Underwater robot state prediction method, system, device and storage medium

Abstract

The present invention discloses a method, system, device and storage medium for predicting the state of an underwater robot. The method first establishes a physically driven dynamic model based on the analysis of the underwater robot system structure; then, a data-driven compensation model is constructed through the dynamic model of the underwater robot and the observed state collected by the onboard sensor; finally, the state prediction of the underwater robot is completed by fusing the physically driven dynamic model and the data-driven compensation model. Therefore, the method of the embodiment of the present invention takes into account the influence of the dynamic environment on the motion state of the underwater robot, combines the physically driven dynamic model and the data-driven compensation model, solves the problem of mismatch between the physically driven model and the observed state of the underwater robot in the dynamic environment, and finally improves the state prediction accuracy and robustness of the underwater robot, and has a strong application prospect.

Description

Technical Field

[0001] The present invention relates to the field of underwater observation technology, and in particular to an underwater robot state prediction method, system, equipment and storage medium. Background Art

[0002] As a new type of ocean observation platform, underwater robots have the advantages of autonomous control, energy saving, and strong mobility. They are widely used in various aspects such as ocean research, marine engineering, and marine development. State prediction requires the ability to accurately predict the motion parameters of underwater robots in unknown environments, which is the basis for underwater robot system optimization, motion control, and route planning.

[0003] Existing underwater robot state prediction methods are mostly based on physical models established by system structure, which rely on basic physical laws and a series of ideal assumptions to establish mathematical expressions. However, these physical models ignore the impact of dynamic environment on underwater robots and have low prediction accuracy. Data-driven models represented by neural networks have the advantages of high accuracy and strong scalability. However, the accuracy of data-driven models depends on the quantity and quality of data, and the robustness of the model is poor. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a method for predicting the state of an underwater robot, which solves the problem that the current pure physical model has low prediction accuracy for the underwater robot and the pure data-driven model has poor robustness in predicting the underwater robot.

[0005] The present invention also provides an underwater robot state prediction system, an underwater robot state prediction device and a computer-readable storage medium.

[0006] The method for predicting the state of an underwater robot according to the first embodiment of the present invention comprises the following steps:

[0007] Based on the onboard sensors, the observation status of the underwater robot is collected;

[0008] Establishing a physically driven dynamic model and inputting control inputs of the underwater robot to obtain a predicted state of the underwater robot;

[0009] Obtaining a state residual according to the observed state and the predicted state;

[0010] Establishing a data-driven compensation model and inputting the predicted state, the state residual and the control input to obtain a predicted residual;

[0011] A fusion state is obtained according to the predicted state and the predicted residual to achieve state prediction of the underwater robot.

[0012] The data-driven underwater robot state prediction method according to the embodiment of the present invention has at least the following beneficial effects:

[0013] First, a physically driven dynamic model is established based on the analysis of the underwater robot system structure; then, a data-driven compensation model is constructed through the underwater robot's dynamic model and the observed state collected by the onboard sensor; finally, the state prediction of the underwater robot is completed by integrating the physically driven dynamic model and the data-driven compensation model. Therefore, the method of the embodiment of the present invention takes into account the influence of the dynamic environment on the motion state of the underwater robot, combines the physically driven dynamic model and the data-driven compensation model, solves the problem of mismatch between the physically driven model and the observed state of the underwater robot in the dynamic environment, and finally improves the state prediction accuracy and robustness of the underwater robot, which has a strong application prospect.

[0014] According to some embodiments of the present invention, the kinetic model is constrained by the following mathematical expression:

[0015]

[0016]

[0017] in, Denotes the derivative of η, η=[η 1 ,η 2 ] represents the state of the underwater robot in the earth coordinate system, that is, the predicted state, η 1 =[x,y,z] T represents the position of the underwater robot, η 2 =[φ,θ,Ψ] T represents the posture of the underwater robot, represents the derivative of ν, ν=[ν 1 ,ν 2 ] represents the state of the underwater robot in the body coordinate system, ν 1 =[u,v,w] T represents the speed of the underwater robot, ν 2 =[p,q,r] T represents the angular velocity of the underwater robot, C(η) is the vector transformation matrix from the earth coordinate system to the body coordinate system, and I m represents the mass matrix of the underwater robot, I c represents the Coriolis and centrifugal matrices, τ(i) represents the external force and torque acting on the underwater robot, and i represents the control input of the underwater robot.

[0018] According to some embodiments of the present invention, the state residual obtained according to the observed state and the predicted state is constrained by the following mathematical expression:

[0019] η r =η s -η,

[0020] Wherein, η represents the predicted state, η s represents the observed state, η r represents the state residual.

[0021] According to some embodiments of the present invention, the compensation model is constrained by the following mathematical expression:

[0022]

[0023] in, represents the prediction residual, f represents the compensation model, η r (t-1) represents the state residual at the previous time, η(t-1) represents the predicted state at the previous time, i(t-1) represents the control input at the previous time, and t represents time.

[0024] According to some embodiments of the present invention, the compensation model is a neural network or a nonlinear fitting equation, and the optimization problem of the compensation model is constrained by the following mathematical expression:

[0025]

[0026] Wherein, f represents the compensation model, argmin represents the variable value when the compensation model takes the minimum value, T represents the time period of the underwater robot's motion state, represents the prediction residual, η r (t) represents the state residual.

[0027] According to some embodiments of the present invention, the fusion state is obtained according to the prediction state and the prediction residual, and is constrained by the following mathematical expression:

[0028]

[0029] in, represents the fusion state, represents the prediction residual, and η(t) represents the prediction state.

[0030] The underwater robot state prediction system according to the second aspect of the present invention includes:

[0031] An observation and acquisition module is used to collect the observation status of the underwater robot based on the onboard sensors;

[0032] A first model processing module is used to establish a physically driven dynamic model and input a control input of the underwater robot to obtain a predicted state of the underwater robot;

[0033] A first calculation module, used for obtaining a state residual according to the observed state and the predicted state;

[0034] A second module processing module is used to establish a data-driven compensation model and input the predicted state, the state residual and the control input to obtain a predicted residual;

[0035] The second calculation module is used to obtain a fusion state according to the predicted state and the predicted residual to achieve state prediction of the underwater robot.

[0036] The underwater robot state prediction system according to the embodiment of the present invention has at least the following beneficial effects:

[0037] First, a physically driven dynamic model is established based on the analysis of the underwater robot system structure; then, a data-driven compensation model is constructed through the underwater robot's dynamic model and the observed state collected by the onboard sensor; finally, the state prediction of the underwater robot is completed by integrating the physically driven dynamic model and the data-driven compensation model. Therefore, the system of the embodiment of the present invention takes into account the influence of the dynamic environment on the motion state of the underwater robot, combines the physically driven dynamic model and the data-driven compensation model, solves the problem of mismatch between the physically driven model and the observed state of the underwater robot in the dynamic environment, and finally improves the state prediction accuracy and robustness of the underwater robot, which has a strong application prospect.

[0038] According to an embodiment of the third aspect of the present invention, an underwater robot state prediction system device includes at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the underwater robot state prediction method as described in any one of the embodiments of the first aspect of the present invention.

[0039] According to the computer-readable storage medium of the fourth aspect embodiment of the present invention, the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the underwater robot state prediction method as described in any one of the embodiments of the first aspect of the present invention.

[0040] It can be understood that the beneficial effects of the third and fourth aspects compared with the relevant technologies are the same as the beneficial effects of the first aspect compared with the relevant technologies. Please refer to the relevant description in the first aspect and will not be repeated here.

[0041] Other features and advantages of the present invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0043] Figure 1 is a flow chart of an underwater robot state prediction method according to an embodiment of the present invention;

[0044] Figure 2 It is a schematic diagram of constructing a reference coordinate system for the motion state of an underwater robot according to an embodiment of the present invention;

[0045] Figure 3 It is a schematic diagram of a model construction of an underwater robot state prediction method according to an embodiment of the present invention;

[0046] Figure 4 It is a schematic diagram of constructing a data-driven compensation model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0048] In the description of the present invention, if there is a description of first, second, etc., it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0049] In the description of the present invention, it should be understood that descriptions involving orientation, such as orientation or positional relationship indicated as up, down, etc., are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0050] In the description of the present invention, it should be noted that, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention in combination with the specific content of the technical solution.

[0051] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described below are only some embodiments of the present invention, not all embodiments.

[0052] See also Figure 1 FIG. 1 is a flow chart of a method for predicting the state of an underwater robot provided by an embodiment of the present invention, and the method comprises the following steps:

[0053] Based on the onboard sensors, the observation status of the underwater robot is collected;

[0054] Establish a physically driven dynamic model and input the control input of the underwater robot to obtain the predicted state of the underwater robot;

[0055] Obtain the state residual according to the observed state and the predicted state;

[0056] Establish a data-driven compensation model and input the predicted state, state residual and control input to obtain the predicted residual;

[0057] The fusion state is obtained according to the predicted state and the predicted residual to realize the state prediction of the underwater robot.

[0058] Specifically, Figure 1 As shown, it can be understood that the observed state of the underwater robot collected by the onboard sensor is the actual state of the underwater robot, and the predicted state processed by the established dynamic model is the ideal state of the underwater robot. Therefore, the deviation between the actual state and the ideal state is obtained, that is, the state residual is obtained.

[0059] Furthermore, the established compensation model is used to continuously compensate and correct the state residual, and the processing output is an optimization quantity, namely the prediction residual. Finally, the prediction residual, the optimization quantity, is fused with the predicted state to obtain the fused state, which is used as the final predicted underwater robot state, thereby completing the underwater robot state prediction.

[0060] In this embodiment, firstly, a physically driven dynamic model is established based on the analysis of the underwater robot system structure; then, a data-driven compensation model is constructed through the dynamic model of the underwater robot and the observed state collected by the onboard sensor; finally, the state prediction of the underwater robot is completed by integrating the physically driven dynamic model and the data-driven compensation model. Therefore, the method of the embodiment of the present invention takes into account the influence of the dynamic environment on the motion state of the underwater robot, combines the physically driven dynamic model and the data-driven compensation model, solves the problem of mismatch between the physically driven model and the observed state of the underwater robot in the dynamic environment, and finally improves the state prediction accuracy and robustness of the underwater robot, which has a strong application prospect.

[0061] In some embodiments, Figure 2 As shown, the dynamic model is constrained by the following mathematical expression:

[0062]

[0063]

[0064] in, Denotes the derivative of η, η=[η 1 ,η 2 ] represents the state of the underwater robot in the earth coordinate system, that is, the predicted state, η 1 =[x,y,z] T represents the position of the underwater robot, η 2 =[φ,θ,Ψ] T represents the posture of the underwater robot, represents the derivative of ν, ν=[ν 1 ,ν 2 ] represents the state of the underwater robot in the body coordinate system, ν 1 =[u,v,w] T represents the speed of the underwater robot, ν 2 =[p,q,r] T represents the angular velocity of the underwater robot, C(η) is the vector transformation matrix from the earth coordinate system to the body coordinate system, and I m represents the mass matrix of the underwater robot, I c represents the Coriolis and centrifugal matrices, τ(i) represents the external force and torque acting on the underwater robot, and i represents the control input of the underwater robot.

[0065] Specifically, it can be understood that the first mathematical expression of this embodiment is a kinematic equation, and the second mathematical expression of this embodiment is a dynamic equation. By inputting the control input of the underwater robot and based on other known parameters, the predicted state of the underwater robot can be obtained according to the above two mathematical expressions.

[0066] Further, refer to Figure 2 For underwater robots, there are two reference coordinate systems in their motion space, namely the earth coordinate system and the body coordinate system. 1 E 2 E 3 The specific components include: origin E, horizontal axis EE 1 , transverse axis EE 2 At sea level, longitudinal axis EE 3Pointing to the center of the earth, each axis of the geodetic coordinate system obeys the right-hand theorem. The specific composition of the body coordinate system O-XYZ includes: the origin O is located at the buoyancy center of the underwater robot, the horizontal axis OX points to the bow direction of the underwater robot, the transverse axis OY points to the side of the underwater robot, and the longitudinal axis OZ points to the bottom of the underwater robot. Each axis of the body coordinate system obeys the right-hand theorem.

[0067] In some embodiments, Figure 3 As shown, the state residual is obtained based on the observed state and the predicted state, which is constrained by the following mathematical expression:

[0068] η r =η s -η,

[0069] Among them, η represents the predicted state, η s represents the observation state, η r Represents the state residual.

[0070] Specifically, refer to Figure 3 It can be understood that, as can be seen from the figure, the state residual is obtained by taking the difference between the observed state and the predicted state, thereby indicating a parameter value that deviates from the actual state.

[0071] In some embodiments, Figure 3 and Figure 4 As shown, the compensation model is constrained by the following mathematical expression:

[0072]

[0073] in, represents the prediction residual, f represents the compensation model, η r (t-1) represents the state residual at the previous time, η(t-1) represents the predicted state at the previous time, i(t-1) represents the control input at the previous time, and t represents time.

[0074] Specifically, refer to Figure 3 and Figure 4 It can be understood that, as can be seen from the figure, by inputting the state residual, predicted state and control input at the last time into the compensation model, after compensation processing, the predicted residual of this time can be output.

[0075] In some embodiments, the compensation model is a neural network or a nonlinear fitting equation, and the optimization problem of the compensation model is constrained by the following mathematical expression:

[0076]

[0077] Where f represents the compensation model, argmin represents the variable value when the compensation model takes the minimum value, T represents the time period of the underwater robot's motion state, represents the prediction residual, η r (t) represents the state residual.

[0078] Specifically, it is understandable that by using a neural network or a nonlinear fitting equation in the compensation model, an optimization solution can be obtained, for example, to satisfy the minimum value of the mathematical expression in the present embodiment, thereby obtaining the corresponding prediction residual.

[0079] In some embodiments, Figure 3 As shown, the fusion state is obtained based on the predicted state and the predicted residual, which is constrained by the following mathematical expression:

[0080]

[0081] in, Indicates the fusion state. represents the prediction residual, and η(t) represents the prediction state.

[0082] Specifically, refer to Figure 3 It can be understood that, as can be seen from the figure, for the fusion between the prediction state and the prediction residual, the fusion state is obtained by summing up.

[0083] In addition, an embodiment of the present invention further provides an underwater robot state prediction system, including: an observation acquisition module, a first model processing module, a first calculation module, a second module processing module, and a second calculation module. The observation acquisition module is used to collect the observation state of the underwater robot based on the onboard sensor; the first model processing module is used to establish a physically driven dynamic model and input the control input of the underwater robot to obtain the predicted state of the underwater robot; the first calculation module is used to obtain the state residual according to the observed state and the predicted state; the second module processing module is used to establish a data-driven compensation model and input the predicted state, state residual and control input to obtain the predicted residual; the second calculation module is used to obtain the fusion state according to the predicted state and the predicted residual to realize the prediction of the underwater robot state.

[0084] Specifically, it can be understood that the underwater robot state prediction system of the embodiment of the present application is used to implement the underwater robot state prediction method. The underwater robot state prediction system of the embodiment of the present application corresponds to the aforementioned underwater robot state prediction method. Please refer to the aforementioned underwater robot state prediction method for the specific processing process, which will not be repeated here.

[0085] In this embodiment, firstly, a physically driven dynamic model is established based on the analysis of the underwater robot system structure; then, a data-driven compensation model is constructed through the dynamic model of the underwater robot and the observed state collected by the onboard sensor; finally, the state prediction of the underwater robot is completed by integrating the physically driven dynamic model and the data-driven compensation model. Therefore, the system of the embodiment of the present invention takes into account the influence of the dynamic environment on the motion state of the underwater robot, combines the physically driven dynamic model and the data-driven compensation model, solves the problem of mismatch between the physically driven model and the observed state of the underwater robot in the dynamic environment, and finally improves the state prediction accuracy and robustness of the underwater robot, which has a strong application prospect.

[0086] In addition, an embodiment of the present invention further provides an underwater robot state prediction device, including: at least one control processor and a memory for communicating with the at least one control processor.

[0087] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0088] The non-transient software program and instructions required to implement the underwater robot state prediction method of the above embodiment are stored in the memory. When executed by the processor, the underwater robot state prediction method of the above embodiment is executed, for example, the above described Figure 1 The method in .

[0089] The system embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. The computer-executable instructions are executed by one or more control processors, so that the one or more control processors can execute a method for predicting the state of an underwater robot in the above method embodiment, for example, executing the above described Figure 1 The method in .

[0091] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0092] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. A method for predicting the state of an underwater robot, characterized in that: The following steps are involved: Based on the onboard sensors, the observation status of the underwater robot is collected; Establishing a physically driven dynamic model and inputting control inputs of the underwater robot to obtain a predicted state of the underwater robot; Obtaining a state residual according to the observed state and the predicted state; Establishing a data-driven compensation model and inputting the predicted state, the state residual and the control input to obtain a predicted residual; Obtaining a fusion state according to the predicted state and the predicted residual to achieve state prediction of the underwater robot; The state residual obtained according to the observed state and the predicted state is constrained by the following mathematical expression: or r =the s -or, η represents the predicted state, η s represents the observed state, η r represents the state residual; The compensation model is constrained by the following mathematical expression: represents the prediction residual, f represents the compensation model, η r (t-1) represents the state residual at the previous time, η(t-1) represents the predicted state at the previous time, i(t-1) represents the control input at the previous time, and t represents time; Wherein, the compensation model is a neural network or a nonlinear fitting equation, and the optimization problem of the compensation model is constrained by the following mathematical expression: f represents the compensation model, argmin represents the variable value when the compensation model takes the minimum value, T represents the time period of the underwater robot's motion state, represents the prediction residual, η r (t) represents the state residual.

2. The underwater robot state prediction method according to claim 1, characterized in that: The kinetic model is constrained by the following mathematical expression: in, represents the derivative of η, η = [η1, η2] represents the state of the underwater robot in the earth coordinate system, that is, the predicted state, η1 = [x, y, z] T represents the position of the underwater robot, η2=[φ,θ,Ψ] T represents the posture of the underwater robot, represents the derivative of ν, ν=[ν1,ν2] represents the state of the underwater robot in the body coordinate system, ν1=[u,v,w] T represents the speed of the underwater robot, ν2=[p,q,r] T represents the angular velocity of the underwater robot, C(η) is the vector transformation matrix from the earth coordinate system to the body coordinate system, and I m represents the mass matrix of the underwater robot, I c represents the Coriolis and centrifugal matrices, τ(i) represents the external force and torque acting on the underwater robot, and i represents the control input of the underwater robot.

3. The underwater robot state prediction method according to claim 1, characterized in that: The fusion state is obtained according to the prediction state and the prediction residual, and is constrained by the following mathematical expression: in, represents the fusion state, represents the prediction residual, and η(t) represents the prediction state.

4. An underwater robot state prediction system, characterized in that: include: An observation and acquisition module is used to collect the observation status of the underwater robot based on the onboard sensors; A first model processing module is used to establish a physically driven dynamic model and input a control input of the underwater robot to obtain a predicted state of the underwater robot; A first calculation module, used for obtaining a state residual according to the observed state and the predicted state; A second module processing module is used to establish a data-driven compensation model and input the predicted state, the state residual and the control input to obtain a predicted residual; A second calculation module is used to obtain a fusion state according to the predicted state and the predicted residual to achieve state prediction of the underwater robot; The state residual obtained according to the observed state and the predicted state is constrained by the following mathematical expression: or r =the s -or, η represents the predicted state, η s represents the observed state, η r represents the state residual; The compensation model is constrained by the following mathematical expression: represents the prediction residual, f represents the compensation model, η r (t-1) represents the state residual at the previous time, η(t-1) represents the predicted state at the previous time, i(t-1) represents the control input at the previous time, and t represents time; Wherein, the compensation model is a neural network or a nonlinear fitting equation, and the optimization problem of the compensation model is constrained by the following mathematical expression: f represents the compensation model, argmin represents the variable value when the compensation model takes the minimum value, T represents the time period of the underwater robot's motion state, represents the prediction residual, η r (t) represents the state residual.

5. An underwater robot state prediction device, characterized in that: It includes at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the underwater robot state prediction method as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the underwater robot state prediction method according to any one of claims 1 to 3.

Images