Cable laying robot submarine cable far-field sensing and positioning method based on multi-modal sensor, electronic equipment and readable storage medium

Abstract

The invention belongs to the technical field of ocean engineering equipment and underwater detection, and particularly discloses a far-field sensing and positioning method for a submarine cable of a cable laying robot based on a multi-modal sensor, electronic equipment and a readable storage medium. The method comprises the following steps: S1, acquiring acoustic, optical, magnetic and electric field original data of the submarine cable by using a multi-mode sensor array; s2, constructing an original data set with time-space synchronization, ensuring that timestamps and space coordinates of acoustics, optics, magnetic force and electric field data are consistent, and forming a standardized four-dimensional tensor data set; s3, performing space-time registration and joint noise reduction processing on the original data to generate standardized sensing data; s4, constructing a dynamic environment fusion model for fusing the real-time sea condition parameters and the prior knowledge base to generate environment constraint conditions; s5, iteratively calculating the position of the cable by adopting a prediction-correction mechanism based on a particle filtering and model prediction control algorithm; and S6, outputting cable positioning coordinates and cable laying robot adaptive path planning parameters.

Description

Technical Field

[0001] The present invention belongs to the technical field of offshore engineering equipment and underwater detection, and particularly relates to a method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, an electronic device, and a readable storage medium. Background Art

[0002] As the core infrastructure for cross-ocean energy and information transmission, the precise detection and positioning technology of submarine cables is a long-standing technical problem in the field of offshore engineering. The current international mainstream technical route mainly relies on a single sensing modality: although the acoustic detection system can achieve large-scale topographic mapping, limited by the underwater acoustic wave propagation characteristics, the long-distance resolution significantly decreases, and it is difficult to clearly identify the boundary characteristics between the cable and the seabed sediment; although the electromagnetic detection scheme is sensitive to the magnetic field characteristics of the cable, it is easily affected by background interference in complex geological environments, and the sensitivity is difficult to meet the requirements of complex seabed environments; although the optical imaging technology can provide intuitive visual information, its effective operating distance and anti-turbidity ability have obvious limitations.

[0003] Due to the systematic defects that generally exist in a single sensing mode in complex operation scenarios; the acoustic system is prone to reverberation interference in a strong scattering environment, resulting in a decrease in the accuracy of cable contour extraction; the detection ability of the electromagnetic device for non-magnetic sheath cables is limited, and it is prone to misjudgment in areas rich in iron minerals; although the existing composite detection schemes attempt to combine different sensing data, due to the lack of an effective spatio-temporal alignment mechanism and dynamic weight allocation strategy, the multi-source data collaboration efficiency is low, and it is difficult to suppress the error accumulation effect under dynamic sea conditions.

[0004] There are three core defects in the existing detection and positioning technologies of submarine cables: 1. The single sensing modality leads to the lack of data dimensions, and none of the acoustic, electromagnetic, or optical means can independently cope with complex seabed environments; among them, sonar is easily affected by water body reverberation interference and it is difficult to distinguish the boundary between the cable and the sediment; the magnetometer fails for non-magnetic sheath cables; the signal-to-noise ratio of optical imaging drops sharply under low illuminance or high turbidity conditions; 2. The data processing architecture is rigid. The traditional system adopts a serial processing process, and the acoustic, electromagnetic, and optical data are independently processed and then simply superimposed, lacking a cross-modal feature association and weight self-adaptive mechanism, resulting in low fusion efficiency; 3. The environmental dynamic adaptation ability is weak. The existing technologies generally adopt a static positioning model, and parameters such as real-time sea current, turbidity, and seabed topography changes are not included in the calculation framework, resulting in the positioning error increasing linearly with the operation time. Especially in sea conditions above level 4, the positioning error of traditional methods can reach 3-5 meters, seriously restricting the accuracy of deep-sea cable laying and maintenance. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention provides a method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, an electronic device, and a readable storage medium.

[0006] For the above purposes, the present invention is achieved through the following technical solutions:

[0007] In a first aspect of the present invention, a method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors is provided, including the following steps:

[0008] S1. Use a multi-modal sensor array to collect the original data of the acoustic, optical, magnetic, and electric fields of the submarine cable.

[0009] S2. Perform spatio-temporal alignment processing on the collected original data to construct a spatio-temporally synchronized original data set, ensuring that the timestamps and spatial coordinates of the acoustic, optical, magnetic, and electric field data are consistent, and forming a standardized four-dimensional tensor data set.

[0010] S3. Perform spatio-temporal registration and joint noise reduction processing on the original data to generate standardized perception data.

[0011] S4. Construct a dynamic environment fusion model for fusing real-time sea condition parameters and a prior knowledge base to generate environmental constraint conditions.

[0012] S5. Adopt a prediction-correction mechanism, based on the particle filter and model predictive control (MPC) algorithms, to iteratively calculate the cable position.

[0013] S6. Output the cable positioning coordinates and the adaptive path planning parameters of the cable-laying robot.

[0014] According to the above method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, preferably, in step S1, the multi-modal sensor array collects through the collaborative collection of an acoustic sensor, an optical sensor, a magnetic sensor, and an electric field sensor array; the acoustic sensor array uses broadband hydrophones to cover the frequency band of 20 Hz - 100 kHz; the optical sensor is equipped with a 532 nm wavelength laser scanner to achieve high-resolution imaging of the seabed topography; the magnetic sensor uses a three-axis fluxgate array to detect magnetic field perturbations of ≤1 nT magnitude; the electric field sensor collects the potential gradient data at a sampling rate of 1 kHz through differential electrode pairs; the nodes of all sensors use the USBL system on the mother ship to periodically transmit acoustic signals containing UTC timestamps, and after the robot receives the signals and compensates for the propagation delay (the sound speed profile needs to be known), the timestamp error is controlled within ±1 ms, and the spatial coordinate error ≤0.05 m.

[0015] According to the above method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, preferably, in step S2, the original data in the original data set is stored as a four-dimensional tensor, and through the spatio-temporal alignment of multi-source heterogeneous data, the problem of data loss and spatio-temporal misalignment of traditional single-modal perception in complex sea conditions can be solved, expressed as:

[0016]

[0017] Among them, represents the multi-modal perception data set after spatio-temporal alignment; N is the number of spatial grids; T is the length of the time series; k represents the acoustic (A) / optical (O) / magnetic (M) / electric field (E) modality; C is the data channel dimension; d represents the multi-modal perception data unit; i represents the spatial grid index; j represents the time series index; l represents the data channel dimension.

[0018] According to the above-mentioned method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, preferably, in step S3, the original data uses a spatio-temporal interpolation algorithm based on feature point matching to achieve multi-modal data registration.

[0019] For the k-th modality data S k , it is aligned to a unified coordinate system through the projection transformation P k = R k ·S k + T k , where the rotation matrix R k and the translation vector T k are iteratively optimized by the least squares method to ensure the registration error.

[0020] The noise reduction link of the original data adopts a hybrid strategy. The acoustic data in the original data is threshold denoised by the Daubechies-8 wavelet basis to eliminate water flow turbulence noise; the magnetic data in the original data uses Kalman filtering to suppress geomagnetic interference, which is expressed as:

[0021] X(k + 1) = A·X(k) + B·U(k);

[0022] Among them, X(k + 1) represents the state vector at the (k + 1)-th moment; X(k) represents the state vector at the k-th moment; U(k) represents the control input vector at the k-th moment; A represents the state transition matrix; B represents the control input matrix.

[0023] The optical data in the original data applies adaptive median filtering, and the window size is dynamically adjusted (from 3×3 to 7×7 pixels) to remove suspended particle scattering noise.

[0024] The normalization processing of the original data uses Min-Max normalization to map each modality data to the interval [-1, 1], and converts it into an HDF5 format spatio-temporal sequence to achieve cross-platform data compatibility.

[0025] According to the above-mentioned method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, preferably, in step S4, the dynamic environment fusion model is used to fuse real-time sea condition parameters and a priori knowledge base to generate dynamic environment constraint conditions, which is to define an environment-position coupling model:

[0026] C safe = f(E t , K p ) = σ(W e ·E t + W k ·K p );

[0027] Among them, C safe is the dynamic safety constraint vector; σ(·) is the non-linear activation function; W k is the weight matrix; E t is the sea condition parameter vector; K p is the a priori knowledge vector; W e is the a priori knowledge weight matrix.

[0028] According to the above-mentioned method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, preferably, in step S5, when iteratively calculating the cable position based on the prediction-correction mechanism, its particle filter state is:

[0029] z k+1 = z k + B·u k + w k ;

[0030] Among them, z k+1 is the robot state vector at time k + 1; z k is the robot state vector at time k; B is the control matrix; u k is the robot trajectory control command; w k ~ N(0, ∑) is the Gaussian process noise.

[0031] Generate a set of particles randomly according to the initial state distribution; then in the prediction stage, predict each particle according to the state update equation to generate a predicted particle set

[0032] Calculate the weight of each particle according to the multi-modal sensor data (acoustic, optical, magnetic, electric field), and normalize it Subsequently, after introducing the weights, use the Monte Carlo method to resample and generate a new particle set

[0033] Calculate the particle mean as the final state estimate:

[0034]

[0035] Calculate the state covariance matrix Σ according to the particle set z :

[0036]

[0037] According to the state covariance matrix Σ z and a 99% confidence level, calculate the parameters of the confidence ellipse, specifically:

[0038]

[0039] where Δz is the position deviation relative to the mean; is the critical value of the chi-square distribution; the degree of freedom df is the state dimension (for example, when the 3D space position is df = 3).

[0040] According to the above-mentioned method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, preferably, the specific steps of step S6 are: output adaptive motion parameters through the geographic coordinate transformation and path planning algorithm, and the path optimization is expressed as:

[0041]

[0042] where p t is the robot pose; α, β, γ are adaptive weight coefficients; P collision is the collision probability model; p goal is the target pose vector.

[0043] The second aspect of the present invention provides an electronic device, including a memory and a processor, and a computer program is stored on the memory, and when the processor executes the computer program, any step in the method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors as described in the first aspect is implemented.

[0044] The third aspect of the present invention provides a computer-readable storage medium, and a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a computer processor, any step in the method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors as described in the first aspect is implemented.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0046] 1. The present invention combines multi-modal sensor collaborative perception, dynamic environment modeling, and prediction-correction optimization mechanism to solve the problems of high-precision positioning of submarine cables and robot adaptive control in complex marine environments. By synchronously collecting and fusing spatio-temporal data from multi-source heterogeneous data such as acoustic, optical, magnetic, and electric fields, and combining real-time sea condition parameters with a prior knowledge base, the present invention constructs a dynamic environment fusion model, and based on particle filter and model predictive control (MPC) algorithms, realizes closed-loop positioning and path planning of the cable, significantly improving the detection efficiency of the cable position and the safety of robot operations.

[0047] 2. The multi-modal sensor of the present invention integrates acoustic, electromagnetic, optical, and electric field sensors, and collaboratively collects multi-dimensional data through a broadband hydrophone, a laser scanner, a triaxial fluxgate array, and differential electrode pairs. Combining the mother ship's periodic emission of acoustic signals containing UTC timestamps through the USBL system, a spatio-temporally aligned original data set is constructed, and a spatio-temporal interpolation algorithm and a hybrid noise reduction strategy are used to achieve standardized processing of multi-source heterogeneous data. Innovatively introducing a dynamic environment fusion model, integrating real-time sea condition parameters and a prior knowledge base to generate environmental constraint conditions, combining particle filter and model predictive control (MPC) algorithms, iteratively optimizing the cable position estimation through a prediction-correction mechanism, and outputting sub-meter-level positioning coordinates and confidence ellipses based on Monte Carlo resampling and covariance analysis, which can solve the problems of long-distance sensing ambiguity, insufficient positioning accuracy, and poor dynamic interference robustness in cable detection in complex seabed environments.

[0048] 3. The present invention realizes adaptive weight allocation and error correction of multi-modal data by defining an environment-position coupling model, significantly improving the anti-interference ability in harsh environments such as strong currents and high turbidity. The present invention can be applied to scenarios such as deep-sea cable laying, seabed energy network operation and maintenance, and communication optical cable fault location, providing a full-chain solution for long-distance and high-precision seabed detection, achieving a generational breakthrough in terms of perception comprehensiveness, positioning accuracy, and environmental adaptability compared with traditional single-modal methods.

[0049] 4. The core advantages of the present invention are reflected in three dimensions: First, a three-level acoustic-magnetic-optical perception system is constructed. Wide-area terrain scanning is achieved using synthetic aperture sonar (SAS), breaking through the physical resolution limitations of traditional sonars. The broadband magnetometer array adopts differential measurement and common-mode rejection techniques to improve the magnetic field detection sensitivity to the 1 nT level, effectively overcoming geomagnetic interference. The low-light optical unit is equipped with laser scanning and dark channel restoration algorithms, which can clearly capture the surface features of the cable at an illuminance of 0.01 lux. The three types of sensors form complementary verification through a spatio-temporal synchronization mechanism, completely solving the data blind spot problem of single-modal perception. Second, a dynamic adaptive data fusion framework is established. An innovative environment-position coupling model is introduced to fuse the data of inertial measurement units (IMUs), Doppler velocity log (DVL), and depth sensors in real time, constructing a spatio-temporal reference with millimeter-level accuracy. At the same time, an environment-position coupling model is constructed based on a deep learning time series prediction algorithm. Through the iterative optimization of particle filtering and model predictive control (MPC), the dynamic allocation of weights and real-time error correction of multi-source heterogeneous data are realized, enabling the present invention to maintain sub-meter-level positioning accuracy in harsh environments such as strong currents and high turbidity. Third, full-chain environmental adaptability enhancement is achieved. The present invention is built with a real-time sea condition parameter perception module, which can dynamically collect environmental variables such as flow velocity, turbidity, and bottom sediment type, and match and analyze them with geological data and cable laying records in the prior knowledge base to generate an environmental constraint matrix to guide algorithm optimization. This "perception-modeling-feedback" closed-loop mechanism enables the present invention to autonomously adjust the working parameters and fusion strategies of sensors, significantly enhancing the robustness under complex working conditions.

[0050] 5. The multi-modal fusion architecture of the present invention breaks through the data dimension limitations and algorithm rigidity bottlenecks of traditional technologies, providing a systematic solution for far-field perception and high-precision positioning of submarine cables through triple innovations in hardware co-design, software intelligent optimization, and environmental dynamic adaptation. Compared with existing domestic and foreign technologies, the present invention has achieved generational improvements in perception comprehensiveness, positioning accuracy, and environmental adaptability, laying a key technical foundation for the construction and maintenance of deep-sea energy networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a schematic flow chart of the present invention;

[0052] Figure 2 is a schematic flow chart of the implementation process for obtaining standardized perception data and dynamic processing strategies in Embodiment 1 of the present invention;

[0053] Figure 3 is a schematic diagram of the confidence ellipse of the cable position obtained by using a prediction-correction mechanism in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0054] The present invention will be further described in detail below through specific embodiments, but the scope of the present invention is not limited thereby.

[0055] Embodiment 1

[0056] A method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors, the process of which is as Figure 1 shown, and includes the following steps:

[0057] S1. Use a multi-modal sensor array to collect the original data of the acoustics, optics, magnetism, and electric field of the submarine cable.

[0058] The multi-modal sensor array collects data through the collaborative collection of an acoustic sensor, an optical sensor, a magnetic sensor, and an electric field sensor array; the acoustic sensor array uses broadband hydrophones to cover the frequency band of 20 Hz - 100 kHz; the optical sensor is equipped with a 532 nm wavelength laser scanner to achieve high-resolution imaging of the seabed topography; the magnetic sensor uses a three-axis fluxgate array to detect magnetic field disturbances of ≤ 1 nT magnitude; the electric field sensor collects potential gradient data with a sampling rate of 1 kHz through differential electrode pairs; the nodes of all sensors use the USBL system on the mother ship to periodically transmit acoustic signals containing UTC timestamps, and the signals received by the robot are compensated for propagation delay (the sound speed profile needs to be known), and the timestamp error is controlled within ±1 ms, and the spatial coordinate error ≤ 0.05 m.

[0059] S2. Perform spatio-temporal alignment processing on the collected original data to construct a spatio-temporally synchronized original data set, ensuring that the timestamps and spatial coordinates of the acoustic, optical, magnetic, and electric field data are consistent, and forming a standardized four-dimensional tensor data set.

[0060] As Figure 2 shown, through the collaborative collection of an acoustic sensor, an optical sensor, a magnetic sensor, and an electric field sensor array, a spatio-temporally synchronized original data set is constructed; the original data in the original data set is stored as a four-dimensional tensor, and through the spatio-temporal alignment of multi-source heterogeneous data, the problems of data loss and spatio-temporal misalignment in traditional single-modal perception under complex sea conditions can be solved, expressed as:

[0061]

[0062] where represents the multi-modal perception data set after spatio-temporal alignment; N is the number of spatial grids; T is the length of the time series; k represents the acoustic (A) / optical (O) / magnetic (M) / electric field (E) modality; C is the data channel dimension; d represents the multi-modal perception data unit; i represents the spatial grid index; j represents the time series index; l represents the data channel dimension.

[0063] Integrate multi-source heterogeneous data through a spatio-temporal alignment mechanism to solve the problems of data loss and spatio-temporal misalignment under complex sea conditions.

[0064] S3. Perform spatio-temporal registration and joint denoising processing on the original data to generate standardized perception data.

[0065] The original data uses a spatio-temporal interpolation algorithm based on feature point matching to achieve multi-modal data registration.

[0066] For the k-th modal data S k , align it to a unified coordinate system through a projection transformation P k = R k ·S k + T k , where the rotation matrix R k and the translation vector T k are iteratively optimized by the least squares method to ensure the registration error.

[0067] The denoising link of the original data adopts a hybrid strategy. The acoustic data in the original data is threshold denoised by the Daubechies-8 wavelet basis to eliminate water flow turbulence noise; the magnetic data in the original data uses Kalman filtering to suppress geomagnetic interference, which is expressed as:

[0068] X(k + 1) = A·X(k) + B·U(k);

[0069] Among them, X(k + 1) represents the state vector at the (k + 1)-th moment; X(k) represents the state vector at the k-th moment; U(k) represents the control input vector at the k-th moment; A represents the state transition matrix; B represents the control input matrix.

[0070] The optical data in the original data applies adaptive median filtering, and the window size is dynamically adjusted (from 3×3 to 7×7 pixels) to remove suspended particle scattering noise.

[0071] The standardization processing of the original data uses Min-Max normalization to map each modal data to the [-1, 1] interval and convert it into an HDF5 format spatio-temporal sequence to achieve cross-platform data compatibility.

[0072] For the dynamic adjustment of the data processing strategy, combine real-time sea condition parameters (flow velocity, turbidity, bottom sediment type) with the prior knowledge base (geological data, cable laying records) to achieve weight adaptability. For example, in a high turbidity environment, increase the fusion weight of acoustic and magnetic data; under low illuminance conditions, preferentially use optical near-field verification data; dynamically adjust the noise covariance parameter of Kalman filtering according to the real-time sea current speed to optimize the denoising effect of magnetic data.

[0073] S4. Construct a dynamic environment fusion model for fusing real-time sea condition parameters and a prior knowledge base to generate environmental constraint conditions.

[0074] The dynamic environment fusion model is used to fuse real-time sea condition parameters and a prior knowledge base to generate dynamic environmental constraint conditions, which is a defined environment-position coupling model:

[0075] C safe = f(E t , K p ) = σ(W e ·E t + W k ·K p );

[0076] Among them, C safe is the dynamic safety constraint vector; σ(·) is the non-linear activation function; W k is the weight matrix; E t is the sea condition parameter vector; K p is the prior knowledge vector; W e is the prior knowledge weight matrix.

[0077] S5. Adopt a prediction-correction mechanism and, based on the particle filter and model predictive control (MPC) algorithms, iteratively calculate the cable position.

[0078] As Figure 3 shown, in the process of submarine cable positioning, the core of the prediction-correction mechanism is to iteratively estimate the cable position through the particle filter. First, for the initial state distribution p(x0), a set of particles is randomly generated, where N is the number of particles. Each particle represents a possible state of the system.

[0079] Then, in the prediction stage, for each particle the prediction is made according to the state update equation:

[0080]

[0081] Among them, is the predicted state of the i-th particle; is the current state of the i-th particle; is the state of the i-th particle at the current moment; B is the control matrix that maps the robot trajectory control instruction to the state space; u k is the robot trajectory control instruction from the model predictive control (MPC) module; is the Gaussian process noise, representing the uncertainty in the system.

[0082] Through the above formula, a set of predicted particle sets

[0083] After the prediction phase is completed, enter the calibration phase; calculate the weight of each particle based on multi-modal sensor data (acoustic, optical, magnetic, electric field); the weight formula is:

[0084]

[0085] where λ m is the confidence weight of the m-th modality (acoustic, optical, magnetic, electric field); s m is the actual observation value of the m-th modality; is the predicted observation value based on the particle state; σ m is the standard deviation of the observation noise of the m-th modality.

[0086] Subsequently, normalize the weights of all particles:

[0087]

[0088] To reduce the particle degeneracy phenomenon, retain high-weight particles and eliminate low-weight particles, and introduce the Monte Carlo resampling step; resample using the Monte Carlo method according to the normalized weights to generate a new particle set

[0089] After resampling is completed, calculate the particle mean as the final state estimate:

[0090]

[0091] At the same time, calculate the state covariance matrix Σ z :

[0092]

[0093] The state covariance matrix Σ z describes the distribution characteristics of the particle set and reflects the uncertainty of the system state.

[0094] According to the state covariance matrix Σ z and the 99% confidence level, calculate the parameters of the confidence ellipse; the specific formula is:

[0095]

[0096] where Δz is the position deviation relative to the mean; is the critical value of the chi-square distribution, and the degrees of freedom df is the state dimension (for example, when the 3D space position is df = 3).

[0097] For the 3D space position, the state covariance matrix Σ zis a 3×3 matrix, whose eigenvalues and eigenvectors represent the semi-axis lengths and directions of the confidence ellipse respectively. The specific steps are as follows:

[0098] Calculate Σ z for its eigenvalues λ1, λ2, λ3 and corresponding eigenvectors v1, v2, v3.

[0099] Calculate the semi-axis lengths of the confidence ellipse according to the eigenvalues:

[0100]

[0101] where a i represents the i-th semi-axis length of the confidence ellipse.

[0102] The eigenvectors v1, v2, v3 define the direction of the confidence ellipse.

[0103] Through the above steps, a three-dimensional confidence ellipsoid can be obtained, whose geometric shape and size intuitively reflect the uncertainty of the system state.

[0104] S6. Output the cable positioning coordinates and the adaptive path planning parameters of the cable-laying robot.

[0105] Output the adaptive motion parameters through the geographic coordinate transformation and the path planning algorithm. The path optimization is expressed as:

[0106]

[0107] where p t is the robot pose; α, β, γ are the adaptive weight coefficients; P collision is the collision probability model; p goal is the target pose vector.

[0108] Embodiment 2

[0109] An electronic device includes a memory and a processor. A computer program is stored on the memory. When the processor executes the computer program, any step in the method for far-field sensing and positioning of submarine cables by a cable-laying robot based on multi-modal sensors described in Embodiment 1 is implemented.

[0110] Furthermore, the process of the method for far-field sensing and positioning of submarine cables by a cable-laying robot based on multi-modal sensors described in Embodiment 1 can be implemented as a computer software program. For example, this embodiment includes a computer program product, which includes a computer program carried on a computer-readable medium. The computer program contains program codes for executing the method. In such an embodiment, the computer program can be downloaded and installed from the network and / or installed from a removable medium. When the computer program is executed by the processor, the above functions defined in the method of the present application are executed.

[0111] Example 3

[0112] A computer-readable storage medium stores a computer program thereon. When the computer program is executed by a processor, any step in implementing a method for far-field sensing and positioning of a submarine cable by a cable-laying robot based on a multi-modal sensor as described in Example 1 is realized.

[0113] The computer-readable medium described in this application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two above. 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 of the above. More specific examples of a computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted by any suitable medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0114] Computer program code for performing the operations of the present application can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Python and C++, and also include conventional procedural programming languages or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0115] The computer-readable storage medium of this embodiment can be accelerated by hardware such as a GPU, and the parallel computing advantage of the GPU is used to accelerate any step in implementing a method for far-field perception and positioning of a submarine cable by a cable-laying robot based on multi-modal sensors as described in Embodiment 1.

[0116] In summary, the present invention effectively overcomes the deficiencies in the prior art and has high industrial utilization value. The role of the above embodiments is to illustrate the substantial content of the present invention, but does not limit the protection scope of the present invention. Those of ordinary skill in the art should understand that the technical solution of the present invention can be modified or equivalently replaced without departing from the essence and protection scope of the technical solution of the present invention.

[0117] The above embodiments are specific implementation manners of the present invention, but the implementation manners of the present invention are not limited by the above embodiments. Any combination, change, modification, substitution, or simplification that does not exceed the design concept of the present invention falls within the protection scope of the present invention.

Claims

1. A method for far-field perception and positioning of submarine cables by a cable-laying robot based on multimodal sensors, characterized in that, Including the following steps: S1. Use a multi-modal sensor array to collect the original data of the acoustic, optical, magnetic, and electric fields of the submarine cable; S2. Perform spatio-temporal alignment processing on the collected original data to construct a spatio-temporally synchronized original data set, ensuring that the timestamps and spatial coordinates of the acoustic, optical, magnetic, and electric field data are consistent, and forming a standardized four-dimensional tensor data set; S3. Perform spatio-temporal registration and joint noise reduction processing on the original data to generate standardized sensing data; S4. Construct a dynamic environment fusion model for fusing real-time sea condition parameters and a prior knowledge base to generate environmental constraint conditions; S5. Adopt a prediction-correction mechanism to iteratively calculate the cable position based on the particle filter and model predictive control algorithms; S6. Output the cable positioning coordinates and the adaptive path planning parameters of the cable-laying robot.

2. The method for far-field sensing and positioning of submarine cables by a cable-laying robot based on multi-modal sensors according to claim 1, wherein, In step S1, the multi-modal sensor array collects data through the collaborative collection of an acoustic sensor, an optical sensor, a magnetic sensor, and an electric field sensor array; the acoustic sensor array uses broadband hydrophones to cover the frequency band of 20 Hz - 100 kHz; the optical sensor is equipped with a 532 nm wavelength laser scanner to achieve high-resolution imaging of the seabed topography; the magnetic sensor uses a three-axis fluxgate array to detect magnetic field disturbances of ≤1 nT magnitude; the electric field sensor collects potential gradient data with a sampling rate of 1 kHz through differential electrode pairs; the nodes of all sensors use the USBL system to periodically transmit acoustic signals containing UTC timestamps, and after receiving the signals, the propagation delay is compensated, and the timestamp error is controlled within ±1 ms, and the spatial coordinate error ≤0.05 m.

3. The method for far-field perception and positioning of submarine cables by a cable-laying robot based on multimodal sensors according to claim 1, wherein In step S2, the original data in the original data set is stored as a four-dimensional tensor, and the spatio-temporal alignment of multi-source heterogeneous data is expressed as: Among them, represents the multi-modal perception data set after spatio-temporal alignment; N is the number of spatial grids; T is the length of the time series; k represents the acoustic (A) / optical (O) / magnetic (M) / electric field (E) modality; C is the data channel dimension; d represents the multi-modal perception data unit; i represents the spatial grid index; j represents the time series index; l represents the data channel dimension.

4. The method for far-field perception and positioning of submarine cables by a cable-laying robot based on multimodal sensors according to claim 1, characterized in that, In step S3, the original data uses a spatio-temporal interpolation algorithm based on feature point matching to achieve multi-modal data registration; For the k-th modal data S k , it is aligned to a unified coordinate system through a projection transformation P k = R k ·S k + T k , where the rotation matrix R k and the translation vector T k are iteratively optimized by the least squares method to ensure the registration error; The noise reduction link of the original data adopts a hybrid strategy. The acoustic data in the original data is threshold denoised by the Daubechies-8 wavelet basis to eliminate water flow turbulence noise; the magnetic data in the original data uses Kalman filtering to suppress geomagnetic interference, which is expressed as: X(k + 1) = A·X(k) + B·U(k); where, X(k + 1) represents the state vector at the (k + 1)-th moment; X(k) represents the state vector at the k-th moment; U(k) represents the control input vector at the k-th moment; A represents the state transition matrix; B represents the control input matrix. The optical data in the original data applies adaptive median filtering, and the window size is dynamically adjusted to remove suspended particle scattering noise; The standardized processing of the original data uses Min-Max normalization to map each modal data to the interval [-1, 1], and converts it into a spatio-temporal sequence in HDF5 format to achieve cross-platform data compatibility.

5. The method for far-field perception and positioning of submarine cables by a cable-laying robot based on multimodal sensors according to claim 1, wherein In step S4, the dynamic environment fusion model is used to fuse real-time sea condition parameters and a prior knowledge base to generate dynamic environmental constraint conditions, which is to define an environment-position coupling model: C safe = f(E t , K p ) = σ(W e ·E t + W k ·K p ); Among them, C safe is the dynamic security constraint vector; σ(·) is the non-linear activation function; W k is the weight matrix; E t is the sea state parameter vector; K p is the prior knowledge vector; W e is the prior knowledge weight matrix.

6. The method for far-field perception and positioning of submarine cables by a cable-laying robot based on multimodal sensors according to claim 1, characterized in that In step S5, when iteratively calculating the cable position based on the prediction-correction mechanism, its particle filter state is: z k+1 = z k + B·u k + w k ; where z k+1 is the robot state vector at time k + 1; z k is the robot state vector at time k; B is the control matrix; u k is the robot trajectory tracking command; w k ~N(0,∑) is Gaussian process noise.

7. The method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors according to claim 2, wherein The specific steps of step S6 are as follows: output adaptive motion parameters through geographic coordinate conversion and path planning algorithms, and the path optimization is expressed as: Among them, p t is the robot pose; α, β, γ are adaptive weight coefficients; P collision is the collision probability model; p goal is the target pose vector.

8. An electronic device, comprising a memory and a processor, wherein a computer program is stored on the memory, characterized in that, When the processor executes the computer program, it implements any step in the method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a computer processor, it implements any step in the method for far-field perception and positioning of submarine cables by a cable-laying robot based on multi-modal sensors as described in any one of claims 1-7.

Images