Underwater robot adaptive iterative learning trajectory tracking system and method for complex environment

Abstract

The invention discloses a complex environment-oriented underwater robot adaptive iterative learning trajectory tracking system and method, which are used for solving the problem that the existing iterative learning control method does not consider non-linear factors and current operation duration fluctuation, and the method comprises the following steps: obtaining input and output data of a system under current iteration through a data collection and preprocessing module; calculating the tracking error of the current iteration according to the actual output and the expected output; a self-adaptive updating law of the underwater robot is designed according to the current iteration tracking error; designing a current iteration underwater robot controller through the current tracking error and the self-adaptive updating law; storing the current data in a memory so as to facilitate the next iteration data extraction; and finally, enabling the controller to act on the underwater robot propeller, and obtaining input and output data of next iteration through an underwater robot dynamic system.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of automatic control, and particularly relates to a complex environment-oriented underwater robot adaptive iterative learning trajectory tracking system and method. BACKGROUND

[0002] The existing iterative learning control technology usually requires that the system model, initial state, reference trajectory and test length remain unchanged in the iteration process when dealing with trajectory tracking tasks. However, in marine exploration, non-uniform operation period challenges are often encountered, specifically, the underwater robot sonar scanning task is forced to terminate early due to equipment failure or sudden environmental interference, and the mechanical arm seabed welding process triggers an emergency stop mechanism due to joint failure or sudden obstacles.

[0003] In the field of adaptive iterative learning control, most existing researches are directed to single-input single-output systems, and for multi-input multi-output systems, the control gain matrix is usually required to be symmetric positive definite (or negative definite) or known. For example, in the traditional method, the control gain matrix needs to satisfy the strict symmetric positive definite condition, which limits its application in complex underwater asymmetric gain environments such as underwater thruster multi-directional thrust control and robot dynamic load adjustment. In addition, the existing technology has limited processing capacity for iterative time-varying reference trajectories.

[0004] In the field of trajectory tracking control, iterative learning control is an effective strategy for improving system tracking performance by utilizing historical control experience. However, the traditional ILC method has the following limitations:

[0005] (1) The system model, initial state, reference trajectory and test length are required to be strictly unchanged, which is difficult to cope with uncertainties in practical applications (such as autonomous operation system of deep diving robot in deep sea exploration task, and emergency stop mechanism triggered by component failure in precise operation process of underwater manipulator).

[0006] (2) For MIMO nonlinear systems, existing adaptive iterative learning control methods usually require the control gain matrix to be symmetric positive definite or known, which limits its application in asymmetric matrix scenarios (such as robot dynamics and industrial stirring system).

[0007] (3) Most algorithms are only applicable to single-input single-output systems or fixed reference trajectories, and cannot handle iterative changes in tracking targets. SUMMARY

[0008] In view of the problems in the prior art, the application provides a complex environment-oriented underwater robot adaptive iterative learning trajectory tracking system and method, solves the strong constraint problem of a control gain matrix of a multiple-input multiple-output nonlinear system in the prior art, realizes high-precision trajectory tracking under a non-uniform test length, supports an iteratively changed reference trajectory and unknown external interference, and thus realizes adaptive iterative learning control of underwater robot trajectory tracking facing an uncertain river basin environment.

[0009] The technical scheme of the application is implemented as follows:

[0010] A complex environment-oriented underwater robot adaptive iterative learning trajectory tracking method comprises the following steps:

[0011] S1, initial state information of an underwater robot and input and output data during operation are collected, and the input and output data are preprocessed; the data before and after processing are stored;

[0012] S2, a dynamic system model of the underwater robot is constructed according to the stored and preprocessed data, actual underwater robot speed states and expected underwater robot speed are obtained according to the dynamic system model of the underwater robot, and a tracking error of the current iteration is calculated;

[0013] S3, an adaptive parameter update law of the underwater robot is designed according to the tracking error of the current iteration;

[0014] S4, an adaptive iterative learning control algorithm of the underwater robot of the current iteration is designed through the tracking error of the current iteration and the adaptive parameter update law;

[0015] S5, an underwater robot controller is applied to a propeller of the underwater robot, and input and output data of the next iteration are obtained through the dynamic system model of the underwater robot.

[0016] Further, in S1, the initial state information of the underwater robot and the input and output data during operation are collected through a data collection and preprocessing module;

[0017] The input data include input force of the propeller in horizontal and vertical directions, and the output data include data collected by various sensors arranged on the underwater robot, speed information of the underwater robot and underwater environment disturbance information; when it is detected that the data collected by the sensors are abnormal, redundant sensor data are used for replacement.

[0018] Further, the data before and after processing are stored through a memory module, and the memory module is used for storing historical data of the underwater robot during operation, preset trajectory data, error data of each iteration and parameters of the adaptive algorithm;

[0019] The historical data include sensor data, control instruction data, and actual running track data collected in each running process.

[0020] The preset track data include track time length of each iteration.

[0021] The parameters of the adaptive algorithm include learning gain in the adaptive iterative learning control algorithm.

[0022] Further, in the S2, a dynamic system model of the underwater robot is constructed, and specifically, the state of the underwater robot is defined as:

[0023]

[0024] wherein M is a mass matrix, C(·) is a Coriolis matrix, D(·) is a damping matrix, w k (t) is a water flow disturbance,

[0025] is a velocity and angular velocity vector, u k (t) is a propeller input, k is an iteration number, and t is a running time length of the underwater robot.

[0026] Further, the state of the underwater robot is converted into a multiple-input multiple-output nonlinear system, that is:

[0027]

[0028] wherein f(x k (t), t) ∈ R n is a bounded nonlinear unknown function; x k (t) ∈ R n is a measurable velocity state of the underwater robot; u k (t) ∈ R n and d k (t) ∈ R n are input and disturbance respectively, t ∈ [0, T], B(t) is a gain matrix, and k = 0, 1, 2, ….

[0029] Further, in the S2, according to initial state information and a preset track route of the underwater robot, a tracking error of the underwater robot in the first iteration running is obtained, and specifically, the method is as follows:

[0030] E k (t) = r d (t) - x k (t);

[0031] wherein E k (t) is a tracking error, r d (t) is an expected underwater robot velocity, and xk (t) is the actual underwater robot speed;

[0032] According to the adaptive parameter update law, the learning gain parameter is adaptively adjusted, so as to further generate the control instruction through the adaptive iterative learning control algorithm, and drive the thruster of the underwater robot to act;

[0033] According to the underwater robot dynamic system model, the tracking error of the current actual underwater robot speed and the preset underwater robot speed is calculated and input to the next iteration process input instruction;

[0034] When the underwater robot is not executed in the full time length in the current iteration, and returns in the middle of the way, the system runs in the non-uniform test length condition, and the duration of each iteration task is set as τ k , and the improved tracking error is defined as:

[0035]

[0036] Wherein, f k (t) is the improved tracking error, and the iteration task duration of the underwater robot is τ k ;

[0037] According to the improved tracking error, it is obtained that:

[0038] f k (t) = λ k E k (t) + (1- λ k ) E k (τ k );

[0039] Wherein, λ k is a conditional parameter, which is 0 or 1;

[0040] When λ k = 1, the system stops returning in the middle of the execution of the task, and does not reach the expected test length;

[0041] When λ k = 0, the system reaches the expected time length.

[0042] Further, according to the constructed underwater robot dynamic system model, the adaptive iterative learning control module is adopted, and the adaptive iterative learning control algorithm is adopted to track and control the changing trajectory length of the underwater robot;

[0043] In each iteration process, according to the error between the actual trajectory of the current iteration and the preset trajectory, the learning gain parameter is adaptively adjusted, so that the control algorithm converges to the optimal control strategy;

[0044] By calculating the tracking error of the speed, the control input is updated by using the adaptive iterative learning control algorithm, the control instruction of the next moment is generated, the thruster of the underwater robot is driven to execute, and the running track of the underwater robot is adjusted.

[0045] Further, in the S3, according to the tracking error of the underwater robot dynamic system model, the control input update law is designed as:

[0046]

[0047] Wherein, u k (t) is an adaptive iterative learning control algorithm, ξ(E k (t))=[E k (t)sign(E k (t))]∈R n×2 , sign(·) is a sign function, is an adaptive parameter, is a gain matrix to be designed;

[0048] According to the tracking error of the underwater robot dynamic system model, the adaptive parameter update law is designed as:

[0049]

[0050] Wherein, Γ is a positive definite symmetric gain matrix, τ k is the task cycle of the kth iteration, which can be non-uniformly changed under the influence of the basin environment, is the adaptive parameter of the previous iteration, is the adaptive parameter of the current iteration.

[0051] A kind of underwater robot adaptive iterative learning track tracking system for complex environment, characterized in that, the underwater robot adaptive iterative learning track tracking method for complex environment of any one described above is applied.

[0052] Compared with the prior art, the present application has the following beneficial effects:

[0053] The present application provides an underwater robot adaptive iterative learning track tracking system and method for complex environment, an adaptive iterative learning controller is designed for the underwater robot system with non-uniform test length, breaking through the limitation of traditional iterative controller on fixed test length and unchanged reference track; it is suitable for the situation that test length changes due to abnormal interruption in industrial manufacturing, robot control and other actual scenes, and improves the system robustness.

[0054] Only the control gain matrix is required to be reversible and have rough prior knowledge, relaxing the strict restriction of traditional adaptive iterative learning method on symmetric positive definite gain matrix. It can be applied to underwater robots and other systems with non-symmetric or unknown control gain matrix, expanding the scope of algorithm application.

[0055] The introduction of composite energy function proves convergence, and the combination of adaptive parameter estimation improves the robustness to system uncertainty. Through theoretical proof and simulation verification, it is ensured that the tracking error converges to zero with the increase of iteration number, and the control precision is improved.

[0056] Relax the constraint of control gain matrix, only require the matrix to be reversible, without symmetric positive definite or known, applicable to more extensive industrial scenarios, such as non-symmetric 2 degree of freedom underwater robot dynamics matrix.

[0057] Adapt to non-uniform test length, dynamically adjust parameters through adaptive law, effectively handle test length changes caused by faults or detection abnormalities.

[0058] Support iterative change reference trajectory: can track different iteration period of differential target trajectory, improve system flexibility.

[0059] Strong robustness: through composite energy function design, suppress external disturbance and system uncertainty, ensure tracking accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 is a flow chart of an adaptive iterative learning trajectory tracking method for underwater robots in complex environments provided in an embodiment of the present application;

[0061] Figure 2 is an adaptive iterative learning control flow chart for underwater robots in complex environments provided in an embodiment of the present application;

[0062] Figure 3 is a test length diagram of an adaptive iterative learning trajectory tracking method for underwater robots in complex environments provided in an embodiment of the present application under different iteration numbers;

[0063] Figure 4 is a trajectory tracking error diagram of an adaptive iterative learning trajectory tracking method for underwater robots in complex environments provided in an embodiment of the present application under different iteration test lengths;

[0064] Figure 5 is an underwater robot control input diagram of an adaptive iterative learning trajectory tracking method for underwater robots in complex environments provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] In order to make the purposes, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of 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 fall within the scope of protection of the present application.

[0066] Embodiments

[0067] As Figures 1 to 5 A self-adaptive iterative learning trajectory tracking method for an underwater robot in a complex environment includes the following steps:

[0068] S1, initial state information of the underwater robot and input and output data at runtime are collected, and the input and output data are preprocessed; the data before and after processing are stored;

[0069] S2, a dynamic system model of the underwater robot is constructed according to the stored and preprocessed data, actual underwater robot speed state and expected underwater robot speed are obtained according to the dynamic system model of the underwater robot, and tracking error of the current iteration is calculated;

[0070] S3, an adaptive parameter update law of the underwater robot is designed according to the tracking error of the current iteration;

[0071] S4, an adaptive iterative learning control algorithm of the underwater robot for the current iteration is designed through the tracking error of the current iteration and the adaptive parameter update law;

[0072] The current data is stored in the memory for data extraction in the next iteration;

[0073] S5, the underwater robot controller acts on the underwater robot thruster, and input and output data for the next iteration are obtained through the dynamic system model of the underwater robot.

[0074] In S1, the initial state information of the underwater robot and the input and output data at runtime are collected through the data collection and preprocessing module; and the data are preprocessed.

[0075] The input data includes the input force of the thruster in horizontal and vertical directions, and the output data includes the data collected by various sensors arranged on the underwater robot, such as an inertial measurement unit (IMU), speed information of the underwater robot, horizontal X-axis and vertical Y-axis speeds, and underwater environmental disturbance information (i.e., surrounding environment, water flow speed, and other disturbance information); accurate data basis is provided for subsequent control. When it is detected that the data collected by the sensor is abnormal, for example, the underwater robot turns back halfway through the operation process, the data will inevitably be missing in the next iteration operation process, in which case redundant sensor data is used to replace it to ensure the integrity and continuity of the data.

[0076] The data before and after processing are stored through a memory module, which is used to store historical data of the operation process of the underwater robot, preset trajectory data, tracking error data of each iteration, and adaptive parameters;

[0077] The historical data includes sensor data, control instruction data, and actual operation trajectory data collected during each operation process; these historical data provide learning samples for adaptive iterative learning control.

[0078] The preset trajectory data is a trajectory planned in advance according to the operation task of the underwater robot, including the trajectory time length of each iteration and other data;

[0079] The parameters of the adaptive algorithm include learning gain in the adaptive iterative learning control algorithm, which can be adjusted and optimized according to the actual situation and stored in the memory module for easy data calling and analysis.

[0080] In S2, a dynamic system model of the underwater robot is constructed, specifically including defining the state of the underwater robot as:

[0081]

[0082] wherein M is a mass matrix, C(·) is a Coriolis matrix, D(·) is a damping matrix, w k (t) is a water flow disturbance,

[0083] is an acceleration and angular acceleration vector, u k (t) is a thruster input, k is the number of iterations, and t is the length of the underwater robot operation time.

[0084] The state of the underwater robot is converted into a multi-input multi-output nonlinear system, that is:

[0085]

[0086] wherein f(x k (t), t) ∈ Rn is a bounded nonlinear unknown function; x k (t)∈R n is the measurable velocity of the underwater robot; u k (t)∈R n and d k (t)∈R n are the thruster input and disturbance respectively, t∈[0,T], B(t) is the gain matrix, k=0,1,2,….

[0087] Through the underwater robot dynamic system construction module, based on the accurate data provided by the data collection and preprocessing module and the related information stored in the memory module, the underwater robot dynamic system model is constructed; considering the influence of hydrodynamic force, flow force, gravity, buoyancy and other factors on the underwater robot in complex environment, the state space model of the underwater robot is established by using dynamics and kinematics principles. The model contains the velocity state variables of the underwater robot, and the control input variables such as the speed of the thruster, the angle of the rudder, etc. Through identification and optimization of the parameters of the model, it can accurately describe the dynamic behavior of the underwater robot in complex environment, and provide reliable system model basis for adaptive iterative learning control.

[0088] In S2, according to the initial state information of the underwater robot and the preset trajectory route, the tracking error of the underwater robot in the first iteration is obtained, and the specific way is:

[0089] E k (t)=r d (t)-x k (t);

[0090] Wherein, E k (t) is the tracking error, r d (t) is the expected underwater robot speed, x k (t) is the actual underwater robot speed;

[0091] According to the adaptive parameter updating law, the learning gain parameter is adaptively adjusted, so as to further generate the control instruction through the adaptive iterative learning control algorithm, and drive the thruster of the underwater robot to act;

[0092] According to the underwater robot dynamic system model, the tracking error between the current actual trajectory and the preset trajectory is calculated and input to the input instruction of the next iteration process;

[0093] When the underwater robot is not executed in the current iteration for the full length of time, and after turning back halfway, the system runs under the condition of non-uniform test length, the duration of each iteration task is set as τ k , and the improved tracking error is defined as:

[0094]

[0095] wherein, f k (t) is the improved tracking error, and the iterative task duration of the underwater robot is τ k ;

[0096] According to the improved tracking error, it is obtained that:

[0097] f k (t) = λ k E k (t) + (1- λ k ) E k (τ k );

[0098] wherein, wherein, λ k is a conditional parameter, and is 0 or 1;

[0099] When λ k = 1, the system occurs the situation of stopping returning in the middle of performing the task, and does not reach the expected test length;

[0100] On the contrary, when λ k = 0, the system reaches the expected time length.

[0101] According to the constructed dynamic system model of the underwater robot, through the adaptive iterative learning control module, the adaptive iterative learning control algorithm is adopted to track and control the changing trajectory length of the underwater robot;

[0102] In each iteration process, according to the error between the actual trajectory of the current iteration and the preset trajectory, the learning gain parameter is adaptively adjusted, so that the control algorithm converges to the optimal control strategy;

[0103] By calculating the error of the speed, the adaptive iterative learning control algorithm is used to update the control input, generate the control instruction at the next moment, drive the thruster of the underwater robot to execute, and adjust the running trajectory of the underwater robot.

[0104] At the same time, considering the time-varying and uncertainty of the underwater environment, in the iterative learning process, the environmental information collected and preprocessed by the data collection and preprocessing module is combined to dynamically adjust the control strategy to adapt to the changes of the complex environment, so as to realize the high-precision tracking of the underwater robot on the trajectory of different lengths in the iterative process.

[0105] In the S3, according to the tracking error of the dynamic system model of the underwater robot, the adaptive iterative learning control algorithm is designed as:

[0106]

[0107] wherein, u k (t) is an adaptive iterative learning control algorithm, ξ(E k (t))=[E k (t) sign(E k (t))]∈R n×2 , sign(·) is a sign function, is an adaptive parameter, is a gain matrix to be designed;

[0108] According to the tracking error of the underwater robot dynamic system model, an adaptive parameter updating law is designed:

[0109]

[0110] wherein, Γ is a positive definite symmetric gain matrix, τ k is the task cycle of the kth iteration, which can be non-uniformly changed under the influence of the river basin environment, is the adaptive parameter of the previous iteration, is the adaptive parameter of the current iteration.

[0111] The updating mode of the parameters contained in the adaptive iterative learning control algorithm is the adaptive parameter updating law.

[0112] A kind of underwater robot adaptive iterative learning trajectory tracking system for complex environment, it is characterized in that, a kind of underwater robot adaptive iterative learning trajectory tracking method for complex environment as described above is applied.

[0113] The present application provides a kind of underwater robot adaptive iterative learning trajectory tracking system and method for complex environment, i.e. a kind of MIMO nonlinear system and control method suitable for non-uniform test length, reversible control gain matrix and supporting iterative change reference trajectory are developed, with important practical significance.

[0114] According to the disclosure and teaching of the above specification, those skilled in the art of the present application can also make changes and modifications to the above embodiments. Therefore, the present application is not limited to the specific embodiments disclosed and described above, and some modifications and changes of the present application should fall within the protection scope of the claims of the present application. In addition, although some specific terms are used in the present specification, these terms are only for convenience of explanation and do not constitute any limitation on the present application.

Claims

1. An adaptive iterative learning trajectory tracking method for underwater robots in complex environments, characterized in that, Includes the following steps: S1. Collect the initial state information and input and output data of the underwater robot during operation, and preprocess the input and output data; store the data before and after processing. S2. Construct an underwater robot dynamic system model based on the stored and preprocessed data. Obtain the actual underwater robot speed state and the expected underwater robot speed based on the underwater robot dynamic system model, and calculate the tracking error of the current iteration. S3. Design an adaptive parameter update law for the underwater robot based on the tracking error of the current iteration; S4. Design an adaptive iterative learning control algorithm for the underwater robot in the current iteration based on the tracking error and the adaptive parameter update law in the current iteration. S5. Apply the underwater robot controller to the underwater robot thruster, and obtain the input and output data for the next iteration through the underwater robot dynamic system model.

2. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 1, characterized in that, In S1, the initial state information and input and output data of the underwater robot during operation are collected through the data collection and preprocessing module. The input data includes the input force of the thrusters in both horizontal and vertical directions. The output data includes data collected by various sensors installed on the underwater robot, the underwater robot's speed information, and underwater environmental disturbance information. When abnormal data collected by the sensors is detected, redundant sensor data is used as a replacement.

3. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 2, characterized in that, The memory module stores the data before and after processing. The memory module is used to store historical data of the underwater robot's operation, preset trajectory data, error data of each iteration, and adaptive parameters. Historical data includes sensor data, control command data, and actual running trajectory data collected during each operation; The preset trajectory data includes the trajectory time length for each iteration; The parameters of the adaptive algorithm include the learning gain in the adaptive iterative learning control algorithm.

4. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 1, characterized in that, In step S2, constructing a dynamic system model for the underwater robot specifically includes defining the state of the underwater robot as follows: Where M is the mass matrix, C(·) is the Coriolis matrix, D(·) is the damping matrix, and w k (t) represents water flow disturbance. Let u be the acceleration and angular acceleration vector. k (t) represents the thruster input, k represents the number of iterations, and t represents the underwater robot's running time.

5. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 4, characterized in that, The state of the underwater robot is transformed into a multi-input multi-output nonlinear system, namely: Where, f(x) k (t), t)∈R n It is a bounded, nonlinear, unknown function; x k (t)∈R n The speed of an underwater robot can be measured; u k (t)∈R n and d k (t)∈R n Let t be the thruster input and disturbance, respectively, t∈[0,T], and B(t) be the gain matrix, k=0,1,2,….

6. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 5, characterized in that, In step S2, the tracking error of the underwater robot in the first iteration is obtained based on the initial state information and preset trajectory of the underwater robot. Specifically, the method is as follows: E k (t)=r d (t)-x k (t); Among them, E k (t) represents the tracking error, r d (t) represents the desired underwater robot velocity, x k (t) represents the actual speed of the underwater robot; Based on the adaptive parameter update law, the learning gain parameter is adaptively adjusted, thereby generating control commands through the adaptive iterative learning control algorithm to drive the underwater robot's thruster action. The tracking error between the current actual underwater robot speed and the preset underwater robot speed is calculated based on the underwater robot dynamic system model and input into the input command of the next iteration process; When the underwater robot does not complete the entire time duration in the current iteration and turns back midway, the system operates under non-uniform test length conditions. The duration of each iteration task is set to τ. k The improved tracking error is defined as: Among them, f k (t) represents the improved tracking error, and the duration of the underwater robot's iterative task is τ. k ; Based on the improved tracking error, we conclude that: f k (t)=λ k E k (t)+(1-λ k )E k (t k ); Where, λ k The parameter is conditional and can be either 0 or 1; When λ k When = 1, the system stops and returns midway through the task, without reaching the expected test length; When λ k When = 0, it is the length of time it takes for the system to reach the desired state.

7. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 6, characterized in that, Based on the constructed underwater robot dynamic system model, the adaptive iterative learning control module and the adaptive iterative learning control algorithm are used to track and control the changing trajectory length of the underwater robot. In each iteration, the learning gain parameter is adaptively adjusted based on the tracking error between the actual underwater robot speed and the preset underwater robot speed in the current iteration, so that the control algorithm converges to the optimal control strategy. By calculating the tracking error of the speed, the control input is updated using an adaptive iterative learning control algorithm to generate the control command for the next moment, which drives the underwater robot's thrusters to execute and adjust the underwater robot's trajectory.

8. The adaptive iterative learning trajectory tracking method for underwater robots in complex environments according to claim 6, characterized in that, In step S3, based on the tracking error of the underwater robot dynamic system model, an adaptive iterative learning control algorithm is designed as follows: Among them, u k (t) represents the adaptive iterative learning control algorithm, ξ(E) k (t))=[E k (t)sign(E k (t))]∈R n×2 sign(·) is the sign function. For adaptive parameters, The gain matrix to be designed; Based on the tracking error of the underwater robot dynamic system model, an adaptive parameter update law is designed: Where Γ is a positive definite symmetric gain matrix, τ k The task cycle for the k-th iteration can vary non-uniformly due to the influence of the watershed environment. These are the adaptive parameters from the previous iteration. These are the adaptive parameters for the current iteration.

9. An adaptive iterative learning trajectory tracking system for underwater robots in complex environments, characterized in that, An adaptive iterative learning trajectory tracking method for underwater robots in complex environments, as described in any one of claims 1 to 8, is applied.

Images