Underwater robot positioning method and system for marine engineering

Abstract

The present invention relates to the field of navigation and positioning technology, and proposes an underwater robot positioning method and system for marine engineering, including: obtaining the position parameters and speed parameters of the current task and the historical tasks for the underwater robot; analyzing the distance between the coordinates of the current task and the historical tasks in the same direction, and determining the offset between the current task and the historical tasks at each time point; referring to the offset corresponding to the time point before each time point when the linear velocity change state of the historical tasks and the current task is similar, and determining the relative offset degree between the current task and the historical tasks; judging the influence of the change of coordinates in each direction on the change of the path and obtaining the overall offset degree; judging the overall offset degree, and positioning and path prediction of the underwater robot under the current task by corresponding to the historical tasks. The present invention aims to solve the problem of inaccurate positioning and path prediction of underwater robots due to changes in motion states.

Description

Technical Field

[0001] The present invention relates to the field of navigation and positioning technology, and in particular to an underwater robot positioning method and system for marine engineering. Background Art

[0002] Kalman filter is an efficient recursive filter that can estimate the state of a dynamic system from a series of incomplete and noisy measurements. By accurately predicting the position and posture of the robot, it can ensure that the sampling tool reaches the target position accurately, improve the success rate of sample collection, and thus improve the accuracy of underwater robot positioning and the accuracy of subsequent path prediction.

[0003] The state motion equation in the Kalman filter algorithm predicts the state at the next moment based on certain motion assumptions. The motion patterns of underwater robots under different paths may be very different. If the motion model is confused due to different paths in the data, then when constructing the state motion equation, the time relationship between the state parameter data cannot be more accurately represented, resulting in a decrease in the accuracy of the Kalman filter algorithm prediction. Summary of the invention

[0004] The present invention provides an underwater robot positioning method and system for marine engineering to solve the problem of inaccurate positioning and predicted path positioning of existing underwater robots due to changes in motion state. The technical solutions adopted are as follows:

[0005] The present invention proposes an underwater robot positioning method for marine engineering, which comprises the following steps:

[0006] Obtaining position parameters and speed parameters of the current task and position parameters and speed parameters of several historical tasks for the underwater robot; the position parameters include coordinates in three directions, and the speed parameters include linear speeds in three directions;

[0007] Analyze the distance between the coordinates of the current task and the historical tasks in the same direction, and determine the offset between the current task and the historical tasks at each time point;

[0008] Refer to the offsets corresponding to the time points when the linear speed change states of the historical tasks and the current task are similar before each time point, and use this to accumulate and judge the offsets to determine the relative offset degree between the current task and the historical tasks;

[0009] Determine the influence of the coordinate changes in each direction on the path change, and determine the overall offset degree between the current task and the historical tasks in combination with the relative offset degree;

[0010] The overall deviation degree is determined, and the underwater robot under the current task is positioned and the path is predicted by corresponding historical tasks.

[0011] Optionally, the analysis of the distance between the coordinates of the current task and the historical tasks in the same direction to determine the offset between the current task and each time point of the historical tasks includes the following specific methods:

[0012] The first The historical tasks are taken as the target historical tasks; the time series of the x-direction coordinates of the target historical tasks is , the target historical tasks include time points; the time series of the x-direction coordinates of the current task is , the current task includes time point;

[0013] Current task Time point and target history task The offset in the x direction at each time point The calculation method is:

[0014]

[0015] in, Indicates the current task The x-coordinate of the time point, Indicates the target history task The x-coordinate of a time point.

[0016] Optionally, the offset corresponding to the time point at which the linear speed change state of the historical task and the current task is similar before each time point is referred to, and the offset is accumulated and judged to determine the relative offset degree between the current task and the historical task, including the specific method of:

[0017] According to the offset between the current task and the target historical task at each time point, an offset matrix between the current task and the target historical task is constructed;

[0018] The offsets of the first time point of the current task and the first time point of the target historical task are accumulated respectively to determine the offsets of the current task and the target historical task and the first row element and the first column element of the matrix;

[0019] For the elements other than the first row element and the first column element in the offset and matrix, the other elements in the offset and matrix are determined according to the similarity relationship between the linear speed changes of the target historical task and the current task at the corresponding time points of the determined adjacent elements;

[0020] The offset sum of each element in the offset sum matrix is ​​determined to determine the relative offset degree between the current task and the target historical task.

[0021] Optionally, the specific method of constructing the offset matrix between the current task and the target historical task includes:

[0022] Build the current task and the target historical task An offset matrix; the elements in the same row in the offset matrix are the offsets in the x direction between the same time point of the current task and the different time point of the target historical task, and the elements in the same column in the offset matrix are the offsets in the x direction between the different time points of the current task and the same time point of the target historical task.

[0023] Optionally, the offsets of the first time point of the current task and the first time point of the target historical task are accumulated respectively to determine the offsets of the current task and the target historical task and the first row element and the first column element of the matrix, including the specific method of:

[0024] Based on the offset matrix of the current task and the target historical task in the x direction, construct the offset sum matrix of the current task and the target historical task in the x direction. Elements The calculation method is:

[0025]

[0026] in, Indicates the first time point of the current task and the first time point of the target historical task The offset in the x direction at a time point, Indicates the first The ordinal value of the element;

[0027] For any element in the first column of the offset matrix, the cumulative sum of the element and all elements above it in the first column at the same position in the offset matrix is ​​used as the offset sum corresponding to the element in the offset matrix.

[0028] Optionally, for the elements other than the first row element and the first column element in the offset and matrix, according to the similarity relationship between the linear speed changes of the target historical task and the current task at the corresponding time points of the determined adjacent elements, the other elements in the offset and matrix are determined, including the specific method of:

[0029] Set the element in the second row and second column as the current element; its adjacent elements are the elements in the first row and first column. , the first row and second column element and the second row and first column element , with the current element and Quantify the linear velocity similarity relationship; preset a time window, the linear velocity of the current element in the x direction corresponding to the second time point in the current task, obtain the linear velocity analysis segments of the current element in the current task and the target historical task respectively; obtain the linear velocity analysis segments of the first row and first column element in the current task and the target historical task;

[0030]

[0031] in, Indicates the similarity of the linear speed change between the current element and the first row and first column element in the current task. It represents the mean value of all linear speeds in the linear speed analysis segment of the current task under the analysis of the second row and second column element and the first row and first column element. It represents the mean value of all linear velocities of the first row and first column in the linear velocity analysis segment of the current task. Indicates the number of linear velocities of the first row and first column element in the linear velocity analysis segment of the current task. Indicates that the second row and second column element and the first row and first column element are analyzed in the linear speed analysis section of the current task. Line speed, Indicates that the first row and first column element is the first in the linear velocity analysis section of the current task. Line speed; represents the absolute value function, represents an exponential function with a natural constant as base;

[0032] Obtain the linear velocity change similarity between the current element and the first row and first column element in the target historical task; take the average of the linear velocity change similarity between the current element and the first row and first column element in the current task and the linear velocity change similarity in the target historical task as the linear velocity confidence of the first row and first column element for the current element;

[0033] Analyze the offset sum of adjacent elements and their linear velocity confidence for the current element to determine the offset sum of the current element;

[0034] By analogy, all elements in the second row and the second column of the offset and matrix are obtained, and then all elements in the third row and the third column are obtained. The elements are obtained row by row and column by column to obtain all other elements in the offset and matrix.

[0035] Optionally, the analyzing the offset sum of adjacent elements and their linear velocity confidence with respect to the current element to determine the offset sum of the current element includes the following specific methods:

[0036] The offset reference of the first row and first column element relative to the current element The calculation method is:

[0037]

[0038] in, Indicates the linear velocity confidence of the first row and first column element for the current element. represents the offset sum of the first row and first column elements in the offset sum matrix;

[0039] The offset references of the elements in the second row and first column and the elements in the first row and second column for the current element are respectively obtained, and the sum of the offsets of the element corresponding to the minimum value of the three offset references and the sum of the offsets corresponding to the same position of the current element in the offset matrix are taken as the sum of the offsets of the current element.

[0040] Optionally, the determining the offset sum of the elements in the offset sum matrix to determine the relative offset degree between the current task and the target historical task includes the following specific methods:

[0041] Linearly normalize all elements in the offset and matrix, and obtain the result as the offset rate of each element. Count the number of elements in the offset and matrix whose offset rate is greater than the offset threshold, and use it as the number of offset elements of the offset and matrix. The ratio of the number of offset elements to the total number of elements of the offset and matrix is ​​used as the relative offset degree between the current task and the target historical task in the x direction.

[0042] Optionally, the determining of the influence of the coordinate changes in each direction on the path change and the determination of the overall offset between the current task and the historical task in combination with the relative offset include the following specific methods:

[0043]

[0044] in, Indicates the degree of position change in the target historical task x direction, Represents the variance of the x-direction coordinates of all time points of the target historical task, Indicates the target history task The x-coordinate of the time point, Indicates the target history task The x-coordinate of the time point, Indicates the number of time points of the target historical task, represents the absolute value function;

[0045] The position change degree of the target historical task in each direction is weighted normalized, and the result is used as the position change weight of the target historical task in each direction; the relative offset degree between the current task and the target historical task in each direction is weighted and summed, and the weight adopts the position change weight corresponding to each direction, and the result is used as the overall offset degree between the current task and the target historical task.

[0046] The present invention also proposes an underwater robot positioning system for marine engineering, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the above method when executing the computer program.

[0047] The beneficial effects of the present invention are as follows: based on the DTW matching concept, the present invention performs deviation analysis on the coordinates of the current task and the historical tasks in the same direction, constructs an offset matrix, and based on the similarity relationship of the speed change at the corresponding time point, accumulates the offset to construct the offset and matrix, thereby ensuring that the similarity of the motion state of the current task and the historical tasks of the underwater robot in complex underwater environments and different operating conditions can be quantified more accurately; on the basis of the relative offset degree in each direction between the current task and the historical tasks, the characteristic performance of the coordinate change in each direction for the task path is considered, and this is used as a weight to perform an overall analysis of the relative offset degree to obtain the overall offset degree; similar historical tasks are screened by the overall offset degree, and the linear velocity of the underwater robot in three directions of the current task is used as a reference to construct the state motion equation and perform positioning and subsequent path prediction, thereby improving the positioning accuracy of the underwater robot in complex environments and motion states. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0049] Figure 1 A schematic flow chart of an underwater robot positioning method for marine engineering provided by one embodiment of the present invention;

[0050] Figure 2 Construct schematics for offsets and matrices. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] See also Figure 1, which shows a flow chart of an underwater robot positioning method for marine engineering provided by an embodiment of the present invention, the method comprising the following steps:

[0053] Step S001: obtaining position parameters and speed parameters of the current task, as well as position parameters and speed parameters of several historical tasks, for the underwater robot.

[0054] The purpose of this embodiment is to compare the motion patterns of the underwater robot through historical tasks and the current task, obtain similar historical tasks and use them to locate the current task of the underwater robot and predict and plan the subsequent path. First, it is necessary to measure the motion state of the underwater robot through an inertial navigation unit, and record the position of the underwater robot in real time, where the position is represented by three-dimensional coordinates.

[0055] Specifically, for the current task performed by the underwater robot, the position parameters and speed parameters of the underwater robot are obtained through the inertial navigation unit (IMU), wherein the position parameters include coordinates in three directions of x, y and z, and the speed parameters include linear velocities in three directions of x, y and z. In this embodiment, the sampling time interval for obtaining speed parameters and position parameters through the inertial navigation unit is set to 1 second, so that the speed parameters and position parameters of several time points are obtained for the current task; wherein the coordinates in three directions are directly obtained through the inertial navigation unit, and the linear velocities in three directions are obtained through the accelerometer in the inertial navigation unit.

[0056] Similarly, in several historical tasks performed by the underwater robot, the inertial navigation unit also records the position parameters and speed parameters in real time, and the sampling time interval is the same, so the position parameters and speed parameters of several historical tasks are obtained; among which this embodiment obtains the position parameters and speed parameters of the underwater robot's most recent 50 historical tasks.

[0057] It should be noted that the path of the underwater robot is the spatial manifestation of its mission execution process; the path includes the movement trajectory of the robot from the starting point to the current position, which involves various operations performed by the robot to complete the task. Therefore, the path (position) of the underwater robot reflects its specific action sequence and spatial position arrangement in the process of performing resource exploration tasks.

[0058] Step S002: Analyze the distance between the coordinates of the current task and the historical tasks in the same direction, and determine the offset between the current task and each time point of the historical tasks.

[0059] It should be noted that in the process of predicting and planning the subsequent path of the underwater robot for the current task through similar historical tasks, it is necessary to obtain historical tasks that are similar to the motion mode and motion state of the current task. The similarity between the historical tasks and the current task can be intuitively reflected through the deviation in position. Therefore, the difference analysis of the coordinates in the three directions at each time point is performed on the current task and the historical tasks respectively. The position offset of the current task and the historical tasks is reflected as a whole through the respective coordinate offsets in the three directions, avoiding the complexity of directly performing the overall position offset analysis process.

[0060] It should be further explained that for the path change of the underwater robot in one direction, the time series can be formed by the coordinates of the time-series changes in this direction. Since the historical task and the current task may have differences in task execution time, in the process of similarity analysis of the coordinate time series of the historical task and the current task, it is necessary to analyze based on the DTW matching idea, and it is necessary to perform difference analysis on the coordinates of each time point of the historical task and each time point of the current task in the same direction, so as to construct the subsequent offset matrix.

[0061] Specifically, this embodiment takes the x direction as an example for analysis. The historical tasks are taken as the target historical tasks, and the time series of the x-direction coordinates of the target historical tasks is: , the target historical tasks include time points; the time series of the x-direction coordinates of the current task is , the current task includes time point; then the current task is Time point and target history task The offset in the x direction at each time point The calculation method is:

[0062]

[0063] in, Indicates the current task The x-coordinate of the time point, Indicates the target history task The x-direction coordinate of a time point; according to the above method, obtain the offset in the x-direction between the current task and any two time points of the target historical task.

[0064] Step S003, referring to the offsets corresponding to the time points before each time point when the linear speed change states of the historical tasks and the current task are similar, the offsets are accumulated and judged to determine the relative offset degree between the current task and the historical tasks.

[0065] It should be noted that during the operation and execution of tasks, the underwater robot is affected by the urgency of the task and the impression of the underwater environment. On the basis of the offset reflecting the path similarity, if the offset is small but there is a difference in speed in the same direction, the offset cannot fully reflect the similarity relationship on the path. It is necessary to accumulate the offsets of the corresponding time point combinations on the basis of considering the similarity of speed, so as to quantify the offset sum. That is, the similarity of the speed change relationship between the time point combinations is, the greater the possibility of subsequent similar paths. In this way, the offset sum matrix is ​​constructed, and the influence of offset accumulation on path similarity analysis is considered to quantify the relative offset degree between the current task and the historical tasks.

[0066] Preferably, in one embodiment of the present invention, the step includes:

[0067] According to the offset between the current task and the target historical task at each time point, an offset matrix between the current task and the target historical task is constructed;

[0068] The offsets of the first time point of the current task and the first time point of the target historical task are accumulated respectively to determine the offsets of the current task and the target historical task and the first row element and the first column element of the matrix;

[0069] For the elements other than the first row element and the first column element in the offset and matrix, the other elements in the offset and matrix are determined according to the similarity relationship between the linear speed changes of the historical task and the current task at the corresponding time points of the determined adjacent elements;

[0070] The offset sum of each element in the offset sum matrix is ​​determined to determine the relative offset degree between the current task and the target historical task.

[0071] As an example, in an embodiment of the present invention, an offset matrix between the current task and the target historical task is constructed according to the offset between each time point of the current task and the target historical task, including:

[0072] Take the target history task as an example, that is, Historical tasks, since the target historical tasks include time points, the current task includes time point, the current task and the target historical task are constructed The offset matrix of the current task and the target historical task is obtained by the elements in the same row in the offset matrix, wherein the elements in the same column in the offset matrix are the offsets in the x direction between the same time point of the current task and the different time point of the target historical task, and the elements in the same column in the offset matrix are the offsets in the x direction between the different time points of the current task and the same time points of the target historical task.

[0073] What needs to be explained is that, by constructing an offset matrix based on the offset of the coordinates in the same direction corresponding to the time point for the current task and the historical tasks, from the idea of ​​the DTW matching algorithm, it is to find the DTW matching matrix corresponding to the shortest path in the DTW matching process, and at the same time provide a basis for the subsequent construction of offsets and matrix analysis of historical tasks and current tasks.

[0074] It should be further explained that, in the process of offset and matrix construction, elements at other positions need to use the DTW matching idea to obtain the combination of time points with the most similar speed change relationship. For the elements in the first row and the first column, since one of the time points is fixed as the first time point of the corresponding task, the first row element or the first column element directly accumulates the offsets of the elements in the same row or column. That is, the first time point is fixed, and the previous combination of time points with similar speed change relationship is the combination corresponding to the change of another time point.

[0075] As an example, in an embodiment of the present invention, the offsets of the first time point of the current task and the first time point of the target historical task are accumulated respectively to determine the offsets of the current task and the target historical task and the first row element and the first column element of the matrix, including:

[0076] Based on the offset matrix of the current task and the target historical task in the x direction, construct the offset sum matrix of the current task and the target historical task in the x direction. Elements The calculation method is:

[0077]

[0078] in, Indicates the first time point of the current task and the first time point of the target historical task The offset in the x direction at a time point, Indicates the first The ordinal value of the element.

[0079] Similarly, for any element in the first column of the offset and matrix, the cumulative sum of the element and all the elements above it in the first column at the same position in the offset matrix is ​​used as the offset sum corresponding to the element in the offset and matrix; the first row element and the first column element of the offset and matrix are determined.

[0080] As an example, in an embodiment of the present invention, for the elements other than the first row element and the first column element in the offset and matrix, according to the similarity relationship between the linear speed changes of the target historical task and the current task at the corresponding time points of the determined adjacent elements, the other elements in the offset and matrix are determined, including:

[0081] It should be noted that in the process of obtaining the offset and other elements in the matrix by accumulating the offset, it is actually a dynamically adjusted accumulation process to find the optimal alignment between the corresponding paths of the two tasks. It is necessary to accumulate the offsets of adjacent elements (time point combinations) with similar linear velocity change relationships, that is, the time point combinations corresponding to the two elements have similar linear velocity changes within a time window of the current task, and similar linear velocity changes within a time window of the historical tasks. In this way, the confidence of the adjacent elements is quantified, and the offset and other elements in the matrix are determined by accumulation combined with the offset of the corresponding adjacent elements.

[0082] It is further necessary to explain that if Figure 2 As shown, it shows a schematic diagram of the offset and matrix construction without considering the similarity relationship of the linear velocity change. The first row elements and the first column elements are all elements in the offset matrix. The left side of the other position elements are elements in the offset matrix, and the right side are elements in the offset and matrix. Figure 2 The offset and matrix directly refer to the minimum offset value in the adjacent elements of the three angles for accumulation. On this basis, it is necessary to analyze the similarity relationship of the linear speed change.

[0083] Specifically, after the offset and the first row element and the first column element of the matrix are determined, the second row and second column element is taken as an example, and the second row and second column element is taken as the current element; its adjacent elements are the first row and first column elements , the first row and second column element and the second row and first column element , with the current element and To quantify the linear velocity similarity relationship, a time window is preset. In this embodiment, the length of the time window is described by taking 5 as an example. The linear velocity of the current element in the x direction at the second time point corresponding to the current task is obtained using the time window. The linear velocity analysis segment of the current element in the current task is obtained, that is, the linear velocity corresponding to the current element is taken as the center, and the linear velocities of the first two adjacent time points and the last two adjacent time points are obtained respectively to form a linear velocity analysis segment. Similarly, the linear velocity analysis segment of the current element in the target historical task is obtained. For the first row and first column element , obtain the linear speed analysis segment in the current task and the target historical task; it is particularly noted that for the first time point in the current task, it cannot completely constitute a linear speed analysis segment, that is, it lacks the linear speed of the adjacent previous time point, then the linear speed corresponding to the time point actually existing in the time window is used to constitute the linear speed analysis segment, and in the similarity analysis of the linear speed change relationship between adjacent elements, it is necessary to ensure that the length of the linear speed analysis segment is equal, then the corresponding linear speed analysis segment is obtained with the minimum length of the linear speed analysis segment, and at the same time, it is ensured that the order values ​​of the linear speeds corresponding to the analyzed adjacent elements in the linear speed analysis segment are equal, for example, the length of the linear speed analysis segment of the first row and first column element in the current task is 3, and the order value of the linear speed corresponding to the element to be analyzed is 1, then the first element of the linear speed analysis segment of the current element in the current task is removed to ensure that the length of the linear speed analysis segment is equal and the order value of the linear speed corresponding to the current element is 1.

[0084] Furthermore, a similarity analysis is performed on the linear velocity analysis segment of the current element and the first row and first column element in the current task to obtain the similarity of the linear velocity change between the current element and the first row and first column element in the current task. , the specific calculation method is:

[0085]

[0086] in, It represents the average value of all linear velocities in the linear velocity analysis segment of the current task under the analysis of the second row and second column element and the first row and first column element (all linear velocities in the x direction). It represents the mean value of all linear velocities of the first row and first column in the linear velocity analysis segment of the current task. Indicates the number of linear velocities of the first row and first column element in the linear velocity analysis segment of the current task. Indicates that the second row and second column element and the first row and first column element are analyzed in the linear speed analysis section of the current task. Line speed, Indicates that the first row and first column element is the first in the linear velocity analysis section of the current task. Line speed; represents the absolute value function, represents an exponential function with a natural constant as the base. In this embodiment, Model to present inverse proportional relationship and normalization, As the input of the model, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0087] It should be noted that in the process of similarity analysis of linear speed change relationships, linear speed analysis segments are constructed to quantify the similarity relationship of linear speed changes of adjacent elements in the current task. The smaller the overall linear speed mean difference in the linear speed analysis segment and the smaller the linear speed difference under the same order value, the more similar the overall change relationship is and the greater the similarity of linear speed changes is.

[0088] Furthermore, a similarity analysis is performed on the linear speed analysis segment of the current element and the first row and first column element in the target historical task to obtain the similarity of the linear speed change between the current element and the first row and first column element in the target historical task. , the specific calculation method is:

[0089]

[0090] in, It represents the mean value of all linear speeds in the linear speed analysis segment of the target historical task under the analysis of the second row and second column elements and the first row and first column elements. It represents the mean value of all linear speeds of the first row and first column elements in the linear speed analysis segment of the target historical task. Indicates the number of linear velocities of the first row and first column element in the linear velocity analysis segment of the target historical task. Indicates the second row and second column element in the linear speed analysis section of the target history task under the analysis of the first row and first column element. Line speed, Indicates that the first row and first column element is the first in the linear speed analysis segment of the target historical task. Line speed; represents the absolute value function, represents an exponential function with a natural constant as the base. In this embodiment, Model to present inverse proportional relationship and normalization processing, As the input of the model, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0091] Furthermore, the average of the linear speed change similarity of the current element and the first row and first column element in the current task and the linear speed change similarity of the target historical task is taken as the linear speed confidence of the first row and first column element for the current element; the linear speed confidence of the second row and first column element and the first row and second column element for the current element are obtained respectively. It should be noted that if the adjacent elements are in the same line or column, the linear speed similarity of the same line is not calculated in the current task, and the linear speed similarity of the target historical task is directly used as the linear speed confidence; the linear speed similarity of the same column is not calculated in the target historical task, and the linear speed similarity of the current task is directly used as the linear speed confidence; then the offset reference of the first row and first column element for the current element is The calculation method is:

[0092]

[0093] in, Indicates the linear velocity confidence of the first row and first column element for the current element (linear velocity in the x direction). Represents the offset sum of the first row and first column elements in the offset sum matrix.

[0094] It should be noted that the offset reference is the linear velocity confidence introduced on the basis of the offset sum of the corresponding adjacent elements. To make the offset reference smaller, it is necessary to ensure that the offset sum of the adjacent elements is smaller and the linear velocity confidence is larger. The offset reference is obtained by the above formula.

[0095] Furthermore, the offset references of the elements in the second row and the first column and the elements in the first row and the second column for the current element are obtained respectively, and the offset sum of the element corresponding to the minimum value of the three offset references and the sum of the offsets corresponding to the same position of the current element in the offset matrix are taken as the offset sum of the current element; and so on, all the elements in the second row and the second column of the offset sum matrix are obtained, and then all the elements in the third row and the third column are obtained, and the elements are obtained row by row and column by column to obtain all the other elements in the offset sum matrix, and combined with the elements in the first row and the first column, the construction of the offset sum matrix between the current task and the target historical task in the x direction is completed.

[0096] As an example, in an embodiment of the present invention, determining the offset sum of each element in the offset sum matrix to determine the relative offset degree between the current task and the target historical task includes:

[0097] Linear normalization is performed on all elements in the offset and matrix, and the result is used as the offset rate of each element. A offset threshold is preset. In this embodiment, the offset threshold is described as 0.5. The number of elements in the offset and matrix whose offset rate is greater than the offset threshold is counted as the number of offset elements of the offset and matrix. The ratio of the number of offset elements to the total number of elements in the offset and matrix is ​​used as the relative offset degree between the current task and the target historical task in the x direction, that is, the relative offset degree between the current task and the target historical task in the x direction. The relative offset of the historical tasks in the x direction.

[0098] It should be noted that after constructing the offset and matrix through the DTW matching idea, the offset and matrix can overall reflect the path offset between the current task and the historical tasks. The more elements that exceed the threshold, excluding the influence of accumulation, the greater the path deviation between the current task and the historical tasks, and the lower the similarity.

[0099] At this point, based on the DTW matching idea, the deviation analysis of the current task and the historical tasks in the same direction of coordinates is carried out to construct an offset matrix, and based on the similarity relationship of the speed change at the corresponding time point, the offset is accumulated to construct the offset and matrix, ensuring that the similarity of the motion state of the current task and the historical tasks of the underwater robot in complex underwater environments and different operating conditions can be quantified more accurately.

[0100] Step S004: determine the influence of the coordinate changes in each direction on the path change, and determine the overall offset degree between the current task and the historical tasks in combination with the relative offset degree.

[0101] It should be noted that after calculating the relative offset degree for the three directions respectively, the relative offset degree reflects the path similarity between the current task and the historical task in a single direction. For the analysis of the overall offset degree, it is necessary to quantify the impact of the path change in each direction on the overall path based on the relative offset degree of each direction. The more frequent the coordinate changes in a single direction, the greater the amplitude of the changes, and the larger the range distribution of the overall coordinates, the more it can reflect the path characteristics of the corresponding task. This is necessary to use this as a weight for an overall analysis of the relative offset degrees in the three directions.

[0102] Specifically, Take the target historical task as an example, the position change degree of the target historical task in the x direction The calculation method is:

[0103]

[0104] in, Represents the variance of the x-direction coordinates of all time points of the target historical task, Indicates the target history task The x-coordinate of the time point, Indicates the target history task The x-coordinate of the time point, Indicates the number of time points of the target historical task, represents the absolute value function.

[0105] It should be noted that the larger the variance of the coordinates in the same direction, the larger the fluctuation range of the coordinates in the corresponding direction. At the same time, the larger the difference between the coordinates of adjacent time points, the more frequent and large the changes in the coordinates in the corresponding direction. The greater the degree of position change in the corresponding direction, the more obvious the path characteristics of the corresponding task.

[0106] Furthermore, the relative offset degree between the current task and the target historical task in each direction is obtained according to the above method, and the position change degree of the target historical task in each direction is obtained, and all position change degrees are softmax normalized (weight normalization function), and the result is used as the position change weight of the target historical task in each direction; the relative offset degree between the current task and the target historical task in each direction is weighted and summed, and the weight adopts the position change weight corresponding to each direction, and the result is used as the overall offset degree between the current task and the target historical task.

[0107] At this point, based on the relative offset degree between the current task and the historical tasks in each direction, the characteristic performance of the coordinate changes in each direction for the task path is considered, and this is used as a weight to perform an overall analysis of the relative offset degree to obtain the overall offset degree.

[0108] Step S005: determine the overall deviation degree, and perform positioning and path prediction for the underwater robot under the current task by corresponding historical tasks.

[0109] Specifically, a preset offset threshold is used. In this embodiment, the offset threshold is described as 0.4. If the overall offset between the current task and the target historical task is less than the offset threshold, the target offset task is used as a similar historical task of the current task; and several similar historical tasks of the current task are obtained.

[0110] It should be noted that in order to predict the motion state and subsequent path of the underwater robot when performing the current task, the Kalman filter algorithm is usually used for prediction. Since the overall path of similar historical tasks is similar to the current task, the state motion matrix is ​​constructed based on this to complete the positioning and path prediction.

[0111] Furthermore, a position matrix is ​​constructed by the coordinates of the underwater robot in the three directions of the current task, and for the linear speed in the three directions, the influence of similar historical tasks needs to be considered; taking the linear speed in the x direction as an example, the result obtained by subtracting the overall offset degree corresponding to each similar historical task from 1 and normalizing the weights is used as the reference weight of each similar historical task, and the linear speed in the x direction of each similar historical task at the last time point is weighted and summed with the reference weight, and the linear speed in the x direction of the current task at the current moment is updated based on this, so as to construct the state motion equation for the linear speed in each direction of the underwater robot of the current task, and then obtain the state motion matrix, and combine the covariance matrix to predict the motion state of the underwater robot at the next time point of the current task, and then use the Kalman gain to update the prior prediction, so as to achieve subsequent positioning and path prediction of the underwater robot; the above process is the existing technology in the Kalman filter algorithm, and will not be repeated in this embodiment.

[0112] At this point, similar historical tasks are screened by the overall deviation degree, and the linear velocity of the underwater robot in three directions of the current task is used as a reference to construct the state motion equation and perform positioning and subsequent path prediction, so as to improve the positioning accuracy of the underwater robot in complex environments and motion states.

[0113] Another embodiment of the present invention provides an underwater robot positioning system for marine engineering, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, the above method steps S001 to S005 are implemented.

[0114] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An underwater robot positioning method for marine engineering, characterized in that: The method comprises the following steps: Obtaining position parameters and speed parameters of the current task, as well as position parameters and speed parameters of several historical tasks, for the underwater robot; the position parameters include coordinates in three directions, and the speed parameters include linear speeds in three directions; Analyze the distance between the coordinates of the current task and the historical tasks in the same direction, and determine the offset between the current task and the historical tasks at each time point; Refer to the offsets corresponding to the time points when the linear speed change states of the historical tasks and the current task are similar before each time point, and use this to accumulate and judge the offsets to determine the relative offset degree between the current task and the historical tasks; Determine the influence of the coordinate changes in each direction on the path change, and determine the overall offset degree between the current task and the historical tasks in combination with the relative offset degree; Determine the overall deviation degree, and locate and predict the path of the underwater robot under the current task by corresponding historical tasks; The specific method of analyzing the distance between the coordinates of the current task and the historical tasks in the same direction and determining the offset between the current task and each time point of the historical tasks includes: The first The historical tasks are taken as the target historical tasks; the time series of the x-direction coordinates of the target historical tasks is , the target historical tasks include time points; the time series of the x-direction coordinates of the current task is , the current task includes time point; Current task Time point and target history task The offset in the x direction at each time point The calculation method is: in, Indicates the current task The x-coordinate of the time point, Indicates the target history task The x-direction coordinate of each time point; The offset corresponding to the time point at which the linear speed change state of the historical task and the current task is similar before each time point is referred to, and the offset is accumulated and judged to determine the relative offset degree between the current task and the historical task, including the specific method as follows: According to the offset between the current task and the target historical task at each time point, an offset matrix between the current task and the target historical task is constructed; The offsets of the first time point of the current task and the first time point of the target historical task are accumulated respectively to determine the offsets of the current task and the target historical task and the first row element and the first column element of the matrix; For the elements other than the first row element and the first column element in the offset and matrix, the other elements in the offset and matrix are determined according to the similarity relationship between the linear speed changes of the target historical task and the current task at the corresponding time points of the determined adjacent elements; The offset sum of each element in the offset sum matrix is ​​determined to determine the relative offset degree between the current task and the target historical task.

2. The underwater robot positioning method for marine engineering according to claim 1, characterized in that: The specific method of constructing the offset matrix between the current task and the target historical task includes: Build the current task and the target historical task An offset matrix; the elements in the same row in the offset matrix are the offsets in the x direction between the same time point of the current task and the different time point of the target historical task, and the elements in the same column in the offset matrix are the offsets in the x direction between the different time points of the current task and the same time point of the target historical task.

3. The underwater robot positioning method for marine engineering according to claim 1, characterized in that: The offsets of the first time point of the current task and the first time point of the target historical task are accumulated respectively to determine the offsets of the current task and the target historical task and the first row element and the first column element of the matrix, including the specific method of: Based on the offset matrix of the current task and the target historical task in the x direction, construct the offset sum matrix of the current task and the target historical task in the x direction. Elements The calculation method is: in, Indicates the first time point of the current task and the first time point of the target historical task The offset in the x direction at a time point, Indicates the first The ordinal value of the element; For any element in the first column of the offset matrix, the cumulative sum of the element and all elements above it in the first column at the same position in the offset matrix is ​​used as the offset sum corresponding to the element in the offset matrix.

4. The underwater robot positioning method for marine engineering according to claim 1, characterized in that: The specific method of determining the other elements in the offset and matrix except the first row element and the first column element according to the similarity relationship between the linear speed changes of the target historical task and the current task at the corresponding time points of the determined adjacent elements is as follows: Set the second row and second column element as the current element; its adjacent elements are the first row and first column elements , the first row and second column element and the second row and first column element , with the current element and Quantify the linear velocity similarity relationship; Preset a time window, the linear velocity of the current element in the x direction at the second time point corresponding to the current task, obtain the linear velocity analysis segments of the current element in the current task and the target historical task respectively; obtain the linear velocity analysis segments of the first row and first column element in the current task and the target historical task; in, Indicates the similarity of the linear speed change between the current element and the first row and first column element in the current task. It represents the average value of all linear speeds in the linear speed analysis segment of the current task under the analysis of the second row and second column element and the first row and first column element. It represents the mean value of all linear velocities of the first row and first column in the linear velocity analysis segment of the current task. Indicates the number of linear velocities of the first row and first column element in the linear velocity analysis segment of the current task. Indicates the analysis of the second row and second column element and the first row and first column element, the linear speed analysis segment of the current task Line speed, Indicates that the first row and first column element is the first in the linear velocity analysis section of the current task. Line speed; represents the absolute value function, represents an exponential function with a natural constant as base; Obtain the linear velocity change similarity between the current element and the first row and first column element in the target historical task; take the average of the linear velocity change similarity between the current element and the first row and first column element in the current task and the linear velocity change similarity in the target historical task as the linear velocity confidence of the first row and first column element for the current element; Analyze the offset sum of adjacent elements and their linear velocity confidence for the current element to determine the offset sum of the current element; By analogy, all elements in the second row and the second column of the offset and matrix are obtained, and then all elements in the third row and the third column are obtained. The elements are obtained row by row and column by column to obtain all other elements in the offset and matrix.

5. The underwater robot positioning method for marine engineering according to claim 4, characterized in that: The method of analyzing the offset sum of adjacent elements and their confidence in the linear velocity of the current element to determine the offset sum of the current element includes: The offset reference of the first row and first column element relative to the current element The calculation method is: in, Indicates the linear velocity confidence of the first row and first column element for the current element. represents the offset sum of the first row and first column elements in the offset sum matrix; The offset references of the elements in the second row and first column and the elements in the first row and second column for the current element are respectively obtained, and the sum of the offsets of the element corresponding to the minimum value of the three offset references and the sum of the offsets corresponding to the same position of the current element in the offset matrix are taken as the sum of the offsets of the current element.

6. The underwater robot positioning method for marine engineering according to claim 1, characterized in that: The method of determining the offset sum of the elements in the offset sum matrix and determining the relative offset degree between the current task and the target historical task includes: Linearly normalize all elements in the offset and matrix, and obtain the result as the offset rate of each element. Count the number of elements in the offset and matrix whose offset rate is greater than the offset threshold, and use it as the number of offset elements of the offset and matrix. The ratio of the number of offset elements to the total number of elements of the offset and matrix is ​​used as the relative offset degree between the current task and the target historical task in the x direction.

7. The underwater robot positioning method for marine engineering according to claim 1, characterized in that: The method of determining the influence of the coordinate changes in each direction on the path change and combining the relative offset degree to determine the overall offset degree between the current task and the historical task includes: in, Indicates the degree of position change in the x direction of the target historical task, Represents the variance of the x-direction coordinates of all time points of the target historical task, Indicates the target history task The x-coordinate of the time point, Indicates the target history task The x-coordinate of the time point, Indicates the number of time points of the target historical task, represents the absolute value function; The position change degree of the target historical task in each direction is weighted normalized, and the result is used as the position change weight of the target historical task in each direction; the relative offset degree between the current task and the target historical task in each direction is weighted and summed, and the weight adopts the position change weight corresponding to each direction, and the result is used as the overall offset degree between the current task and the target historical task.

8. An underwater robot positioning system for marine engineering, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the underwater robot positioning method for marine engineering as described in any one of claims 1-7 are implemented.

Images