Navigation positioning method, system and device for underwater robot

Abstract

This invention relates to the technical field of robot navigation, specifically to a navigation and positioning method, system, and device for underwater robots. The invention acquires inertial navigation data and raw acoustic data of the robot in the underwater environment. Based on the inertial navigation data, it analyzes the robot's operational posture characteristics to obtain motion feature vectors for different underwater terrains. Based on the raw acoustic data, it analyzes the differences in sound wave feedback generated by the robot in different underwater terrains to obtain acoustic feature vectors for the corresponding terrains. The motion feature vectors and acoustic feature vectors are fused and clustered to obtain correction compensation amounts for the Doppler log under different terrain conditions. When the robot performs its actual task, it performs navigation fusion positioning based on the correction compensation amounts for different terrain environments. This technical solution improves the accuracy of robot navigation in different underwater terrain environments.

Description

Technical Field

[0001] This invention relates to the technical field of robot navigation, and specifically to a navigation and positioning method, system, and device for underwater robots. Background Technology

[0002] With the rapid development of marine resource exploration, marine scientific research, and underwater engineering operations, robots are playing an increasingly prominent role as key tools for human exploration and development of the ocean. Precise, reliable, and continuous navigation and positioning capabilities enable robots to successfully complete predetermined tasks, such as area search, target tracking, precise navigation, fixed-point operations, and data georeferencing. Unlike terrestrial and aerial environments, the underwater environment possesses unique complexity and challenges, causing traditional global satellite navigation system signals to almost completely attenuate and become ineffective underwater. Therefore, developing high-precision, high-reliability autonomous navigation and positioning methods that do not rely on external satellite signals and are suitable for the unique underwater environment has become one of the core key technologies in robotics development.

[0003] In existing technologies, robot navigation consists of an inertial navigation system and a Doppler Velocity Log (DVL). Dead reckoning is performed based on the coupled navigation system. However, in complex seabed topography environments, the navigation performance of the Doppler Velocity Log will significantly decrease, resulting in insufficient positioning accuracy of the robot by the combined navigation system. Summary of the Invention

[0004] To address the technical challenge of improving the navigation accuracy of robots in diverse underwater terrain environments, this invention aims to provide a navigation and positioning method, system, and device for underwater robots. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of the present invention provide a navigation and positioning method for underwater robots, the method comprising:

[0006] Acquire inertial navigation data of the robot in the underwater environment, as well as raw acoustic data from the Doppler log installed on the robot;

[0007] The robot's operational posture characteristics are analyzed based on inertial navigation data to obtain motion feature vectors of the robot in different underwater terrains.

[0008] The differences in acoustic feedback generated by the robot in different underwater terrains are analyzed based on the original acoustic data to obtain the acoustic feature vector of the robot in the corresponding underwater environment.

[0009] The motion feature vector and acoustic feature vector are fused and clustered to obtain the correction compensation amount of the Doppler log under different terrain environments.

[0010] When the robot performs tasks underwater, it performs navigation fusion positioning based on the correction compensation amount for different terrain environments.

[0011] In one optional embodiment, acquiring inertial navigation data of the robot in the underwater environment, as well as raw acoustic data from the Doppler log installed on the robot, includes:

[0012] When the robot is in an underwater environment, the robot is sent an execution command for a target underwater navigation mission, which is a underwater navigation mission with a constant altitude and constant speed near the bottom.

[0013] When the robot is performing a target underwater mission, it synchronously collects inertial navigation data and raw acoustic data according to a preset frequency.

[0014] In one optional embodiment, the robot's operational posture characteristics are analyzed based on inertial navigation data to obtain motion feature vectors of the robot in different underwater terrains, including:

[0015] Extract corresponding data from the inertial navigation data according to the preset sliding time window, and input it into the preset first processing model;

[0016] Based on the output of the first processing model, the robot's vertical acceleration standard deviation, vertical acceleration peak value, first absolute average value of pitch angular velocity, second absolute average value of roll angular velocity, cross-correlation coefficient between vertical acceleration and pitch angular velocity, center-of-gravity frequency of vertical acceleration power spectrum, first range of pitch angle variation, and second range of roll angle variation are obtained.

[0017] Underwater terrain analysis is performed based on the standard deviation of vertical acceleration, peak vertical acceleration, average of the first absolute value, average of the second absolute value, cross-correlation coefficient, center of gravity frequency, first range of variation, and second range of variation to obtain the robot's motion feature vector in the corresponding underwater terrain.

[0018] In one optional embodiment, the differences in acoustic feedback generated by the robot under different underwater terrains are analyzed based on the original acoustic data to obtain the acoustic feature vector of the robot under the corresponding terrain in the underwater environment, including:

[0019] Based on the sliding time window for calculating the motion feature vector, acoustic feature data of multiple beams are extracted from the original acoustic data.

[0020] The acoustic characteristic data of each beam are used to evaluate beam quality and the impact of terrain features to obtain the effectiveness index of each beam.

[0021] Based on the effectiveness index of each beam and the interference index of acoustic multipath effect, the acoustic feature vector of the robot in the underwater environment corresponding to the terrain is obtained.

[0022] In one optional embodiment, the acoustic feature data includes the echo intensity, reference intensity, and correlation index for each beam; beam quality evaluation and terrain feature impact evaluation are performed on the acoustic feature data of each beam to obtain an effectiveness index for each beam, including:

[0023] Based on the echo intensity and reference intensity of each beam, the intensity-to-quality ratio of the corresponding beam is obtained;

[0024] Based on the intensity-to-quality ratio and correlation index of each beam, the quality evaluation value of the corresponding beam is obtained;

[0025] The terrain feature influence values ​​of all beams are obtained based on the standard deviation and mean rate of change of the quality evaluation values ​​of each beam in the same period.

[0026] The effectiveness index of the corresponding beam is obtained based on the influence value of terrain features, the quality evaluation value of each beam, and the standard deviation of the evaluation value for a preset time period.

[0027] In one optional embodiment, the effectiveness index of the corresponding beam is obtained based on the terrain feature influence value, the quality evaluation value of each beam, and the standard deviation of the evaluation values ​​over a preset time period, including:

[0028] The quality reliability of each beam is obtained by the ratio of its quality evaluation value to the influence value of terrain features.

[0029] The effectiveness index of each beam is obtained based on the stability of its quality reliability and the standard deviation of its evaluation value over time.

[0030] In one optional embodiment, before obtaining the acoustic feature vector of the robot corresponding to the terrain in the underwater environment, the method further includes:

[0031] Based on the amount and range of echo intensity change for each beam, the degree of intensity change for the corresponding beam can be obtained.

[0032] Based on the intensity variation and correlation variation rate of each beam, the interference index of the corresponding beam under the acoustic multipath effect is obtained.

[0033] In one optional embodiment, the motion feature vector and acoustic feature vector are fused and clustered to obtain the correction compensation amount of the Doppler log under different terrain environments, including:

[0034] The motion feature vector and acoustic feature vector are standardized and weighted and concatenated to obtain a fused feature vector;

[0035] Cluster all fused feature vectors to establish a correspondence between each cluster and the underwater terrain;

[0036] Based on the lateral velocity error and forward velocity error of Doppler logs in the same cluster, the correction compensation amount of Doppler logs under different terrain conditions is obtained.

[0037] Secondly, embodiments of the present invention also provide a navigation and positioning system for underwater robots, the system being any navigation and positioning method for underwater robots described in the first aspect, the system comprising:

[0038] The acquisition module is used to acquire inertial navigation data of the robot in the underwater environment, as well as raw acoustic data from the Doppler log installed on the robot.

[0039] The first acquisition module is used to perform operational posture feature analysis on the robot based on inertial navigation data, so as to obtain the motion feature vector of the robot in different underwater terrains.

[0040] The second acquisition module is used to analyze the differences in acoustic feedback generated by the robot in different underwater terrains based on the original acoustic data, so as to obtain the acoustic feature vector of the robot in the underwater environment corresponding to the terrain.

[0041] The third acquisition module is used to fuse and cluster the motion feature vector and acoustic feature vector to obtain the correction compensation amount of the Doppler log under different terrain environments.

[0042] The positioning module is used to perform navigation fusion positioning of the robot when it is performing tasks underwater, based on the correction and compensation amount for different terrain environments.

[0043] Thirdly, embodiments of the present invention also provide a navigation and positioning device for underwater robots, the device operating based on any of the navigation and positioning methods for underwater robots in the first aspect.

[0044] The present invention has the following beneficial effects:

[0045] The technical solution of this invention acquires inertial navigation data and raw acoustic data of a robot in an underwater environment. The inertial navigation data continuously and stably reflects the robot's motion state underwater, while the raw acoustic data retains detailed features of environmental information such as underwater sound wave propagation, reflection, and multipath propagation. Since different underwater terrains act on the robot through hydrodynamic and attitude control feedback, causing statistically significant differences in its motion responses such as pitch, roll, and vertical acceleration, the robot's operational attitude characteristics can be analyzed based on the inertial navigation data to obtain motion feature vectors for different underwater terrains. To characterize the propagation and reflection characteristics of sound waves under different terrain and sediment conditions, the differences in sound wave feedback generated by the robot under different underwater terrains are further analyzed based on the raw acoustic data to obtain acoustic feature vectors corresponding to the underwater environment. The motion feature vectors and acoustic feature vectors are fused and clustered to obtain Doppler log correction compensation amounts for different terrain environments. When the robot performs actual tasks, navigation fusion and positioning are performed based on the correction compensation amounts for different terrain environments. Based on this technical solution, the measurement error of the Doppler log can be corrected in real time and adaptively according to the actual terrain environment in which the robot is located. This effectively suppresses the influence of factors such as complex terrain and multipath interference on the accuracy of velocity calculation, significantly improves the robustness of the robot's navigation in different underwater terrain environments, and thus improves the accuracy of the robot's navigation in different underwater terrain environments. Attached Figure Description

[0046] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 A flowchart illustrating a navigation and positioning method for an underwater robot according to an embodiment of the present invention;

[0048] Figure 2 This is a schematic diagram illustrating the principle of clustering all fused feature vectors into multiple clusters according to an embodiment of the present invention;

[0049] Figure 3 This is a schematic diagram illustrating the principle of underwater single-type terrain calculation correction compensation amount provided in an embodiment of the present invention;

[0050] Figure 4 This is a schematic diagram of a navigation and positioning system for underwater robots provided in one embodiment of the present invention. Detailed Implementation

[0051] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a navigation and positioning method, system, and device for underwater robots proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0053] Doppler logs measure a ship's speed and cumulative distance relative to the seabed by utilizing the Doppler frequency shift between the emitted sound waves and the received reflected waves from the seabed. Robots equipped with Doppler logs can obtain their speed relative to the seabed / water layer, making them a core component of integrated navigation systems. However, in complex seabed topography, the performance of Doppler logs degrades significantly, and traditional Doppler log compensation relies on post-processing adjustments based on output parameters, affecting the smoothness of the integrated navigation system's adaptation to environmental changes. The following, in conjunction with the accompanying drawings, will specifically describe a navigation and positioning method, system, and device for underwater robots provided by this invention.

[0054] Please see Figure 1 This document illustrates a flowchart of a navigation and positioning method for underwater robots according to an embodiment of the present invention. This method can be applied to the operation of the robot's control terminal. The control terminal can be a control board with an integrated CPU (Central Processing Unit) or other types of controllers, as long as they can run this navigation and positioning method; no specific limitations are imposed here. The target control scenario can be the seabed or underwater, and the navigation and positioning method includes:

[0055] S11: Acquire inertial navigation data of the robot in the underwater environment, as well as raw acoustic data from the Doppler log installed on the robot.

[0056] Specifically, inertial navigation data can be obtained by reading the output data of an inertial measurement unit (IMU). The IMU allows for real-time sensing of the robot's underwater motion state, which is then represented based on the inertial navigation data. Similarly, raw acoustic data can be acquired simultaneously with the inertial navigation data, and both types of data are stored synchronously using the acquisition time as a tag. It is understood that when the robot navigates near the bottom underwater, its motion state interacts with the seabed topography, resulting in specific dynamic responses. Simultaneously, topographic features are related to sediment types, which in turn affect the reflection and scattering characteristics of sound waves, thus impacting the performance of the Doppler log. Therefore, this invention infers the acoustic characteristics of the seabed topography by analyzing the robot's motion state and raw acoustic data, thereby compensating for errors caused by changes in sound speed due to topography.

[0057] Due to the complexity of the seabed environment and the potential for various terrain features to be interspersed, data collection based on a fixed water depth may result in data that does not accurately represent the corresponding terrain environment. Therefore, in one specific implementation, step S11 includes sub-steps S11-1 to S11-2, as detailed below:

[0058] S11-1: When the robot is in an underwater environment, send execution commands for the target underwater navigation task, which is for the robot to navigate near the seabed at a constant altitude and speed. Execution commands can be sent to the robot from a shore-based user. After receiving the execution commands underwater, the robot navigates at a constant altitude and speed, and adaptively adjusts its distance from the seabed in real time.

[0059] S11-2: When the robot is performing a target underwater mission, the Doppler log emits four beams and receives the returned acoustic signals, using the acoustic signals of each beam as the corresponding raw acoustic data; similarly, inertial navigation data is collected synchronously according to a preset frequency, and the two types of data are stored synchronously.

[0060] At this point, inertial navigation data and raw acoustic data have been acquired simultaneously using the above method, and we proceed to step S12.

[0061] S12: Analyze the robot's operating posture characteristics based on inertial navigation data to obtain the robot's motion feature vectors in different underwater terrains.

[0062] Specifically, different seabed topography will generate different hydrodynamic interactions with the robot, thus affecting the robot's motion response. When the seabed topography changes, in order to maintain a constant altitude or maintain a horizontal navigation attitude, the robot's attitude control system will generate corresponding pitch and roll adjustments. Therefore, these subtle but regular dynamic responses can be captured by calculating the rate of change of pitch and roll angles (i.e., angular velocity) and fluctuations in vertical acceleration.

[0063] For example, when climbing slopes, the robot will tilt its head up, causing a continuous increase in pitch angle; while descending slopes, it will exhibit the opposite pitch response. When the robot traverses rough terrain, the irregular seabed surface will induce turbulence and pressure changes, which will directly manifest as high-frequency fluctuations in vertical acceleration and angular velocity. For regular terrain features, such as sand ripples, specific periodic patterns will be observed in the motion response. Therefore, the robot's dynamic response in the vertical direction can be extracted to reflect the different terrains the robot is in, and this can be represented by motion feature vectors.

[0064] For example, step S12 includes sub-steps S12-1 to S12-3, which are described in detail below:

[0065] S12-1: Extract corresponding data from the inertial navigation data according to the preset sliding time window and input it into the preset first processing model. The sliding time window can be set according to actual needs, such as 5-10 seconds based on the degree of terrain change, and calculate the statistics of motion characteristics within the sliding time window.

[0066] S12-2: Based on the output of the first processing model, obtain the robot's vertical acceleration standard deviation, vertical acceleration peak value, first absolute average value of pitch angular velocity, second absolute average value of roll angular velocity, cross-correlation coefficient between vertical acceleration and pitch angular velocity, center-of-gravity frequency of vertical acceleration power spectrum, first range of pitch angle variation, and second range of roll angle variation.

[0067] S12-3: Underwater terrain analysis is performed based on the standard deviation of vertical acceleration, peak vertical acceleration, average of the first absolute value, average of the second absolute value, cross-correlation coefficient, center of gravity frequency, first range of variation, and second range of variation to obtain the robot's motion feature vectors for the corresponding underwater terrain. The constructed motion feature vectors can effectively distinguish different terrain features. In this way, different terrains form separable clusters in the feature space, providing a basis for subsequent clustering and compensation.

[0068] At this point, the motion feature vectors of the robot in different underwater terrains have been obtained based on the above method, and we proceed to step S13.

[0069] S13: Analyze the differences in acoustic feedback generated by the robot in different underwater terrains based on the original acoustic data to obtain the acoustic feature vector of the robot in the corresponding underwater environment.

[0070] Specifically, the Doppler log uses four tilted beams to calculate the robot's three-dimensional velocity relative to the seabed, with each beam providing raw data such as echo intensity and correlation. Since the raw acoustic data frames are typically output at a high frequency, a sliding time window of equal length to the motion feature vector can be selected to calculate the features within that window, along with the statistics of the acoustic characteristics of these beams, forming an acoustic feature vector. Specifically, the following features are selected: average echo intensity, echo intensity uniformity, average correlation, correlation uniformity, effective beam strength, solution consistency index, beam stability index, and interference index under acoustic multipath effects.

[0071] In practical applications, due to the differences in the emission angles of different beams, the resulting effectiveness varies. Calculating acoustic feature vectors based on the same effectiveness would lead to insufficient accuracy in the calculation. Therefore, in one specific implementation, step S13 includes sub-steps S13-1 to S13-3, as detailed below:

[0072] S13-1: Based on the sliding time window for calculating the motion feature vector, extract the acoustic feature data of multiple beams from the original acoustic data. The acoustic feature data includes the echo intensity, reference intensity, and correlation index of each beam. Data variable names can be constructed based on each type of data, and the corresponding data can be extracted and stored through the data variable names.

[0073] S13-2: For each beam, perform beam quality evaluation and terrain feature impact assessment on the acoustic characteristic data to obtain an effectiveness index for each beam. Since some beams are blocked when propagating in complex seabed environments, and beams reach the receiver via different reflection paths, it is necessary to extract features that reflect the Doppler log measurement quality and environmental acoustics to determine the interference of terrain features on beam effectiveness. The effectiveness index quantifies the effectiveness of different beams.

[0074] The following section will elaborate on the calculation of the effectiveness index for each beam, including:

[0075] The first step is to obtain the intensity-to-quality ratio (I / M ratio) of each beam based on its echo intensity and reference intensity. Taking four beams as an example, let's denote the beam number as i, where i is a positive integer, 1 ≤ i ≤ 4. The echo intensity of beam i is... Reference intensity is The intensity-to-mass ratio of beam i is .

[0076] The second step involves obtaining the quality assessment value for each beam based on its intensity-to-quality ratio and correlation index. This quality assessment value quantifies the characteristic quality of the corresponding beam. Beam quality depends on echo intensity and correlation index. Echo intensity reflects the attenuation along the sound wave propagation path and the seabed's reflection capability, while the correlation index reflects the stability and reliability of the measurement signal. For each beam, the quality can be assessed using the following formula: Obtain the quality evaluation value of beam i. In the formula, This indicates the echo intensity of beam i; Indicates reference strength; This represents the correlation index of beam i. Reference strength. Used to characterize the echo intensity level under normal reference operating conditions, it is essentially an intensity normalized reference quantity; correlation index It is used to quantify the consistency between the received echo and the transmitted signal during the matching process. By performing correlation calculations on the transmitted signal and the received echo, the maximum correlation peak is found to estimate the Doppler frequency shift. The correlation index is a quantitative result of the sharpness, concentration, and reliability of the correlation peak.

[0077] Among them, reference strength The value is the standard echo intensity value of the Doppler log at the factory calibration, or the average echo intensity statistical value under undisturbed conditions in the current water area, which can be set according to the actual scenario.

[0078] The greater the echo intensity and correlation of a beam in the current data frame, the higher its beam quality; conversely, a weaker echo intensity and correlation indicate lower beam quality. The quality rating value characterizes the quality of any beam within a Doppler log data frame, reflecting the degree of quality of the underlying beam based on its echo intensity and correlation.

[0079] The third step involves obtaining the terrain feature impact value for all beams based on the standard deviation and mean rate of change of the quality evaluation values ​​for each beam within the same period. Terrain inhomogeneity can lead to differences in measurement results between different beams. The terrain feature impact value characterizes the degree to which beam ineffectiveness and terrain features affect beam quality. By calculating the differences in echo intensity and correlation between beams, the uniformity of the terrain can be assessed, thus deriving the terrain feature impact value. For example, if the echo intensity and correlation of all beams are very similar, it indicates that the terrain is relatively uniform; conversely, the terrain may be non-uniform. Based on this characteristic, the effectiveness of each beam within the data frame of the original acoustic data for the same period can be quantified, yielding an effectiveness index for each beam.

[0080] It should be noted that when the measurement results of a certain beam differ significantly from those of other beams, it does not necessarily indicate a decrease in the effectiveness of the beam. It may be due to uneven terrain pointing to the surface. Some beams may exhibit different echo characteristics because they are pointing to specific terrain features. This does not necessarily mean that the beam is invalid, but rather reflects the degree of influence of terrain features on the measurement signal.

[0081] Therefore, it is necessary to distinguish between beamouts that are invalid and beam quality that is low due to terrain features. If the beam quality is low due to terrain features, the quality of the beam may change continuously during the robot's movement. For example, as the robot enters a region with rough terrain, the beam quality gradually decreases. However, if it is a case of sudden obstruction or a malfunction that results in low beam quality, the beam quality may drop abruptly. Furthermore, the decrease in beam quality caused by terrain features is usually characterized by a cliff-like difference between beams.

[0082] When determining the influence value of terrain features in a specific frame of the raw acoustic data, the formula can be used: Calculate the terrain feature influence value of the original acoustic data in the current frame, where, The standard deviation of the four beam quality evaluation values ​​in the same frame; The average value representing the rate of change of quality evaluation values; with added coefficients. To prevent calculation errors caused by a zero denominator, you can set... It is 0.01. This indicates that the maximum and minimum values ​​have been normalized to limit the value range to [0,1] and eliminate the influence of dimensions. From the above formula, we can conclude that the smaller the standard deviation of the beam quality evaluation values ​​of a beam within the same frame, the less abrupt the changes in beam quality evaluation values ​​between beams are, and the more likely the differences are caused by corresponding terrain features; furthermore, the smaller the rate of change of beam quality over time, the more likely it is caused by terrain features. The larger the value, the smoother the change in beam quality and the higher the consistency, which is more likely to be caused by continuous terrain features; The smaller the value (closer to 0), the greater the abrupt change or dispersion in beam quality, which is more likely to be caused by beam obstruction or failure.

[0083] The fourth step involves obtaining the effectiveness index of the corresponding beam based on the influence value of terrain features, the quality evaluation value of each beam, and the standard deviation of the evaluation value over a preset time period. The preset time period can be set based on actual needs, such as 5-10 consecutive frames of data preceding the current frame; alternatively, it can be set to other values. On a flat and uniform seabed, the measurement results emitted by the Doppler log beam should be highly consistent. Therefore, beam effectiveness is quantified through inter-beam difference analysis. This analysis must exclude significant inter-beam differences caused by uneven terrain features. Combined with the continuous stability of the beam over time, the effectiveness of the beam is quantified to obtain an accurate effectiveness index. A calculation model can be constructed based on the correlation between the influence value of terrain features, the quality evaluation value, the standard deviation of the evaluation value, and the effectiveness index. The effectiveness index of each beam is derived based on the output of the calculation model.

[0084] For example, the effectiveness index for each beam is calculated based on the following method.

[0085] The first step is to obtain the quality confidence level of each beam based on the ratio of its quality evaluation value to the terrain feature influence value. The quality evaluation value of beam i is... The influence value of terrain features is The quality reliability of beam i is Based on the quality reliability, we can determine the degree to which the beam is trustworthy in the current frame of data.

[0086] The second step involves obtaining the effectiveness index of each beam based on the stability of its quality reliability and the standard deviation of its evaluation value over time. This is achieved using the formula: Calculate the effectiveness index of beam i. In the formula, This represents the quality evaluation value of beam i; This indicates the impact value of terrain features in the current frame; This represents the standard deviation of the beam quality evaluation value of beam i over a consecutive number of frames; (includes coefficients) Similarly, prevent the denominator from being zero. Characterizing the stability and consistency of the beam in the time dimension, by taking the reciprocal in the calculation, stable beams are given higher weight in the effectiveness calculation, while unstable beams are naturally suppressed.

[0087] Based on the above formula, it can be concluded that when the quality of beam i's own signal is higher and it has a higher terrain feature influence value (indicating higher terrain flatness), the confidence level of the current beam quality performance is higher, and a decrease in beam quality at this time can indicate a beam ineffectiveness; furthermore, the higher the continuous stability of beam quality over time, that is... The smaller the value, the higher the beam effectiveness; conversely, the larger the value, the lower the beam effectiveness.

[0088] Furthermore, in complex terrain areas, when sound waves reach the receiver via reflection from the seabed or other obstacles besides the direct path, interference occurs, known as multipath effect. This leads to reduced correlation and deviations in velocity calculation. In cluster analysis, if multipath effect is not identified, acoustic features with the same terrain but different degrees of multipath interference may be misclassified into different clusters, reducing clustering accuracy. Multipath effect typically manifests as reduced correlation and potentially rapid fluctuations in echo intensity. Therefore, in one specific implementation, it is also necessary to calculate the interference index for each beam under acoustic multipath effect.

[0089] The first step is to obtain the degree of intensity change for each beam based on the amount and range of change in echo intensity for each beam. The amount of echo intensity change for beam i is denoted as... This characterizes the change in echo intensity of beam i between the current frame and the previous frame; the range of echo intensity change is... , The data length used to characterize this variation range; the degree of intensity variation of beam i is .

[0090] The second step is to obtain the interference index of the corresponding beam under acoustic multipath effects based on the intensity variation and correlation rate of each beam. This can be done using the formula: The interference index of beam i under acoustic multipath effect was calculated. ; This represents the rate of change of the correlation between beam i in the current frame and the previous frame, i.e., the rate of change of the correlation of beam i. Based on the above formula, it can be seen that when calculating the interference index based on the fluctuation of echo intensity in adjacent original acoustic data frames, a larger fluctuation may indicate multipath interference, and multipath interference is characterized by a decrease in correlation. When the value is negative and the absolute value is larger, the multipath interference index is greater.

[0091] It should be noted that the interference index quantifies the degree of multipath interference experienced by a single beam in the current environment by characterizing the abnormal fluctuation of echo intensity in adjacent frames and the degradation trend of correlation over time. When the echo intensity changes abnormally and is accompanied by a significant decrease in correlation, the value of this index increases significantly, thereby effectively distinguishing multipath interference from acoustic changes caused by normal terrain.

[0092] S13-3: Based on the effectiveness index of each beam and the interference index of acoustic multipath effect, obtain the acoustic feature vector of the robot in the underwater environment corresponding to the terrain.

[0093] The effectiveness index of each beam and the interference index under the acoustic multipath effect can be vectorized and concatenated as independent feature elements to construct an acoustic feature vector. This vector is then used to represent similar terrain features in the subsequent clustering process, placing the same terrain feature into the same cluster. This enables effective clustering of new terrain features, resulting in effective adaptive compensation of the original acoustic data. This ensures that the acoustic feature vector effectively represents the impact of terrain on Doppler log measurements.

[0094] For example, for a Doppler speedometer containing four beams, the four validity indicators can be calculated in order of beam sequence or index category. and four interference indicators The features are sequentially filled into the feature array to construct an acoustic feature vector containing at least 8 dimensions (e.g., ...). This independent splicing method preserves the original distribution information of beam quality and interference level, enabling clustering algorithms to learn the nonlinear relationship between the two from a high-dimensional space.

[0095] At this point, the acoustic feature vector of the robot in the underwater environment corresponding to the terrain has been calculated based on the above calculations, and we proceed to step S14.

[0096] S14: Perform fusion and clustering processing on motion feature vectors and acoustic feature vectors to obtain the correction compensation amount of the Doppler log under different terrain environments.

[0097] Specifically, since similar topographic features have similar seabed sediments or structures, they also have similar acoustic reflection characteristics, that is, similar acoustic feature vectors. Therefore, the data quality of Doppler logs will indicate similar measurement errors. Thus, by correlating and fusing the synchronized motion feature vectors and acoustic feature vectors, and then performing clustering, clusters corresponding to different topographic features can be obtained. Statistical analysis based on the cluster data can then yield the correction compensation amount for each topographic environment.

[0098] For example, step S14 includes sub-steps S14-1 to S14-3, which are described in detail below:

[0099] S14-1: Standardize and weight the motion feature vector and acoustic feature vector to obtain a fused feature vector. Standardize the terrain feature vector and acoustic feature vector separately to eliminate dimensional differences; based on the robot's running timestamp, concatenate the motion feature vector and acoustic feature vector corresponding to the time. The fused features can be weighted according to the effective beam, thus obtaining the fused feature vector.

[0100] S14-2: Cluster all fused feature vectors to establish a correspondence between each cluster and underwater terrain. Using the standardized fused feature vectors as data, perform K-means clustering to divide all samples into several clusters. Each cluster represents a different terrain feature, and each cluster represents a typical terrain-acoustic environment combination. (See also...) Figure 2 After clustering all fused feature vectors in the figure, four clusters are obtained, specifically including flat terrain, gentle slope terrain, steep slope terrain and rough terrain. Each cluster has a corresponding cluster center and decision boundary.

[0101] S14-3: Based on the lateral and forward velocity errors of Doppler logs within the same cluster, obtain the correction compensation for Doppler logs under different terrain conditions. Under different combinations of terrain features, analyze the changes in Doppler log performance indicators to identify terrain feature combination patterns that lead to performance degradation. Within each cluster, calculate the mean of the Doppler log velocity differences for all samples based on the difference between the measured velocity data and the reference velocity for each sample data, thus obtaining the correction compensation for the corresponding terrain measurement error of the Doppler log within that cluster.

[0102] The reference speed refers to a high-precision external reference speed (such as long baseline LBL, ultra-short baseline USBL, or high-precision navigation data processed offline). The reference speed is based on the final speed provided by the final system after integration with INS and correction by other sensors.

[0103] Please see Figure 3 , Figure 3 This is a schematic diagram illustrating the principle of calculating correction compensation for a single type of terrain. Based on all data of this cluster, the mean and standard deviation are calculated, and the final correction compensation for forward velocity error is -0.09 m / s, and the correction compensation for lateral velocity error is -0.038 m / s.

[0104] At this point, the correction compensation amount of the Doppler log under different terrain conditions has been calculated based on the above method, and we proceed to step S15.

[0105] S15: When the robot is performing a task underwater, it performs navigation fusion positioning based on the correction compensation amount for different terrain environments.

[0106] Specifically, in real-time underwater navigation, the robot calculates motion and acoustic feature vectors within the current time window. These vectors are then matched with pre-built cluster centers to find the most similar cluster. Based on the matched clusters, the Doppler log error correction compensation is obtained for the specific terrain environment. Therefore, during navigation and positioning decisions for new terrain features, the Doppler log velocity data is compensated for that specific terrain feature, resulting in accurate and effective position data. For example, the measured Doppler log velocity data within the current time window is supplemented with the Doppler log error correction compensation before being fed into the integrated navigation filter, thereby improving the robot's overall positioning accuracy and robustness in complex terrain.

[0107] Based on the same technical concept as the positioning method, this embodiment of the invention also provides a navigation and positioning system for underwater robots. This navigation and positioning system is the system corresponding to any navigation and positioning method for underwater robots. Please refer to... Figure 4 , Figure 4 This is a schematic diagram of the structure of a navigation and positioning system, which includes an acquisition module 41, a first acquisition module 42, a second acquisition module 43, a third acquisition module 44, and a positioning module 45.

[0108] The acquisition module 41 is used to acquire the robot's inertial navigation data in the underwater environment, as well as the raw acoustic data from the Doppler log installed on the robot.

[0109] The first acquisition module 42 is used to perform operational posture feature analysis on the robot based on inertial navigation data, so as to obtain the motion feature vector of the robot in different underwater terrains.

[0110] The second acquisition module 43 is used to analyze the differences in sound wave feedback generated by the robot in different underwater terrains based on the original acoustic data, so as to obtain the acoustic feature vector of the robot in the underwater environment corresponding to the terrain.

[0111] The third acquisition module 44 is used to perform fusion and clustering processing on motion feature vectors and acoustic feature vectors to obtain the correction compensation amount of the Doppler log under different terrain environments.

[0112] The positioning module 45 is used to perform navigation fusion positioning of the robot based on the correction compensation amount of different terrain environments when the robot is performing tasks underwater.

[0113] Based on the same technical concept as the positioning method, this embodiment of the invention also provides a navigation and positioning device for underwater robots, which operates based on any navigation and positioning method for underwater robots. The navigation and positioning device for underwater robots can be a control terminal suitable for the aforementioned robots, or it can be a robot with an integrated control terminal.

[0114] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0115] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A navigation and positioning method for underwater robots, characterized in that, The method includes: Acquire inertial navigation data of the robot in the underwater environment, as well as raw acoustic data from the Doppler log installed on the robot; Based on the inertial navigation data, the robot's operational posture characteristics are analyzed to obtain the robot's motion feature vectors in different underwater terrains. The differences in acoustic feedback generated by the robot under different underwater terrains are analyzed based on the original acoustic data to obtain the acoustic feature vector of the robot under the corresponding terrain in the underwater environment. The motion feature vector and the acoustic feature vector are fused and clustered to obtain the correction compensation amount of the Doppler log under different terrain environments; When the robot performs a task underwater, it performs navigation fusion positioning based on the correction compensation amount for different terrain environments.

2. The navigation and positioning method for underwater robots according to claim 1, characterized in that, The acquisition of inertial navigation data of the robot in the underwater environment, and raw acoustic data from the Doppler log installed on the robot, includes: When the robot is in an underwater environment, the robot is given an execution command for a target underwater mission, wherein the target underwater mission is a underwater mission in which the robot navigates near the bottom at a constant altitude and constant speed. When the robot performs the target underwater mission, the inertial navigation data and the raw acoustic data are collected synchronously according to a preset frequency.

3. The navigation and positioning method for underwater robots according to claim 1, characterized in that, The step of performing motion posture feature analysis on the robot based on the inertial navigation data to obtain motion feature vectors of the robot in different underwater terrains includes: According to the preset sliding time window, the corresponding data is extracted from the inertial navigation data and input into the preset first processing model; Based on the output of the first processing model, the robot's vertical acceleration standard deviation, vertical acceleration peak value, first absolute average value of pitch angular velocity, second absolute average value of roll angular velocity, cross-correlation coefficient between vertical acceleration and pitch angular velocity, centroid frequency of vertical acceleration power spectrum, first range of pitch angle variation, and second range of roll angle variation are obtained. Underwater terrain analysis is performed based on the vertical acceleration standard deviation, the vertical acceleration peak value, the first absolute value average value, the second absolute value average value, the cross-correlation coefficient, the center of gravity frequency, the first variation range, and the second variation range to obtain the motion feature vector of the robot in the corresponding underwater terrain.

4. The navigation and positioning method for underwater robots according to claim 1, characterized in that, The step of analyzing the differences in acoustic feedback generated by the robot under different underwater terrains based on the original acoustic data to obtain the acoustic feature vector of the robot under the corresponding terrain in the underwater environment includes: Based on the sliding time window used to calculate the motion feature vector, acoustic feature data of multiple beams are extracted from the original acoustic data. The acoustic characteristic data of each beam are used to evaluate beam quality and the impact of terrain features to obtain the effectiveness index of each beam. Based on the effectiveness index of each beam and the interference index of acoustic multipath effect, the acoustic feature vector of the robot in the underwater environment corresponding to the terrain is obtained.

5. The navigation and positioning method for underwater robots according to claim 4, characterized in that, The acoustic feature data includes the echo intensity, reference intensity, and correlation index of each beam; the acoustic feature data of each beam are subjected to beam quality evaluation and terrain feature impact evaluation to obtain the effectiveness index of each beam, including: Based on the echo intensity and reference intensity of each beam, the intensity-to-quality ratio of the corresponding beam is obtained; Based on the intensity-to-quality ratio and correlation index of each beam, the quality evaluation value of the corresponding beam is obtained; The terrain feature influence values ​​of all beams are obtained based on the standard deviation and mean rate of change of the quality evaluation values ​​of each beam in the same period. The effectiveness index of the corresponding beam is obtained based on the influence value of the terrain features, the quality evaluation value of each beam, and the standard deviation of the evaluation value for a preset time period.

6. The navigation and positioning method for underwater robots according to claim 5, characterized in that, Based on the terrain feature impact value, the quality evaluation value of each beam, and the standard deviation of the evaluation values ​​for a preset time period, the effectiveness index of the corresponding beam is obtained, including: The quality reliability of each beam is obtained by comparing the quality evaluation value of each beam with the influence value of the terrain feature. The effectiveness index of the corresponding beam is obtained based on the quality reliability of each beam and the stability of the standard deviation of the evaluation value in the time dimension.

7. The navigation and positioning method for underwater robots according to claim 4, characterized in that, Before obtaining the acoustic feature vector of the robot in the underwater environment corresponding to the terrain, the method further includes: Based on the amount and range of echo intensity change for each beam, the degree of intensity change for the corresponding beam can be obtained. Based on the intensity variation and correlation variation rate of each beam, the interference index of the corresponding beam under the acoustic multipath effect is obtained.

8. The navigation and positioning method for underwater robots according to claim 1, characterized in that, The process of fusing and clustering the motion feature vector and the acoustic feature vector to obtain the correction compensation amount of the Doppler log under different terrain environments includes: The motion feature vector and the acoustic feature vector are standardized and weighted and concatenated to obtain a fused feature vector; Cluster all fused feature vectors to establish a correspondence between each cluster and the underwater terrain; Based on the lateral velocity error and forward velocity error of the Doppler logs in the same cluster, the correction compensation amount of the Doppler logs under different terrain environments is obtained.

9. A navigation and positioning system for underwater robots, characterized in that, The system is the system corresponding to any one of the navigation and positioning methods for underwater robots described in claims 1-8, and the system includes: The acquisition module is used to acquire inertial navigation data of the robot in the underwater environment, as well as raw acoustic data from the Doppler log installed on the robot. The first acquisition module is used to perform operational posture feature analysis on the robot based on the inertial navigation data, so as to obtain the motion feature vector of the robot in different underwater terrains. The second acquisition module is used to analyze the differences in sound wave feedback generated by the robot in different underwater terrains based on the original acoustic data, so as to obtain the acoustic feature vector of the robot in the underwater environment corresponding to the terrain. The third acquisition module is used to perform fusion and clustering processing on the motion feature vector and the acoustic feature vector to obtain the correction compensation amount of the Doppler log under different terrain environments. The positioning module is used to perform navigation fusion positioning of the robot based on the correction compensation amount for different terrain environments when the robot is performing tasks underwater.

10. A navigation and positioning device for underwater robots, characterized in that, The device operates based on any one of the navigation and positioning methods for underwater robots as described in claims 1-8.

Images