Underwater robot collision avoidance control method based on multi-sensor data fusion, electronic equipment and storage medium

Abstract

The invention belongs to the technical field of underwater robot collision avoidance control, and particularly discloses an underwater robot collision avoidance control method based on multi-sensor data fusion, electronic equipment and a storage medium, and the method comprises the steps: S1, carrying out the preprocessing based on the data obtained by an optical camera and through combining with an image enhancement technology, and obtaining an image data set; s2, processing in a data preprocessing stage based on data acquired by sonar and a sonar signal processing algorithm to obtain a distance data set; s3, based on the data acquired by the pipeline instrument, performing data preprocessing to obtain a pipeline data set; s4, performing data fusion processing according to the obtained image data set, distance data set and pipeline data set; and S5, when it is detected that an obstacle exists in front, intelligent decision making is carried out according to a preset collision avoidance control strategy, and the robot is guided to bypass the obstacle.

Description

Technical Field

[0001] The present invention belongs to the technical field of underwater robot collision avoidance control, and in particular relates to an underwater robot collision avoidance control method based on multi-sensor data fusion, electronic equipment and storage medium. Background Art

[0002] When performing tasks in underwater environments, robots face various complex obstacles and environmental factors. Therefore, collision avoidance control of robots is crucial. However, due to the lack of light, special sound wave propagation characteristics, and magnetic field interference in the underwater environment, a single sensor is difficult to provide reliable environmental perception and collision avoidance control.

[0003] Existing collision avoidance control methods for underwater robots mainly include collision avoidance technologies based on sonar and optical cameras and collision avoidance algorithms based on machine learning; among them, sonar technology relies on the propagation of sound waves and can detect obstacles in turbid water bodies, but its resolution is low; optical camera technology can provide high-resolution images, but it is not effective in turbid water bodies; electromagnetic data obtained based on pipeline meters can determine the position and status of the robot by detecting the electromagnetic signals of underground metal pipelines, but it cannot provide comprehensive environmental perception information.

[0004] In addition, the collision avoidance control method based on a single sensor has problems with insufficient accuracy and reliability, cannot effectively identify and track dynamic objects, and has poor adaptability in different underwater environments. Summary of the invention

[0005] Based on the above problems, the present invention provides an underwater robot collision avoidance control method, electronic device and storage medium based on multi-sensor data fusion, which can perceive more reliable and accurate environmental information.

[0006] Based on the above purpose, the present invention is achieved through the following technical solutions:

[0007] A first aspect of the present invention provides an underwater robot collision avoidance control method based on multi-sensor data fusion, comprising the following steps:

[0008] S1. Preprocessing the data acquired by the optical camera in combination with image enhancement technology to obtain an image data set.

[0009] S2. Based on the data acquired by sonar, the data is processed in the data preprocessing stage based on the sonar signal processing algorithm to obtain a distance data set.

[0010] S3. Based on the data acquired by the pipeline instrument, data preprocessing is performed to obtain a pipeline data set.

[0011] S4. Perform data fusion processing based on the obtained image data set, distance data set and pipeline data set.

[0012] S5. When an obstacle is detected ahead, an intelligent decision is made based on the preset collision avoidance control strategy to guide the robot around the obstacle.

[0013] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, the specific steps of step S1 include: firstly, using an optical camera to collect original image data of the underwater environment, and then using dark channel prior image enhancement technology to pre-process the original image data, and the pre-processing includes contrast adjustment, noise removal and detail enhancement of the original image data to obtain an optimized image data set.

[0014] The expression for calculating the dark channel image using image enhancement technology is:

[0015]

[0016] Among them, J dark (x) represents the dark channel image; P(x) represents a local window of pixel x; I c Represents the pixel value of color channel c.

[0017] The dark channel image J is obtained by calculation dark (x), select the highest brightness area to estimate the atmospheric light value A.

[0018] Calculate the relevant parameter transmittance t(x), the expression is:

[0019]

[0020] Among them, ω represents the adjustment parameter and its value is 0.95.

[0021] The scene radiation value J(x) is restored using the estimated atmospheric light value A and transmittance t(x). The expression of J(x) is:

[0022]

[0023] Where I(x) represents the pixel value of color channel c; t 0 Represents a constant that prevents the denominator from being zero.

[0024] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, the specific steps of step S2 include: collecting sound wave data in the underwater environment through sonar, and preprocessing the sound wave data based on the sonar signal processing algorithm, which preprocessing includes signal bandpass filtering, matched filtering and distance measurement to obtain an accurate distance data set.

[0025] Use a bandpass filter to filter the sonar signal to remove low-frequency noise and high-frequency interference; the expression of the bandpass filter is:

[0026]

[0027] Among them, y(t) represents the filtered signal; x(t) represents the original signal; h(t) represents the impulse response of the filter; τ represents the time delay or offset.

[0028] By matching the known pulse signal with the target signal, the target signal is enhanced. The expression of the matched filter is:

[0029]

[0030] Where z(t) represents the filtered signal; h m (t) represents the impulse response of the matched filter.

[0031] Use the constant false alarm rate algorithm to detect the echo signal and calculate the distance of the target; the constant false alarm rate algorithm is the CFAR algorithm, and its expression is:

[0032]

[0033] Among them, N represents the window size; T represents the detection threshold; window represents the sliding window; x(i) represents the i-th x signal.

[0034] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, the specific steps of step S3 include: using a pipeline instrument to collect electromagnetic data of the underwater pipeline, using wavelet transform to denoise the electromagnetic signal, and obtaining a pipeline data set.

[0035] De-noising is used to remove high-frequency noise and interference. Its denoising expression is:

[0036]

[0037] Where x(t) represents the denoised signal; c k represents the wavelet coefficient; ψ k (t) represents the wavelet basis function.

[0038] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, the specific steps of step S4 include: aligning and fusing the correlation data between the correlations between the data acquired by the optical camera, sonar and pipeline meter through multi-sensor data fusion technology; and realizing effective identification of robot-related parameters through multi-sensor data fusion.

[0039] The measurement value is determined by using the information of the target at the previous moment to predict the data at the current moment, and comparing the predicted result with the current detection result.

[0040] First, Kalman filtering is used to predict the robot's state parameters at the next moment, and the selected state X(k) is expressed as:

[0041]

[0042] Among them, x(k) represents the relative position in the x direction; v x (k) represents the velocity in the x direction; y(k) represents the relative position in the y direction; v y (k) represents the velocity in the y direction; a x (k) represents the acceleration in the x direction; a y (k) represents the acceleration in the y direction.

[0043] Then the state transfer equation is expressed as:

[0044] X(k+1)=A·X(k)+B·U(k);

[0045] Among them, A represents the state transfer matrix; B represents the control matrix; U(k) represents the control input vector.

[0046] When the detection time interval is T, the state transfer matrix A and the control matrix B are expressed as:

[0047]

[0048] The control input vector U(k) is expressed as:

[0049]

[0050] in, and denote the control input acceleration in the x and y directions respectively.

[0051] The observation equation is expressed as:

[0052] Z(k)=H·X(k)+W(k);

[0053] Where H represents the observation matrix and W(k) represents the observation noise.

[0054] The observation matrix H is expressed as:

[0055]

[0056] The observation noise W(k) is a Gaussian noise with a mean of zero, and its covariance matrix R is:

[0057]

[0058] in, and represent the standard deviation of the measurement noise in the x and y directions, respectively.

[0059] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, in step S4, the image data set, the distance data set and the pipeline data set are combined with the data fusion technology, and the inverse and weight of the joint covariance matrix are used to perform data alignment and fusion.

[0060] First, the inverse and weight of the joint covariance matrix are calculated and then applied to the fusion of state estimation; the data of the optical camera, sonar and pipeline instrument are respectively: D camera , D sonar and D pipeline ; Calculate the covariance matrix of each sensor data, expressed as:

[0061] The covariance matrix of the optical camera data:

[0062] P 1 =∑ camera = cov(D camera );

[0063] The covariance matrix of the sonar data:

[0064] P 2 =∑ sonar = cov(D sonar );

[0065] The covariance matrix of the pipeline instrument data is:

[0066] P 3 =∑ pipeline = cov(D pipeline );

[0067] The joint covariance matrix is ​​the sum of the covariance matrices of the three sensors: optical camera, sonar and pipeline meter. The joint covariance matrix is ​​expressed as:

[0068]

[0069] Where P represents the covariance of the three sensors.

[0070] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, the inverse P of the joint covariance matrix is ​​obtained by calculating -1 ; And according to the inverse of the joint covariance matrix, the gain matrix K is obtained, and the gain matrix K is expressed as:

[0071] K=P 1 P -1 ;

[0072] According to K update, the fused state estimate is obtained:

[0073]

[0074] The updated covariance fusion formula is expressed as:

[0075] P f =P 1 -KP 1 .

[0076] According to the above-mentioned underwater robot collision avoidance control method based on multi-sensor data fusion, preferably, the specific steps of step S5 include: when an obstacle is detected ahead, making intelligent decisions based on a preset collision avoidance control strategy, and guiding the robot to bypass the obstacle in an appropriate manner through a precise control mechanism, thereby ensuring that it continues to move forward safely and efficiently, and avoiding the robot from colliding with obstacles when moving in an underwater environment.

[0077] A second aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements any step in the underwater robot collision avoidance control method based on multi-sensor data fusion as described in the first aspect.

[0078] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a computer processor, the computer program implements any step in the underwater robot collision avoidance control method based on multi-sensor data fusion as described in the first aspect.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] 1. The present invention uses multiple sensors such as optical cameras, sonars and pipeline meters to identify obstacles and perceive the environment in the underwater environment. After preprocessing the acquired data, a collision avoidance control method based on multi-sensor data fusion is adopted, which can perceive more reliable and accurate environmental information, making the underwater robot more sensitive and efficient, and able to give early warning before colliding with obstacles, thereby reducing development costs and false alarm rates, and improving the reliability and practicality of the overall system; the present invention can also realize the recognition of dynamic targets, thereby improving the accuracy of collision avoidance prediction and the adaptability of the algorithm; the present invention uses multi-sensor fusion technology to facilitate the underwater robot to operate more safely and stably in the underwater environment, and significantly improve the ability of the underwater robot to adapt to complex underwater environments.

[0081] 2. Due to the noise and false detection in the original data collected by optical cameras, sonars and pipeline meters, the perception results are inaccurate and unstable. The present invention can improve the data quality by preprocessing the acquired electromagnetic signal data, avoid the influence of environmental noise and interference on feature extraction, and conveniently eliminate false signals to improve the accuracy of target identification. At the same time, the present invention uses wavelet transform technology to denoise the signal, which can effectively remove high-frequency noise while retaining key features, and is suitable for the preprocessing of electromagnetic signals. The signal after denoising will retain more detailed information, which is conducive to subsequent feature extraction and improving the accuracy of signal analysis. By using the information of the target at the previous moment to predict the information at the current moment, and then comparing the predicted result with the detection result at the current moment, a more reliable measurement value can be obtained. The advantages of optical cameras, sonars and pipeline meters are complementary to each other to achieve effective target identification. Finally, by combining advanced data fusion technology, high-level decision fusion of optical cameras, sonars and pipeline meters is achieved.

[0082] 3. The present invention obtains an image data set by preprocessing the data obtained by an optical camera in combination with image enhancement technology; obtains a distance data set by processing the data obtained by sonar based on a sonar signal processing algorithm in the data preprocessing stage; obtains a pipeline data set by preprocessing the data obtained by a pipeline meter; performs data fusion processing based on the obtained image data set, distance data set and pipeline data set; when an obstacle is detected ahead, an intelligent decision is made according to a preset collision avoidance control strategy, and the robot is guided to bypass the obstacle in an appropriate manner through a precise control mechanism, thereby ensuring that it continues to move forward safely and efficiently; the present invention uses optical cameras, sonars and pipeline meters for environmental perception and obstacle detection, and uses multi-sensor fusion technology to make collision avoidance control more accurate and efficient, identify and avoid potential risks in advance, and improve the reliability and practicality of the overall system.

[0083] 4. The present invention adopts "acoustic, optical, electromagnetic" multi-sensor data fusion technology to integrate data from different sensors, utilize the complementary advantages of each sensor, and provide higher perception accuracy and reliability; an optical camera is used to provide high-resolution image data, sonar can provide distance information in turbid water bodies, and electromagnetic data obtained based on pipeline meters can determine the position and status of the robot by detecting the electromagnetic signals of underground metal pipelines; through the data fusion algorithm, the data of different sensors are comprehensively processed, which can effectively improve the collision avoidance ability and environmental adaptability of the underwater robot. Therefore, the present invention can effectively solve the shortcomings of the single sensor method, provide more reliable and accurate environmental perception information, and is of great significance to improving the safety and efficiency of the collision avoidance control of the underwater robot.

[0084] 5. The present invention can provide more comprehensive, reliable and accurate environmental perception information in complex underwater environments by using multi-sensor data fusion technology that integrates optical camera, sonar and pipeline instrument data, making it easier for underwater robots to operate more safely and stably in underwater environments, and significantly improving the ability of underwater robots to adapt to complex underwater environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 It is a schematic diagram of the process of the present invention;

[0086] Figure 2 is a schematic diagram of the robot's related coordinate system;

[0087] Figure 3 It is a schematic diagram of the overall framework of the multi-sensor data fusion perception control of the present invention. DETAILED DESCRIPTION

[0088] The present invention is further described in detail below through specific examples, but the scope of the present invention is not limited thereto.

[0089] Example 1

[0090] A collision avoidance control method for underwater robots based on multi-sensor data fusion, the process is as follows Figure 1 As shown, its overall framework is as follows Figure 3 As shown, the following steps are included:

[0091] S1. Preprocessing the data acquired by the optical camera in combination with image enhancement technology to obtain an image data set.

[0092] The specific steps of step S1 include: firstly, using an optical camera to collect the original image data of the underwater environment, and then using dark channel prior image enhancement technology to preprocess the original image data, wherein the preprocessing includes contrast adjustment, noise removal and detail enhancement of the original image data to obtain an optimized image data set.

[0093] In this embodiment, in order to improve the quality of the original image data of the underwater environment collected by the optical camera, it is necessary to perform enhancement processing in the image preprocessing stage to minimize the adverse effects of uneven external environmental illumination and environmental noise on feature extraction. The original image is preprocessed using image enhancement technology, specifically including calculating the dark channel image, estimating the atmospheric light value, the transmittance of related parameters, and restoring the scene radiation value to obtain an optimized image data set.

[0094] The expression for calculating the dark channel image in this embodiment is:

[0095] J dark (x) = min y∈P(x) (min c∈{r,g,b} Ic (y));

[0096] Among them, J dark (x) represents the dark channel image; P(x) represents a local window of pixel x; I c Represents the pixel value of color channel c.

[0097] The dark channel image J is obtained by calculation dark (x), select the highest brightness area to estimate the atmospheric light value A.

[0098] The expression for calculating the relevant parameter transmittance t(x) in this embodiment is:

[0099]

[0100] Among them, ω represents the adjustment parameter and its value is 0.95.

[0101] The scene radiation value J(x) is restored using the estimated atmospheric light value A and transmittance t(x). The expression of J(x) is:

[0102]

[0103] Where I(x) represents the pixel value of color channel c; t 0 Represents a constant that prevents the denominator from being zero.

[0104] S2. Based on the data acquired by sonar, the data is processed in the data preprocessing stage based on the sonar signal processing algorithm to obtain a distance data set.

[0105] The specific steps of step S2 include: collecting sound wave data in the underwater environment through sonar, and preprocessing the sound wave data based on the sonar signal processing algorithm, wherein the preprocessing includes signal bandpass filtering, matched filtering and distance measurement to obtain an accurate distance data set.

[0106] In this embodiment, since the raw data collected by the sonar contains a large amount of negative sample content, if no corresponding processing is performed, it will not only interfere with the subsequent feature extraction, but also reduce the computing efficiency, increase the amount of calculation and the computing cost. Therefore, the sonar signal processing algorithm is applied to pre-process the sound wave data, specifically including signal bandpass filtering, matched filtering and distance measurement, to obtain an accurate set of related distance data.

[0107] The sonar signal is filtered using a bandpass filter to remove low-frequency noise and high-frequency interference; the expression of the bandpass filter used in this embodiment is:

[0108]

[0109] Among them, y(t) represents the filtered signal; x(t) represents the original signal; h(t) represents the impulse response of the filter; τ represents the time delay or offset.

[0110] The target signal is enhanced by performing matched filtering with the known pulse signal. The expression of the matched filter used in this embodiment is:

[0111]

[0112] Where z(t) represents the filtered signal; h m (t) represents the impulse response of the matched filter.

[0113] A constant false alarm rate algorithm is used to detect the echo signal and calculate the distance of the target; the constant false alarm rate algorithm is the CFAR algorithm. The expression of the CFAR algorithm used in this embodiment is:

[0114]

[0115] Among them, N represents the window size; T represents the detection threshold; window represents the sliding window; x(i) represents the i-th x signal.

[0116] S3. Based on the data acquired by the pipeline instrument, data preprocessing is performed to obtain a pipeline data set.

[0117] The specific steps of step S3 include: using a pipeline instrument to collect electromagnetic data of underwater pipelines, and using wavelet transform to perform denoising on the electromagnetic signals to obtain a pipeline data set.

[0118] In this embodiment, the original electromagnetic signal collected by the pipeline instrument contains a large amount of high-frequency noise and interference. If it is not properly processed, it will not only interfere with the subsequent data analysis, but also reduce the computing efficiency and increase the computing cost; therefore, the electromagnetic signal is denoised by wavelet transform. The specific steps include wavelet decomposition of the signal, application of threshold denoising technology to remove high-frequency noise, and wavelet reconstruction of the signal to obtain a clear and accurate pipeline data set.

[0119] The expression for denoising the electromagnetic signal using wavelet transform in this embodiment is:

[0120]

[0121] Where x(t) represents the denoised signal; c k represents the wavelet coefficient; ψ k (t) represents the wavelet basis function.

[0122] S4. Perform data fusion processing based on the obtained image data set, distance data set and pipeline data set.

[0123] The specific steps of step S4 include: aligning and fusing the correlation between the data acquired by the optical camera, sonar and pipeline meter through multi-sensor data fusion technology; and realizing effective identification of robot-related parameters through multi-sensor data fusion.

[0124] The measurement value is determined by using the information of the target at the previous moment to predict the data at the current moment, and comparing the predicted result with the current detection result.

[0125] This embodiment uses Kalman filtering to predict the robot's relevant state parameters at the next moment;

[0126] (I) Determine the state X(k), such as Figure 2 As shown, the state X(k) is composed of the relative position x(k) in the x direction and the velocity v in the x direction. x (k), relative position y(k) in the y direction, acceleration a in the x direction x (k) and the acceleration a in the y direction y (k) is composed, and the specific expression of state X(k) is:

[0127]

[0128] Among them, x(k) represents the relative position in the x direction; v x (k) represents the velocity in the x direction; y(k) represents the relative position in the y direction; v y (k) represents the velocity in the y direction; a x (k) represents the acceleration in the x direction; a y (k) represents the acceleration in the y direction.

[0129] The expression of the state transfer equation is determined as:

[0130] X(k+1)=A·X(k)+B·U(k);

[0131] Among them, A represents the state transfer matrix; B represents the control matrix; U(k) represents the control input vector.

[0132] The expression of the control input vector U(k) is:

[0133]

[0134] in, and denote the control input acceleration in the x and y directions respectively.

[0135] When the detection time interval is T, according to the kinematic equation, the expressions of the state transfer matrix A and the control matrix B are:

[0136]

[0137] The observation equation is determined as:

[0138] Z(k)=H·X(k)+W(k);

[0139] Where H represents the observation matrix and W(k) represents the observation noise.

[0140] The observation matrix H is expressed as:

[0141]

[0142] The observation noise W(k) is a Gaussian noise with a mean of zero, and its covariance matrix R is:

[0143]

[0144] in, and represent the standard deviation of the measurement noise in the x and y directions, respectively.

[0145] (II) The three types of data information, namely, image data set, distance data set and pipeline data set, are combined with data fusion technology, and the inverse sum of the joint covariance matrix and weights are used for data alignment and fusion; first, the inverse sum of the joint covariance matrix is ​​calculated, and then it is applied to the fusion of state estimation.

[0146] The data of optical camera, sonar and pipeline instrument are: D camera , D sonar and D pipeline ; Calculate the covariance matrix of each sensor data, expressed as:

[0147] The covariance matrix of the optical camera data:

[0148] P 1 =∑ camera = cov(D camera );

[0149] The covariance matrix of the sonar data:

[0150] P 2 =∑ sonar = cov(D sonar );

[0151] The covariance matrix of the pipeline instrument data is:

[0152] P 3 =∑ pipeline = cov(D pipeline );

[0153] The joint covariance matrix is ​​the sum of the covariance matrices of the three sensors, namely, the optical camera, the sonar and the pipeline meter. The expression for determining the joint covariance matrix is:

[0154]

[0155] Where P represents the covariance of the three sensors.

[0156] The inverse of the joint covariance matrix P is obtained by calculating -1 ; And according to the inverse of the joint covariance matrix, the gain matrix K is obtained, and the expression of the gain matrix K is determined as:

[0157] K=P 1 P -1 ;

[0158] The expression of the fused state estimation obtained according to K update is:

[0159]

[0160] The expression for determining the covariance fusion formula is:

[0161] P F =P 1 -KP 1 .

[0162] S5. When an obstacle is detected ahead, an intelligent decision is made based on a preset collision avoidance control strategy to guide the robot to bypass the obstacle. The specific steps include: when an obstacle is detected ahead, an intelligent decision is made based on a preset collision avoidance control strategy, and the robot is guided to bypass the obstacle in an appropriate manner through a precise control mechanism, thereby ensuring that the robot continues to move forward safely and efficiently, and avoiding the robot from colliding with the obstacle when moving in an underwater environment.

[0163] Example 2

[0164] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, any step of the underwater robot collision avoidance control method based on multi-sensor data fusion as described in Example 1 is implemented.

[0165] Further, the underwater robot collision avoidance control method process based on multi-sensor data fusion described in Example 1 can be implemented as a computer software program. For example, this embodiment includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a program code for executing the method. In such an embodiment, the computer program can be downloaded and installed from a network, and / or installed from a removable medium. When the computer program is executed by a processor, the above functions defined in the method of the present application are executed.

[0166] Example 3

[0167] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step in a method for collision avoidance control of an underwater robot based on multi-sensor data fusion as described in Example 1.

[0168] The computer-readable medium described in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0169] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0170] In summary, the present invention effectively overcomes the deficiencies in the prior art and has a high industrial utilization value. The above embodiments serve to illustrate the substantive content of the present invention, but do not limit the protection scope of the present invention. Those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and protection scope of the technical solution of the present invention.

[0171] The above embodiments are specific implementation methods of the present invention, but the implementation methods of the present invention are not limited to the above embodiments. Any other combination, change, modification, substitution and simplification that does not exceed the design concept of the present invention shall fall within the protection scope of the present invention.

Claims

1. A collision avoidance control method for an underwater robot based on multi-sensor data fusion, characterized in that: The following steps are involved: S1, preprocessing the data acquired by the optical camera in combination with image enhancement technology to obtain an image data set; S2, based on the data acquired by sonar, and processed in the data preprocessing stage based on the sonar signal processing algorithm, to obtain a distance data set; S3, based on the data obtained by the pipeline instrument, the data is preprocessed to obtain a pipeline data set; S4, performing data fusion processing according to the obtained image data set, distance data set and pipeline data set; S5. When an obstacle is detected ahead, an intelligent decision is made based on the preset collision avoidance control strategy to guide the robot around the obstacle.

2. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 1 is characterized in that: The specific steps of step S1 include: firstly, collecting original image data of the underwater environment by using an optical camera, and then preprocessing the original image data by using a dark channel prior image enhancement technology to obtain an optimized image data set; The expression for calculating the dark channel image using image enhancement technology is: J dark (x)=min y∈P(x) (my c∈{r,g,b} IN c (y)); Among them, J dark (x) represents the dark channel image; P(x) represents a local window of pixel x; I c Represents the pixel value of color channel c; The dark channel image J is obtained by calculation dark (x), select the highest brightness area to estimate the atmospheric light value A; Calculate the relevant parameter transmittance t(x), the expression is: Among them, ω represents the adjustment parameter; The scene radiation value J(x) is restored using the estimated atmospheric light value A and transmittance t(x). The expression of J(x) is: Wherein, I(x) represents the pixel value of color channel c; t0 represents a constant to prevent the denominator from being zero.

3. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 1 is characterized in that: The specific steps of step S2 include: collecting sound wave data in the underwater environment through sonar, pre-processing the sound wave data based on the sonar signal processing algorithm, and obtaining a distance data set; Use a bandpass filter to filter the sonar signal to remove low-frequency noise and high-frequency interference; the expression of the bandpass filter is: Where y(t) represents the filtered signal; x(t) represents the original signal; h(t) represents the impulse response of the filter; τ represents the time delay or offset; By matching the known pulse signal with the target signal, the target signal is enhanced. The expression of the matched filter is: Where z(t) represents the filtered signal; h m (t) represents the impulse response of the matched filter; Use the constant false alarm rate algorithm to detect the echo signal and calculate the distance of the target; the constant false alarm rate algorithm is the CFAR algorithm, and its expression is: Among them, N represents the window size; T represents the detection threshold; window represents the sliding window; x(i) represents the i-th x signal.

4. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 1 is characterized in that: The specific steps of step S3 include: using a pipeline instrument to collect electromagnetic data of underwater pipelines, using wavelet transform to perform denoising on the electromagnetic signals, and obtaining a pipeline data set; De-noising is used to remove high-frequency noise and interference. Its denoising expression is: Where x(t) represents the denoised signal; c k represents the wavelet coefficient; ψ k (t) represents the wavelet basis function.

5. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 1 is characterized in that: The specific steps of step S4 include: aligning and fusing the correlation between the data acquired by the optical camera, sonar and pipeline instrument through multi-sensor data fusion technology; The measurement value is determined by using the information of the target at the previous moment to predict the data at the current moment and comparing the prediction result with the current detection result; First, Kalman filtering is used to predict the robot's state parameters at the next moment, and the selected state X(k) is expressed as: Among them, x(k) represents the relative position in the x direction; v x (k) represents the velocity in the x direction; y(k) represents the relative position in the y direction; v y (k) represents the velocity in the y direction; a x (k) represents the acceleration in the x direction; a y (k) represents the acceleration in the y direction; Then the state transfer equation is expressed as: X(k+1)=A·X(k)+B·U(k); Where A represents the state transfer matrix; B represents the control matrix; U(k) represents the control input vector; When the detection time interval is T, the state transfer matrix A and the control matrix B are expressed as: The control input vector U(k) is expressed as: in, and denote the control input acceleration in the x and y directions respectively; The observation equation is expressed as: Z(k)=H·X(k)+W(k); Where H represents the observation matrix; W(k) represents the observation noise; The observation matrix H is expressed as: The observation noise W(k) is a Gaussian noise with a mean of zero, and its covariance matrix R is: in, and represent the standard deviation of the measurement noise in the x and y directions, respectively.

6. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 5 is characterized in that: In step S4, the image data set, the distance data set and the pipeline data set are combined with a data fusion technique, and the inverse and weight of the joint covariance matrix are used to perform data alignment and fusion; First, the inverse and weight of the joint covariance matrix are calculated and then applied to the fusion of state estimation; the data of the optical camera, sonar and pipeline instrument are respectively: D camera , D sonar and D pipeline ; Calculate the covariance matrix of each sensor data, expressed as: The covariance matrix of the optical camera data: P1=∑ camera =cov(D camera ); The covariance matrix of the sonar data: P2=∑ sonar =cov(D sonar ); The covariance matrix of the pipeline instrument data is: P3=∑ pipeline =cov(D pipeline ); The joint covariance matrix is ​​the sum of the covariance matrices of the three sensors: optical camera, sonar and pipeline meter. The joint covariance matrix is ​​expressed as: Where P represents the covariance of the three sensors.

7. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 6 is characterized in that: The inverse of the joint covariance matrix P is obtained by calculating -1 ; And according to the inverse of the joint covariance matrix, the gain matrix K is obtained, and the gain matrix K is expressed as: K=P1P -1 ; According to K update, the fused state estimate is obtained: The updated covariance fusion formula is expressed as: P f =P1-KP1.

8. The underwater robot collision avoidance control method based on multi-sensor data fusion according to claim 1 is characterized in that: The specific steps of step S5 include: when an obstacle is detected ahead, making intelligent decisions according to a preset collision avoidance control strategy, and guiding the robot to bypass the obstacle in an appropriate manner through a precise control mechanism.

9. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, any step in the underwater robot collision avoidance control method based on multi-sensor data fusion as described in any one of claims 1-8 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a computer processor, it implements any step in the underwater robot collision avoidance control method based on multi-sensor data fusion as described in any one of claims 1-8.

Images