An unmanned ship water area monitoring method based on an improved SLAM algorithm

Abstract

The application provides an unmanned ship water area monitoring method based on an improved SLAM algorithm, comprising: detecting obstacles using a sonar system of the unmanned ship, extracting obstacle features, determining obstacle types and dynamic characteristics thereof, and calculating a shortest safe distance of the unmanned ship or adjusting a route; comparing target recognition results with historical data according to obstacle features extracted from sonar echo data, and judging common behavior patterns of the obstacles; if the obstacles are dynamic objects, analyzing speeds and directions thereof, constructing an obstacle position prediction model, predicting future positions, and adjusting obstacle avoidance strategies to shorten obstacle avoidance time; determining an expected position of the unmanned ship according to characteristic data of the dynamic obstacles and operation characteristics of the unmanned ship, and planning a new route; and generating a dynamic path graph of the obstacles and optimizing obstacle avoidance strategies by applying the obstacle position prediction model according to obstacle monitoring and environmental analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to an unmanned ship water area monitoring method based on an improved SLAM algorithm. BACKGROUND

[0002] With the increase of global marine activities and the expansion of marine resource utilization, unmanned ships play an increasingly important role in marine monitoring, scientific research, resource exploration and other fields. Especially in complex and variable marine environments, efficient and accurate marine monitoring becomes the key to ensuring the safety of marine activities and improving the efficiency of resource utilization. However, the autonomous navigation and obstacle avoidance capabilities of unmanned ships still face many challenges. First, the existing unmanned ship navigation system is difficult to accurately predict and avoid dynamic and static obstacles in complex marine environments, especially in multi-obstacle environments. Traditional obstacle avoidance methods are usually based on preset rules and static environmental models, which are difficult to cope with variable marine environments and the emergence of unknown obstacles. Second, the energy efficiency of unmanned ships is another important technical challenge. Unmanned ships usually rely on limited power or fuel, so optimizing energy use, extending range and performing more tasks become important considerations. Existing technologies are insufficient in dynamically adjusting speed and route to improve energy efficiency. In addition, the environmental adaptability of unmanned ships in complex marine environments is also a key issue. The marine environment is unpredictable, including weather, currents, waves and other factors, which affect the safety and efficiency of unmanned ship navigation. Existing technologies need to be improved in real-time perception of environmental changes and adjustment of navigation strategies accordingly. Furthermore, existing sonar systems can identify underwater obstacles to some extent, but they still have limitations in distinguishing between marine organisms and non-biological obstacles, and in dealing with noise interference in complex marine acoustic fields. This limits the application of unmanned ships in areas with rich biodiversity. Finally, the fusion and analysis of multi-source data is a shortcoming in existing technologies. Unmanned ships need to combine GPS, sonar, vision and other multi-sensing data to achieve more accurate environmental perception and obstacle identification, but how to effectively integrate these heterogeneous data and how to extract useful information from them to guide navigation is still a challenge. Therefore, the existing unmanned ship technology has certain limitations in obstacle avoidance capability, energy efficiency, environmental adaptability, obstacle identification accuracy and multi-source data fusion and analysis capability in multi-obstacle environments. Solving these problems is crucial to improving the safety and efficiency of unmanned ship navigation. SUMMARY

[0003] The present application provides an unmanned ship water area monitoring method based on an improved SLAM algorithm, mainly comprising:

[0004] The sonar system of the unmanned ship detects obstacles, extracts obstacle features, determines the type and dynamic characteristics of the obstacles, and calculates the shortest safe distance or adjusts the route of the unmanned ship; according to the obstacle features extracted from the sonar echo data, the target recognition result is compared with the historical data to judge the common behavior mode of the obstacle; if the obstacle is a dynamic object, its speed and direction are analyzed, a position prediction model of the obstacle is constructed, the future position is predicted, and the obstacle avoidance strategy is adjusted to shorten the obstacle avoidance time; according to the characteristic data of the dynamic obstacle and the operation characteristics of the unmanned ship, the expected position of the unmanned ship is determined, and a new route is planned; according to the GPS positioning of the unmanned ship and the sonar data, the potential collision point is determined, the dynamic route planning is carried out, and the effect of the route adjustment strategy is evaluated by comparing the actual obstacle avoidance time with the predicted obstacle avoidance time; according to the obstacle data detected by the sonar system, the risk of the obstacle is evaluated using a decision tree algorithm, and according to the comparison of the calculated obstacle avoidance time and the scheduled task time, the route planning of the unmanned ship is adjusted to ensure the safety and efficiency of the task; according to the obstacle monitoring and environmental analysis, the obstacle position prediction model is applied to generate the dynamic path diagram of the obstacle and optimize the obstacle avoidance strategy; according to the adjustment of the obstacle avoidance strategy and the route planning, the task execution progress of the unmanned ship is evaluated, and if the predicted delay is possible, the route planning is optimized to ensure the best efficiency of the task execution of the unmanned ship.

[0005] In some embodiments, the sonar system of the unmanned ship detects obstacles, extracts obstacle features, determines the type and dynamic characteristics of the obstacles, and calculates the shortest safe distance or adjusts the route of the unmanned ship, comprising:

[0006] The sonar system of the unmanned ship is used to detect the surrounding environment and obtain sonar echo data; when the sonar system of the unmanned ship detects a fixed obstacle, GPS data is obtained, and based on the GPS and sonar data, the initial position and direction of the unmanned ship and the position of the obstacle are determined; the fast Fourier transform is used to extract the obstacle features from the sonar echo data, including estimating the size and shape of the obstacle based on the echo intensity and duration, inferring the possible material of the obstacle through the propagation speed and attenuation characteristics of sound waves in different materials, and determining the movement speed and direction of the obstacle according to the Doppler shift of the echo signal; the SLAM algorithm is used to construct a SLAM map containing the position and features of the obstacle, with the current position of the unmanned ship as the center, in combination with the sonar and GPS data; the iterative closest point algorithm is used to register the data of continuous scanning and accurately map the position of the obstacle; according to the extracted movement speed of the obstacle, it is judged whether the obstacle is a fixed or dynamic object; if the obstacle is a fixed object, the shortest safe distance between the unmanned ship and the obstacle is calculated according to the current position of the unmanned ship and the target point, and the optimal angle for bypassing the obstacle is determined; if the obstacle is a dynamic object, the position and speed information of the obstacle is updated in real time; further comprising: using the SLAM algorithm, constructing a SLAM map in real time and calculating the shortest safe distance between the unmanned ship and the fixed obstacle according to the sonar data and GPS data.

[0007] The SLAM algorithm is used to construct a SLAM map in real time and calculate the shortest safe distance between the unmanned ship and the fixed obstacle according to the sonar data and GPS data, specifically including:

[0008] The SLAM algorithm is used to construct a real-time SLAM map of the surrounding environment in combination with the sonar and GPS data, and the map contains the accurately positioned position and size of the fixed obstacle. Through the SLAM algorithm, the position and direction of the unmanned ship relative to the obstacle are updated in real time. Through continuous data updating, the SLAM algorithm is used to track the movement of the obstacle relative to the unmanned ship. Based on the SLAM map and the dynamic data of the unmanned ship, including speed, turning radius, in combination with the stopping distance and maneuvering ability of the unmanned ship in emergency situations, the SLAM algorithm is used to calculate the shortest safe distance between the unmanned ship and the obstacle. According to the shortest safe distance between the unmanned ship and the obstacle, the optimal angle for bypassing the obstacle is determined.

[0009] In some embodiments, the obstacle features extracted from the sonar echo data are compared with the target recognition results and historical data to judge the common behavior patterns of the obstacle, including:

[0010] Obstacle features extracted from sonar echo data are obtained, and the current obstacle features are compared with obstacle features recorded in historical data. If the current obstacle features match the obstacle features in the historical data, behavior pattern data of the obstacle, including moving speed, path, and dwell time, are extracted. The behavior pattern data are integrated into the data structure of the map, and the obstacle information in the SLAM map is updated. Through comparison of new and old data, the operation of checking the logical consistency of the data is performed to check the updated SLAM map and ensure the accuracy and consistency of the information.

[0011] In some embodiments, if the obstacle is a dynamic object, the speed and direction of the obstacle are analyzed, a prediction model of the position of the obstacle is constructed, a future position is predicted, and an obstacle avoidance strategy is adjusted to shorten the obstacle avoidance time, including:

[0012] According to the moving speed of the obstacle extracted from the sonar echo data, it is judged whether the obstacle is a dynamic object. According to the historical speed and position data of the obstacle, a prediction model of the position of the obstacle is constructed by using an ARIMA algorithm for model training, and a future position of the obstacle is predicted. According to the predicted position, a preliminary obstacle avoidance strategy is designed and a route is planned. After the route planning is performed, it is monitored whether the actual obstacle avoidance time meets the expectation. If the actual obstacle avoidance time is greater than a preset threshold, the position of the obstacle is predicted again, and the parameters of the prediction model of the position of the obstacle are adjusted. According to the adjusted position prediction, the obstacle avoidance route is replanned. The new obstacle avoidance route is simulated to verify the efficiency and safety of the obstacle avoidance. If the simulation result shows that the new route meets the safety requirement, the obstacle avoidance route is solidified, and the obstacle avoidance strategy is updated in the system. After the obstacle avoidance strategy is deployed, real-time data is continuously monitored to ensure that the obstacle avoidance route is still effective under changing environmental conditions.

[0013] In some embodiments, the expected position of the unmanned ship is determined according to the feature data of the dynamic obstacle and the operating characteristics of the unmanned ship, and a new route is planned, including:

[0014] According to the feature data of the current dynamic obstacle, the prediction model of the position of the obstacle is used to obtain the predicted position and moving trajectory of the obstacle. According to the current position and speed of the unmanned ship, the operating characteristics of the unmanned ship, the predicted position and trajectory data of the obstacle, the expected positions of the unmanned ship and the obstacle at different time points are calculated, the minimum distance between the unmanned ship and the obstacle is obtained, and the minimum reaction distance required in an emergency is evaluated. The operating characteristics of the unmanned ship include braking distance and turning radius. If the calculated minimum distance is less than the minimum reaction distance required in an emergency, a new route is automatically planned. According to the new route, the heading of the unmanned ship is adjusted, and the distance to the surrounding obstacles is reevaluated. Environmental changes are continuously monitored, and the position of the obstacle and the route data are updated in real time. If the environmental changes result in the emergence of new risks, the risks are immediately reevaluated and the route is adjusted. If the risks continue to rise, an emergency braking program is started.

[0015] In some embodiments, the method of determining potential collision points based on GPS positioning and sonar data, dynamic route planning, and evaluating the effectiveness of route adjustment strategies by comparing actual obstacle avoidance time with predicted obstacle avoidance time includes:

[0016] Using SLAM algorithm, combining GPS positioning and sonar data, real-time updating SLAM map, including the location and movement trajectory of potential obstacles; based on SLAM map, predicting the coordinates and predicted collision time of potential collision points; using Dijkstra algorithm, planning a new route to avoid collision based on the coordinates and time of potential collision points; obtaining new route planning data, including predicted obstacle avoidance time, path length and turning angle; implementing new route through automatic navigation system, while using GPS and sonar system to monitor the motion trajectory of unmanned ship in actual environment; obtaining and recording the actual obstacle avoidance time after the unmanned ship executes the new route; judging the effectiveness of route adjustment strategy by calculating the time difference between predicted obstacle avoidance time and actual obstacle avoidance time; using the method of calculating standard deviation and variance, quantitatively analyzing the obstacle avoidance time data before and after implementation; evaluating the efficiency of dynamic path planning based on Dijkstra algorithm through quantitative analysis results; if the analysis shows that the obstacle avoidance time reduction value is greater than the preset threshold, the algorithm is determined to be effective, otherwise, it is determined that the route needs to be further adjusted; if quantitative analysis shows that the route of the unmanned ship needs to be further optimized, the environment map is updated again using SLAM algorithm, combined with the current state of the unmanned ship and obstacle data, and the improved Dijkstra algorithm is used to calculate the new route planning again; also including: based on the position, speed and energy efficiency data of the unmanned ship, combined with environmental monitoring and obstacle identification, evaluating the potential impact on navigation, real-time adjusting the navigation strategy of the unmanned ship and monitoring environmental changes.

[0017] The method of evaluating the potential impact on navigation based on the position, speed and energy efficiency data of the unmanned ship, combined with environmental monitoring and obstacle identification, real-time adjusting the navigation strategy of the unmanned ship and monitoring environmental changes, specifically includes:

[0018] Obtaining the current position, speed, and energy consumption rate data of the unmanned ship. At the same time, monitoring the surrounding environment, including sea current conditions, wind speed. Using sonar and radar data, identifying the types of surrounding obstacles, including ships, buoys, and natural obstacles. Evaluating the potential impact of obstacles on navigation, including possible collision risks. Using the formula The current energy efficiency is calculated. Wherein d is the distance between the unmanned ship and the obstacle, k is the influence degree of speed on energy efficiency, s is the speed, and O is the influence coefficient of obstacle type on energy efficiency. According to the calculation result of energy efficiency, the speed of the unmanned ship is dynamically adjusted. The dynamic programming algorithm is used to determine the optimal route according to the current position, the predetermined destination, the energy efficiency and the obstacle information. The planned route is input into the navigation system of the unmanned ship. The sailing performance is monitored in real time, including energy consumption, speed and heading, to ensure that the optimal route is followed. According to the data obtained in actual sailing, the parameters in the energy efficiency formula are adjusted to adapt to the changing marine environment and actual energy consumption.

[0019] In some embodiments, the obstacle data detected by the sonar system is used to assess the risk of the obstacle using a decision tree algorithm, and the route planning of the unmanned ship is adjusted to ensure task safety and task efficiency according to the comparison of the calculated obstacle avoidance time and the scheduled task time, including:

[0020] The sonar system of the unmanned ship is used to detect the surrounding environment, obtain sonar echo data, identify the existence of obstacles and their positions; if the sonar system detects a new obstacle, fast Fourier transform is used to extract obstacle features from the sonar echo data, including size, shape and material, speed and direction; according to the obstacle feature data, a decision tree algorithm is used for model training to assess the risk level of the obstacle, including high risk, medium risk and low risk; according to the risk level of the obstacle, the safe distance of the obstacle is determined, and the time required to avoid the obstacle and maintain a safe distance is calculated according to the current sailing data of the unmanned ship and the obstacle information; the calculation result of the obstacle avoidance time is obtained and compared with the scheduled task time; if the calculated obstacle avoidance time is within the range of the scheduled task time, a new route is planned to ensure that the obstacle is avoided and the task is not delayed, and it is checked whether the route meets the safety protocol standards, including the minimum safe distance from other ships and compliance with maritime rules; if the calculated obstacle avoidance time exceeds the scheduled task time, the obstacle avoidance time is recalculated according to the improved obstacle data and the state data of the unmanned ship, and the route planning is adjusted to reduce the obstacle avoidance time; further including: adjusting the obstacle avoidance strategy and emergency response of the unmanned ship according to multi-obstacle monitoring and obstacle behavior prediction; optimizing the obstacle avoidance efficiency of the unmanned ship according to multi-source data analysis and dynamic environment assessment.

[0021] The obstacle avoidance strategy and emergency response of the unmanned ship are adjusted according to multi-obstacle monitoring and obstacle behavior prediction, specifically including:

[0022] Integrate multi-modal perception systems on the unmanned ship, including enhanced sonar, radar, and optical sensors. Real-time acquisition and analysis of surrounding environment data, including the relative position, size, speed, and predicted trajectory of obstacles. Use Kalman filter to evaluate the dynamic interaction between multiple obstacles and potential collision risks, identify the intersection of obstacle motion trajectories and potential dangerous areas. Develop a priority-based cooperative obstacle avoidance strategy, assign a priority to each obstacle, sort them according to their potential impact on navigation safety, determine which obstacles need to be avoided first and which obstacles need to be processed later. Develop an obstacle avoidance strategy for complex scenarios, such as multi-ship dense areas. Apply A-Star algorithm based on obstacle constraints to calculate and adjust the optimal route to avoid all obstacles in real time, according to the size, turning ability and current speed of the unmanned ship, as well as the predicted changes in sea current and wind speed. Adjust the route according to the current state of the unmanned ship and dynamic environmental factors, including changes in sea current and wind speed. Based on historical data and real-time data, use ARIMA algorithm for model training to predict the future behavior and path of obstacles, regularly update and train the model to adapt to different types of marine environment and obstacle behavior patterns. Set a safety margin in route planning to ensure a safe distance even in the case of prediction errors or sudden changes in obstacle behavior. Set up an emergency response mechanism, including automatic deceleration or change of direction, to deal with unpredictable situations.

[0023] The optimization of unmanned ship obstacle avoidance efficiency based on multi-source data analysis and dynamic environment assessment includes:

[0024] Obtain the current position D, speed V and surrounding environment information of the unmanned ship through the sonar system, GPS, radar, identify and record the detected obstacle type and movement characteristics M. Obtain additional information R from radar and other sensors, including marine environmental conditions and other factors that may affect navigation. Use the formula Calculate the time required for obstacle avoidance. According to the time estimation result, use dynamic programming algorithm to determine the optimal obstacle avoidance route according to the operation limitations of the unmanned ship and environmental constraints, including avoiding high-risk areas and complying with maritime rules, to minimize the obstacle avoidance time and maintain a safe distance. Apply the optimized route to the navigation system of the unmanned ship and start execution. Monitor the navigation performance of the unmanned ship in real time, including speed, heading and energy consumption, and compare it with the expected obstacle avoidance time. According to the real-time monitoring data and environmental changes, continuously optimize the parameters in the obstacle avoidance time calculation formula.

[0025] In some embodiments, the application of obstacle position prediction model for obstacle dynamic path generation and obstacle avoidance strategy optimization based on obstacle monitoring and environmental analysis includes:

[0026] Obtain the position of obstacles through GPS and sonar systems and generate coordinate data of the obstacles; use an obstacle position prediction model to predict the position of obstacles in the future according to obstacle feature data; monitor the position changes of obstacles in real time and obtain real-time position data of obstacles through sensors; compare with the prediction data and ensure the accuracy of the prediction model by calculating the difference between the predicted position and the real-time monitoring position; if the prediction and real-time data are consistent, use the predicted position data to generate a dynamic path graph of the obstacles; periodically scan the environment to identify new obstacles and update the position and quantity information of the obstacles; adjust the obstacle avoidance strategy by evaluating the position, motion direction and speed of the obstacles; apply a decision tree algorithm for model training according to the current environment and obstacle data to determine the optimal obstacle avoidance path and determine the type and degree of decision-making action, and the environmental data includes water flow, wind speed, and the position of other ships; continuously monitor environmental changes and dynamically adjust the obstacle avoidance strategy; further comprising: identifying marine organisms according to sonar signal processing and spectral analysis, and adjusting the obstacle avoidance strategy of the unmanned ship based on the obstacle position prediction model.

[0027] The obstacle avoidance strategy of the unmanned ship based on the obstacle position prediction model is specifically:

[0028] Implement a sonar signal processing flow, including depth filtering and spectral analysis, to identify and separate the sonar features of marine organisms. Create a marine organism sonar feature library to obtain sonar signal samples of different types of marine organisms. According to the sonar signal samples of different types of marine organisms, use a long short-term memory network algorithm for model pre-training to train classification and identification models for different types of marine organisms. Integrate the trained model into the unmanned ship navigation system to classify and identify the obtained sonar data in real time. Match the sonar signal with the biological sonar feature library to distinguish between obstacle and biological signals. Apply the obstacle position prediction model to the sonar signal identified as non-biological to predict the future position and movement trend of the obstacle. Adjust the obstacle avoidance strategy of the unmanned ship based on the predicted obstacle trajectory, including adjusting the route and speed.

[0029] In some embodiments, the adjustment of the obstacle avoidance strategy and the route planning evaluates the task execution progress of the unmanned ship, and if a delay is expected, optimizes the route planning to ensure the best efficiency of the task execution of the unmanned ship, including:

[0030] The unmanned ship task execution program is started, and navigation data is obtained in real time, including the current position, speed, path taken, and estimated arrival time; based on the obtained navigation data, an ARIMA algorithm is used for model training to predict the progress of the unmanned ship task execution, assess the progress of the current task, and generate a progress report including the estimated remaining navigation time and predicted arrival time; if the progress report indicates potential delay, based on the current position of the unmanned ship, the scheduled destination, the remaining distance, the current speed, and environmental factors, a scheduled arrival time is set, a dynamic programming algorithm is used to determine the optimal route; based on the data of the adjusted route and the route before adjustment, the execution efficiency of the adjusted route is evaluated, including the time efficiency and safety of the new and old routes; if the efficiency evaluation shows that the new route needs further optimization, the route planning is adjusted to further shorten the path or improve safety, and the route planning adjustment includes adjusting the path selection and navigation speed; the optimized route is confirmed and input into the navigation system of the unmanned ship to start task execution update.

[0031] The technical scheme provided by the embodiment of the present application can include the following beneficial effects:

[0032] The present application provides an unmanned ship water area monitoring method based on an improved SLAM algorithm. It significantly improves the autonomous navigation and obstacle avoidance capabilities of the unmanned ship in complex marine environments. Through the improved monitoring system, the unmanned ship can more accurately detect and identify various obstacles, whether static or dynamic. This capability enables the unmanned ship to safely and efficiently avoid potential dangers while optimizing the route. The present application can identify and adapt to various environmental changes, including weather, ocean currents, and other marine activities, ensuring that the unmanned ship can safely navigate even in complex and unpredictable conditions. In addition, by optimizing the route planning, the unmanned ship can complete the task while improving energy efficiency, extending the operation time and range. The present application not only enhances the obstacle avoidance capability of the unmanned ship, but also improves its adaptability and task execution efficiency in a changing marine environment. These improvements have important practical application value in a wide range of marine activities, such as scientific research, resource exploration, and environmental monitoring. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 A flowchart of an unmanned ship water area monitoring method based on an improved SLAM algorithm of the present application.

[0034] Figure 2 A schematic diagram of an unmanned ship water area monitoring method based on an improved SLAM algorithm of the present application.

[0035] Figure 3 Another schematic diagram of an unmanned ship water area monitoring method based on an improved SLAM algorithm of the present application. DETAILED DESCRIPTION

[0036] For a further understanding of the present application, reference will be made to the following detailed description of the application taken in conjunction with the accompanying drawings. The following detailed description of the application is made in connection with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the application and are not intended to limit the application in any way. It is further understood that the drawings are not drawn to scale and that the specific dimensions shown are only illustrative and not limiting.

[0037] The unmanned ship water area monitoring method based on the improved SLAM algorithm can specifically include the following steps.

[0038] S101, using the sonar system of the unmanned ship to detect obstacles, extracting obstacle features, determining the type and dynamic characteristics of the obstacles, and calculating the shortest safe distance of the unmanned ship or adjusting the route.

[0039] The sonar system of the unmanned ship is used to detect the surrounding environment and obtain sonar echo data. When the sonar system of the unmanned ship detects a fixed obstacle, GPS data is obtained, and based on the GPS and sonar data, the initial position and direction of the unmanned ship and the position of the obstacle are determined. Fast Fourier transform is used to extract obstacle features from sonar echo data, including estimating the size and shape of the obstacle based on echo intensity and duration, inferring the possible material of the obstacle by the propagation speed and attenuation characteristics of sound waves in different materials, and determining the motion speed and direction of the obstacle according to the Doppler shift of the echo signal; using the SLAM algorithm, combining the sonar and GPS data, a SLAM map containing the position and features of the obstacle is constructed with the current position of the unmanned ship as the center. The iterative closest point algorithm is used to register the continuously scanned data to accurately map the position of the obstacle. According to the extracted motion speed of the obstacle, it is judged whether the obstacle is a fixed or dynamic object. If the obstacle is a fixed object, the shortest safe distance between the unmanned ship and the obstacle is calculated according to the current position of the unmanned ship and the target point, and the optimal angle to bypass the obstacle is determined. If the obstacle is a dynamic object, the position and speed information of the obstacle is updated in real time.

[0040] For example, an unmanned ship is conducting a marine monitoring mission, the current speed of the unmanned ship is 5 knots, the position is determined by GPS as 30.1234 ° N, 50.5678 ° E, its sonar system detects a fixed obstacle ahead. The sonar system detects an echo signal with a duration of 0.5 seconds and an intensity higher than the environmental noise threshold. Combining sonar and GPS data, it is determined that the obstacle is located 100 meters ahead of the unmanned ship. Fast Fourier transform analysis shows that the frequency composition and decay pattern of the echo signal indicate that the obstacle may be a large metal buoy with an estimated size of 5 meters in diameter. The SLAM algorithm uses these data to construct a real-time map centered on the current position of the unmanned ship, containing the location and characteristics of the obstacle. Since there is no Doppler shift in the echo signal, it is judged that the obstacle is a fixed object. Based on the speed and maneuvering performance of the unmanned ship, and taking into account environmental factors such as sea currents and wind speed, the SLAM algorithm calculates that it is appropriate to maintain a safety distance of at least 20 meters from the obstacle. Then, the algorithm plans a new route, deviating 15 degrees to the right to bypass the obstacle while ensuring a safety distance of at least 20 meters from the obstacle. The unmanned ship automatically adjusts its heading and follows the new route, successfully avoiding the obstacle and continuing its marine monitoring mission.

[0041] Using the SLAM algorithm, a SLAM map is constructed in real time based on sonar data and GPS data, and the shortest safe distance between the unmanned ship and the fixed obstacle is calculated.

[0042] Using the SLAM algorithm, a real-time SLAM map of the surrounding environment is constructed based on sonar and GPS data, and the map contains the precise location and size of the fixed obstacle. Through the SLAM algorithm, the position and direction of the unmanned ship relative to the obstacle are updated in real time. Through continuous data updates, the SLAM algorithm is used to track the movement of the obstacle relative to the unmanned ship. Based on the SLAM map and dynamic data of the unmanned ship, including speed, turning radius, combined with the stopping distance and maneuvering ability of the unmanned ship in emergency situations, the SLAM algorithm is used to calculate the shortest safe distance between the unmanned ship and the obstacle. According to the shortest safe distance between the unmanned ship and the obstacle, the optimal angle to bypass the obstacle is determined.

[0043] For example, an unmanned surface vessel (USV) is conducting a river survey and needs to avoid floating obstacles such as pontoons and trees. Using a SLAM algorithm combined with sonar and GPS data, a real-time SLAM map is constructed, containing the precise location and size of fixed obstacles. One fixed tree obstacle is located at (-10,5) in the coordinate system and measures 2m x 2m. The USV's current position is (0,0) in the coordinate system, its speed is 1m / s, and its turning radius is 3m. The USV also needs to stop within 2m in an emergency and possesses good maneuverability. Based on the SLAM map and the USV's dynamic data, the SLAM algorithm can calculate the shortest safe distance between the USV and the obstacle. The USV continues forward at its current speed and turning radius. The SLAM algorithm predicts the shortest distance between the USV and the obstacle to be 5m. However, given the USV's maneuverability and the 2m stopping distance in an emergency, this predicted distance is less than the predicted stopping distance, meaning the USV cannot stop within the predicted distance and obstacle avoidance measures are necessary. Based on the calculation of the shortest safe distance, the optimal angle for bypassing the obstacle can be determined, and it is found to be 30 degrees. By adjusting the unmanned surface vessel's (USV) steering angle to 30 degrees, the USV can bypass the obstacle while maintaining a sufficient safe distance. Through continuous data updates and the application of the SLAM algorithm, the USV can track the relative position and movement of obstacles in real time, thereby achieving safe obstacle avoidance and navigation.

[0044] S102. Based on the obstacle features extracted from the sonar echo data, compare the target recognition results with historical data to determine the common behavior patterns of the obstacles.

[0045] Obstacle features are extracted from sonar echo data and compared with those recorded in historical data. If a match is found, the obstacle's behavior pattern data, including movement speed, path, and dwell time, is extracted. This behavior pattern data is then integrated into the map's data structure to update the obstacle information in the SLAM map. By comparing the new and old data and checking the logical consistency of the data, the updated SLAM map is validated to ensure the accuracy and consistency of the information.

[0046] For example, an unmanned vessel is navigating a busy waterway. The sonar system detects an object and extracts its features from the echo data, such as size, shape, and echo reflection intensity. The data shows the object is approximately 10 meters long and cylindrical. The detected features are matched against obstacle features in a historical database, revealing a match for a common small yacht model. Behavioral pattern data for this type of yacht is extracted from the database, such as average speed of 8 knots, typical path, and dwell time. This behavioral pattern data is then integrated into the SLAM map's data structure to update obstacle information on the map.

[0047] S103、If the obstacle is a dynamic object, analyze its speed and direction, construct an obstacle position prediction model, predict future position, and adjust the obstacle avoidance strategy to shorten the obstacle avoidance time.

[0048] According to the motion speed of the obstacle extracted from the sonar echo data, determine whether the obstacle is a dynamic object. According to the historical speed and position data of the obstacle, use ARIMA algorithm for model training, construct obstacle position prediction model, predict future position of the obstacle. According to the predicted position, design preliminary obstacle avoidance strategy and route planning. After executing route planning, monitor whether the actual obstacle avoidance time meets the expectation, if the actual obstacle avoidance time is greater than the preset threshold, re-predict the position of the obstacle, and adjust the parameters of the obstacle position prediction model. According to the adjusted position prediction, re-plan the obstacle avoidance route. Simulate the new obstacle avoidance route to verify the efficiency and safety of obstacle avoidance. If the simulation result shows that the new route meets the safety requirements, solidify the obstacle avoidance route, and update the obstacle avoidance strategy in the system. After the deployment of the obstacle avoidance strategy, continuously monitor the real-time data to ensure that the obstacle avoidance route is still effective under changing environmental conditions.

[0049] For example, an unmanned ship is sailing along the coastline, its current position is 34.00° N, 120.00° E, and its speed is 20 knots. The sonar system of the unmanned ship detects the echo of an object, through analyzing the sonar data, it is found that the object is a small speedboat that is moving. Collect the speed and position data of the speedboat in the past hour, find that its average speed is 25 knots. Use ARIMA algorithm to train a model based on these data, predict the position of the speedboat in the next 30 minutes, the prediction result shows that the speedboat will cross the current route of the unmanned ship. Design a new route, change the heading by 15° to avoid the predicted collision point. After executing the new route, monitor the actual obstacle avoidance time, the estimated obstacle avoidance time is 10 minutes. The actual obstacle avoidance time is recorded as 12 minutes, which exceeds the preset threshold. Re-analyze the position and speed data of the speedboat, adjust the parameters of the ARIMA model. Recalculate the route, reduce the turning angle to 10°. Test the obstacle avoidance efficiency and safety of the new route using simulation software, the simulation result shows that the new route meets the safety requirements, and the obstacle avoidance time is reduced to 8 minutes. Deploy the new route to the navigation system of the unmanned ship. Continuously monitor the real-time data to ensure that the new route is still effective when the environmental conditions change.

[0050] S104、According to the characteristic data of the dynamic obstacle and the operating characteristics of the unmanned ship, determine the expected position of the unmanned ship, and plan a new route.

[0051] According to the characteristic data of the current dynamic obstacle, the predicted position and trajectory of the obstacle are obtained using an obstacle position prediction model. According to the current position and speed of the unmanned ship, the operating characteristics of the unmanned ship, the predicted position and trajectory data of the obstacle, the expected positions of the unmanned ship and the obstacle at different time points are calculated, the minimum distance between the unmanned ship and the obstacle is obtained, and the minimum reaction distance required in an emergency is evaluated. The operating characteristics of the unmanned ship include braking distance and turning radius. If the calculated minimum distance is less than the minimum reaction distance required in an emergency, a new route is automatically planned. According to the new route, the heading of the unmanned ship is adjusted, and the distance to the surrounding obstacles is re-evaluated. The environment is continuously monitored, and the obstacle position and route data are updated in real time. If the environment changes result in new risks, the risks are immediately re-evaluated and the route is adjusted. If the risk continues to rise, an emergency braking program is started.

[0052] For example, an unmanned ship is sailing in open waters, its current position is 45.00° N, 30.00° E, the speed is 25 knots, the braking distance of the unmanned ship is 500 meters, and the turning radius is 200 meters. The sonar system of the unmanned ship detects a small boat approaching the unmanned ship from 45.0020° N, 30.0030° E at a speed of 15 knots. Using the obstacle position prediction model, the movement trajectory of the small boat in the next 15 minutes is predicted. According to the speed of the unmanned ship and the small boat and the predicted trajectory, the expected positions of the two ships at the next time points are calculated. The calculated minimum distance is 400 meters, which is the closest distance between the two ships in the next 15 minutes. Since the calculated minimum distance of 400 meters is less than the minimum reaction distance of 500 meters required in an emergency, the system automatically plans a new route that increases the minimum distance between the two ships to 600 meters. The new route requires the unmanned ship to deviate from the original heading by 30°. The unmanned ship adjusts its heading and sails according to the new planned route. The system continuously monitors the actual positions of the unmanned ship and the small boat using GPS and sonar to ensure compliance with the new route. During the voyage, if the environment changes or new obstacles appear, the system will update the route data in real time. If a sudden risk leads to an increased risk of collision, the system will start an emergency braking program to ensure the safety of the ships.

[0053] S105, according to the GPS positioning and sonar data of the unmanned ship, determine the potential collision point, plan the dynamic route, and evaluate the effect of the route adjustment strategy by comparing the actual obstacle avoidance time with the predicted obstacle avoidance time.

[0054] The SLAM algorithm is used in combination with GPS positioning and sonar data to update the SLAM map in real time, including the location and movement trajectory of potential obstacles. Based on the SLAM map, the coordinates of potential collision points and the predicted collision time are predicted. The Dijkstra algorithm is adopted to plan a new collision-avoiding route based on the coordinates and time of potential collision points. The new route planning data, including the predicted obstacle avoidance time, path length, and turning angle, are obtained. The new route is implemented through the automatic navigation system, while the GPS and sonar systems are used to monitor the actual movement trajectory of the unmanned ship in the real environment. The actual obstacle avoidance time after the unmanned ship executes the new route is obtained and recorded. The effect of the route adjustment strategy is judged by calculating the time difference between the predicted and actual obstacle avoidance times. The standard deviation and variance are calculated to quantitatively analyze the obstacle avoidance time data before and after implementation based on the Dijkstra algorithm. The efficiency of the dynamic path planning based on the Dijkstra algorithm is evaluated through the quantitative analysis results. If the analysis shows that the obstacle avoidance time reduction is greater than the preset threshold, the algorithm is determined to be effective, otherwise, further adjustment of the route is required. If the quantitative analysis indicates that the route of the unmanned ship needs to be further optimized, the SLAM algorithm is used to update the environment map again, and the improved Dijkstra algorithm is used to calculate a new route planning based on the current state of the unmanned ship and obstacle data.

[0055] For example, using the SLAM algorithm in combination with GPS positioning and sonar data, the SLAM map is updated in real time, and based on the GPS and sonar data, it is found that a ship with a speed of 15 knots is moving from 35.0050° N, 25.0020° E towards the unmanned ship. Based on their speed and direction, it is predicted that the two ships will meet at 35.0030° N, 25.0010° E in about 10 minutes. Using the Dijkstra algorithm, a new route is calculated for the unmanned ship to avoid collision, considering the current position and the predicted collision point. The predicted obstacle avoidance time for the new route is 8 minutes, the path length increases to 5 kilometers, and the heading needs to be changed by 15°. The automatic navigation system of the unmanned ship adjusts the heading according to the new route. The GPS and sonar systems continuously monitor the actual movement trajectory of the unmanned ship. After implementing the new route, the unmanned ship successfully avoids the other ship, and the actual obstacle avoidance time is recorded as 7 minutes. Comparing the predicted obstacle avoidance time of 8 minutes with the actual obstacle avoidance time of 7 minutes, it is found that the actual time is shorter. By calculating the standard deviation and variance of the obstacle avoidance time before and after implementation, a low value is obtained, indicating that the route adjustment is very effective. The quantitative analysis results show that the Dijkstra algorithm successfully reduces the obstacle avoidance time. There is no need for further adjustment, confirming that the current route planning algorithm is efficient.

[0056] Based on the position, speed, and energy efficiency data of the unmanned ship, in combination with environmental monitoring and obstacle identification, the potential impact on navigation is evaluated, and the navigation strategy of the unmanned ship is adjusted in real time and the environmental changes are monitored.

[0057] The current position, speed, and energy consumption rate of the unmanned ship are obtained. The surrounding environment, including sea current conditions and wind speed, is monitored. Using sonar and radar data, the types of surrounding obstacles, including ships, buoys, and natural obstacles, are identified. The potential impact of obstacles on navigation, including possible collision risks, is evaluated. The current energy efficiency is calculated using the formula where d is the distance between the unmanned ship and the obstacle, k is the degree of influence of speed on energy efficiency, s is the speed, and O is the influence coefficient of obstacle type on energy efficiency. Based on the calculation results of energy efficiency, the speed of the unmanned ship is dynamically adjusted. Using a dynamic programming algorithm, the optimal route is determined based on the current position, the predetermined destination, energy efficiency, and obstacle information. The planned route is input into the navigation system of the unmanned ship. Real-time monitoring of navigation performance, including energy consumption, speed, and heading, ensures compliance with the optimal route. Based on the data obtained during actual navigation, the parameters in the energy efficiency formula are adjusted to adapt to the changing marine environment and actual energy consumption.

[0058] For example, an unmanned ship is heading to a destination 100 kilometers away. The current position of the unmanned ship is determined to be 20 kilometers from the starting point, with a speed of 15 knots and an energy consumption rate of 2 kilowatt-hours per kilometer. Environmental monitoring shows that the current sea current speed is 2 knots and the wind speed is 10 knots. Sonar and radar detect a fishing boat 15 kilometers away from the current position of the unmanned ship. The size, shape, and speed of the fishing boat indicate that it may cross the route of the unmanned ship, posing a collision risk. Using the formula the energy efficiency is calculated, where d is 15 kilometers, s is 15 knots, and O is 1.2. If k is 0.05, E = 0.06 is calculated. Based on the calculation results of energy efficiency, the system determines that the speed needs to be reduced to improve efficiency, and the speed is adjusted to 12 knots. Using a dynamic programming algorithm, a new route is recalculated to avoid the fishing boat, taking into account the new speed and energy efficiency. The new route is input into the navigation system of the unmanned ship, and the ship begins to navigate according to the new route. Real-time monitoring of energy consumption, speed, and heading ensures that the unmanned ship navigates according to the optimal route. Based on actual navigation data, the parameters in the energy efficiency formula are continuously adjusted to adapt to changing marine environments and energy consumption. Continuous monitoring of environmental changes and navigation performance ensures the safety and efficiency of the entire voyage.

[0059] S106, based on the obstacle data detected by the sonar system, a decision tree algorithm is used to evaluate the risk of the obstacle, and based on the comparison of the calculated obstacle avoidance time and the scheduled time of the task, the route planning of the unmanned ship is adjusted to ensure the safety and efficiency of the task.

[0060] The sonar system of the unmanned ship detects the surrounding environment, acquires sonar echo data, and identifies the presence and location of obstacles. If the sonar system detects a new obstacle, a fast Fourier transform is used to extract obstacle characteristics from the sonar echo data, including size, shape, and material, as well as movement speed and direction. Based on the obstacle characteristic data, a decision tree algorithm is used for model training to assess the risk level of the obstacle, which includes high risk, medium risk, and low risk. According to the risk level of the obstacle, a safe distance is determined, and based on the current navigation data of the unmanned ship and obstacle information, the time required to avoid the obstacle and maintain a safe distance is calculated. The calculation result of the time required to avoid obstacles is obtained and compared with the scheduled task time. If the calculated obstacle avoidance time is within the scheduled task time range, a new route is planned to ensure that the obstacle is bypassed without delaying the task, and the route is checked for compliance with safety protocol standards, including minimum safe distance from other ships and adherence to maritime rules. If the calculated obstacle avoidance time exceeds the scheduled task time, the obstacle avoidance time calculation is performed again based on the improved obstacle data and unmanned ship state data, and the route planning is adjusted to reduce the obstacle avoidance time.

[0061] For example, an unmanned ship is conducting a sea measurement task, currently located at 30.00° N, 45.00° E, and its task is to sail along a predetermined route to another point 100 kilometers away for data collection, with an estimated total sailing time of 5 hours. The ship's sonar system detects an object 2 kilometers ahead of the route, and through analysis of the sonar echo data, the presence and location of the obstacle are identified. The sonar echo data is processed using a fast Fourier transform to extract obstacle characteristics, which are approximately 10 meters x 5 meters in size, cylindrical in shape, made of metal, and stationary. A decision tree algorithm is applied to assess the risk level of the obstacle, and the result shows that it is of medium risk. Based on the ship's speed of 15 knots and the risk level of the obstacle, a minimum safe distance of 500 meters is determined. The additional time required to bypass the obstacle and maintain a safe distance is calculated to be 10 minutes. Comparing the additional obstacle avoidance time of 10 minutes with the scheduled task time of 5 hours, it is found that there will be no delay in the task. A new route is planned to ensure that the obstacle is bypassed without delaying the task. The new route requires a change in direction of 15°. The new route is checked for compliance with safety protocol standards, including maintaining a minimum safe distance from other ships and adhering to maritime rules. It is confirmed that the new route complies with all safety protocols. The ship adjusts its course according to the new route and bypasses the obstacle. Throughout the process, the system continues to monitor the environment and update the obstacle location and route data in real time to ensure safe navigation.

[0062] Based on multi-obstacle monitoring and obstacle behavior prediction, the obstacle avoidance strategy and emergency response of the unmanned ship are adjusted.

[0063] Integrate multi-modal perception systems on the unmanned ship, including enhanced sonar, radar, and optical sensors. Real-time acquisition and analysis of surrounding environment data, including the relative position, size, speed, and predicted trajectory of obstacles. Use Kalman filter to evaluate the dynamic interaction between multiple obstacles and potential collision risks, identify the intersection of obstacle motion trajectories and potential dangerous areas. Develop a priority-based cooperative obstacle avoidance strategy, assign a priority to each obstacle, sort them according to their potential impact on navigation safety, determine which obstacles need to be avoided first and which obstacles need to be processed later. Develop an obstacle avoidance strategy for complex scenarios, such as multi-ship dense areas. Apply A-Star algorithm based on obstacle constraints to calculate and adjust the optimal route to avoid all obstacles in real time according to the size, turning ability and current speed of the unmanned ship, as well as the predicted changes in sea current and wind speed. Adjust the route according to the current state of the unmanned ship and dynamic environmental factors, including changes in sea current and wind speed. Based on historical data and real-time data, use ARIMA algorithm for model training to predict the future behavior and path of obstacles, regularly update and train the model to adapt to different types of marine environment and obstacle behavior patterns. Set a safety margin in route planning to ensure a safe distance even in the case of prediction errors or sudden changes in obstacle behavior. Set up an emergency response mechanism, including automatic deceleration or change of direction, to deal with unpredictable situations.

[0064] For example, integrating a multi-modal perception system on an unmanned ship, including enhanced sonar, radar, and optical sensors, can provide real-time data on the surrounding environment. If the sonar detects an obstacle with a relative position of (10, 5) meters, a size of 2 meters, and a speed of 1 meter per second, and predicts that it will continue to move to the right. Through the Kalman filter, the dynamic interaction and potential collision risk between multiple obstacles are evaluated. If there is another obstacle with a relative position of (12, 6) meters, a size of 1 meter, and a speed of -5 meters per second, i.e. moving to the left. The Kalman filter can estimate the future positions of the two obstacles based on this information and evaluate the potential collision risk between them. According to the priority-based cooperative obstacle avoidance strategy, in order to ensure the safety of navigation, obstacle 1 is set as priority 1 and obstacle 2 is set as priority 2. The obstacle with priority 1 needs to be avoided first, while the obstacle with priority 2 is considered in subsequent processing. According to the A-Star algorithm and the size of the unmanned ship, turning ability, current speed, predicted sea current and wind speed changes, the optimal navigation route to avoid all obstacles is calculated to be a heading angle of 30 degrees. Use the ARIMA algorithm to predict the behavior and path of the obstacle. Based on historical data and real-time data, the ARIMA model trained by the model predicts that the future position of obstacle 1 will move to the right to (12, 6) meters, and obstacle 2 will continue to move to the left to (11, 6) meters. Set a safety margin in the route planning, if set to 2 meters. Even in the case of prediction errors or sudden changes in obstacle behavior, the unmanned ship can maintain a safe distance of 2 meters from the obstacle. Set up an emergency response mechanism, for example, when the unmanned ship approaches obstacle 1, automatically slow down and change direction to avoid collision.

[0065] According to multi-source data analysis and dynamic environment evaluation, the obstacle avoidance efficiency of the unmanned ship is optimized.

[0066] Through the sonar system, GPS, radar, the current position D and speed V of the unmanned ship and the information of the surrounding environment are obtained, and the type and movement characteristics M of the detected obstacles are identified and recorded. Additional information R is obtained from radar and other sensors, including marine environmental conditions and other factors that may affect navigation. Use the formula Calculate the time required for obstacle avoidance. According to the time estimation result, use the dynamic programming algorithm to determine the optimal obstacle avoidance route according to the operation restrictions and environmental constraints of the unmanned ship, including avoiding high-risk areas and complying with maritime rules, to minimize the obstacle avoidance time and maintain a safe distance. Apply the optimized route to the navigation system of the unmanned ship and start execution. Monitor the performance of the unmanned ship in real time, including speed, heading and energy consumption, and compare it with the expected obstacle avoidance time. According to the real-time monitoring data and environmental changes, continuously optimize the parameters in the obstacle avoidance time calculation formula.

[0067] For example, an unmanned surface vessel (USV) is conducting a scientific exploration mission in open waters. Its current location is 20.00°N, 130.00°E, and its speed is 12 knots. The USV's sonar, GPS, and radar systems detect a large cargo ship moving slowly at 5 knots, 1 kilometer ahead. Additional information obtained from radar and other sensors indicates that the visibility R around the USV is 0.935. Using the formula... The time required for obstacle avoidance is calculated, where D is the distance to the obstacle (1 km), V is the speed of the unmanned surface vessel (USV) (12 knots), and M is the speed of the obstacle (5 knots). The calculated time is T ≈ 0.055 hours, or approximately 3.3 minutes. Using a dynamic programming algorithm, considering the operational limitations and environmental constraints of the USV, a new route is planned to avoid the cargo ship while maintaining a safe distance. This new route requires the USV to deviate 30° from its original course. The new route is input into the USV's navigation system and execution begins. The USV's navigation performance, including speed, course, and energy consumption, is monitored in real time. The actual distance between the USV and the cargo ship is continuously monitored to ensure the obstacle avoidance action meets expectations. Based on real-time data, the parameters in the obstacle avoidance time calculation formula are adjusted to adapt to possible environmental changes.

[0068] S107. Based on obstacle monitoring and environmental analysis, apply the obstacle location prediction model to generate a dynamic obstacle path map and optimize the obstacle avoidance strategy.

[0069] Obstacle locations are acquired using GPS and sonar systems, generating their coordinate data. An obstacle location prediction model is used to predict the obstacle's position over a future period based on its characteristic data. Real-time obstacle location changes are monitored using sensors to acquire real-time obstacle location data. This data is compared with the predicted data; the difference between the predicted and real-time monitored positions is calculated to ensure the accuracy of the prediction model. If the prediction matches the real-time data, a dynamic path map of the obstacle is generated using the predicted location data. The environment is periodically scanned to identify new obstacles and update their location and quantity information. Obstacle avoidance strategies are adjusted by evaluating the obstacle's position, direction of movement, and speed. Based on current environmental and obstacle data, a decision tree algorithm is applied to train the model, determining the optimal obstacle avoidance path and the type and severity of the obstacle avoidance action. Environmental data includes water flow, wind speed, and the positions of other vessels. The obstacle avoidance strategy is dynamically adjusted based on continuous monitoring of environmental changes.

[0070] For example, an unmanned ship is navigating along the coastline, with the current position at 35.00° N, 140.00° E, and the ship's speed is 15 knots. The ship's GPS and sonar systems detect a small fishing boat 500 meters ahead, with the current coordinates at 35.0010° N, 140.0015° E. Using the obstacle position prediction model, analyze the fishing boat's characteristic data such as size and speed, and if the fishing boat's speed is 8 knots, predict the fishing boat's position in the next 30 minutes. Continuously monitor the fishing boat's position in real-time through sonar and GPS, compare the predicted position with the real-time monitored position, and find that the error is within 50 meters, the prediction model is relatively accurate. According to the prediction data, generate a dynamic path diagram of the fishing boat, showing its predicted route. Regularly scan the environment to confirm that there are no new obstacles nearby. Considering the direction and speed of the fishing boat's movement, as well as the current environmental conditions of the unmanned ship, such as wind speed of 5 knots and water flow of 2 knots, adjust the obstacle avoidance strategy. Use the decision tree algorithm to analyze the data and plan a new route that avoids the fishing boat and maintains a safe distance of at least 200 meters. Execute the new route while continuously monitoring environmental changes, including the dynamics of other ships and weather changes. Update the route in real-time to adapt to the complex marine environment and ensure the safe navigation of the unmanned ship.

[0071] According to the sonar signal processing and spectral analysis, identify marine organisms and adjust the obstacle avoidance strategy of the unmanned ship based on the obstacle position prediction model.

[0072] Implement the sonar signal processing flow, including depth filtering and spectral analysis, to identify and separate the sonar features of marine organisms. Create a marine organism sonar feature library to obtain sonar signal samples of different types of marine organisms. According to the sonar signal samples of different types of marine organisms, use the long short-term memory network algorithm for model pre-training, and train the classification and identification model for different types of marine organisms. Integrate the trained model into the unmanned ship navigation system to classify and identify the obtained sonar data in real-time. Match the sonar signal with the biological sonar feature library to distinguish between obstacle and biological signals. Apply the obstacle position prediction model to the sonar signals identified as non-biological to predict the future position and movement trend of the obstacle. Based on the predicted obstacle trajectory, adjust the obstacle avoidance strategy of the unmanned ship, including adjusting the route and speed.

[0073] For example, an unmanned ship is conducting environmental monitoring in a biodiversity-rich sea area. The ship's sonar system detects multiple echo signals, and performs depth filtering and spectral analysis to identify and separate the signals. From previous investigations and data collection, a marine bio-sonar feature library has been established, containing sonar signal samples of various fish, dolphins, and whales. Based on the sonar signal samples of different marine organisms, a long short-term memory network algorithm is used to train a classification and recognition model for these marine organisms. The trained long short-term memory network model is integrated into the ship's navigation system, which processes sonar data in real time to distinguish between obstacles and biological signals. A group of approaching dolphins and a small boat are identified, and the sonar signals are matched with the feature library to confirm the presence of the dolphin group, which is excluded from obstacles. At the same time, the small boat is confirmed as a non-biological obstacle. The sonar signal of the small boat is applied to the obstacle position prediction model to predict its future position and movement trend, predicting that the small boat will continue to move along its current path at a speed of 5 knots. Based on the predicted obstacle trajectory, the ship's obstacle avoidance strategy is adjusted, such as changing the route to avoid the small boat while maintaining a sufficient distance to observe the dolphin group.

[0074] S108, according to the adjustment of the obstacle avoidance strategy and the route planning, the progress of the unmanned ship's task execution is evaluated, and if the predicted delay is possible, the route planning is optimized to ensure the best efficiency of the unmanned ship's task execution.

[0075] The unmanned ship task execution program is started, and real-time navigation data is obtained, including current position, speed, path taken, and estimated arrival time. Based on the obtained navigation data, an ARIMA algorithm is used to train a model to predict the progress of the unmanned ship's task execution, evaluate the progress of the current task, and generate a progress report, including the estimated remaining navigation time and the predicted arrival time. If the progress report indicates potential delays, based on the current position of the unmanned ship, the scheduled destination, the remaining distance, the current speed, and environmental factors, a scheduled arrival time is set, and a dynamic programming algorithm is used to determine the optimal route. Based on the data of the adjusted route and the original route, the execution efficiency of the adjusted route is evaluated, including the time efficiency and safety of the new and old routes. If the efficiency evaluation shows that the new route needs further optimization, the route planning is adjusted to further shorten the path or improve safety, and the route planning adjustment includes adjusting the path selection and navigation speed. The optimized route is confirmed and input into the navigation system of the unmanned ship, and the task execution update is started.

[0076] For example, an unmanned ship is conducting a marine research mission, and needs to travel from the current position of 25.00° N, 150.00° E to a destination of 30.00° N, 160.00° E, with an estimated travel distance of 500 km, a speed of 15 knots, and an estimated travel time of 18 hours. The unmanned ship starts the mission execution program and collects real-time travel data. The current position is 25.00° N, 150.00° E, the speed is 15 knots, and 200 km has already been traveled, with an estimated time of arrival of 10.7 hours. Using the ARIMA algorithm to analyze the travel data and predict the progress of the mission, the progress report shows that due to the unfavorable sea currents, the estimated remaining travel time is 12 hours, which is 1.3 hours later than the original plan. Since the progress report indicates potential delays, a new route is planned using the dynamic programming algorithm based on the current position, destination, remaining distance, current speed affected by sea currents reduced to 12 knots, and environmental factors. The new route avoids unfavorable sea currents and estimates the travel time to be reduced to 11 hours, allowing the destination to be reached on time. Comparing the routes before and after adjustment, the new route improves the time efficiency, but increases the travel distance to 520 km to avoid unfavorable sea currents. The safety assessment shows that the new route avoids high-risk areas. According to the efficiency evaluation results, it is determined that the new route does not need further optimization. Confirming the optimized route, the new route is input into the navigation system of the unmanned ship. The unmanned ship starts to travel according to the updated route, avoiding unfavorable sea currents, and is expected to arrive at the destination on time.

[0077] The above description is only a preferred embodiment of one or more embodiments of the present specification, and does not limit one or more embodiments of the present specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of one or more embodiments of the present specification should be included in the scope of protection of one or more embodiments of the present specification.

Claims

1. A method for unmanned surface vessel (USV) water monitoring based on an improved SLAM algorithm, characterized in that, The method includes: The process involves using an unmanned surface vessel's (USV) sonar system to detect obstacles, extract obstacle features, determine obstacle types and dynamic characteristics, and calculate the USV's shortest safe distance or adjust its course. This includes: using a SLAM algorithm to construct a SLAM map in real-time based on sonar and GPS data and calculating the shortest safe distance between the USV and fixed obstacles; comparing the obstacle features extracted from sonar echo data with target recognition results and historical data to determine common obstacle behavior patterns; if the obstacle is a dynamic object, analyzing its speed and direction, constructing an obstacle position prediction model to predict its future position, and adjusting obstacle avoidance strategies to shorten avoidance time; and determining the USV's expected position and planning based on the dynamic obstacle's feature data and the USV's operational characteristics. New route planning; potential collision points are identified based on the unmanned surface vessel's (USV) GPS positioning and sonar data, and dynamic route planning is performed. The effectiveness of the route adjustment strategy is evaluated by comparing the actual obstacle avoidance time with the estimated obstacle avoidance time. Based on obstacle data detected by the sonar system, a decision tree algorithm is used to assess obstacle risks, and the USV's route planning is adjusted based on the comparison between the calculated obstacle avoidance time and the scheduled mission time to ensure mission safety and efficiency. Based on obstacle monitoring and environmental analysis, an obstacle location prediction model is applied to generate a dynamic obstacle path map and optimize the obstacle avoidance strategy. Based on the adjustments to the obstacle avoidance strategy and route planning, the mission execution progress of the USV is evaluated. If a potential delay is anticipated, the route planning is optimized to ensure the best efficiency of the USV's mission execution. The optimized route planning also includes assessing the potential impact on navigation based on the unmanned surface vessel's (USV) position, speed, and energy efficiency data, combined with environmental monitoring and obstacle identification. This involves adjusting the USV's navigation strategy in real time and monitoring environmental changes. Specifically, this includes: acquiring the USV's current position, speed, and energy consumption rate data; simultaneously monitoring the surrounding environment, including ocean currents and wind speed; using sonar and radar data to identify the types of surrounding obstacles, including vessels, buoys, and natural obstacles; assessing the potential impact of obstacles on navigation, including possible collision risks; and utilizing formulas... Calculate the current energy efficiency; where d is the distance between the unmanned vessel and obstacles, k is the influence of speed on energy efficiency, s is the speed, and O is the influence coefficient of obstacle type on energy efficiency; dynamically adjust the speed of the unmanned vessel based on the calculated energy efficiency; use a dynamic programming algorithm to determine the optimal route based on the current position, predetermined destination, energy efficiency, and obstacle information; input the planned route into the navigation system of the unmanned vessel; monitor navigation performance in real time, including energy consumption, speed, and heading, to ensure adherence to the optimal route; adjust the parameters in the energy efficiency formula based on data obtained during actual navigation to adapt to the constantly changing marine environment and actual energy consumption.

2. The method according to claim 1, wherein, The process of using the unmanned surface vessel's sonar system to detect obstacles, extract obstacle features, determine obstacle types and their dynamic characteristics, and calculate the shortest safe distance for the unmanned surface vessel or adjust its course includes: The unmanned surface vessel (USV) uses its sonar system to detect the surrounding environment and acquire sonar echo data. When the USV's sonar system detects a fixed obstacle, it acquires GPS data and, based on the GPS and sonar data, determines the USV's initial position and orientation, as well as the obstacle's location. A Fast Fourier Transform (FFT) is used to extract obstacle features from the sonar echo data, including estimating the obstacle's size and shape based on echo intensity and duration, inferring the obstacle's possible material through the propagation speed and attenuation characteristics of sound waves in different materials, and determining the obstacle's speed and direction based on the Doppler changes in the echo signal. A SLAM algorithm is used, combining sonar and GPS data, to construct a SLAM map centered on the USV's current location, containing the obstacle's position and features. An iterative nearest-point algorithm is used to register continuously scanned data, accurately mapping the obstacle's location. Based on the extracted obstacle's speed, it is determined whether the obstacle is a fixed or dynamic object. If the obstacle is a fixed object, calculate the shortest safe distance between the unmanned vessel and the obstacle based on the current position of the unmanned vessel and the target point, and determine the optimal angle to bypass the obstacle; If the obstacle is a dynamic object, update the obstacle's position and speed information in real time; The process of utilizing SLAM algorithms to construct a real-time SLAM map and calculate the shortest safe distance between the unmanned surface vessel (USV) and fixed obstacles based on sonar and GPS data includes: constructing a real-time SLAM map of the surrounding environment using SLAM algorithms combined with sonar and GPS data, the map containing the precise location and size of fixed obstacles; updating the USV's position and orientation relative to the obstacles in real-time using SLAM algorithms; tracking the movement of obstacles relative to the USV using SLAM algorithms through continuous data updates; calculating the shortest safe distance between the USV and obstacles using SLAM algorithms based on the SLAM map and the USV's dynamic data, including speed and turning radius, combined with the USV's stopping distance and maneuverability in emergency situations; and determining the optimal angle to bypass the obstacles based on the shortest safe distance between the USV and obstacles.

3. The method according to claim 1, wherein, The step of comparing the obstacle features extracted from sonar echo data with target recognition results and historical data to determine the common behavior patterns of the obstacle includes: Obtain obstacle features extracted from sonar echo data and compare the current obstacle features with obstacle features recorded in historical data; If the current obstacle features match the obstacle features in historical data, then extract the obstacle's behavior pattern data, including movement speed, path, and dwell time; integrate the behavior pattern data into the map's data structure, and update the obstacle information in the SLAM map; By comparing old and new data and checking the logical consistency of the data, the updated SLAM map is validated to ensure the accuracy and consistency of the information.

4. The method according to claim 1, wherein, If the obstacle is a dynamic object, its speed and direction are analyzed, an obstacle position prediction model is constructed, its future position is predicted, and the obstacle avoidance strategy is adjusted to shorten the obstacle avoidance time, including: Based on the speed of movement of the obstacle extracted from the sonar echo data, determine whether the obstacle is a dynamic object; Based on historical speed and location data of obstacles, an obstacle location prediction model is built using the ARIMA algorithm to predict the future location of obstacles. Based on the predicted location, a preliminary obstacle avoidance strategy is designed and route planning is performed. After route planning, the actual obstacle avoidance time is monitored to see if it meets expectations. If the actual obstacle avoidance time exceeds a preset threshold, the obstacle location prediction is re-performed, and the parameters of the obstacle location prediction model are adjusted. Based on the adjusted location prediction, a new obstacle avoidance route is planned. The new obstacle avoidance route is simulated to verify its efficiency and safety. If the simulation results show that the new route meets safety requirements, the obstacle avoidance route is fixed, and the obstacle avoidance strategy is updated in the system. After the obstacle avoidance strategy is deployed, real-time data is continuously monitored to ensure that the obstacle avoidance route remains effective under changing environmental conditions.

5. The method according to claim 1, wherein, The process of determining the expected position of the unmanned surface vessel (USV) and planning a new route based on the characteristic data of dynamic obstacles and the operational characteristics of the USV includes: Based on the current dynamic obstacle characteristic data, an obstacle position prediction model is used to obtain the predicted position and trajectory of the obstacle. Based on the current position and speed of the unmanned surface vessel (USV), its operational characteristics, and the predicted position and trajectory data of the obstacle, the expected positions of the USV and the obstacle at different time points are calculated to obtain the minimum distance between the USV and the obstacle. The minimum reaction distance required in an emergency is also assessed. The USV's operational characteristics include braking distance and turning radius. If the calculated minimum distance is less than the minimum reaction distance required in an emergency, a new route is automatically planned. Based on the new route, the USV's heading is adjusted, and the distance to surrounding obstacles is reassessed. Environmental changes are continuously monitored, and obstacle position and route data are updated in real time. If environmental changes lead to new risks, the risks are immediately reassessed, and the route is adjusted. If the risk continues to escalate, initiate emergency braking procedures.

6. The method according to claim 1, wherein, The process involves determining potential collision points based on the unmanned surface vessel's GPS positioning and sonar data, performing dynamic route planning, and evaluating the effectiveness of the route adjustment strategy by comparing the actual obstacle avoidance time with the estimated obstacle avoidance time. Using the SLAM algorithm, combined with GPS positioning and sonar data, the SLAM map is updated in real time, including the location and movement trajectory of potential obstacles; Based on SLAM maps, predict the coordinates of potential collision points and the collision time; Using Dijkstra's algorithm, a new collision-avoidance route is planned based on the coordinates and time of potential collision points. New route planning data, including estimated obstacle avoidance time, path length, and turning angle, is acquired. The new route is implemented via an automatic navigation system, while GPS and sonar systems monitor the unmanned surface vessel's (USV) trajectory in the actual environment. The actual obstacle avoidance time after the USV executes the new route is acquired and recorded. The effectiveness of the route adjustment strategy is assessed by calculating the time difference between the estimated and actual obstacle avoidance times. Quantitative analysis is performed based on the obstacle avoidance time data before and after implementation, using standard deviation and variance calculations. The efficiency of the dynamic path planning based on Dijkstra's algorithm is evaluated using the quantitative analysis results. If the analysis shows a reduction in obstacle avoidance time greater than a preset threshold, the algorithm is deemed effective; otherwise, further route adjustments are deemed necessary. If quantitative analysis indicates that further optimization of the unmanned surface vessel (USV) route planning is needed, the environmental map is updated again using the SLAM algorithm. Combining the current state of the USV and obstacle data, the improved Dijkstra algorithm is used to recalculate the new route planning.

7. The method according to claim 1, wherein, The process involves using a decision tree algorithm to assess obstacle risks based on obstacle data detected by the sonar system, and adjusting the unmanned surface vessel's (USV) course planning based on a comparison between the calculated obstacle avoidance time and the scheduled mission time to ensure mission safety and efficiency. This includes: The sonar system of the unmanned vessel is used to detect the surrounding environment, obtain sonar echo data, and identify the presence and location of obstacles. If the sonar system detects a new obstacle, it uses a Fast Fourier Transform to extract obstacle features from the sonar echo data, including size, shape, material, speed, and direction of movement. Based on the obstacle feature data, it uses a decision tree algorithm to train a model and assess the obstacle's risk level, which includes high, medium, and low risk. Based on the obstacle's risk level, it determines the safe distance from the obstacle and calculates the time required to avoid the obstacle and maintain a safe distance based on the unmanned vessel's current navigation data and obstacle information. The calculation result of the time required for obstacle avoidance is obtained and compared with the scheduled time of the mission; If the calculated obstacle avoidance time is within the mission's scheduled time, a new route is planned to ensure that obstacles are avoided without delaying the mission. The route is then checked to ensure compliance with safety protocols, including minimum safe distances from other vessels and adherence to maritime regulations. If the calculated obstacle avoidance time exceeds the mission's scheduled time, the time required for obstacle avoidance is recalculated based on improved obstacle data and unmanned vessel status data. The route is then adjusted to reduce the obstacle avoidance time. It also includes: adjusting the obstacle avoidance strategy and emergency response of unmanned surface vessels based on multi-obstacle monitoring and obstacle behavior prediction; and optimizing the obstacle avoidance efficiency of unmanned surface vessels based on multi-source data analysis and dynamic environmental assessment. The process of adjusting the obstacle avoidance strategy and emergency response of the unmanned surface vessel (USV) based on multi-obstacle monitoring and obstacle behavior prediction specifically includes: integrating a multimodal perception system on the USV, including enhanced sonar, radar, and optical sensors; acquiring and analyzing surrounding environmental data in real time, including the relative position, size, speed, and predicted trajectory of obstacles; using a Kalman filter to assess the dynamic interactions and potential collision risks between multiple obstacles, identifying intersections of obstacle trajectories and potential hazard areas; developing a priority-based cooperative obstacle avoidance strategy, assigning a priority to each obstacle, ranking them according to their potential impact on navigation safety, and determining obstacles that need to be avoided first and those requiring subsequent handling; and formulating obstacle avoidance strategies for complex scenarios, such as densely populated areas with multiple vessels. The system employs the A-Star algorithm with obstacle constraints to calculate and adjust the optimal route to avoid all obstacles in real time, based on the size, turning ability, and current speed of the unmanned surface vessel (USV), as well as predicted changes in ocean currents and wind speeds. The route is adjusted according to the USV's current state and dynamic environmental factors, including changes in ocean currents and wind speeds. Based on historical and real-time data, the system uses the ARIMA algorithm for model training to predict the future behavior and paths of obstacles, and regularly updates and trains the model to adapt to different types of marine environments and obstacle behavior patterns. A safety margin is set in the route planning to ensure a safe distance is maintained even in the event of prediction errors or sudden changes in obstacle behavior. An emergency response mechanism is established, including automatic deceleration or course changes, to cope with unpredictable situations. The optimization of obstacle avoidance efficiency for unmanned surface vessels (USVs) based on multi-source data analysis and dynamic environmental assessment specifically includes: acquiring the USV's current position (D), velocity (V), and surrounding environment information via sonar, GPS, and radar; identifying and recording the types and movement characteristics (M) of detected obstacles; acquiring additional information (R) from radar and other sensors, including marine environmental conditions and other factors that may affect navigation; and using formulas... Calculate the time required for obstacle avoidance; based on the time estimation results, use a dynamic programming algorithm to determine the optimal obstacle avoidance route according to the unmanned vessel's operational constraints and environmental constraints, in order to minimize obstacle avoidance time and maintain a safe distance. The unmanned vessel's operational constraints and environmental constraints include avoiding high-risk areas and complying with maritime regulations; apply the optimized route to the unmanned vessel's navigation system and begin execution; monitor the unmanned vessel's navigation performance in real time, including speed, heading, and energy consumption, and compare it with the expected obstacle avoidance time; continuously optimize the parameters in the obstacle avoidance time calculation formula based on real-time monitoring data and environmental changes.

8. The method according to claim 1, wherein, The process of generating dynamic obstacle path maps and optimizing obstacle avoidance strategies based on obstacle monitoring and environmental analysis, using an obstacle location prediction model, includes: The location of obstacles is obtained through GPS and sonar systems, and the coordinate data of the obstacles is generated. Using an obstacle location prediction model, the system predicts the location of obstacles over a future period based on obstacle feature data; it monitors obstacle location changes in real time by acquiring real-time obstacle location data through sensors; it compares the predicted location with the real-time monitored location to ensure the accuracy of the prediction model; if the prediction matches the real-time data, it generates a dynamic path map of the obstacles using the predicted location data; it periodically scans the environment to identify new obstacles and updates the location and quantity information of obstacles; it adjusts the obstacle avoidance strategy by evaluating the location, direction of movement, and speed of obstacles; and it applies a decision tree algorithm to train the model based on the current environment and obstacle data to determine the optimal obstacle avoidance path, while also determining the type and degree of obstacle avoidance action. Environmental data includes water flow, wind speed, and the positions of other vessels. Continuous monitoring of environmental changes and dynamic adjustment of obstacle avoidance strategies; also includes: identifying marine life based on sonar signal processing and spectrum analysis, and adjusting the obstacle avoidance strategies of unmanned vessels based on obstacle location prediction models; The process of identifying marine life based on sonar signal processing and spectral analysis, and adjusting the obstacle avoidance strategy of the unmanned surface vessel (USV) based on an obstacle location prediction model, specifically includes: implementing a sonar signal processing workflow, including depth filtering and spectral analysis, to identify and separate the sonar features of marine life; creating a marine biological sonar feature library to acquire sonar signal samples from different species of marine life; using a long short-term memory network algorithm to pre-train a model based on the sonar signal samples from different species of marine life, and training a classification and recognition model for the sonar signal samples from different species of marine life; integrating the trained model into the USV navigation system to classify and identify the acquired sonar data in real time; matching the sonar signals with the biological sonar feature library to distinguish between obstacles and biological signals; applying an obstacle location prediction model to sonar signals identified as non-biological to predict the future position and movement trend of obstacles; and adjusting the obstacle avoidance strategy of the USV based on the predicted obstacle trajectory, including adjusting the course and speed.

9. The method according to claim 1, wherein, The process involves adjusting obstacle avoidance strategies and route planning to assess the mission progress of the unmanned surface vessel (USV). If delays are anticipated, the route planning is optimized to ensure the best efficiency of the USV mission execution. This includes: The unmanned surface vessel (USV) mission execution program is initiated, acquiring real-time navigation data, including current position, speed, current path, and estimated arrival time. Based on the acquired navigation data, the ARIMA algorithm is used to train a model, predict the mission execution progress, evaluate the current mission progress, and generate a progress report, including the estimated remaining travel time and predicted arrival time. If the progress report indicates potential delays, a predetermined arrival time is set based on the USV's current position, destination, remaining range, current speed, and environmental factors. A dynamic programming algorithm is used to determine the optimal route. The execution efficiency of the adjusted route is evaluated based on data from the original route and the adjusted route, including time efficiency and safety. If the efficiency evaluation shows that the new route needs further optimization, the route planning is adjusted to further shorten the path or improve safety. The route planning adjustment includes adjusting the path selection and sailing speed. The optimized route is confirmed and input into the USV's navigation system, initiating the mission execution update.

Images