided in this embodiment of the invention and the results of conventional solutions; Figure 4 This is a comparison chart of the induced resistance of the control scheme provided in this embodiment of the invention and the induced resistance of the conventional scheme. Detailed Implementation

[0016] The structure of the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that these embodiments are descriptive and not limiting.

[0017] For an autonomous control method for underwater vehicles with high speed and low energy consumption based on predictive model optimization, please refer to [link / reference]. Figures 1-4 Its inventive point is: including the following steps: S1: Establish a longitudinal prediction model for the underwater vehicle. The longitudinal prediction model integrates the vehicle's kinematic relationships, longitudinal dynamic equations, and hydrodynamic parameters that vary with speed and depth. S2: In each control cycle, acquire the current state information of the underwater vehicle; using the current state information as initial conditions, based on the longitudinal prediction model, and with minimizing the preset objective function as the optimization objective, continuously solve for the optimal control sequence for a future period; specifically: Based on the aforementioned longitudinal prediction model, the state trajectory of an underwater vehicle under different control sequences within a certain time period is predicted. For each predicted state trajectory, calculate its corresponding performance index value according to the objective function; The optimal control sequence is found by using an optimization algorithm to minimize the performance index value.

[0018] S3: Output the first control quantity in the optimal control sequence to the actuator of the underwater vehicle to control the underwater vehicle's attitude and thrust.

[0019] The specific implementation methods for the above three steps are as follows: S1. Establish a longitudinal prediction model for the underwater vehicle: The longitudinal motion of the vehicle includes forward and backward motion, heave motion, and pitch motion. This embodiment combines kinematic and dynamic equations, and the specific form is as follows: In the formula, For the additional quality coefficient, For depth, The pitch angle is the angle of inclination. For forward speed, Vertical velocity, The pitch angular velocity, The coefficient of viscosity resistance. For the mass of the aircraft, For pitch rotation inertia, For the angle of attack.

[0020] In this embodiment, the angle of attack The following formula is used to obtain: To adapt the model to the drastically changing hydrodynamic environment at high speeds, key hydrodynamic parameters related to the angle of attack were designed to vary with the main influencing factor (speed). and depth The system adapts to changes rather than remaining constant; the specific patterns of change are obtained through computational fluid dynamics (CFD), tank testing, or system identification methods. In this embodiment, CFD is used to calculate the vehicle's speed and depth at different times in advance. A corresponding continuous function is formed by fitting the data. Based on this, the prediction model is updated in real time according to the aircraft's state during model operation. The longitudinal prediction model (a continuous-time model) is discretized at the sampling period to obtain a discrete state-space model, as follows: In the formula, For discrete time steps, A function to describe the changes in the model; Parameters and speed and depth This indicates that the hydrodynamic parameters are updated online.

[0021] S2. Construct an optimization objective function; establish an objective function that includes at least a depth tracking error term, a navigation resistance term, and a control variable change rate penalty term. The navigation resistance estimation term includes an induced resistance term, which is calculated based on the real-time angle of attack and rudder angle of the underwater vehicle. The rudder angle change rate penalty term is used to limit the amplitude and rate of control actions and ensure navigation stability. The specific form of the objective function is as follows: In the formula, the symbol Indicates in Always The predicted value of the state at time step. To predict the length of the time domain, To control the length of the time domain, Depth tracking error weights, Target depth value, Resistance term weight, Total resistance during navigation Weight of rudder angle change rate.

[0022] In this embodiment, take , , , , , .

[0023] Total resistance Calculate according to the following formula: In the formula, The induced drag coefficient, This represents the rudder angle drag coefficient.

[0024] In this embodiment, take , .

[0025] During the rolling solution process in step S2, constraints are applied to the real-time angle of attack of the underwater vehicle. The range of these constraints is determined based on the current speed and depth to ensure that the vehicle is in a safe and efficient attitude range.

[0026] The constraints are set as follows: In this embodiment, , , , , ,but: S3. In this embodiment, the quadratic programming solution is performed using the quadprog() function in MATLAB to obtain the optimal command sequence for rudder angle and thrust. ,extract The first element serves as the optimal control command for the current moment, controlling the vehicle. This yields the motion state change curve of the vehicle under the optimal control command, where the angle of attack change curve is shown below. Figure 2 As shown, the depth variation curve is as follows: Figure 3 As shown. Based on the aforementioned induced drag calculation formula, the induced drag curve of the vehicle under optimal control commands can be calculated using the vehicle's angle of attack and stern elevator angle, as shown below. Figure 4 As shown, it is significantly smaller than the conventional approach.

[0027] In summary, the prediction model construction process of this invention fully considers the impact of angle of attack on the hydrodynamics of the vehicle. During the control process, the hydrodynamic parameters related to the angle of attack are updated in real time based on the real-time speed and depth, thereby improving the accuracy of the prediction model. The objective function incorporates a drag estimation term related to the real-time angle of attack and rudder angle, which can proactively plan the optimal sailing attitude with the lowest energy consumption while ensuring control accuracy, and simultaneously suppressing high-speed bow-down phenomena, thus improving sailing stability.

[0028] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization, characterized in that: Includes the following steps: S1: Establish a longitudinal prediction model for underwater vehicles; S2: In each control cycle, acquire the current status information of the underwater vehicle; Using the current state information as the initial condition, based on the longitudinal prediction model, and with minimizing the preset objective function as the optimization objective, the optimal control sequence for a future period of time is solved in a rolling manner. S3: Output the first control quantity in the optimal control sequence to the actuator of the underwater vehicle to control the navigation attitude and thrust of the underwater vehicle.

2. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 1, characterized in that: In step S1, the set of state variables involved in the longitudinal prediction model Adopt the following form: In the formula, For depth, The pitch angle is the angle of inclination. Forward speed, Vertical velocity, It is the pitch angular velocity; The set of control variables involved in the longitudinal prediction model Adopt the following form: In the formula, For the stern horizontal rudder angle, For thruster thrust; The longitudinal prediction model consists of a set of differential equations, including kinematic equations and longitudinal dynamic equations: The kinematic equations are: The longitudinal dynamic equation is: In the formula, For the additional quality coefficient, The coefficient of viscosity resistance. For the mass of the aircraft, For pitch rotation inertia, For the angle of attack.

3. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 2, characterized in that: In step S1, the angle of attack The following formula is used to obtain: 。 4. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 2, characterized in that: In step S1, the CFD method is used to calculate the vehicle's speed and depth at different times in advance. A corresponding continuous function is formed by fitting the model, and the longitudinal prediction model is updated in real time according to the vehicle's state during model operation. The expression is as follows: in, These are hydrodynamic parameters.

5. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 2, characterized in that: In step S1, the longitudinal prediction model is discretized under the sampling period to obtain a discrete state-space model, as follows: ； In the formula, For discrete time steps, A function to describe the changes in the model; Parameters and speed and depth This indicates that the hydrodynamic parameters are updated online.

6. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 1, characterized in that: In step S2, the objective function includes at least a depth tracking error term, a sailing resistance estimation term, a rudder angle change rate limit term, and an angle of attack tracking error term. The objective function expression is as follows: In the formula, the symbol Indicates in Always The predicted value of the state at time step. To predict the length of the time domain, To control the length of the time domain, Depth tracking error weights, Target depth value, Resistance term weight, Total resistance during navigation Weight of rudder angle change rate The rudder angle is for the stern horizontal rudder. Total resistance Calculate according to the following formula: In the formula, The induced drag coefficient, This represents the rudder angle drag coefficient.

7. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 2, characterized in that: During the rolling solution process in step S2, constraints are applied to the real-time angle of attack of the underwater vehicle. The range of these constraints is determined based on the current speed and depth. The constraint conditions are set as follows: 。 8. The high-speed, low-energy autonomous control method for underwater vehicles based on predictive model optimization according to claim 1, characterized in that: The optimal control sequence obtained from step S2 Extract its first element, which corresponds to the current time. The optimal control command, used for aircraft control, is expressed as follows: In the formula, for Optimal control command for stern rudder angle at all times. for Optimal thrust control command for the constant thruster.

Images