A navigation method, system, electronic device and storage medium for an underwater robot

Abstract

This invention discloses an underwater robot navigation method, system, electronic device, and storage medium. The method includes: generating first driving data based on an autonomous underwater robot path navigation platform and sending it to the autonomous underwater robot; the autonomous underwater robot collecting real-time status data, real-time environmental data, and real-time terrain data; and selecting a first micro-navigation model or a second micro-navigation model based on depth data to process the collected data and generate first micro-driving data or second micro-driving data; adjusting the first driving data using the first micro-driving data or second micro-driving data to obtain second driving data; and using the second driving data to drive the autonomous underwater robot to perform displacement. This invention improves the data processing efficiency of autonomous underwater robots, effectively reduces the robot's computing resources, and improves the accuracy of path navigation during underwater operations, and can be widely applied in the field of robot navigation technology.

Description

Technical Field

[0001] This invention relates to the field of robot navigation technology, and in particular to an underwater robot navigation method, system, electronic device and storage medium. Background Technology

[0002] Autonomous underwater vehicles (AUVs) can accurately explore marine environments and seabed geology according to specific mission requirements. Their unique flexibility and high accuracy have attracted increasing attention from researchers. However, due to the special nature of their operating environment, path navigation during operation remains a challenge for AUVs. Currently, path navigation for AUVs mainly includes traditional methods such as graph-based, genetic, or evolutionary algorithms, as well as artificial intelligence methods based on machine learning and deep learning algorithms. These existing methods primarily generate collision-free navigation paths within a predetermined environment, considering only data from target detection points and obstacles along the path. They fail to consider real-time operating environment data and local terrain feature data, resulting in navigation decisions that are difficult to apply to real-time detection environments and have low accuracy. Furthermore, existing methods heavily rely on the computational resources of the autonomous robot, requiring real-time processing of global data, leading to low path navigation efficiency and high cost. Summary of the Invention

[0003] This invention aims to at least partially address the limitations of related technologies. To this end, this invention proposes an underwater robot navigation method, system, electronic device, and storage medium, capable of accurately navigating underwater robots.

[0004] On one hand, embodiments of the present invention provide an underwater robot navigation method, comprising the following steps:

[0005] Obtain the first driving data generated by the path navigation platform;

[0006] Acquire real-time data of the target underwater robot; real-time data includes real-time status data, real-time environmental data, and real-time terrain data;

[0007] The target micro-navigation model is matched with the depth data in the real-time data, and the target micro-navigation model is used to generate micro-drive data based on the real-time data.

[0008] The first drive data is adjusted based on the micro-drive data to obtain the second drive data;

[0009] The second driving data is sent to the drive unit of the target underwater robot so that the target underwater robot can move according to the second driving data;

[0010] The second driving data is used as the first driving data, and the process returns to the step of acquiring real-time data of the target underwater robot until the target underwater robot reaches the detection point.

[0011] Optionally, obtaining the first driving data generated by the path navigation platform includes the following steps:

[0012] Input the physical parameter data of the target underwater robot and the target data into the path navigation platform;

[0013] The physical parameter data includes weight and three-dimensional dimensions; the target data includes at least one of target location point data, terrain feature point data, and historical environmental data; the target location point data includes location data of detection point locations and obstacle point locations.

[0014] The path navigation platform calls a pre-trained path planning model to generate the first driving data based on physical parameter data and target data;

[0015] The path planning model is trained using an action and evaluation architecture based on an encoding network and a deep reinforcement learning network.

[0016] Optionally, real-time data includes real-time status data, real-time environmental data, and real-time terrain data; acquiring real-time data of the target underwater robot includes the following steps:

[0017] The real-time status data of the target underwater robot is acquired based on a preset first sensor; the real-time status data includes depth data and motion parameter data.

[0018] The target underwater robot acquires real-time environmental data based on a pre-set second sensor; the real-time environmental data includes ocean current velocity and direction.

[0019] The real-time terrain data of the area where the target underwater robot is located is acquired based on the preset terrain detection device; the real-time terrain data includes multibeam sonar underwater terrain elevation image data and side-scan sonar data.

[0020] Optionally, the real-time terrain data includes multibeam sonar underwater terrain elevation image data and side-scan sonar data; the target micro-navigation model is matched based on the depth data in the real-time data, and micro-drive data is generated based on the real-time data using the target micro-navigation model, including the following steps:

[0021] Acquire target terrain feature point data for the target underwater robot;

[0022] When the depth data is greater than or equal to a preset first threshold, the first micro-navigation model generates first micro-drive data based on real-time status data, real-time environmental data, multibeam sonar underwater terrain elevation image data, and target terrain feature point data; otherwise,

[0023] The second micro-navigation model is used to generate second micro-drive data based on real-time status data, real-time environmental data, side-scan sonar data, and target terrain feature point data.

[0024] Both the first and second micro-navigation models include an encoder, a fusion unit, and a decoder; the encoder includes multiple sub-encoders.

[0025] Optionally, the first micro-navigation model includes a first encoder, a first fusion unit, and a first decoder. The first encoder includes a first sub-encoder, a second sub-encoder, a third sub-encoder, and a fourth sub-encoder. Generating first micro-drive data using the first micro-navigation model based on real-time state data, real-time environmental data, multibeam sonar underwater terrain elevation image data, and target terrain feature point data includes the following steps:

[0026] The first sub-encoder is used to encode the real-time state data to obtain the first sub-vector; the second sub-encoder is used to encode the real-time environment data to obtain the second sub-vector; the third sub-encoder is used to encode the multibeam sonar underwater terrain elevation image data to obtain the third sub-vector; and the fourth sub-encoder is used to encode the target terrain feature point data to obtain the fourth sub-vector.

[0027] The first and second sub-encoders employ any one of the following: a domain dictionary model, a BERT model, and an LSTM network; the third sub-encoder employs a convolutional neural network based on an attention mechanism; and the fourth sub-encoder employs an LSTM neural network based on an attention mechanism.

[0028] The first sub-vector, the second sub-vector, the third sub-vector, and the fourth sub-vector are fused using the first fusion processor to obtain the first vector;

[0029] The first vector is decoded using the first decoder to obtain the first micro-drive data;

[0030] The first decoder can be either an LSTM network or a BI_LSTM network.

[0031] Optionally, the second micro-navigation model includes a second encoder, a second fusion unit, and a second decoder. The second encoder includes a fifth sub-encoder, a sixth sub-encoder, a seventh sub-encoder, and an eighth sub-encoder. Generating second micro-drive data using the second micro-navigation model based on real-time state data, real-time environmental data, side-scan sonar data, and target terrain feature point data includes the following steps:

[0032] The fifth sub-encoder is used to encode the real-time status data to obtain the fifth sub-vector; the sixth sub-encoder is used to encode the real-time environment data to obtain the sixth sub-vector; the seventh sub-encoder is used to encode the side-scan sonar data to obtain the seventh sub-vector; and the eighth sub-encoder is used to encode the target terrain feature point data to obtain the eighth sub-vector.

[0033] Among them, the fifth and sixth sub-encoders adopt any one of the domain dictionary model, BERT model and LSTM network; the seventh sub-encoder adopts stacked convolutional neural network; and the eighth sub-encoder adopts LSTM network.

[0034] The second fusion process is used to perform a second fusion process on the fifth, sixth, seventh, and eighth sub-vectors to obtain the second vector;

[0035] The second vector is processed by the second decoder to obtain the second micro-drive data;

[0036] The second decoder can be either an LSTM network or a BI_LSTM network.

[0037] Optionally, acquiring target terrain feature point data for the target underwater robot includes the following steps:

[0038] Based on the positioning data of the target underwater robot, multiple terrain feature point data within a preset distance of the target underwater robot are obtained and processed to obtain the target terrain feature point data;

[0039] The positioning data includes at least one of long baseline positioning data, short baseline positioning data, and ultra-short baseline positioning data.

[0040] On the other hand, embodiments of the present invention provide an underwater robot navigation system, comprising:

[0041] The first module is used to obtain the first driving data generated by the path navigation platform;

[0042] The second module is used to acquire real-time data of the target underwater robot; the real-time data includes real-time status data, real-time environmental data, and real-time terrain data.

[0043] The third module is used to match the target micro-navigation model with the depth data in the real-time data, and use the target micro-navigation model to generate micro-drive data based on the real-time data.

[0044] The fourth module is used to adjust the first drive data based on the micro-drive data to obtain the second drive data;

[0045] The fifth module is used to send the second driving data to the drive unit of the target underwater robot so that the target underwater robot can move according to the second driving data.

[0046] The sixth module is used to take the second driving data as the first driving data and return to execute the second module until the target underwater robot reaches the detection point.

[0047] On the other hand, embodiments of the present invention provide an electronic device, including: a processor and a memory; the memory is used to store a program; the processor executes the program to implement the above-described underwater robot navigation method.

[0048] On the other hand, embodiments of the present invention provide a computer storage medium storing a processor-executable program, which, when executed by a processor, is used to implement the above-described underwater robot navigation method.

[0049] This invention involves: acquiring first driving data generated by a path navigation platform; acquiring real-time data of a target underwater robot; the real-time data including real-time status data, real-time environmental data, and real-time terrain data; matching a target micro-navigation model with depth data from the real-time data, and generating micro-driving data based on the real-time data using the target micro-navigation model; adjusting the first driving data according to the micro-driving data to obtain second driving data; sending the second driving data to the drive unit of the target underwater robot so that the target underwater robot can move according to the second driving data; using the second driving data as the first driving data, and returning to execute the step of acquiring real-time data of the target underwater robot until the target underwater robot reaches the detection point position. This invention obtains the first driving data from a path navigation platform. The autonomous underwater robot only needs to process real-time environmental, terrain, and status data at local moments based on a micro-navigation model, without needing to process global environmental, terrain, or status data, or historical environmental and terrain data. This improves data processing efficiency and effectively reduces the robot's computational resources. Simultaneously, this invention obtains micro-driving data based on real-time status, environmental, and terrain data, and uses this micro-driving data to adjust the first driving data, improving the accuracy of the autonomous underwater robot's path navigation. Furthermore, this invention uses different micro-navigation models and different data feature spaces based on different operating depths, making the autonomous underwater robot's path navigation more closely match the operating environment, further improving path navigation accuracy. This invention enables efficient and accurate underwater robot navigation. Attached Figure Description

[0050] The accompanying drawings are provided to further understand the technical solutions of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the technical solutions of the present invention, and do not constitute a limitation on the technical solutions of the present invention.

[0051] Figure 1 This is a schematic diagram of an implementation environment for the underwater robot navigation method provided in this embodiment of the invention;

[0052] Figure 2 This is a flowchart illustrating an underwater robot navigation method provided in an embodiment of the present invention;

[0053] Figure 3 A schematic diagram illustrating the unfolding process of step S100 provided in an embodiment of the present invention;

[0054] Figure 4 A schematic diagram illustrating the unfolding process of step S200 provided in an embodiment of the present invention;

[0055] Figure 5 A schematic diagram illustrating the unfolding process of step S300 provided in an embodiment of the present invention;

[0056] Figure 6 A schematic diagram illustrating the overall process of the underwater robot navigation method provided in an embodiment of the present invention;

[0057] Figure 7 A schematic diagram illustrating the overall process principle of the underwater robot navigation method provided in this embodiment of the invention;

[0058] Figure 8 A schematic diagram illustrating the architecture principle of underwater robot navigation provided in an embodiment of the present invention;

[0059] Figure 9 This is a schematic diagram of the structure of an underwater robot navigation system provided in an embodiment of the present invention;

[0060] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0062] It should be noted that although functional modules are divided in the system diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system or the order in the flowchart. The terms "first / S100," "second / S200," etc., in the specification, claims, and the aforementioned figures are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0063] In this invention, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this invention can be combined with other embodiments.

[0064] It is understood that the underwater robot navigation method provided in this embodiment of the invention can be applied to any computer device with data processing and computing capabilities, and this computer device can be various types of terminals or servers. When the computer device in the embodiment is a server, the server is an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Optionally, the terminal can be a smartphone, tablet computer, laptop computer, or desktop computer, but it is not limited to these.

[0065] To facilitate understanding of the technical solution of this invention, the technical features and proper nouns that may appear in the embodiments of this invention will first be explained:

[0066] like Figure 1 The diagram shown is a schematic representation of an implementation environment provided by an embodiment of the present invention. (Refer to...) Figure 1 The implementation environment includes at least one terminal 102 and a server 101. The terminal 102 and the server 101 can be connected via a network, either wirelessly or via a wired connection, to complete data transmission and exchange.

[0067] Server 101 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0068] Additionally, server 101 can also be a node server in a blockchain network. Blockchain is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms.

[0069] Terminal 102 can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. Terminal 102 and server 101 can be directly or indirectly connected via wired or wireless communication, and this embodiment of the invention does not impose any limitations.

[0070] Exemplary based on Figure 1 The implementation environment shown in this embodiment of the invention provides an underwater robot navigation method. The following description uses the application of this underwater robot navigation method in server 101 as an example. It can be understood that the underwater robot navigation method can also be applied in terminal 102.

[0071] Reference Figure 2 , Figure 2 The flowchart illustrates an underwater robot navigation method applied to a server, as provided in an embodiment of the present invention. The executing entity of this underwater robot navigation method can be any of the aforementioned computer devices (including a server or a terminal). (Refer to...) Figure 2 The method includes the following steps:

[0072] S100: Obtain the first driving data generated by the path navigation platform;

[0073] It should be noted that in some embodiments, such as Figure 3 As shown, step S100 may include the following steps: S101, inputting the physical parameter data and target data of the target underwater robot into the path navigation platform; wherein, the physical parameter data includes weight and three-dimensional dimensions; the target data includes at least one of target location point data, terrain feature point data and historical environmental data; the target location point data includes location data of detection point location and obstacle point location; S102, using the path navigation platform to call a pre-trained path planning model to generate first driving data based on the physical parameter data and target data; wherein, the path planning model is trained using an action and evaluation architecture based on an encoding network and a deep reinforcement learning network.

[0074] For example, in some specific embodiments, the present invention employs an autonomous underwater robot path navigation platform and an autonomous underwater robot body to achieve precise path navigation for the underwater robot. The path navigation platform is equipped with an intelligent path planning model, which generates first driving data. This first driving data is used to drive the autonomous underwater robot to complete underwater operations according to the planned path. Before launching the operation, the underwater robot connects to the path navigation platform via wired or wireless communication mechanisms and obtains the data required for the operation from the platform. This data includes at least the first driving data.

[0075] In some embodiments, the autonomous underwater robot path navigation platform generates first driving data using an intelligent path planning model, specifically by acquiring the underwater robot's physical parameter data, which includes at least the underwater robot's weight and three-dimensional dimensions, and can be represented as PARA. G = {WEIGHT, SIZE}, where WEIGHT represents the weight of the robot and SIZE represents the size of the robot, which can be characterized by length, width, and height, or by other methods.

[0076] Acquire target point location data, which includes probe point location data and / or obstacle point location data. Probe points are the locations the underwater robot needs to reach, and obstacle points are the points the underwater robot needs to avoid. Target point location data consists of a set of triples, each containing the target point's type label, latitude (lat), and longitude (lon). When the target point is a probe point, its type label is 1, and the corresponding triple can be represented as follows: When the target point is an obstacle point, its type label is 0, and the corresponding triple can be represented as: The target point location data can be represented as:

[0077]

[0078] Finally, the first driving data is generated based on the physical parameter data, target point location data, and the pre-trained path planning model.

[0079] In some specific application scenarios, a path planning model is pre-trained. This model employs an encoding network and a deep reinforcement learning network (DQN). The DQN uses an action-evaluation architecture: a positive Q-value is used when the robot performs an action that avoids an obstacle or reaches a predicted point; a negative Q-value is used when the robot performs an action that moves closer to an obstacle or further away from the predicted point. Physical parameter data and target point position data are input into the encoding network to obtain an encoded vector, which is then input into the DQN to generate the first driving data. This first driving data is used to drive the autonomous underwater robot to move along the planned path.

[0080] In other embodiments, terrain feature point data and historical environmental data can also be acquired. The terrain feature point data can be obtained by identifying historically acquired multibeam sonar underwater terrain elevation image data and side-scan sonar data. The historical environmental data can be environmental data from the same time last year or average environmental data from the same time over several years. The historical environmental data includes at least ocean current velocity and ocean current direction. Furthermore, the first driving data can be generated based on physical parameter data, target point location data, terrain feature point data, historical environmental data, and path planning model.

[0081] S200: Acquire real-time data of the target underwater robot;

[0082] Real-time data includes real-time status data, real-time environmental data, and real-time terrain data.

[0083] It should be noted that in some embodiments, such as Figure 4 As shown, step S200 may include the following steps: S201, acquiring real-time status data of the target underwater robot based on a preset first sensor; the real-time status data includes depth data and motion parameter data; S202, acquiring real-time environmental data of the target underwater robot based on a preset second sensor; the real-time environmental data includes ocean current velocity and ocean current direction; S203, acquiring real-time terrain data of the area where the target underwater robot is located based on a preset terrain detection device; the real-time terrain data includes multibeam sonar underwater terrain elevation image data and side-scan sonar data.

[0084] For example, in some specific implementations, real-time status data of the autonomous underwater vehicle (AUV) during operation is acquired. This real-time status data includes at least the robot's depth data and motion parameter data, denoted as AUVstate = {H, PARA}. m}, where H represents real-time depth data, which characterizes the distance between the robot and the seabed. Represents motion parameter data, x m ,Φ m These represent the robot's position along the x-axis and heading angle in the Earth coordinate system, respectively; y m ,θ m These represent the robot's position and pitch angle along the y-axis in the Earth coordinate system, respectively; z m , These represent the robot's position along the z-axis and yaw angle in the Earth coordinate system, respectively; u m ,p m Let v represent the robot's velocity along the X-axis of the body coordinate system and the angle between the X-axis of the body coordinate system and the horizontal plane, respectively. m ,q m These represent the robot's velocity along the Y-axis of the body coordinate system and the angle between the Y-axis of the body coordinate system and the horizontal plane, respectively; wm ,r m These represent the robot's velocity along the Z-axis of the body coordinate system and the angle between the Z-axis and north, respectively. Real-time status data of the autonomous underwater robot can be obtained through sensors and / or computational models.

[0085] Acquire real-time environmental data for autonomous underwater vehicles (AUVs). During operation, AUVs collect real-time environmental data via sensors. This real-time environmental data includes at least ocean current velocity and direction. Where u e ,Φ e Let v represent the robot's velocity along the X-axis and the angle between v and the X-axis in the Earth coordinate system. e ,θ e Let w represent the robot's velocity along the Y-axis and the angle between the current and the Y-axis in the Earth coordinate system. e , These represent the robot's velocity along the Z-axis and the angle between the current and the Z-axis in the Earth coordinate system, respectively.

[0086] This invention acquires real-time terrain data collected by an autonomous underwater robot. The autonomous underwater robot is equipped with various types of terrain detection devices, including but not limited to multibeam bathymetry and sonar. Based on the characteristics of these terrain detection devices, this invention employs different data spaces at different depths to achieve micro-navigation for the autonomous robot. Specifically: when H ≥ ε 0 At that time, multibeam bathymetry topographic data is used, that is, multibeam sonar underwater topographic elevation image data (IMBS) is used, when H < ε 0 At that time, the terrain data used is from side-scan sonar, i.e., ISSS data, where H represents the distance between the autonomous underwater vehicle and the seabed, which can be obtained through multibeam bathymetry, ε 0 These are preset values ​​that can be set based on historical experience. By obtaining depth data between the autonomous underwater robot and the seabed, and using different seabed topography data for micro-navigation based on different depth data, the accuracy of robot navigation is effectively improved.

[0087] S300: Match the target micro-navigation model with the depth data in the real-time data, and use the target micro-navigation model to generate micro-drive data based on the real-time data;

[0088] It should be noted that real-time terrain data includes multibeam sonar underwater terrain elevation image data and side-scan sonar data; in some embodiments, such as Figure 5As shown, step S300 may include the following steps: S301, acquiring target terrain feature point data of the target underwater robot; S302, when the depth data is greater than or equal to a preset first threshold, generating first micro-drive data using a first micro-navigation model based on real-time state data, real-time environment data, multi-beam sonar underwater terrain elevation image data, and target terrain feature point data; otherwise, executing subsequent steps; S303, generating second micro-drive data using a second micro-navigation model based on real-time state data, real-time environment data, side-scan sonar data, and target terrain feature point data; wherein, both the first micro-navigation model and the second micro-navigation model include an encoder, a fusion unit, and a decoder; the encoder includes multiple sub-encoders.

[0089] For example, in some specific embodiments, since different data spaces are used for micro-navigation based on the depth data between the underwater robot and the seabed, the present invention designs and trains two micro-navigation models. When the depth data H ≥ ε 0 When using the first micro-navigation model, if the depth data H < ε 0 At that time, a second micro-navigation model is adopted. The first micro-navigation model uses real-time status data, real-time environmental data, multi-beam sonar underwater terrain elevation image data, and target terrain feature point data as input data, and outputs the first micro-drive data. The second micro-navigation model uses real-time status data, real-time environmental data, side-scan sonar data, and target terrain feature point data as input data, and outputs the second micro-drive data.

[0090] In some embodiments, step S301 may include the following steps: obtaining multiple terrain feature point data within a preset distance of the target underwater robot based on the positioning data of the target underwater robot and organizing them to obtain target terrain feature point data; wherein, the positioning data includes at least one of long baseline positioning data, short baseline positioning data and ultra-short baseline positioning data.

[0091] For example, in some specific embodiments, the target terrain feature point data are seabed terrain feature points within the current operating area of ​​the autonomous underwater robot, which can be obtained through the underwater robot's long baseline positioning data and / or short baseline positioning data, and / or ultra-short baseline positioning data, as well as terrain feature point data, specifically:

[0092] Acquire terrain feature point data. Before performing the operation, the autonomous underwater robot acquires terrain feature point data from the path navigation platform, in addition to the initial drive data. The acquired terrain feature point data is represented as follows:

[0093] Acquire positioning data for autonomous underwater vehicles (AUVs). Positioning data for AUVs includes long baseline positioning data and / or short baseline positioning data.

[0094] Target terrain feature point data is obtained based on long baseline positioning data, and / or short baseline positioning data, and / or ultra-short baseline positioning data, and terrain feature point data. Multiple terrain feature points within a preset distance are selected from the terrain feature point data using the long baseline and / or short baseline positioning data as the target terrain feature point data; alternatively, a preset number of terrain feature points can be selected from the terrain feature point data using the long baseline and / or short baseline positioning data as the target terrain feature point data. Target terrain feature point data. The location is within a preset distance of the long baseline positioning data and / or short baseline positioning data, or The location is close to long baseline positioning data and / or short baseline positioning data, and The number is within a preset limit. For example, target terrain feature point data is obtained based on terrain feature point data LOC2, long baseline positioning data, and / or short baseline positioning data.

[0095] In some embodiments, the first micro-navigation model includes a first encoder, a first fusion unit, and a first decoder. The first encoder includes a first sub-encoder, a second sub-encoder, a third sub-encoder, and a fourth sub-encoder. Generating first micro-drive data using the first micro-navigation model based on real-time state data, real-time environment data, multibeam sonar underwater terrain elevation image data, and target terrain feature point data may include the following steps: performing a first encoding on the real-time state data using the first sub-encoder to obtain a first sub-vector; performing a second encoding on the real-time environment data using the second sub-encoder to obtain a second sub-vector; and performing a third encoding on the multibeam sonar underwater terrain elevation image data using the third sub-encoder to obtain a third sub-vector. The target terrain feature point data is encoded using a fourth sub-encoder to obtain a fourth sub-vector. The first and second sub-encoders employ any one of a domain dictionary model, a BERT model, and an LSTM network. The third sub-encoder employs a convolutional neural network based on an attention mechanism. The fourth sub-encoder employs an LSTM neural network based on an attention mechanism. A first fusion processor performs a first fusion process on the first, second, third, and fourth sub-vectors to obtain a first vector. A first decoder performs a first decoding process on the first vector to obtain first micro-driving data. The first decoder employs any one of an LSTM network and a BI_LSTM network.

[0096] For example, in some specific embodiments, the first micro-navigation model includes a first encoder, a first fusion unit, and a first decoder. The first encoder includes a first sub-encoder, a second sub-encoder, a third sub-encoder, and a fourth sub-encoder. The first sub-encoder is used to encode the real-time state data AUVstate to obtain a first sub-vector. Specifically:

[0097] The second sub-encoder is used to encode real-time environmental data to obtain the second sub-vector. Specifically: First sub-encoder Second sub-encoder Domain dictionary models, BERT models, and LSTM networks can be used. This invention uses the BERT model as the first sub-encoder and the second sub-encoder.

[0098] The third sub-encoder is used to encode the multibeam sonar underwater topographic elevation image data to obtain the third sub-vector. Because multibeam sonar bathymetry is more accurate for larger terrain features but less accurate for smaller ones, the third sub-encoder employs a convolutional neural network with an attention mechanism. It assigns greater weight to large target features such as seabed protrusions and depressions, and less weight to small target features such as seabed cracks. Specifically:

[0099] The fourth sub-encoder is used to encode the target terrain feature point data to obtain the fourth sub-vector. The fourth sub-encoder employs an attention-based LSTM neural network, assigning greater weights to large-sized topographic features such as seabed protrusions and depressions, and less weights to small-sized topographic features such as seabed cracks. Specifically:

[0100] The first fusion unit is used to fuse the first sub-vector, the second sub-vector, the third sub-vector, and the fourth sub-vector to obtain the first vector, specifically as follows:

[0101] The first decoder is used to decode the first vector to obtain the first micro-drive data, specifically v = DECODER1(v1). The first decoder can use an LSTM network (i.e., a long short-term memory network), a BI_LSTM network (i.e., a bidirectional long short-term memory network), etc. This invention uses a BI_LSTM network, specifically v = DECODER2(v1) = BI_LSTM(v1).

[0102] In this invention, the distance H between the autonomous underwater robot and the seabed is greater than or equal to a preset value ε.0 At the same time, multibeam sonar underwater terrain elevation image data is used as real-time terrain data, and the first micro-navigation model is used to obtain the first micro-drive data. Meanwhile, considering the characteristics of multibeam sonar underwater terrain elevation image data, the first micro-navigation model uses a third encoder and a fourth encoder based on an attention mechanism. High weights are assigned to large-sized terrain feature points such as seabed protrusions and seabed depressions, while low weights are assigned to small-sized terrain feature points such as seabed cracks, which further improves the accuracy of micro-navigation.

[0103] In some embodiments, the second micro-navigation model includes a second encoder, a second fusion unit, and a second decoder. The second encoder includes a fifth sub-encoder, a sixth sub-encoder, a seventh sub-encoder, and an eighth sub-encoder. Generating second micro-drive data using the second micro-navigation model based on real-time state data, real-time environmental data, side-scan sonar data, and target terrain feature point data may include the following steps: using the fifth sub-encoder to perform a fifth encoding on the real-time state data to obtain a fifth sub-vector; using the sixth sub-encoder to perform a sixth encoding on the real-time environmental data to obtain a sixth sub-vector; using the seventh sub-encoder to perform a seventh encoding on the side-scan sonar data to obtain a seventh sub-vector; and using the eighth sub-encoder to perform an eighth encoding on the target terrain feature point data to obtain an eighth sub-vector. The fifth and sixth sub-encoders employ any one of a domain dictionary model, a BERT model, and an LSTM network; the seventh sub-encoder employs a stacked convolutional neural network; and the eighth sub-encoder employs an LSTM network. The second fusion unit performs a second fusion process on the fifth, sixth, seventh, and eighth sub-vectors to obtain a second vector. The second decoder performs a second decoding process on the second vector to obtain the second micro-drive data. The second decoder employs any one of an LSTM network and a BI_LSTM network.

[0104] For example, in some specific embodiments, the second micro-navigation model includes a second encoder, a second fusion unit, and a second decoder. The second encoder includes a fifth sub-encoder, a sixth sub-encoder, a seventh sub-encoder, and an eighth sub-encoder. The fifth sub-encoder is used to encode the real-time state data AUVstate to obtain a fifth sub-vector. Specifically: The second sub-encoder is used to encode real-time environmental data to obtain the sixth sub-vector. Specifically: Fifth sub-encoder and the sixth sub-encoder Similarly, the domain dictionary model, BERT model and LSTM network can be used. This invention uses the BERT model as the fifth sub-encoder and the sixth sub-encoder.

[0105] The seventh sub-encoder is used to encode the side-scan sonar data to obtain the seventh sub-vector. Because side-scan sonar probes close to the seabed and has good detection accuracy even for small-sized seabed topographic features, the seventh sub-encoder does not need to distinguish between large and small topographic features. Therefore, the seventh sub-encoder uses a stacked convolutional neural network, i.e.

[0106] The eighth sub-encoder is used to encode the target terrain feature point data to obtain the eighth sub-vector. The eighth sub-encoder also does not require setting different weights for small and large seabed topographic feature points, and therefore does not require an attention mechanism. This invention uses an LSTM network as the eighth sub-encoder, i.e.:

[0107] The second fusion unit is used to fuse the fifth, sixth, seventh, and eighth sub-vectors to obtain the second vector, specifically as follows:

[0108] The second decoder is used to decode the second vector to obtain the second micro-drive data, specifically v = DECODER2(v2). The second decoder can be an LSTM network, a BI_LSTM network, etc. This invention uses a BI_LSTM network, specifically v = DECODER2(v2) = BI_LSTM(v2).

[0109] This invention relates to an autonomous underwater robot whose distance H from the seabed is less than a preset value ε. 0 At the same time, side-scan sonar data is used as real-time terrain data, and the second micro-navigation model is used to obtain the second micro-drive data. Meanwhile, considering the characteristics of side-scan sonar data, the second micro-navigation model uses a seventh sub-encoder and an eighth sub-encoder that are different from the first micro-navigation model. It no longer distinguishes between large-size and small-size seabed feature points, which further improves the accuracy of micro-navigation.

[0110] S400: Adjust the first drive data according to the micro drive data to obtain the second drive data;

[0111] For example, in some specific embodiments, the first driving data is adjusted based on the micro-drive data to obtain the second driving data. Specifically, the second driving data is obtained when the distance H between the autonomous underwater robot and the seabed is greater than or equal to a preset value ε. 0 The first micro-drive data is used to adjust the first drive data to obtain the second drive data, when the distance H between the autonomous underwater robot and the seabed is less than the preset value ε. 0At this time, the first driving data is adjusted using the second micro-drive data to obtain the second driving data. Specifically, the adjustment method involves superimposing either the first or second micro-drive data onto the first driving data within the current detection range to obtain the second driving data.

[0112] S500: Send the second drive data to the drive unit of the target underwater robot so that the target underwater robot can move according to the second drive data;

[0113] For example, in some embodiments, second driving data is sent to the driving unit to cause the autonomous underwater robot to move according to the second driving data. The second driving data is sent to the driving unit of the autonomous underwater robot, and the autonomous underwater robot moves according to the second driving data.

[0114] S600: Use the second driving data as the first driving data, and return to the step of obtaining real-time data of the target underwater robot until the target underwater robot reaches the detection point.

[0115] For example, in some specific embodiments, steps S200 to S500 in this invention are performed iteratively according to the operating area. That is, in the next operating area, the autonomous underwater robot still executes steps S200 to S500 until the task is completed. Since the autonomous underwater robot adjusts the first driving data with micro-drive data in each operating area and uses the adjusted second driving data to drive the operation of the autonomous underwater robot, the accuracy of the robot's path navigation during underwater operation is improved; at the same time, the autonomous underwater robot only needs to process real-time environmental data, real-time terrain data, and real-time status data within the operating area, and does not need to process global environmental data, terrain data, and status data, nor does it process historical environmental data and terrain data, thus improving operational efficiency and saving the underwater robot's computing resources.

[0116] To explain in detail the principle of the technical solution of the present invention, the overall process of the present invention will be described below with reference to some specific embodiments. It is easy to understand that the following is an explanation of the technical principle of the present invention and should not be regarded as a limitation of the present invention.

[0117] First, it should be noted that, given the shortcomings of existing technologies, there is an urgent need for a lightweight, efficient, and highly accurate navigation method for autonomous underwater vehicles (AUVs) to improve the accuracy of path navigation during operation and reduce the computational resource requirements for path navigation. Therefore, as... Figure 6 and Figure 7 As shown, this invention provides an underwater robot navigation method that can be applied to autonomous underwater robots. The method of this invention can be implemented as follows:

[0118] S1: Obtain the first driving data.

[0119] See Figure 8 This invention employs an autonomous underwater robot path navigation platform and an autonomous underwater robot body to achieve precise path navigation for the underwater robot. The path navigation platform is equipped with an intelligent path planning model, which generates first driving data. This first driving data is used to drive the autonomous underwater robot to complete underwater operations according to the planned path. Before launching the operation, the underwater robot connects to the path navigation platform via wired or wireless communication mechanisms and obtains the necessary data from the platform, including at least the first driving data.

[0120] In some embodiments, the autonomous underwater robot path navigation platform generates first driving data using an intelligent path planning model, specifically by acquiring the underwater robot's physical parameter data, which includes at least the underwater robot's weight and three-dimensional dimensions, and can be represented as PARA. G = {WEIGHT, SIZE}, where WEIGHT represents the weight of the robot and SIZE represents the size of the robot, which can be characterized by length, width, and height, or by other methods.

[0121] Acquire target point location data, which includes probe point location data and / or obstacle point location data. Probe points are the locations the underwater robot needs to reach, and obstacle points are the points the underwater robot needs to avoid. Target point location data consists of a set of triples, each containing the target point's type label, latitude (lat), and longitude (lon). When the target point is a probe point, its type label is 1, and the corresponding triple can be represented as follows:

[0122] When the target point is an obstacle point, its type label is 0, and the corresponding triple can be represented as: The target point location data can be represented as:

[0123]

[0124] The first driving data is generated based on physical parameter data, target point location data, and a pre-trained path planning model.

[0125] A path planning model is pre-trained, employing an encoding network and a deep reinforcement learning network (DQN). The DQN uses an action-evaluation architecture: a positive Q-value is used when the robot performs an action that avoids an obstacle or reaches a predicted point; a negative Q-value is used when the robot moves closer to an obstacle or further away from the predicted point. Encoded vectors are obtained by inputting physical parameter data and target point position data into the encoding network, and these vectors are then input into the DQN to generate the first driving data. This first driving data is used to propel the autonomous underwater robot along the planned path.

[0126] In other embodiments, the autonomous underwater robot path navigation platform generates first driving data using an intelligent path planning model. This can also include: acquiring physical parameter data of the underwater robot, which at least includes the underwater robot's weight and three-dimensional dimensions, and can be represented as PARA. G = {WEIGHT, SIZE}, where WEIGHT represents the weight of the robot and SIZE represents the size of the robot, which can be characterized by length, width, and height, or by other methods.

[0127] Acquire target point location data, which includes probe point location data and / or obstacle point location data. Probe points are the locations the underwater robot needs to reach, and obstacle points are the points the underwater robot needs to avoid. Target point location data consists of a set of triples, each containing the target point's type label, latitude (lat), and longitude (lon). When the target point is a probe point, its type label is 1, and the corresponding triple can be represented as follows: When the target point is an obstacle point, its type label is 0, and the corresponding triple can be represented as: The target point location data can be represented as:

[0128]

[0129] Acquire topographic feature point data. Topographic feature points are location points in the operational area that possess unique topographic features. These can be selected from seabed protrusions, depressions, fissures, or points with significant seabed topographic undulations. Topographic feature point data can be obtained through identification using historically acquired multibeam sonar underwater topographic elevation images and side-scan sonar data. Identification can be performed based on a constructed recognition network. Topographic feature point data also consists of sets of triples. Each triple contains the type label of the topographic feature point, its latitude (lat), and its longitude (lon). When the topographic feature point is a seabed protrusion, its type label is 2, and the corresponding triple can be represented as follows: When the topographic feature point is a submarine depression, its type is 3, and the corresponding triplet can be represented as: And so on. The final terrain feature point data is represented as follows:

[0130]

[0131] Acquire historical environmental data. Historical environmental data for the operational area can be obtained, using data from the same time last year or average environmental data from the same time over several years. Historical environmental data must include at least ocean current velocity and direction. It means that among them These represent the velocity of the ocean current along the X-axis and the angle between the current and the X-axis in the Earth coordinate system, respectively. These represent the velocity of the ocean current along the Y-axis and the angle between the current and the Y-axis in the Earth coordinate system, respectively. These represent the velocity of the ocean current along the Z-axis and the angle between the current and the Z-axis in the Earth coordinate system, respectively.

[0132] The first driving data is generated based on physical parameter data, target point location data, terrain feature point data, historical environmental data, and path planning model.

[0133] A path planning model is pre-trained, employing both an encoding network and a deep reinforcement learning network (DQN). The DQN uses an action-evaluation mechanism: a positive Q-value is used when the robot's actions lead it around obstacles, certain types of terrain features, or to the predicted point or certain types of terrain features; a negative Q-value is used when the robot's actions move it closer to or further away from the predicted point or certain types of terrain features. Encoded vectors are obtained by inputting physical parameter data, target point location data, terrain feature data, and historical environment data into the encoding network. These encoded vectors are then input into the DQN to generate the first driving data. This first driving data is used to propel the autonomous underwater robot along the planned path.

[0134] After generating the first driving data based on the autonomous underwater robot path navigation platform, the autonomous underwater robot can obtain the first driving data through the autonomous underwater robot path navigation platform or client.

[0135] S2: Acquire real-time status data, real-time environmental data, and real-time terrain data of the autonomous underwater robot.

[0136] To acquire real-time status data of an autonomous underwater vehicle (AUV) during operation, the real-time status data should include at least the robot's depth data and motion parameter data, using AUVstate = {H, PARA}. M}, where H represents real-time depth data, which characterizes the distance between the robot and the seabed. Represents motion parameter data, x m ,Φ m These represent the robot's position along the x-axis and heading angle in the Earth coordinate system, respectively; y m ,θ m These represent the robot's position and pitch angle along the y-axis in the Earth coordinate system, respectively; z m , These represent the robot's position along the z-axis and yaw angle in the Earth coordinate system, respectively; u m ,p m Let v represent the robot's velocity along the X-axis of the body coordinate system and the angle between the X-axis of the body coordinate system and the horizontal plane, respectively. m ,q m These represent the robot's velocity along the Y-axis of the body coordinate system and the angle between the Y-axis of the body coordinate system and the horizontal plane, respectively; w m ,r m These represent the robot's velocity along the Z-axis of the body coordinate system and the angle between the Z-axis and north, respectively. Real-time status data of the autonomous underwater robot can be obtained through sensors and / or computational models.

[0137] Acquire real-time environmental data for autonomous underwater vehicles (AUVs). During operation, AUVs collect real-time environmental data via sensors. This real-time environmental data includes at least ocean current velocity and direction. Where u e ,Φ e Let v represent the robot's velocity along the X-axis and the angle between v and the X-axis in the Earth coordinate system. e ,θ e Let w represent the robot's velocity along the Y-axis and the angle between the current and the Y-axis in the Earth coordinate system. e , These represent the robot's velocity along the Z-axis and the angle between the current and the Z-axis in the Earth coordinate system, respectively.

[0138] This invention acquires real-time terrain data collected by an autonomous underwater robot. The autonomous underwater robot is equipped with various types of terrain detection devices, including but not limited to multibeam bathymetry and sonar. Based on the characteristics of these terrain detection devices, this invention employs different data spaces at different depths to achieve micro-navigation for the autonomous robot. Specifically: when H ≥ ε 0 At that time, multibeam bathymetry topographic data is used, that is, multibeam sonar underwater topographic elevation image data (IMBS) is used, when H < ε 0At that time, the terrain data used is from side-scan sonar, i.e., ISSS data, where H represents the distance between the autonomous underwater vehicle and the seabed, which can be obtained through multibeam bathymetry, ε 0 These are preset values ​​that can be set based on historical experience. By obtaining depth data between the autonomous underwater robot and the seabed, and using different seabed topography data for micro-navigation based on different depth data, the accuracy of robot navigation is effectively improved.

[0139] S3: Generate micro-drive data based on real-time status data, real-time environment data, real-time terrain data, and a pre-trained micro-navigation model.

[0140] Because different data spaces are used for micro-navigation based on the depth data between the underwater robot and the seabed, this invention designs and trains two micro-navigation models. When the depth data H ≥ ε 0 When using the first micro-navigation model, if the depth data H < ε 0 At that time, the second micro-navigation model was adopted.

[0141] The first micro-navigation model uses real-time state data, real-time environmental data, multi-beam sonar underwater terrain elevation image data, and target terrain feature point data as input data, and outputs the first micro-drive data. The target terrain feature point data refers to the seabed terrain feature points within the current operating area of ​​the autonomous underwater robot. This data can be obtained through the underwater robot's long baseline positioning data and / or short baseline positioning data, and / or ultra-short baseline positioning data, along with the terrain feature point data. Specifically:

[0142] Acquire terrain feature point data. Before performing the operation, the autonomous underwater robot acquires terrain feature point data from the path navigation platform, in addition to the initial drive data. The acquired terrain feature point data is represented as follows:

[0143] Acquire positioning data for autonomous underwater vehicles (AUVs). Positioning data for AUVs includes long baseline positioning data and / or short baseline positioning data.

[0144] Target terrain feature point data is obtained based on long baseline positioning data, and / or short baseline positioning data, and / or ultra-short baseline positioning data, and terrain feature point data. Multiple terrain feature points within a preset distance are selected from the terrain feature point data using the long baseline and / or short baseline positioning data as the target terrain feature point data; alternatively, a preset number of terrain feature points can be selected from the terrain feature point data using the long baseline and / or short baseline positioning data as the target terrain feature point data. Target terrain feature point data. The location is within a preset distance of the long baseline positioning data and / or short baseline positioning data, or The location is close to long baseline positioning data and / or short baseline positioning data, and The number is within a preset limit. For example, target terrain feature point data is obtained based on terrain feature point data LOC2, long baseline positioning data, and / or short baseline positioning data.

[0145] The first micro-navigation model includes a first encoder, a first fusion unit, and a first decoder. The first encoder includes a first sub-encoder, a second sub-encoder, a third sub-encoder, and a fourth sub-encoder. The first sub-encoder is used to encode the real-time state data (AUVstate) to obtain a first sub-vector. Specifically: The second sub-encoder is used to encode real-time environmental data to obtain the second sub-vector. Specifically: First sub-encoder Second sub-encoder Domain dictionary models, BERT models, and LSTM networks can be used. This invention uses the BERT model as the first sub-encoder and the second sub-encoder.

[0146] The third sub-encoder is used to encode the multibeam sonar underwater topographic elevation image data to obtain the third sub-vector. Because multibeam sonar bathymetry is more accurate for larger terrain features but less accurate for smaller ones, the third sub-encoder employs a convolutional neural network with an attention mechanism. It assigns greater weight to large target features such as seabed protrusions and depressions, and less weight to small target features such as seabed cracks. Specifically: ATTENTION_CNN(IMBS).

[0147] The fourth sub-encoder is used to encode the target terrain feature point data to obtain the fourth sub-vector. The fourth sub-encoder employs an attention-based LSTM neural network, assigning greater weights to large-sized topographic features such as seabed protrusions and depressions, and less weights to small-sized topographic features such as seabed cracks. Specifically:

[0148] The first fusion unit is used to fuse the first sub-vector, the second sub-vector, the third sub-vector, and the fourth sub-vector to obtain the first vector, specifically as follows:

[0149] The first decoder is used to decode the first vector to obtain the first micro-drive data, specifically v = DECODER1(v1). The first decoder can use an LSTM network (i.e., a long short-term memory network), a BI_LSTM network (i.e., a bidirectional long short-term memory network), etc. This invention uses a BI_LSTM network, specifically v = DECODER1(v1) = BI_LSTM(v1).

[0150] In this invention, the distance H between the autonomous underwater robot and the seabed is greater than or equal to a preset value ε. 0 At the same time, multibeam sonar underwater terrain elevation image data is used as real-time terrain data, and the first micro-navigation model is used to obtain the first micro-drive data. Meanwhile, considering the characteristics of multibeam sonar underwater terrain elevation image data, the first micro-navigation model uses a third encoder and a fourth encoder based on an attention mechanism. High weights are assigned to large-sized terrain feature points such as seabed protrusions and seabed depressions, while low weights are assigned to small-sized terrain feature points such as seabed cracks, which further improves the accuracy of micro-navigation.

[0151] The second micro-navigation model uses real-time state data, real-time environmental data, side-scan sonar data, and target terrain feature point data as input data, and outputs second micro-drive data. The method for obtaining the target terrain feature point data is the same as the previously described method, and the obtained target terrain feature point data is...

[0152] The second micro-navigation model includes a second encoder, a second fusion unit, and a second decoder. The second encoder comprises a fifth sub-encoder, a sixth sub-encoder, a seventh sub-encoder, and an eighth sub-encoder. The fifth sub-encoder is used to encode the real-time state data (AUVstate) to obtain the fifth sub-vector. Specifically: The second sub-encoder is used to encode real-time environmental data to obtain the sixth sub-vector. Specifically: Fifth sub-encoder and the sixth sub-encoder Similarly, the domain dictionary model, BERT model and LSTM network can be used. This invention uses the BERT model as the fifth sub-encoder and the sixth sub-encoder.

[0153] The seventh sub-encoder is used to encode the side-scan sonar data to obtain the seventh sub-vector. Because side-scan sonar probes close to the seabed and has good detection accuracy even for small-sized seabed topographic features, the seventh sub-encoder does not need to distinguish between large and small topographic features. Therefore, the seventh sub-encoder uses a stacked convolutional neural network, i.e.

[0154] The eighth sub-encoder is used to encode the target terrain feature point data to obtain the eighth sub-vector. The eighth sub-encoder also does not require setting different weights for small and large seabed topographic feature points, and therefore does not require an attention mechanism. This invention uses an LSTM network as the eighth sub-encoder, i.e.:

[0155] The second fusion unit is used to fuse the fifth, sixth, seventh, and eighth sub-vectors to obtain the second vector, specifically as follows:

[0156] The second decoder is used to decode the second vector to obtain the second micro-drive data, specifically v = DECODER2(v2). The second decoder can be an LSTM network, a BI_LSTM network, etc. This invention uses a BI_LSTM network, specifically v = DECODER2(v2) = BI_LSTM(v2).

[0157] This invention relates to an autonomous underwater robot whose distance H from the seabed is less than a preset value ε. 0 At the same time, side-scan sonar data is used as real-time terrain data, and the second micro-navigation model is used to obtain the second micro-drive data. Meanwhile, considering the characteristics of side-scan sonar data, the second micro-navigation model uses a seventh sub-encoder and an eighth sub-encoder that are different from the first micro-navigation model. It no longer distinguishes between large-size and small-size seabed feature points, which further improves the accuracy of micro-navigation.

[0158] S4: Adjust the first drive data based on the micro-drive data to obtain the second drive data. Specifically, the autonomous underwater robot is at a distance H from the seabed that is greater than or equal to a preset value ε. 0 The first micro-drive data is used to adjust the first drive data to obtain the second drive data, when the distance H between the autonomous underwater robot and the seabed is less than the preset value ε. 0 At this time, the first driving data is adjusted using the second micro-drive data to obtain the second driving data. Specifically, the adjustment method involves superimposing either the first or second micro-drive data onto the first driving data within the current detection range to obtain the second driving data.

[0159] S5: Send the second drive data to the drive unit so that the autonomous underwater robot can move according to the second drive data. The second drive data is sent to the drive unit of the autonomous underwater robot, and the autonomous underwater robot moves according to the second drive data.

[0160] It is understood that steps S2-S5 in this invention are performed iteratively according to the operating area. That is, in the next operating area, the autonomous underwater robot still executes steps S2-S5 until the task is completed. Since the autonomous underwater robot adjusts the first drive data with micro-drive data in each operating area and uses the adjusted second drive data to drive the operation of the autonomous underwater robot, the accuracy of the robot's path navigation during underwater operation is improved. At the same time, the autonomous underwater robot only needs to process real-time environmental data, real-time terrain data, and real-time status data within the operating area, and does not need to process global environmental data, terrain data, and status data, nor does it process historical environmental data and terrain data, thus improving operational efficiency and saving the underwater robot's computing resources.

[0161] In summary, this invention obtains the first driving data based on historical environmental and terrain data in the autonomous underwater robot path navigation platform. The autonomous underwater robot only needs to process real-time environmental, terrain, and status data at local moments based on the micro-navigation model, without needing to process global environmental, terrain, or status data, or historical environmental and terrain data, thus improving data processing efficiency and effectively reducing the robot's computational resources. Simultaneously, this invention obtains micro-driving data based on real-time status, environmental, and terrain data, and uses this micro-driving data to adjust the first driving data, improving the accuracy of the autonomous underwater robot's path navigation. Furthermore, this invention uses different micro-navigation models and different data feature spaces based on different operating depths, making the autonomous underwater robot's path navigation more closely aligned with the operating environment, further improving path navigation accuracy. Moreover, when generating the first driving data, this invention comprehensively considers detector point data, obstacle point data, terrain feature point data, and historical environmental data, along with multiple data constraints and a specific path planning model, making the autonomous underwater robot's path planning more accurate.

[0162] On the other hand, such as Figure 9 As shown, an embodiment of the present invention provides an underwater robot navigation system 900, which may include:

[0163] The first module 901 is used to obtain the first driving data generated by the path navigation platform;

[0164] The second module 902 is used to acquire real-time data of the target underwater robot; the real-time data includes real-time status data, real-time environmental data, and real-time terrain data.

[0165] The third module 903 is used to match the target micro-navigation model with the depth data in the real-time data, and use the target micro-navigation model to generate micro-drive data based on the real-time data.

[0166] The fourth module 904 is used to adjust the first drive data based on the micro drive data to obtain the second drive data;

[0167] The fifth module 905 is used to send the second driving data to the drive unit of the target underwater robot so that the target underwater robot can move according to the second driving data.

[0168] The sixth module 906 is used to take the second driving data as the first driving data and return to execute the second module until the target underwater robot reaches the detection point position.

[0169] The content of the method embodiments of the present invention is applicable to the system embodiments. The specific functions implemented in the system embodiments are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above methods.

[0170] On the other hand, embodiments of the present invention also provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned underwater robot navigation method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0171] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0172] like Figure 10 As shown, Figure 10 The hardware structure of an electronic device 1000 according to another embodiment is illustrated. The electronic device 1000 includes:

[0173] The processor 1001 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (aSIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention.

[0174] The memory 1002 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RaM). The memory 1002 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1002 and is called and executed by the processor 1001 to execute the network node population optimization method of the embodiments of this invention.

[0175] Input / output interface 1003 is used to implement information input and output;

[0176] The communication interface 1004 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0177] Bus 1005 transmits information between various components of the device (e.g., processor 1001, memory 1002, input / output interface 1003, and communication interface 1004);

[0178] The processor 1001, memory 1002, input / output interface 1003 and communication interface 1004 are connected to each other within the device via bus 1005.

[0179] The electronic device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0180] The content of the method embodiments of the present invention is applicable to the embodiments of the present electronic device. The specific functions implemented by the embodiments of the present electronic device are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above methods.

[0181] Another aspect of this invention provides a computer-readable storage medium storing a program that is executed by a processor to implement the aforementioned method.

[0182] It should be noted that the computer-readable medium shown in the embodiments of the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD to ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, wherein computer-readable program code is carried. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0183] The content of the method embodiments of the present invention is applicable to the computer-readable storage medium embodiments. The specific functions implemented by the computer-readable storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above methods.

[0184] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned method.

[0185] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0186] It should be noted that although several modules for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0187] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the method according to the embodiments of the present invention.

[0188] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented in this invention. Alternative embodiments are contemplated, in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0189] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the system disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0190] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0191] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, system, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, system, or device). For the purposes of this specification, "computer-readable medium" can mean any system that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, system, or device.

[0192] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0193] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0194] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0195] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0196] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A navigation method for an underwater robot, characterized in that, Includes the following steps: Obtain the first driving data generated by the path navigation platform; Acquire real-time data of the target underwater robot; the real-time data includes real-time status data, real-time environmental data, and real-time terrain data, the real-time terrain data including multibeam sonar underwater terrain elevation image data and side-scan sonar data; Match the target micro-navigation model with the depth data in the real-time data, and use the target micro-navigation model to generate micro-drive data based on the real-time data, specifically: obtain the target terrain feature point data of the target underwater robot; When the depth data is greater than or equal to a preset first threshold, the first micro-navigation model generates first micro-drive data based on the real-time status data, the real-time environment data, the multibeam sonar underwater terrain elevation image data, and the target terrain feature point data; Otherwise, the second micro-navigation model is used to generate second micro-drive data based on the real-time status data, the real-time environment data, the side-scan sonar data, and the target terrain feature point data; The first drive data is adjusted according to the micro-drive data to obtain the second drive data; The second driving data is sent to the drive unit of the target underwater robot so that the target underwater robot can move according to the second driving data; Using the second driving data as the first driving data, the process returns to the step of acquiring real-time data of the target underwater robot until the target underwater robot reaches the detection point.

2. The underwater robot navigation method according to claim 1, characterized in that, The process of obtaining the first driving data generated by the path navigation platform includes the following steps: Input the physical parameter data and target data of the target underwater robot into the path navigation platform; The physical parameter data includes weight and three-dimensional dimensions; the target data includes at least one of target location point data, terrain feature point data, and historical environmental data; the target location point data includes the location data of the detection point and the obstacle point. The path navigation platform invokes a pre-trained path planning model to generate the first driving data based on the physical parameter data and the target data; The path planning model is trained using an action and evaluation architecture based on an encoding network and a deep reinforcement learning network.

3. The underwater robot navigation method according to claim 1, characterized in that, The process of acquiring real-time data of the target underwater robot includes the following steps: The real-time status data of the target underwater robot is acquired based on a preset first sensor; the real-time status data includes depth data and motion parameter data. The real-time environmental data of the target underwater robot is acquired based on a preset second sensor; the real-time environmental data includes ocean current velocity and ocean current direction; The real-time terrain data of the area where the target underwater robot is located is acquired based on a preset terrain detection device; the real-time terrain data includes multibeam sonar underwater terrain elevation image data and side-scan sonar data.

4. The underwater robot navigation method according to claim 1, characterized in that, Both the first micro-navigation model and the second micro-navigation model include an encoder, a fusion unit, and a decoder; the encoder includes multiple sub-encoders.

5. The underwater robot navigation method according to claim 4, characterized in that, The first micro-navigation model includes a first encoder, a first fusion unit, and a first decoder. The first encoder includes a first sub-encoder, a second sub-encoder, a third sub-encoder, and a fourth sub-encoder. The step of generating first micro-drive data using the first micro-navigation model based on the real-time state data, the real-time environment data, the multibeam sonar underwater terrain elevation image data, and the target terrain feature point data includes the following steps: The first sub-encoder is used to encode the real-time status data to obtain a first sub-vector; the second sub-encoder is used to encode the real-time environment data to obtain a second sub-vector; the third sub-encoder is used to encode the multibeam sonar underwater terrain elevation image data to obtain a third sub-vector; and the fourth sub-encoder is used to encode the target terrain feature point data to obtain a fourth sub-vector. The first and second sub-encoders employ any one of the following: a domain dictionary model, a BERT model, and an LSTM network; the third sub-encoder employs a convolutional neural network based on an attention mechanism; and the fourth sub-encoder employs an LSTM neural network based on an attention mechanism. The first fusion processor is used to perform a first fusion process on the first sub-vector, the second sub-vector, the third sub-vector, and the fourth sub-vector to obtain a first vector; The first vector is processed by the first decoder to obtain the first micro-drive data; The first decoder uses either an LSTM network or a BI_LSTM network.

6. The underwater robot navigation method according to claim 4, characterized in that, The second micro-navigation model includes a second encoder, a second fusion unit, and a second decoder. The second encoder includes a fifth sub-encoder, a sixth sub-encoder, a seventh sub-encoder, and an eighth sub-encoder. The step of generating second micro-drive data using the second micro-navigation model based on the real-time state data, the real-time environmental data, the side-scan sonar data, and the target terrain feature point data includes the following steps: The real-time status data is encoded using the fifth sub-encoder to obtain the fifth sub-vector; the real-time environment data is encoded using the sixth sub-encoder to obtain the sixth sub-vector; the side-scan sonar data is encoded using the seventh sub-encoder to obtain the seventh sub-vector; and the target terrain feature point data is encoded using the eighth sub-encoder to obtain the eighth sub-vector. The fifth and sixth sub-encoders employ any one of the following: a domain dictionary model, a BERT model, and an LSTM network; the seventh sub-encoder employs a stacked convolutional neural network; and the eighth sub-encoder employs an LSTM network. The second fusion processor is used to perform a second fusion process on the fifth sub-vector, the sixth sub-vector, the seventh sub-vector, and the eighth sub-vector to obtain a second vector; The second vector is subjected to a second decoding process using the second decoder to obtain the second micro-drive data; The second decoder uses either an LSTM network or a BI_LSTM network.

7. The underwater robot navigation method according to claim 4, characterized in that, The process of acquiring the target terrain feature point data of the target underwater robot includes the following steps: Based on the positioning data of the target underwater robot, multiple terrain feature point data within a preset distance of the target underwater robot are obtained and processed to obtain the target terrain feature point data; The positioning data includes at least one of long baseline positioning data, short baseline positioning data, and ultra-short baseline positioning data.

8. An underwater robot navigation system, characterized in that, include: The first module is used to obtain the first driving data generated by the path navigation platform; The second module is used to acquire real-time data of the target underwater robot; the real-time data includes real-time status data, real-time environmental data and real-time terrain data, and the real-time terrain data includes multibeam sonar underwater terrain elevation image data and side-scan sonar data; The third module is used to match the target micro-navigation model based on the depth data in the real-time data, and use the target micro-navigation model to generate micro-drive data based on the real-time data. Specifically, it is used to obtain the target terrain feature point data of the target underwater robot. When the depth data is greater than or equal to a preset first threshold, the first micro-navigation model generates first micro-drive data based on the real-time status data, the real-time environment data, the multibeam sonar underwater terrain elevation image data, and the target terrain feature point data; Otherwise, the second micro-navigation model is used to generate second micro-drive data based on the real-time status data, the real-time environment data, the side-scan sonar data, and the target terrain feature point data; The fourth module is used to adjust the first drive data according to the micro-drive data to obtain the second drive data; The fifth module is used to send the second driving data to the drive unit of the target underwater robot so that the target underwater robot can move according to the second driving data; The sixth module is used to take the second driving data as the first driving data and return to execute the second module until the target underwater robot reaches the detection point position.

9. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method as described in any one of claims 1 to 7.

10. A computer storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to implement the method as described in any one of claims 1 to 7.

Images