A real-time path replanning unmanned sailboat obstacle avoidance method and device

Abstract

The present invention discloses an unmanned sailboat obstacle avoidance method and device with real-time path replanning, which belongs to the field of path planning, wherein the method includes: continuously collecting sensor monitoring information sequence and visual monitoring information sequence, and monitoring to obtain wind condition information; performing obstacle identification, obtaining obstacle position information sequence, and analyzing to obtain obstacle movement path information; when it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat, obtaining an obstacle avoidance area according to the obstacle avoidance range based on the intersection point; in the obstacle avoidance area, optimizing the obstacle avoidance path, obtaining the optimal obstacle avoidance path, performing path replanning and controlling the unmanned sailboat to perform obstacle avoidance navigation. Through computer vision and sensor fusion, the technical effect of intelligent perception and accurate prediction of obstacles in complex sea conditions, and real-time optimization of obstacle avoidance paths, and improving the real-time and reliability of obstacle avoidance of unmanned sailboats is achieved.

Description

Technical Field

[0001] The invention relates to the field of path planning, and in particular to an unmanned sailboat obstacle avoidance method and device for real-time path replanning. Background Art

[0002] Unmanned sailboats have broad application prospects, but in the actual navigation process, the complex and changeable sea conditions pose a challenge to their safety. In order to ensure that unmanned sailboats can complete their tasks safely and efficiently, it is very important to realize intelligent obstacle avoidance. The existing unmanned sailboat obstacle avoidance methods mainly use sensors such as radar and sonar to detect obstacles, and then plan obstacle avoidance paths based on the location information of obstacles. However, these methods have the following shortcomings: First, in harsh sea conditions, complex environmental factors such as waves and sea fog will affect the detection accuracy of sensors, reducing the reliability of obstacle identification; second, it is difficult for existing methods to accurately predict the movement state of obstacles and make reasonable obstacle avoidance decisions in time; third, the planning of obstacle avoidance paths often does not fully consider factors such as the sailing distance and the state of the sailboat itself, resulting in low obstacle avoidance efficiency. Therefore, the real-time and reliability of obstacle avoidance of unmanned sailboats in the existing technology are poor. Summary of the invention

[0003] The present application provides a method and device for unmanned sailboat obstacle avoidance with real-time path replanning, aiming to solve the technical problems that the existing unmanned sailboat obstacle avoidance has low obstacle detection accuracy in complex sea conditions, cannot accurately predict the movement state of obstacles, and has a poor obstacle avoidance path, resulting in poor real-time and reliability of unmanned sailboat obstacle avoidance.

[0004] In view of the above problems, the present application provides an unmanned sailboat obstacle avoidance method and device with real-time path replanning.

[0005] The first aspect disclosed in the present application provides an unmanned sailboat obstacle avoidance method with real-time path replanning, the method comprising: continuously collecting sensor monitoring information sequences and visual monitoring information sequences during the navigation of the unmanned sailboat through obstacle sensors and visual equipment, and monitoring and obtaining wind condition information through wind condition sensors; performing obstacle identification according to the sensor monitoring information sequences and visual monitoring information sequences, obtaining obstacle position information sequences, and analyzing and obtaining obstacle movement path information in combination with the navigation path of the unmanned sailboat; when it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat, obtaining an obstacle avoidance area according to the obstacle avoidance range based on the intersection point; optimizing the obstacle avoidance path within the obstacle avoidance area to obtain the optimal obstacle avoidance path, performing path replanning and controlling the unmanned sailboat to perform obstacle avoidance navigation, wherein, The purpose is to increase the sailing distance of the obstacle avoidance path of the unmanned sailboat and reduce the change amplitude of the angle between the sail surface and the wind condition information in the obstacle avoidance path and the angle between the sail surface and the wind condition information in the planned sailing path, calculate the fitness of the obstacle avoidance path, and optimize the obstacle avoidance path; wherein, the obstacle avoidance path distance of the obstacle avoidance path is obtained, and the planned path distance of the planned sailing path in the obstacle avoidance area is obtained, and the difference between the obstacle avoidance path distance and the planned path distance is calculated as the increased sailing distance, and the sailing distance increase amplitude is calculated; the average planned angle between the sail surface in the planned sailing path and the wind condition information is obtained, and the average obstacle avoidance angle between the sail surface in the obstacle avoidance path and the wind condition information is obtained, and the angle change amplitude is calculated; according to the sailing distance increase amplitude and the angle change amplitude, the obstacle avoidance fitness of the obstacle avoidance path is calculated to optimize the obstacle avoidance path.

[0006] Another aspect disclosed in the present application provides an unmanned sailboat obstacle avoidance device with real-time path replanning, the device comprising: an information monitoring unit, used to continuously collect sensor monitoring information sequences and visual monitoring information sequences during the navigation of the unmanned sailboat through obstacle sensors and visual equipment, and to monitor and obtain wind condition information through wind condition sensors; an identification and analysis unit, used to identify obstacles based on the sensor monitoring information sequences and visual monitoring information sequences, obtain obstacle position information sequences, and analyze and obtain obstacle movement path information in combination with the navigation path of the unmanned sailboat; an area division unit, used to divide the obstacle movement path information into obstacle avoidance areas based on the intersection points according to the obstacle avoidance range when it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat; a path optimization unit, used to optimize the obstacle avoidance path in the obstacle avoidance area, obtain the optimal obstacle avoidance path, perform path replanning and An unmanned sailboat is controlled to perform obstacle avoidance navigation, wherein the fitness of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized for the purpose of reducing the increased sailing distance of the unmanned sailboat's obstacle avoidance path and reducing the change amplitude of the angle between the sail surface in the obstacle avoidance path and the wind condition information and the angle between the sail surface and the wind condition information in the planned sailing path; wherein the obstacle avoidance path distance of the obstacle avoidance path is obtained, and the planned path distance of the planned sailing path in the obstacle avoidance area is obtained, the difference between the obstacle avoidance path distance and the planned path distance is calculated as the increased sailing distance, and the sailing distance increase amplitude is calculated; the average planned angle between the sail surface in the planned sailing path and the wind condition information is obtained, and the average obstacle avoidance angle between the sail surface in the obstacle avoidance path and the wind condition information is obtained, and the angle change amplitude is calculated; according to the sailing distance increase amplitude and the angle change amplitude, the obstacle avoidance fitness of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] Due to the use of obstacle sensors and visual equipment, the sensor monitoring information sequence and the visual monitoring information sequence are continuously collected during the navigation of the unmanned sailboat, and the wind condition information is monitored and obtained through the wind condition sensor, providing a data basis for subsequent obstacle identification and path planning; according to the sensor monitoring information sequence and the visual monitoring information sequence, the obstacle identification is carried out to obtain the obstacle position information sequence, and the obstacle movement path information is analyzed and obtained in combination with the navigation path of the unmanned sailboat. Through the fusion analysis of multi-source information, the accurate identification of obstacles and the prediction of movement paths are realized, providing a basis for obstacle avoidance decisions; when it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat, the obstacle avoidance area is obtained according to the obstacle avoidance range based on the intersection point, and the area where obstacle avoidance is required is divided by judging the intersection of the obstacle movement path and the navigation path, and the obstacle avoidance range is determined; in the obstacle avoidance area, the obstacle avoidance path is optimized to obtain the optimal obstacle avoidance path. The path is replanned and the unmanned sailboat is controlled to perform obstacle avoidance navigation, wherein the purpose is to reduce the increased sailing distance of the unmanned sailboat's obstacle avoidance path and reduce the change amplitude of the angle between the sail surface and the wind condition information in the obstacle avoidance path and the angle between the sail surface and the wind condition information in the planned sailing path, calculate the fitness of the obstacle avoidance path, optimize the obstacle avoidance path, and plan the optimal obstacle avoidance path by comprehensively considering factors such as the sailing distance and the sailboat's own state in the obstacle avoidance area. The navigation path is adjusted in real time accordingly to complete the technical solution of the obstacle avoidance process, solve the technical problems of low obstacle detection accuracy, inability to accurately predict the movement state of obstacles, poor obstacle avoidance path, and poor real-time and reliability of unmanned sailboat obstacle avoidance in complex sea conditions in existing unmanned sailboat obstacle avoidance. Through computer vision and sensor fusion, the technical effect of intelligent perception and accurate prediction of obstacles in complex sea conditions, real-time optimization of obstacle avoidance path, and improvement of real-time and reliability of unmanned sailboat obstacle avoidance is achieved.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A flowchart of an unmanned sailboat obstacle avoidance method with real-time path replanning is provided for an embodiment of the present application;

[0011] Figure 2 A structural schematic diagram of an unmanned sailboat obstacle avoidance device with real-time path replanning is provided for an embodiment of the present application.

[0012] Explanation of the reference numerals: information monitoring unit 11 , identification and analysis unit 12 , area division unit 13 , path optimization unit 14 . DETAILED DESCRIPTION

[0013] The overall idea of ​​the technical solution provided by this application is as follows:

[0014] The embodiment of the present application provides an unmanned sailboat obstacle avoidance method and device with real-time path replanning. First, the sensor monitoring information sequence, visual monitoring information sequence and wind condition information of the unmanned sailboat's navigation environment are continuously collected through obstacle sensors and visual equipment to provide a data basis for obstacle identification and path planning. Then, computer vision and sensor fusion technology are used to analyze the collected multi-source information to achieve accurate identification of obstacles and prediction of moving paths, and to determine the intersection of the obstacle moving path and the unmanned sailboat's navigation path, and accordingly divide the area that needs to be avoided. Afterwards, the optimal obstacle avoidance path is planned in the obstacle avoidance area by comprehensively considering factors such as the navigation distance and the state of the sailboat, and the navigation path is adjusted in real time to guide the unmanned sailboat to safely pass through the obstacle area, thereby realizing intelligent perception and accurate prediction of obstacles under complex sea conditions, and improving the real-time and reliability of the unmanned sailboat's obstacle avoidance.

[0015] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0016] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides an unmanned sailboat obstacle avoidance method with real-time path replanning, the method comprising:

[0017] S1: Through obstacle sensors and visual equipment, the sensor monitoring information sequence and visual monitoring information sequence are continuously collected during the navigation of the unmanned sailboat, and the wind condition information is monitored and obtained through the wind condition sensor.

[0018] Specifically, obstacle sensors (such as radar devices) collect physical characteristic data such as the distance and orientation of surrounding obstacles by emitting detection signals and receiving reflected signals. These data reflect the spatial position relationship of obstacles relative to the unmanned sailboat. Visual equipment (such as camera devices) collect image data of the navigation environment. These image data contain visual characteristic information such as the shape and size of obstacles. The sensor monitoring information collected at different times is organized in chronological order to form a sensor monitoring information sequence; similarly, the collected visual monitoring information is organized in chronological order to form a visual monitoring information sequence. These information sequences can reflect the movement characteristics of obstacles over time and provide data support for subsequent obstacle identification and trajectory prediction. At the same time, the wind condition sensor monitors the wind condition information in the navigation environment in real time to provide information for the subsequent optimization of obstacle avoidance paths.

[0019] By collecting multi-source information about the navigation environment, a data foundation is laid for subsequent intelligent obstacle avoidance decisions. Compared with existing technologies, the use of multi-source information fusion can improve the reliability of environmental perception and effectively overcome the detection limitations of a single type of sensor in harsh sea conditions.

[0020] S2: According to the sensor monitoring information sequence and the visual monitoring information sequence, obstacle identification is performed to obtain an obstacle position information sequence, and combined with the navigation path of the unmanned sailboat, obstacle movement path information is analyzed and obtained.

[0021] Specifically, after obtaining the sensor monitoring information sequence, the visual monitoring information sequence and the wind condition information, first, the historical navigation position sequence of the unmanned sailboat at multiple moments in the past is obtained, and the historical navigation position sequence records the navigation trajectory of the unmanned sailboat. Then, based on the sensor monitoring information sequence and the visual monitoring information sequence, the relative position of the obstacle and the unmanned sailboat is identified to obtain the obstacle relative direction sequence and the obstacle relative distance sequence. These sequences reflect the change in the spatial relationship of the obstacle relative to the unmanned sailboat. Next, according to the historical navigation position sequence of the unmanned sailboat, combined with the obstacle relative direction sequence and the obstacle relative distance sequence, the position information sequence of the obstacle in the global coordinate system is calculated. Based on the position information sequence and the corresponding time information, the moving speed information of the obstacle is calculated, and then the moving path information of the obstacle is fitted by analyzing the position and speed of the obstacle.

[0022] Through the above method, the accurate prediction of the movement state of the obstacle is achieved, which provides the necessary obstacle trajectory information for subsequent obstacle avoidance decisions, making the obstacle avoidance decisions more forward-looking.

[0023] S3: when it is determined that the obstacle moving path information will intersect with the planned navigation path of the unmanned sailboat, an obstacle avoidance area is obtained according to the obstacle avoidance range based on the intersection point.

[0024] Specifically, first, the navigation speed information and planned navigation path of the unmanned sailboat are obtained. Then, based on the navigation speed information and planned navigation path of the unmanned sailboat, combined with the obstacle movement path information obtained in the previous steps, the intersection risk is judged. The judgment process predicts whether the unmanned sailboat and the obstacle will intersect at some point in the future by analyzing the positional relationship between the unmanned sailboat and the obstacle at different times. When the judgment result shows that there is no intersection risk, the unmanned sailboat continues to sail according to the original planned navigation path; when the judgment result shows that there is an intersection risk, the predicted intersection point position information is obtained. With the intersection point as the center, the obstacle avoidance area is divided according to the pre-set obstacle avoidance range (for example, a specific radius range with the intersection point as the center).

[0025] Through the above methods, early warning of potential collision risks is achieved, and the reasonable division of obstacle avoidance areas provides spatial range constraints for subsequent obstacle avoidance path planning. By predicting and dividing the obstacle avoidance areas in advance, sufficient obstacle avoidance decision space is provided for the unmanned sailboat, improving the safety and reliability of obstacle avoidance.

[0026] S4: In the obstacle avoidance area, optimize the obstacle avoidance path, obtain the optimal obstacle avoidance path, replan the path and control the unmanned sailboat to perform obstacle avoidance navigation, wherein the fitness of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized for the purpose of reducing the increased navigation distance of the obstacle avoidance path of the unmanned sailboat and reducing the change amplitude of the angle between the sail surface in the obstacle avoidance path and the wind condition information and the angle between the sail surface and the wind condition information in the planned navigation path. Among them, the obstacle avoidance path distance of the obstacle avoidance path is obtained, and the planned path distance of the planned navigation path in the obstacle avoidance area is obtained, and the difference between the obstacle avoidance path distance and the planned path distance is calculated as the increased navigation distance, and the navigation distance increase amplitude is calculated; the average planned angle between the sail surface in the planned navigation path and the wind condition information is obtained, and the average obstacle avoidance angle between the sail surface in the obstacle avoidance path and the wind condition information is obtained, and the angle change amplitude is calculated; according to the navigation distance increase amplitude and the angle change amplitude, the obstacle avoidance fitness of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized.

[0027] Specifically, after the obstacle avoidance area is determined, the obstacle avoidance path is optimized in the area to obtain the optimal obstacle avoidance path, and the unmanned sailboat is controlled to perform obstacle avoidance navigation.

[0028] First, in the designated obstacle avoidance area, an initial obstacle avoidance path that does not intersect with the obstacle movement path is randomly generated. For the generated obstacle avoidance path, the distance increase of the path relative to the original planned navigation path is calculated to obtain the navigation distance increase amplitude. At the same time, the angle change between the sail surface and the wind condition information on the obstacle avoidance path is calculated to obtain the angle change amplitude. With the optimization goal of reducing the increased navigation distance of the unmanned sailboat obstacle avoidance path and reducing the change amplitude of the angle between the sail surface and the wind condition information in the obstacle avoidance path and the sail surface and the wind condition information in the planned navigation path, the obstacle avoidance path is iteratively optimized. During the optimization process, the obstacle avoidance fitness is calculated based on the navigation distance increase amplitude and the angle change amplitude. By continuously generating new obstacle avoidance paths and calculating their fitness until the optimization process converges, the path with the maximum obstacle avoidance fitness is output as the optimal obstacle avoidance path. Afterwards, the unmanned sailboat is controlled to perform obstacle avoidance navigation according to the optimal obstacle avoidance path.

[0029] Through the above method, under the premise of ensuring the safety of obstacle avoidance, the dual-objective optimization of the obstacle avoidance path is achieved. Not only the increase in the sailing distance, but also the influence of wind conditions on the sailing of the sailboat are considered, making the obstacle avoidance process more in line with the actual navigation characteristics of the unmanned sailboat, and improving the obstacle avoidance efficiency and reliability.

[0030] Furthermore, the embodiment of the present application also includes:

[0031] Through obstacle sensors and visual equipment, sensor monitoring information and visual monitoring information are continuously collected during the navigation of the unmanned sailboat; based on the sensor monitoring information and visual monitoring information at multiple moments in the past, sensor monitoring information sequences and visual monitoring information sequences are constructed; through wind condition sensors, wind direction information at multiple moments in the past is collected, and the average wind direction information is processed to obtain the wind condition information.

[0032] In a feasible implementation, during the navigation of the unmanned sailboat, the obstacle sensors and visual monitoring information are continuously collected through the obstacle sensors and visual devices installed on the unmanned sailboat. Among them, the obstacle sensor can be selected from radar, laser radar, ultrasonic sensor, etc., to obtain information such as the distance and direction of the obstacle; the visual device can be selected from visible light camera, infrared camera, etc., to obtain image information of the surrounding environment. The obstacle sensor and the visual device work continuously at a certain time interval (such as 0.1 seconds) to obtain the sensor monitoring information and visual monitoring information reflecting the environmental state in real time. Then, the sensor monitoring information and visual monitoring information collected by the obstacle sensor and the visual device at multiple moments in the past are integrated to construct a complete sensor monitoring information sequence and visual monitoring information sequence. The sensor monitoring information sequence reflects the information of the obstacle position, motion state, etc. that changes over time, and the visual monitoring information sequence reflects the information of the surrounding environment image that changes over time. By constructing the information sequence, the dynamic change process of the navigation environment is reflected. At the same time, the wind direction information in the past period of time is collected through the wind condition sensor installed on the unmanned sailboat. Among them, the wind condition sensor can be a wind vane, anemometer, etc. The wind direction information collected at multiple times is processed, the average wind direction is calculated, and the wind condition information is obtained. The wind condition information reflects the overall situation of the wind direction over a period of time and provides a reference for the navigation control of the unmanned sailboat.

[0033] Through information collection and processing, the sensor monitoring information sequence, visual monitoring information sequence and wind condition information reflecting the navigation environment are obtained, providing the necessary data basis for subsequent obstacle identification and path planning. At the same time, during the navigation of the unmanned sailboat, this information collection continues, and by constantly updating the information sequence and wind condition information, the unmanned sailboat can obtain environmental changes in a timely manner and provide support for dynamic obstacle avoidance.

[0034] Furthermore, the embodiment of the present application also includes:

[0035] The method comprises obtaining a historical navigation position sequence of the unmanned sailboat at multiple moments in the past; identifying the relative position of the obstacle and the unmanned sailboat according to the sensor monitoring information sequence and the visual monitoring information sequence, and obtaining an obstacle relative direction sequence and an obstacle relative distance sequence; processing and obtaining an obstacle position information sequence according to the historical navigation position sequence, in combination with the obstacle relative direction sequence and the obstacle relative distance sequence; calculating and obtaining obstacle movement speed information according to the obstacle position information sequence and multiple moments, and obtaining obstacle movement path information by fitting in combination with the obstacle position information sequence.

[0036] In a preferred embodiment, after obtaining the sensor monitoring information sequence and the visual monitoring information sequence, the information is processed and analyzed to realize obstacle recognition, obtain the obstacle position information sequence, and combine with the navigation path of the unmanned sailboat to analyze and obtain the obstacle movement path information.

[0037] First, the historical navigation position sequence of the unmanned sailboat at multiple moments in the past is obtained. The navigation position of the unmanned sailboat can be obtained through positioning devices such as GPS and inertial navigation systems. By recording the position coordinates of the unmanned sailboat over a period of time, a historical navigation position sequence reflecting the motion trajectory of the unmanned sailboat is constructed. Then, the relative position of the obstacle and the unmanned sailboat is identified based on the sensor monitoring information sequence and the visual monitoring information sequence. The information obtained by the obstacle sensor reflects the direction and distance of the obstacle relative to the sensor, and the image information obtained by the visual device also contains the pixel position of the obstacle relative to the camera. By analyzing and processing this information, the pixel coordinates are converted into the actual direction and distance, and fused with the sensor data to obtain the direction and distance sequence of the obstacle relative to the unmanned sailboat, and generate the obstacle relative direction sequence and obstacle relative distance sequence. After that, the obstacle relative direction sequence and obstacle relative distance sequence are fused with the historical navigation position sequence of the unmanned sailboat. By transforming the relative position information with the absolute position of the unmanned sailboat, the position coordinate sequence of the obstacle in the geographic coordinate system, that is, the obstacle position information sequence, is calculated. The obstacle position information sequence reflects the position change of the obstacle relative to the geographic environment. Then, according to the obstacle position information sequence, the obstacle movement speed information is calculated. By analyzing the change of the obstacle position over time and using the relationship between displacement and time, the average movement speed of the obstacle over a period of time is estimated. Combining the obstacle position information sequence and movement speed information, curve fitting and other methods are used to obtain the obstacle movement path information, that is, the possible movement trajectory of the obstacle in the future, to provide support for the subsequent unmanned sailboat obstacle avoidance.

[0038] Furthermore, the embodiment of the present application also includes:

[0039] According to the historical data of obstacle recognition, a sample sensor monitoring information set, a sample visual monitoring information set, a sample relative direction set and a sample relative distance set are collected; the sample sensor monitoring information set and the sample visual monitoring information set are respectively used as input supervised training data, and the sample relative direction set and the sample relative distance set are used as output supervised training data to train the sensor recognition branch and the visual recognition branch respectively; the multiple sensor monitoring information and the multiple visual monitoring information in the sensor monitoring information sequence and the visual monitoring information sequence are respectively input into the sensor recognition branch and the visual recognition branch, and multiple sensor relative directions, multiple sensor relative distances, multiple visual relative directions and multiple visual relative distances are obtained as output; the multiple sensor relative directions and the multiple visual relative directions are weighted and calculated to obtain multiple relative directions, and the obstacle relative direction sequence is constructed; the multiple sensor relative distances and the multiple visual relative distances are weighted and calculated to obtain multiple relative distances, and the obstacle relative distance sequence is constructed.

[0040] In a preferred embodiment, first, based on the historical obstacle recognition data, a sample sensor monitoring information set, a sample visual monitoring information set, a sample relative direction set and a sample relative distance set are collected. Among them, the sample sensor monitoring information set includes obstacle information data collected by sensors such as radar, laser radar, and ultrasonic waves at different times; the sample visual monitoring information set includes obstacle image data collected by visual devices such as visible light cameras and infrared cameras at different times; the sample relative direction set and the sample relative distance set include the azimuth angle and distance value of the obstacle relative to the unmanned sailboat at the corresponding time. These sample data are obtained by manual or automatic annotation and are used to construct a training data set for supervised learning. Then, the sample sensor monitoring information set and the sample visual monitoring information set are used as inputs of the neural network model, and the sample relative direction set and the sample relative distance set are used as outputs of the neural network model, respectively, to construct a sensor recognition branch and a visual recognition branch. Among them, the sensor recognition branch adopts a structure such as a feedforward neural network or a convolutional neural network, takes the sample sensor monitoring information as input, and takes the sample relative direction and the sample relative distance as output, and performs parameter optimization and model training through a back propagation algorithm. The visual recognition branch adopts structures such as convolutional neural networks or target detection networks, takes sample visual monitoring information as input, and takes sample relative direction and sample relative distance as output. It also uses the back propagation algorithm for parameter optimization and model training.

[0041] Subsequently, the sensor monitoring information sequence and visual monitoring information sequence collected in real time are input into the trained sensor recognition branch and visual recognition branch respectively. The sensor recognition branch performs forward calculation on each sensor monitoring information in the sensor monitoring information sequence to obtain the corresponding sensor relative direction and sensor relative distance. Similarly, the visual recognition branch performs forward calculation on each visual monitoring information in the visual monitoring information sequence to obtain the corresponding visual relative direction and visual relative distance. Through the prediction of the neural network model, the relative position information of the obstacles identified from different data sources is obtained. Afterwards, the multiple sensor relative directions output by the sensor recognition branch and the multiple visual relative directions output by the visual recognition branch are weighted averaged to obtain the fused obstacle relative direction. Similarly, the multiple sensor relative distances and the multiple visual relative distances are weighted averaged to obtain the fused obstacle relative distance. Among them, the weight of the weighted average is set according to the credibility or recognition accuracy of the sensor recognition branch and the visual recognition branch to make full use of the recognition results of the two branches. Arranging the fused relative directions and relative distances of obstacles in chronological order, we can obtain the obstacle relative direction sequence and obstacle relative distance sequence, thus completing the serialized representation of the relative positions of obstacles.

[0042] Through machine learning methods, the relative position recognition of obstacles based on sensor data and visual image data is realized, which improves the accuracy and adaptability of recognition and lays the foundation for subsequent obstacle spatial position calculation and obstacle avoidance decision-making.

[0043] Furthermore, the embodiment of the present application also includes:

[0044] The navigation speed information and the planned navigation path of the unmanned sailboat are obtained; according to the navigation speed information and the planned navigation path, an intersection judgment is performed with the obstacle movement path information; when it is judged that there is no intersection, obstacle avoidance is not performed; when it is judged that there is an intersection, an intersection point is obtained, and an obstacle avoidance area is obtained by dividing the intersection point as the center according to a preset obstacle avoidance range.

[0045] In a feasible implementation, after obtaining the obstacle movement path information, it is necessary to determine whether it intersects with the planned navigation path of the unmanned sailboat, and divide the obstacle avoidance area when there is an intersection, providing a reference for planning the obstacle avoidance path.

[0046] First, obtain the navigation speed information and planned navigation path of the unmanned sailboat. The navigation speed information is measured in real time by the speed sensor, which reflects the current movement speed of the unmanned sailboat; the planned navigation path is the navigation trajectory pre-set by the unmanned sailboat, which is composed of a series of waypoint coordinates and reflects the target route of the unmanned sailboat. Combining the navigation speed information with the planned navigation path, the movement trajectory of the unmanned sailboat in the future can be predicted. Then, according to the navigation speed information and the planned navigation path of the unmanned sailboat, the intersection judgment with the obstacle movement path information is performed. Specifically, first, according to the navigation speed information and the planned navigation path, the movement trajectory of the unmanned sailboat in the future is predicted; then, the predicted movement trajectory of the unmanned sailboat is compared with the obstacle movement path information in space and time to determine whether the two trajectories have an intersection or are close to each other; if there is an intersection or the shortest distance between the two trajectories is less than the set safety threshold, it is considered that there is an intersection risk and obstacle avoidance is required. The purpose of intersection judgment is to predict whether the unmanned sailboat may collide with the obstacle, and provide a basis for timely starting the obstacle avoidance strategy. After that, according to the result of the intersection judgment, it is decided whether to avoid the obstacle and divide the obstacle avoidance area. If it is found in the intersection judgment that there is no intersection risk between the planned navigation path of the unmanned sailboat and the obstacle movement path information, there is no need to avoid obstacles, and the unmanned sailboat can continue to sail along the original navigation path; if it is judged that there is an intersection risk, obstacle avoidance is required and the obstacle avoidance area is divided. The division of the obstacle avoidance area is centered on the intersection point, which is the intersection point or the closest approach point between the motion trajectory of the unmanned sailboat and the obstacle movement path information. With the intersection point as the center and the preset obstacle avoidance range as the radius, a circular obstacle avoidance area is divided. The size of the obstacle avoidance range is set according to the safety requirements and motion performance of the unmanned sailboat. The division of the obstacle avoidance area provides spatial constraints for the subsequent obstacle avoidance path planning. The unmanned sailboat needs to plan a safe navigation path in the obstacle avoidance area that does not collide with obstacles.

[0047] Furthermore, the embodiment of the present application also includes:

[0048] In the obstacle avoidance area, randomly select a first obstacle avoidance path that does not intersect with the obstacle movement path information; obtain a first obstacle avoidance path distance of the first obstacle avoidance path, and obtain a first planned path distance of the planned navigation path in the obstacle avoidance area, calculate the difference between the first obstacle avoidance path distance and the first planned path distance as a first increased navigation distance, and calculate to obtain a first navigation distance increase amplitude; obtain an average planned angle between the sail surface in the planned navigation path and the wind condition information, and obtain an average obstacle avoidance angle between the sail surface in the first obstacle avoidance path and the wind condition information, and calculate to obtain a first angle change amplitude; calculate to obtain a first obstacle avoidance fitness of the first obstacle avoidance path according to the first navigation distance increase amplitude and the first angle change amplitude; continue to optimize the obstacle avoidance path in the obstacle avoidance area until the obstacle avoidance path optimization converges, and output the obstacle avoidance path with the largest obstacle avoidance fitness as the optimal obstacle avoidance path.

[0049] In a preferred embodiment, after the obstacle avoidance area is determined, the obstacle avoidance path is optimized in the obstacle avoidance area to obtain an optimal obstacle avoidance path.

[0050] First, a random obstacle avoidance path that does not intersect with the obstacle moving path information is selected in the obstacle avoidance area as the first obstacle avoidance path. The selection of the first obstacle avoidance path can be achieved by randomly generating path points, and the safety of the selected path is ensured by the intersection judgment with the obstacle moving path information. Then, the impact of the first obstacle avoidance path on the navigation distance is evaluated. Specifically, the length of the first obstacle avoidance path is obtained as the first obstacle avoidance path distance; the length of the planned navigation path in the obstacle avoidance area is obtained as the first planned path distance; the difference between the first obstacle avoidance path distance and the first planned path distance is calculated to obtain the first increased navigation distance, which represents the increased navigation distance by adopting the first obstacle avoidance path; in order to further quantify the increase in the navigation distance, the ratio of the first increased navigation distance to the first planned path distance is calculated as the first navigation distance increase amplitude. The smaller the first navigation distance increase amplitude is, the smaller the increase in the navigation distance caused by the first obstacle avoidance path is, and the higher the obstacle avoidance efficiency is. At the same time, the impact of the first obstacle avoidance path on the sail angle control of the sailboat is evaluated. Specifically, the angles between the sail surface and the wind direction at multiple positions on the planned navigation path are obtained, and the average value thereof is calculated as the average planned angle; the angles between the sail surface and the wind direction at multiple positions on the first obstacle avoidance path are obtained, and the average value thereof is calculated as the average obstacle avoidance angle; then, the difference between the average obstacle avoidance angle and the average planned angle is calculated, and then divided by the average planned angle to obtain the first angle change amplitude, which indicates the degree of change of the angle between the sail surface and the wind direction relative to the planned navigation path after the first obstacle avoidance path is adopted. The smaller the first angle change amplitude, the smaller the adjustment of the sail surface angle caused by the first obstacle avoidance path, and the smaller the impact on navigation control.

[0051] Subsequently, the first obstacle avoidance fitness of the first obstacle avoidance path is calculated according to the first sailing distance increase amplitude and the first angle change amplitude. Obstacle avoidance fitness is a comprehensive evaluation index used to measure the degree of optimization of the obstacle avoidance path in terms of sailing distance and sailboat control. Obstacle avoidance fitness can be calculated by weighted summation. The sailing distance increase amplitude and the angle change amplitude are multiplied by the corresponding weight coefficients respectively, and then added to obtain the obstacle avoidance fitness. The larger the obstacle avoidance fitness, the better the comprehensive optimization effect of the obstacle avoidance path. Repeat the above process to continue searching and optimizing the obstacle avoidance path in the obstacle avoidance area. Through iterative optimization, the obstacle avoidance path is continuously updated and the obstacle avoidance fitness is calculated until the convergence condition is met (such as the number of iterations reaches the upper limit or the improvement of the obstacle avoidance fitness is less than the threshold). Finally, the obstacle avoidance path with the largest obstacle avoidance fitness is selected as the optimal obstacle avoidance path.

[0052] By fully considering the impact of the obstacle avoidance path on the sailing distance and sailboat control, two indicators, the sailing distance increase amplitude and the angle change amplitude, are introduced to quantitatively evaluate the advantages and disadvantages of the obstacle avoidance path. The optimal obstacle avoidance path is searched by iterative optimization. While avoiding obstacles, the sailing distance is shortened as much as possible, the sail angle adjustment is reduced, and the comprehensive optimization of the obstacle avoidance path is achieved. The acquisition of the optimal obstacle avoidance path provides the optimal path selection for the obstacle avoidance decision of the unmanned sailboat, which helps to improve the safety and efficiency of obstacle avoidance.

[0053] Furthermore, according to the first navigation distance increase amplitude and the first angle change amplitude, a first obstacle avoidance fitness of the first obstacle avoidance path is calculated and obtained as follows:

[0054] ;

[0055] in, For obstacle avoidance adaptation, and is the weight, Add amplitude to the sailing distance, is the angle change amplitude.

[0056] In a preferred embodiment, when calculating the first obstacle avoidance fitness of the first obstacle avoidance path according to the first navigation distance increase amplitude and the first angle change amplitude, the following formula is used for calculation: .in, For obstacle avoidance adaptation, and is the weight, which is used to balance the contribution of the increase in navigation distance and the change in angle to the obstacle avoidance adaptability. is the increase in the navigation distance, which indicates the increase in the navigation distance of the first obstacle avoidance path relative to the original planned navigation path. is the angle change amplitude, which indicates the change amplitude of the angle between the sail and the wind direction of the first obstacle avoidance path relative to the original planned navigation path. The smaller the angle, the greater the change in the angle The smaller it is, the better the obstacle avoidance adaptability is. The larger it is, the more it meets the goal of optimizing the obstacle avoidance path, that is, while avoiding obstacles, shorten the sailing distance of the obstacle avoidance path as much as possible, reduce the changes in the sailboat's heading and sail angle during the obstacle avoidance process, and reduce the impact of obstacle avoidance on navigation control.

[0057] By calculating the obstacle avoidance fitness, the candidate obstacle avoidance paths are quantitatively evaluated and compared. In the process of optimizing the obstacle avoidance path, the path with the largest obstacle avoidance fitness is searched as the optimal obstacle avoidance path. The introduction of obstacle avoidance fitness provides a clear optimization goal and evaluation index for the optimization of the obstacle avoidance path, which helps to improve the rationality and effectiveness of obstacle avoidance decisions.

[0058] In summary, the unmanned sailboat obstacle avoidance method with real-time path replanning provided in the embodiment of the present application has the following technical effects:

[0059] Through obstacle sensors and visual equipment, the sensor monitoring information sequence and visual monitoring information sequence are continuously collected during the navigation of the unmanned sailboat, and the wind condition information is monitored and obtained through the wind condition sensor, providing the necessary data basis for obstacle identification, movement path prediction and obstacle avoidance path planning. According to the sensor monitoring information sequence and the visual monitoring information sequence, the obstacle identification is carried out to obtain the obstacle position information sequence, and the obstacle movement path information is analyzed and obtained in combination with the navigation path of the unmanned sailboat. Through multi-source information fusion, the obstacle is accurately perceived and its movement trend is predicted, providing a basis for obstacle avoidance decision-making. When it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat, the obstacle avoidance area is obtained based on the intersection point according to the obstacle avoidance range, so that the subsequent path optimization can be carried out in a targeted manner. In the obstacle avoidance area, the obstacle avoidance path is optimized to obtain the optimal obstacle avoidance path, and the path is replanned and the unmanned sailboat is controlled to perform obstacle avoidance navigation. Among them, the obstacle avoidance path is optimized with the purpose of reducing the increased navigation distance of the unmanned sailboat obstacle avoidance path and reducing the angle change between the sail surface and the wind condition information, so as to achieve real-time and reliable obstacle avoidance of the unmanned sailboat.

[0060] Embodiment 2, based on the same inventive concept as the unmanned sailboat obstacle avoidance method of real-time path replanning in the above embodiment, Figure 2 As shown, the embodiment of the present application provides an unmanned sailboat obstacle avoidance device with real-time path replanning, the device comprising:

[0061] The information monitoring unit 11 is used to continuously collect sensor monitoring information sequences and visual monitoring information sequences during the navigation of the unmanned sailboat through obstacle sensors and visual equipment, and to monitor and obtain wind condition information through wind condition sensors;

[0062] The identification and analysis unit 12 is used to identify obstacles according to the sensor monitoring information sequence and the visual monitoring information sequence, obtain the obstacle position information sequence, and analyze and obtain the obstacle movement path information in combination with the navigation path of the unmanned sailboat;

[0063] The area division unit 13 is used to obtain an obstacle avoidance area by dividing the obstacle avoidance range based on the intersection point when it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat;

[0064] The path optimization unit 14 is used to optimize the obstacle avoidance path in the obstacle avoidance area, obtain the optimal obstacle avoidance path, re-plan the path and control the unmanned sailboat to perform obstacle avoidance navigation, wherein the obstacle avoidance path is optimized for the purpose of reducing the increased sailing distance of the unmanned sailboat's obstacle avoidance path and reducing the angle change between the sail surface and the wind condition information. Among them, the obstacle avoidance path distance of the obstacle avoidance path is obtained, and the planned path distance of the planned navigation path in the obstacle avoidance area is obtained, the difference between the obstacle avoidance path distance and the planned path distance is calculated as the increased navigation distance, and the navigation distance increase amplitude is calculated; the average planned angle between the sail surface in the planned navigation path and the wind condition information is obtained, and the average obstacle avoidance angle between the sail surface in the obstacle avoidance path and the wind condition information is obtained, and the angle change amplitude is calculated; according to the navigation distance increase amplitude and the angle change amplitude, the obstacle avoidance adaptability of the obstacle avoidance path is calculated to optimize the obstacle avoidance path.

[0065] Furthermore, the information monitoring unit 11 includes the following execution steps:

[0066] Through obstacle sensors and visual equipment, sensor monitoring information and visual monitoring information are continuously collected during the navigation of the unmanned sailboat; based on the sensor monitoring information and visual monitoring information at multiple moments in the past, sensor monitoring information sequences and visual monitoring information sequences are constructed; through wind condition sensors, wind direction information at multiple moments in the past is collected, and the average wind direction information is processed to obtain the wind condition information.

[0067] Furthermore, the identification and analysis unit 12 includes the following execution steps:

[0068] The method comprises obtaining a historical navigation position sequence of the unmanned sailboat at multiple moments in the past; identifying the relative position of the obstacle and the unmanned sailboat according to the sensor monitoring information sequence and the visual monitoring information sequence, and obtaining an obstacle relative direction sequence and an obstacle relative distance sequence; processing and obtaining an obstacle position information sequence according to the historical navigation position sequence, in combination with the obstacle relative direction sequence and the obstacle relative distance sequence; calculating and obtaining obstacle movement speed information according to the obstacle position information sequence and multiple moments, and obtaining obstacle movement path information by fitting in combination with the obstacle position information sequence.

[0069] Furthermore, the identification and analysis unit 12 further includes the following execution steps:

[0070] According to the historical data of obstacle recognition, a sample sensor monitoring information set, a sample visual monitoring information set, a sample relative direction set and a sample relative distance set are collected; the sample sensor monitoring information set and the sample visual monitoring information set are respectively used as input supervised training data, and the sample relative direction set and the sample relative distance set are used as output supervised training data to train the sensor recognition branch and the visual recognition branch respectively; the multiple sensor monitoring information and the multiple visual monitoring information in the sensor monitoring information sequence and the visual monitoring information sequence are respectively input into the sensor recognition branch and the visual recognition branch, and multiple sensor relative directions, multiple sensor relative distances, multiple visual relative directions and multiple visual relative distances are obtained as output; the multiple sensor relative directions and the multiple visual relative directions are weighted and calculated to obtain multiple relative directions, and the obstacle relative direction sequence is constructed; the multiple sensor relative distances and the multiple visual relative distances are weighted and calculated to obtain multiple relative distances, and the obstacle relative distance sequence is constructed.

[0071] Furthermore, the area division unit 13 includes the following execution steps:

[0072] The navigation speed information and the planned navigation path of the unmanned sailboat are obtained; according to the navigation speed information and the planned navigation path, an intersection judgment is performed with the obstacle movement path information; when it is judged that there is no intersection, obstacle avoidance is not performed; when it is judged that there is an intersection, an intersection point is obtained, and an obstacle avoidance area is obtained by dividing the intersection point as the center according to a preset obstacle avoidance range.

[0073] Furthermore, the path optimization unit 14 includes the following execution steps:

[0074] In the obstacle avoidance area, randomly select a first obstacle avoidance path that does not intersect with the obstacle movement path information; obtain a first obstacle avoidance path distance of the first obstacle avoidance path, and obtain a first planned path distance of the planned navigation path in the obstacle avoidance area, calculate the difference between the first obstacle avoidance path distance and the first planned path distance as a first increased navigation distance, and calculate to obtain a first navigation distance increase amplitude; obtain an average planned angle between the sail surface in the planned navigation path and the wind condition information, and obtain an average obstacle avoidance angle between the sail surface in the first obstacle avoidance path and the wind condition information, and calculate to obtain a first angle change amplitude; calculate to obtain a first obstacle avoidance fitness of the first obstacle avoidance path according to the first navigation distance increase amplitude and the first angle change amplitude; continue to optimize the obstacle avoidance path in the obstacle avoidance area until the obstacle avoidance path optimization converges, and output the obstacle avoidance path with the largest obstacle avoidance fitness as the optimal obstacle avoidance path.

[0075] Furthermore, according to the first navigation distance increase amplitude and the first angle change amplitude, a first obstacle avoidance fitness of the first obstacle avoidance path is calculated and obtained as follows:

[0076] ;

[0077] in, For obstacle avoidance adaptation, and is the weight, Add amplitude to the sailing distance, is the angle change amplitude.

[0078] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0079] Furthermore, the first or second mentioned above may not only represent an order relationship, but may also represent a specific concept, and / or refer to multiple elements that can be selected individually or in full. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these changes and variations.

Claims

1. A real-time path replanning unmanned sailboat obstacle avoidance method, characterized in that: The method comprises: Through obstacle sensors and visual equipment, the sensor monitoring information sequence and visual monitoring information sequence are continuously collected during the navigation of the unmanned sailboat, and the wind condition information is monitored and obtained through the wind condition sensor; According to the sensor monitoring information sequence and the visual monitoring information sequence, obstacle identification is performed to obtain an obstacle position information sequence, and combined with the navigation path of the unmanned sailboat, obstacle movement path information is analyzed and obtained; When it is determined that the obstacle moving path information will intersect with the planned navigation path of the unmanned sailboat, an obstacle avoidance area is obtained according to the obstacle avoidance range based on the intersection point; In the obstacle avoidance area, the obstacle avoidance path is optimized to obtain the optimal obstacle avoidance path, the path is replanned and the unmanned sailboat is controlled to perform obstacle avoidance navigation, wherein the fitness of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized for the purpose of reducing the increased sailing distance of the unmanned sailboat's obstacle avoidance path and reducing the change amplitude of the angle between the sail surface in the obstacle avoidance path and the wind condition information and the angle between the sail surface and the wind condition information in the planned navigation path; The obstacle avoidance path distance of the obstacle avoidance path is obtained, and the planned path distance of the planned navigation path in the obstacle avoidance area is obtained, the difference between the obstacle avoidance path distance and the planned path distance is calculated as the increased navigation distance, and the navigation distance increase amplitude is calculated; Obtaining an average planned angle between the sail surface in the planned navigation path and the wind condition information, and obtaining an average obstacle avoidance angle between the sail surface in the obstacle avoidance path and the wind condition information, and calculating an angle change amplitude; According to the increase in the navigation distance and the change in the angle, the obstacle avoidance adaptability of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized; Among them, according to the sensor monitoring information sequence and the visual monitoring information sequence, obstacle identification is performed to obtain the obstacle position information sequence, and combined with the navigation path of the unmanned sailboat, the obstacle movement path information is analyzed and obtained, including: Acquire a historical sailing position sequence of the unmanned sailboat at multiple moments in the past; According to the sensor monitoring information sequence and the visual monitoring information sequence, the relative position of the obstacle and the unmanned sailboat is identified to obtain the obstacle relative direction sequence and the obstacle relative distance sequence; According to the historical navigation position sequence, combined with the obstacle relative direction sequence and the obstacle relative distance sequence, an obstacle position information sequence is obtained by processing; According to the obstacle position information sequence and multiple moments, the obstacle moving speed information is calculated and obtained, and the obstacle moving path information is fitted by combining the obstacle position information sequence.

2. The unmanned sailboat obstacle avoidance method with real-time path replanning according to claim 1, characterized in that: Through obstacle sensors and visual equipment, the sensor monitoring information sequence and visual monitoring information sequence are continuously collected during the navigation of the unmanned sailboat, and the wind condition information is monitored and obtained through the wind condition sensor, including: Through obstacle sensors and visual equipment, the sensor monitoring information and visual monitoring information are continuously collected during the navigation of the unmanned sailboat; Constructing a sensor monitoring information sequence and a visual monitoring information sequence according to the sensor monitoring information and the visual monitoring information at multiple moments in the past; The wind direction information at multiple times in the past is collected through the wind condition sensor, and the average wind direction information is processed to obtain the wind condition information.

3. The unmanned sailboat obstacle avoidance method with real-time path replanning according to claim 1, characterized in that: According to the sensor monitoring information sequence and the visual monitoring information sequence, the relative position of the obstacle and the unmanned sailboat is identified to obtain the obstacle relative direction sequence and the obstacle relative distance sequence, including: According to the historical data of obstacle identification, a sample sensor monitoring information set, a sample visual monitoring information set, a sample relative direction set and a sample relative distance set are collected; The sample sensor monitoring information set and the sample visual monitoring information set are respectively used as input supervision training data, and the sample relative direction set and the sample relative distance set are used as output supervision training data to train the sensor recognition branch and the visual recognition branch respectively; Inputting the plurality of sensor monitoring information and the plurality of visual monitoring information in the sensor monitoring information sequence and the visual monitoring information sequence into the sensor recognition branch and the visual recognition branch respectively, and outputting a plurality of sensor relative directions, a plurality of sensor relative distances, a plurality of visual relative directions, and a plurality of visual relative distances; The plurality of sensing relative directions and the plurality of visual relative directions are weightedly calculated to obtain a plurality of relative directions, and the obstacle relative direction sequence is constructed; the plurality of sensing relative distances and the plurality of visual relative distances are weightedly calculated to obtain a plurality of relative distances, and the obstacle relative distance sequence is constructed.

4. The unmanned sailboat obstacle avoidance method with real-time path replanning according to claim 1, characterized in that: When it is determined that the obstacle moving path information will intersect with the planned navigation path of the unmanned sailboat, obtaining an obstacle avoidance area according to the obstacle avoidance range based on the intersection point includes: Obtaining the sailing speed information of the unmanned sailboat and planning the sailing path; According to the navigation speed information and the planned navigation path, an intersection determination is performed with the obstacle movement path information; When it is determined that there is no intersection, obstacle avoidance is not performed. When it is determined that there is an intersection, the intersection point is obtained, and the obstacle avoidance area is obtained by dividing the intersection point as the center according to the preset obstacle avoidance range.

5. The unmanned sailboat obstacle avoidance method with real-time path replanning according to claim 1, characterized in that: In the obstacle avoidance area, optimizing the obstacle avoidance path to obtain the optimal obstacle avoidance path includes: In the obstacle avoidance area, randomly selecting a first obstacle avoidance path that does not intersect with the obstacle movement path information; Acquire a first obstacle avoidance path distance of the first obstacle avoidance path, and acquire a first planned path distance of the planned navigation path in the obstacle avoidance area, calculate a difference between the first obstacle avoidance path distance and the first planned path distance as a first increased navigation distance, and calculate to obtain a first navigation distance increase amplitude; Obtaining an average planned angle between the sail surface in the planned navigation path and the wind condition information, and obtaining an average obstacle avoidance angle between the sail surface in the first obstacle avoidance path and the wind condition information, and calculating a first angle change amplitude; Calculating a first obstacle avoidance fitness of the first obstacle avoidance path according to the first navigation distance increase amplitude and the first angle change amplitude; Continue to optimize the obstacle avoidance path in the obstacle avoidance area until the obstacle avoidance path optimization converges, and output the obstacle avoidance path with the maximum obstacle avoidance fitness as the optimal obstacle avoidance path.

6. The unmanned sailboat obstacle avoidance method with real-time path replanning according to claim 5, characterized in that: According to the first navigation distance increase amplitude and the first angle change amplitude, the first obstacle avoidance fitness of the first obstacle avoidance path is calculated and obtained as follows: ； in, For obstacle avoidance adaptation, and is the weight, Add amplitude to the sailing distance, is the angle change amplitude.

7. An unmanned sailboat obstacle avoidance device with real-time path replanning, characterized in that: An unmanned sailboat obstacle avoidance method for implementing a real-time path replanning as described in any one of claims 1 to 6, the device comprising: An information monitoring unit, the information monitoring unit is used to continuously collect sensor monitoring information sequences and visual monitoring information sequences during the navigation of the unmanned sailboat through obstacle sensors and visual equipment, and to monitor and obtain wind condition information through a wind condition sensor; An identification and analysis unit, the identification and analysis unit is used to identify obstacles according to the sensor monitoring information sequence and the visual monitoring information sequence, obtain an obstacle position information sequence, and analyze and obtain obstacle movement path information in combination with the navigation path of the unmanned sailboat; A region division unit, the region division unit is used to obtain an obstacle avoidance region by dividing the obstacle avoidance range based on the intersection point when it is determined that the obstacle movement path information will intersect with the planned navigation path of the unmanned sailboat; A path optimization unit, the path optimization unit is used to optimize the obstacle avoidance path in the obstacle avoidance area, obtain the optimal obstacle avoidance path, re-plan the path and control the unmanned sailboat to perform obstacle avoidance navigation, wherein the fitness of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized for the purpose of reducing the increased sailing distance of the unmanned sailboat's obstacle avoidance path and reducing the change amplitude of the angle between the sail surface in the obstacle avoidance path and the wind condition information and the angle between the sail surface and the wind condition information in the planned navigation path; The obstacle avoidance path distance of the obstacle avoidance path is obtained, and the planned path distance of the planned navigation path in the obstacle avoidance area is obtained, the difference between the obstacle avoidance path distance and the planned path distance is calculated as the increased navigation distance, and the navigation distance increase amplitude is calculated; Obtaining an average planned angle between the sail surface in the planned navigation path and the wind condition information, and obtaining an average obstacle avoidance angle between the sail surface in the obstacle avoidance path and the wind condition information, and calculating an angle change amplitude; According to the increase in the navigation distance and the change in the angle, the obstacle avoidance adaptability of the obstacle avoidance path is calculated and the obstacle avoidance path is optimized.

Images