Real-time intelligent sensing and real-time obstacle avoidance system of unmanned ship at sea and operation method

Abstract

The invention relates to the technical field of navigation and obstacle avoidance of unmanned ships, in particular to a real-time intelligent sensing and real-time obstacle avoidance system and an operation method of an offshore unmanned ship, which comprises a sensor module, a data fusion module, a dynamic obstacle avoidance module and a self-adaptive PID (Proportion Integration Differentiation) control module, according to the technical scheme, through cooperation of multiple algorithms of Kalman filtering, DWA and self-adaptive PID control, real-time obstacle avoidance can be achieved in a complex marine environment, the flexibility of path selection is kept, and the driving stability is improved. The multi-algorithm combination enables navigation and obstacle avoidance of the unmanned ship to be more intelligent and accurate, and is suitable for dynamic and complex application scenes.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned boat navigation and obstacle avoidance, and in particular to a real-time intelligent perception and real-time obstacle avoidance system and an operation method for an unmanned boat at sea. Background Art

[0002] With the widespread application of unmanned boats in the fields of marine environmental monitoring, rescue, logistics and transportation, resource exploration, etc., their autonomous obstacle avoidance capability in dynamic and complex environments has become one of the key technologies. Unmanned boats in marine environments need to be able to make real-time obstacle avoidance decisions when facing variable obstacles (such as floating objects, other ships, marine life, etc.) to ensure the safety of navigation and the reliability of missions.

[0003] Most of the existing unmanned boat obstacle avoidance technologies rely on global path planning or local path selection methods with fixed parameters. Global path planning is usually based on prior environmental information or static maps, and lacks adaptability to real-time dynamic changes. Although local path selection methods have certain real-time response capabilities, their fixed cost function parameters are often difficult to make fast and effective adjustments when facing changes in obstacle density or large deviations in target direction. In addition, traditional path planning methods rarely combine technical means of multi-sensor data fusion, resulting in errors in obstacle position perception, further affecting the obstacle avoidance effect.

[0004] Therefore, in order to improve the dynamic obstacle avoidance capability of unmanned boats in complex marine environments, an unmanned boat obstacle avoidance system that combines multi-sensor data fusion, dynamic window method (DWA) path planning and adaptive PID control is urgently needed. The system can dynamically adjust the path selection parameters based on real-time monitoring of the environment to ensure that the unmanned boat can achieve flexible and accurate obstacle avoidance in areas with dense obstacles or complex paths. Summary of the invention

[0005] The present invention provides a real-time intelligent perception and real-time obstacle avoidance system and operation method for unmanned boats at sea that can achieve real-time obstacle avoidance for unmanned boats in complex dynamic environments through multi-sensor data fusion, path planning and dynamic control, has high environmental adaptability and navigation stability, and is suitable for use in the fields of water transportation, environmental monitoring, marine resource exploration and rescue, etc.

[0006] The technical solution of the present invention is as follows: A real-time intelligent perception and real-time obstacle avoidance system for unmanned marine boats, comprising: Sensor module, used to collect environmental data, including the current location of obstacles ( ) and speed information; A data fusion module, which fuses the multi-sensor data collected by the sensor module based on a Kalman filter algorithm to generate a real-time environment model; The dynamic obstacle avoidance module plans the path of obstacles in the environment model based on the dynamic window method (DWA) algorithm, selects the optimal path through the cost function, and generates the target position ( ) and target speed , so that the unmanned boat can avoid obstacles and move towards the target direction; The adaptive PID control module, according to the target position ( ) and target speed Current speed of the unmanned boat and direction are adjusted in real time to ensure smooth driving and execute the optimal path.

[0007] Preferably, the data fusion module fuses the multi-sensor data of the laser radar and the millimeter-wave radar based on a Kalman filter algorithm to reduce noise and improve the accuracy of obstacle information; the Kalman filter algorithm includes the following steps: Status prediction: ; ; in, is the state prediction value, is the forecast error covariance, is the state transfer matrix, is the control input matrix, is the control input, is the process noise covariance.

[0008] Update steps: Calculate the Kalman gain: ; in, is the Kalman gain, is the observation matrix, is the observation noise covariance.

[0009] Status Update: ); Covariance update: .

[0010] Preferably, the dynamic window method algorithm of the dynamic obstacle avoidance module selects the optimal path based on the following cost function: Distance Cost : Used to evaluate the minimum distance to obstacles in the path to ensure that the unmanned boat avoids obstacles within a safe range; ; in, Indicates the current position of the unmanned boat ( , ) and obstacle position ( , ) between the two.

[0011] Target direction cost : It is used to evaluate the degree to which the current path deviates from the target direction in order to keep the unmanned boat heading towards the target; ; Comprehensive cost function : Considering the above costs, the weights are adaptively adjusted according to the environment density, and the path with the lowest cost is selected.

[0012] ; in, and is the weight coefficient of the cost function.

[0013] Preferably, the PID control module adjusts the speed and direction of the unmanned boat in real time according to the proportional integral differential control formula to ensure smooth driving: e(t)+ ; in, To control the output, e(t)= is the current error, , and are the proportional, integral and derivative gain parameters respectively.

[0014] Preferably, the adaptive PID control module automatically adjusts the PID gain parameters according to the real-time navigation status and path smoothness of the unmanned boat, ensuring that the response speed is improved in complex path areas and remains stable in smooth areas.

[0015] Preferably, the dynamic window method in the dynamic obstacle avoidance module dynamically adjusts the window size according to the obstacle density and the speed of the unmanned boat, narrowing the window range when the obstacles are dense and expanding the window range when the obstacles are sparse, so as to improve obstacle avoidance accuracy and driving efficiency.

[0016] The operating method of the real-time intelligent perception and real-time obstacle avoidance system of the unmanned boat at sea of ​​the present invention comprises the following steps: Step 1: Environmental data collection and integration 1.1 Start the sensor module: The laser radar and millimeter wave radar on the unmanned boat begin to collect obstacle data in the surrounding environment, including location ( , ), distance and speed information;

[0017] 1.2 Data fusion: Multi-sensor data is fused through the Kalman filter algorithm to remove noise errors, generate an accurate real-time environmental model, and provide the latest position and motion status of obstacles.

[0018] Step 2: Adaptive adjustment of weight coefficients 2.1 Environmental density analysis: Real-time detection of the density of obstacles around the unmanned boat and calculation of the number of obstacles in a unit area; 2.2 Dynamically adjust weight coefficient: formula: ; in, For the distance cost, The target direction cost.

[0019] Step 3: Path planning and dynamic window method calculation 3.1 Setting the dynamic window: According to the current position of the unmanned boat ( , ),speed and direction, setting the dynamic window range to limit all feasible paths within the current speed and acceleration range; 3.2 Calculate the cost function: Distance Cost : Evaluate the minimum distance to obstacles in each path to ensure the safety of the path; ; in, Indicates the distance between the current position of the unmanned boat and the obstacle; Target direction cost : Evaluate the deviation of the current path from the target direction and try to keep the unmanned boat facing the target; ; Comprehensive cost calculation: Based on the weight coefficient adjusted in real time, the path with the lowest cost is selected as the optimal obstacle avoidance path.

[0020] Step 4: Optimal Path Execution and Adaptive PID Control 4.1 Path execution: DWA algorithm outputs the target speed of the optimal path and the target location ( , ), the unmanned boat performs obstacle avoidance along the optimal path; 4.2 Adaptive PID control: The PID control module dynamically adjusts the ratio based on the target speed and current position speed planned by DWA. ,integral and differential Gain parameters, which adjust the speed and direction of the unmanned boat in real time to smoothly execute the path; Error calculation: Calculate the speed error e(t)= ; PID control formula: e(t)+ ; When the path is complex, the proportional and differential gains are increased to quickly respond to path changes; when the path is stable, the gain parameters are reduced to keep the unmanned boat stable.

[0021] Step 5: Real-time feedback and path updates 5.1 Real-time monitoring and feedback loop: The system continuously monitors the position and obstacle distribution of the unmanned boat during its driving process. When the position or movement of the obstacle is detected to change, the feedback loop returns to step 2 to readjust the weight coefficient and update the optimal path; 5.2 Adaptive window adjustment: Dynamically adjust the DWA window range according to the density of obstacles around the unmanned boat and the driving speed. When obstacles are dense, the window is reduced to improve obstacle avoidance accuracy; when obstacles are sparse, the window is expanded to improve driving efficiency.

[0022] Step 6: Task completion and system reset 6.1 Arrival at the target: After the unmanned boat completes obstacle avoidance and reaches the target location, the system enters standby mode and is ready to receive new mission instructions; 6.2 System reset and self-check: The unmanned boat resets PID control parameters and cleans up sensor data to prepare for the next operation.

[0023] Preferably, in step 2 and Adaptive adjustment is made based on obstacle density and direction deviation. The adjustment rules are as follows: When obstacles are dense, increase the distance cost weight , ensuring that the unmanned boat has priority in avoiding obstacles; When obstacles are sparse and deviate from the target direction, increase the target direction cost weight , ensuring that the unmanned boat moves towards the target direction first; The algorithm logic is as follows: ; and Represents the adaptive weight coefficient based on obstacle density and direction deviation.

[0024] Technical Effects The real-time intelligent obstacle avoidance system for unmanned boats at sea based on dynamic window method (DWA), Kalman filtering and adaptive PID control, through the innovative combination of multi-sensor data fusion, dynamic path planning and adaptive control technology, significantly improves the obstacle avoidance capability, environmental adaptability and path execution stability of unmanned boats in complex marine environments. The specific technical effects are as follows: 1. Real-time and high-precision environmental perception The present invention uses multi-sensor data fusion of laser radar and millimeter-wave radar, combined with the Kalman filter algorithm, to accurately locate obstacles in a dynamically changing environment. Kalman filtering effectively removes sensor noise, making the real-time position, speed and other information of obstacles more accurate, and providing a reliable environmental model for subsequent path planning. This effect enables the unmanned boat to more accurately judge the distribution and dynamic characteristics of obstacles in complex environments, greatly reducing the risk of collision caused by perception errors.

[0025] 2. Intelligent obstacle avoidance path planning with adaptive weights The system's dynamic window method (DWA) path planning module introduces an adaptive weighted cost function, which automatically adjusts the weight coefficient of path selection according to the density of obstacles and the degree of deviation from the target direction. Such adaptive adjustments ensure that the unmanned boat gives priority to safety distance in areas with dense obstacles, and gives priority to heading towards the target when the path is sparse or the target deviates greatly, thus realizing intelligent path selection. Compared with the traditional fixed cost function, the adaptive weight mechanism of the present invention can dynamically optimize the path according to real-time environmental changes, greatly improving the flexibility and obstacle avoidance efficiency of the unmanned boat in complex marine environments.

[0026] 3. Smooth and precise adaptive PID control The present invention combines an adaptive PID control algorithm to adjust the PID gain parameters in real time according to the complexity of the path, so that the unmanned boat can travel more smoothly in a dynamic environment. In areas with complex paths, the system automatically increases the proportional and differential gains to improve the response speed and obstacle avoidance sensitivity; in areas with stable paths, the gain parameters are reduced to ensure the stability and energy efficiency of the unmanned boat. This technical effect enables the unmanned boat to travel stably along the planned path in a changing environment, effectively reducing path jitter and deviation, and improving the accuracy of task execution and driving efficiency.

[0027] 4. Efficient feedback closed-loop structure Through closed-loop feedback between data fusion, path planning, and control modules, the system realizes an efficient feedback loop for real-time obstacle avoidance of the unmanned boat. When the position of an obstacle or environmental changes are detected, the Kalman filter and DWA path planning modules immediately adjust the new path planning, and the PID control module responds to new speed and direction adjustment instructions in real time to ensure that the unmanned boat can always be on the optimal path. This closed-loop feedback structure effectively improves the response speed of the system, allowing the unmanned boat to quickly adapt to and avoid obstacles in the event of sudden environmental changes, further enhancing the reliability and adaptability of the system.

[0028] 5. Ability to adapt to a wide range of complex environments The obstacle avoidance system of the present invention has high environmental adaptability and is particularly suitable for complex waters with dense obstacles and unfixed positions. The system can not only deal with static obstacles, but also predict and avoid dynamic obstacles. It is suitable for mission scenarios with high requirements for real-time obstacle avoidance, such as maritime rescue, ocean monitoring, and resource exploration. Compared with traditional methods, the present invention significantly improves the mission success rate and safety of unmanned boats, and provides technical support for unmanned boats to achieve efficient and autonomous obstacle avoidance in diverse marine environments.

[0029] 6. Combination of Kalman Filter and Dynamic Window Method (DWA) The Kalman filter realizes multi-sensor data fusion and provides a real-time and accurate environmental model, which provides stable and high-precision obstacle position and speed information for the dynamic window method (DWA) path planning. Since DWA relies on the precise position of obstacles for dynamic path selection, the noise filtering and data fusion functions of the Kalman filter significantly improve the obstacle avoidance accuracy of DWA, allowing the unmanned boat to more flexibly choose a safe path in a dynamic and complex environment.

[0030] 7. Synergy between Dynamic Window Method (DWA) and Adaptive PID Control After identifying obstacles, the DWA path planning module calculates the optimal path and target speed and passes them to the adaptive PID control module. The PID control adjusts the speed and direction of the unmanned boat in real time according to the target path generated by DWA, so that it can travel smoothly along the optimal path. This combination of algorithms ensures that the unmanned boat can move smoothly and efficiently towards the target direction while avoiding obstacles, greatly improving the system's control accuracy and path execution stability.

[0031] 8. Triple combination of Kalman filter, DWA and adaptive PID control By combining the environmental perception of Kalman filtering, the path planning of DWA and the closed-loop control of adaptive PID control, this system realizes an adaptive, intelligent real-time obstacle avoidance closed loop. Kalman filtering ensures the high accuracy of environmental data, DWA selects the best path for the unmanned boat, and the adaptive PID control module dynamically adjusts the driving direction and speed of the unmanned boat. The combination of the three forms an efficient feedback closed loop, enabling the unmanned boat to autonomously adapt to sudden changes in complex environments and achieve continuous and stable obstacle avoidance effects.

[0032] Through the multi-algorithm collaboration of Kalman filtering, DWA and adaptive PID control, the technical solution of the present invention can achieve real-time obstacle avoidance in complex marine environments, maintaining the flexibility of path selection and improving the stability of driving. This combination of multiple algorithms makes the navigation and obstacle avoidance of the unmanned boat more intelligent and accurate, and is suitable for dynamic and complex application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a control flow chart of the present invention. DETAILED DESCRIPTION

[0034] Embodiment 1: The real-time intelligent perception and real-time obstacle avoidance system for unmanned boats at sea of ​​this embodiment includes: Sensor module, used to collect environmental data, including the current location of obstacles ( ) and speed information; A data fusion module, which fuses the multi-sensor data collected by the sensor module based on a Kalman filter algorithm to generate a real-time environment model; The dynamic obstacle avoidance module plans the path of obstacles in the environment model based on the dynamic window method (DWA) algorithm, selects the optimal path through the cost function, and generates the target position ( ) and target speed , so that the unmanned boat can avoid obstacles and move towards the target direction; The adaptive PID control module, according to the target position ( ) and target speed Current speed of the unmanned boat and direction are adjusted in real time to ensure smooth driving and execute the optimal path.

[0035] Embodiment 2: The difference between this embodiment and embodiment 1 is that, in this embodiment, the data fusion module fuses the multi-sensor data of the laser radar and the millimeter-wave radar based on the Kalman filter algorithm to reduce noise and improve the accuracy of obstacle information; the Kalman filter algorithm includes the following steps:

[0036] Status prediction: ; ; in, is the state prediction value, is the forecast error covariance, is the state transfer matrix, is the control input matrix, is the control input, is the process noise covariance.

[0037] Update steps: Calculate the Kalman gain: ; in, is the Kalman gain, is the observation matrix, is the observation noise covariance.

[0038] Status Update: ); Covariance update: .

[0039] Preferably, the dynamic window method algorithm of the dynamic obstacle avoidance module selects the optimal path based on the following cost function: Distance Cost : Used to evaluate the minimum distance to obstacles in the path to ensure that the unmanned boat avoids obstacles within a safe range; ; in, Indicates the current position of the unmanned boat ( , ) and obstacle position ( , ) between the two.

[0040] Target direction cost : It is used to evaluate the degree to which the current path deviates from the target direction in order to keep the unmanned boat heading towards the target; ; Comprehensive cost function : Considering the above costs, the weights are adaptively adjusted according to the environment density, and the path with the lowest cost is selected; ; in, and is the weight coefficient of the cost function.

[0041] Preferably, the PID control module adjusts the speed and direction of the unmanned boat in real time according to the proportional integral differential control formula to ensure smooth driving: e(t)+ ; in, To control the output, e(t)= is the current error, , and are the proportional, integral and derivative gain parameters respectively.

[0042] Preferably, the adaptive PID control module automatically adjusts the PID gain parameters according to the real-time navigation status and path smoothness of the unmanned boat, ensuring that the response speed is improved in complex path areas and remains stable in smooth areas.

[0043] Preferably, the dynamic window method in the dynamic obstacle avoidance module dynamically adjusts the window size according to the obstacle density and the speed of the unmanned boat, narrowing the window range when the obstacles are dense and expanding the window range when the obstacles are sparse, so as to improve obstacle avoidance accuracy and driving efficiency.

[0044] Embodiment 3: This embodiment is an operation method of a real-time intelligent perception and real-time obstacle avoidance system for an unmanned marine vehicle, which includes the following steps: Step 1: Environmental data collection and integration 1.1 Start the sensor module: The laser radar and millimeter wave radar on the unmanned boat begin to collect obstacle data in the surrounding environment, including location ( , ), distance and speed information;

[0045] 1.2 Data fusion: Multi-sensor data is fused through the Kalman filter algorithm to remove noise errors, generate an accurate real-time environmental model, and provide the latest position and motion status of obstacles.

[0046] Step 2: Adaptive adjustment of weight coefficients 2.1 Environmental density analysis: Real-time detection of the density of obstacles around the unmanned boat and calculation of the number of obstacles in a unit area; 2.2 Dynamically adjust weight coefficient: formula: ; in, For the distance cost, The target direction cost.

[0047] Step 3: Path planning and dynamic window method calculation 3.1 Setting the dynamic window: According to the current position of the unmanned boat ( , ),speed and direction, setting the dynamic window range to limit all feasible paths within the current speed and acceleration range; 3.2 Calculate the cost function: Distance Cost : Evaluate the minimum distance to obstacles in each path to ensure the safety of the path; ; in, Indicates the distance between the current position of the unmanned boat and the obstacle; Target direction cost : Evaluate the deviation of the current path from the target direction and try to keep the unmanned boat facing the target; ; Comprehensive cost calculation: Based on the weight coefficient adjusted in real time, the path with the lowest cost is selected as the optimal obstacle avoidance path.

[0048] Step 4: Optimal Path Execution and Adaptive PID Control 4.1 Path execution: DWA algorithm outputs the target speed of the optimal path and the target location ( , ), the unmanned boat performs obstacle avoidance along the optimal path.

[0049] 4.2 Adaptive PID control: The PID control module dynamically adjusts the ratio based on the target speed and current position speed planned by DWA. ,integral and differential Gain parameters, adjust the speed and direction of the unmanned boat in real time to smoothly execute the path.

[0050] Error calculation: Calculate the speed error e(t)= ; PID control formula: e(t)+ ; When the path is complex, the proportional and differential gains are increased to quickly respond to path changes; when the path is stable, the gain parameters are reduced to keep the unmanned boat stable.

[0051] Step 5: Real-time feedback and path updates 5.1 Real-time monitoring and feedback loop: The system continuously monitors the position and obstacle distribution of the unmanned boat during its driving process. When the position or movement of the obstacle is detected to change, the feedback loop returns to step 2 to readjust the weight coefficient and update the optimal path; 5.2 Adaptive window adjustment: Dynamically adjust the DWA window range according to the density of obstacles around the unmanned boat and the driving speed. When obstacles are dense, the window is reduced to improve obstacle avoidance accuracy; when obstacles are sparse, the window is expanded to improve driving efficiency.

[0052] Step 6: Task completion and system reset 6.1 Arrival at the target: After the unmanned boat completes obstacle avoidance and reaches the target location, the system enters standby mode and is ready to receive new mission instructions; 6.2 System reset and self-check: The unmanned boat resets PID control parameters and cleans up sensor data to prepare for the next operation.

[0053] Embodiment 4: The difference between this embodiment and embodiment 3 is that in step 2 of this embodiment and Adaptive adjustment is made based on obstacle density and direction deviation. The adjustment rules are as follows: When obstacles are dense, increase the distance cost weight , ensuring that the unmanned boat has priority in avoiding obstacles; When obstacles are sparse and deviate from the target direction, increase the target direction cost weight , ensuring that the unmanned boat moves towards the target direction first; The algorithm logic is as follows: ; and Represents the adaptive weight coefficient based on obstacle density and direction deviation.

Claims

1. A real-time intelligent perception and real-time obstacle avoidance system for unmanned boats at sea, characterized in that: include: Sensor module, used to collect environmental data, including the current location of obstacles ( ) and current speed information ; A data fusion module, which fuses the multi-sensor data collected by the sensor module based on a Kalman filter algorithm to generate a real-time environment model; The dynamic obstacle avoidance module plans the path of obstacles in the environment model based on the DWA dynamic window method algorithm, selects the optimal path through the cost function, and generates the target position ( ) and target speed , so that the unmanned boat can avoid obstacles and move towards the target direction; The adaptive PID control module, according to the target position ( ) and target speed Current speed of the unmanned boat and direction are adjusted in real time to ensure smooth driving and execute the optimal path.

2. The real-time intelligent perception and real-time obstacle avoidance system for unmanned marine vehicles according to claim 1 is characterized in that: in, The data fusion module fuses the multi-sensor data of the laser radar and the millimeter-wave radar based on the Kalman filter algorithm; the Kalman filter algorithm includes the following steps: Status prediction: ； ； in, is the state prediction value, is the forecast error covariance, is the state transfer matrix, is the control input matrix, is the control input, is the process noise covariance; Update steps: Calculate the Kalman gain: ； in, is the Kalman gain, is the observation matrix, is the observation noise covariance; Status Update: )； Covariance update: 。 3. The real-time intelligent perception and real-time obstacle avoidance system for unmanned marine vehicles according to claim 1 is characterized in that: in, The dynamic window method algorithm of the dynamic obstacle avoidance module selects the optimal path based on the following cost function: Distance Cost : Used to evaluate the minimum distance to obstacles in the path to ensure that the unmanned boat avoids obstacles within a safe range; ； in, Indicates the current position of the unmanned boat ( , ) and obstacle position ( , ) between the two sides; Target direction cost : It is used to evaluate the degree to which the current path deviates from the target direction in order to keep the unmanned boat heading towards the target; ； Comprehensive cost function : Considering the above costs, the weights are adaptively adjusted according to the environment density, and the path with the lowest cost is selected; ； in, and is the weight coefficient of the cost function.

4. The real-time intelligent perception and real-time obstacle avoidance system for unmanned marine vehicles according to claim 1 is characterized in that: in, The PID control module adjusts the speed and direction of the unmanned boat in real time according to the proportional-integral-differential control formula: e(t)+ ; in, To control the output, e(t)= is the current error, , and are the proportional, integral and derivative gain parameters respectively.

5. The real-time intelligent perception and real-time obstacle avoidance system for unmanned marine boats according to claim 1 or 2, characterized in that the adaptive PID control module automatically adjusts the PID gain parameters according to the real-time navigation status and path smoothness of the unmanned boat.

6. The real-time intelligent perception and real-time obstacle avoidance system for unmanned boats at sea according to claim 1 is characterized in that the dynamic window method in the dynamic obstacle avoidance module dynamically adjusts the window size according to the obstacle density and the speed of the unmanned boat, narrows the window range when the obstacles are dense, and expands the window range when the obstacles are sparse.

7. An operating method of the real-time intelligent perception and real-time obstacle avoidance system for unmanned marine vehicles according to any one of claims 1 to 6, characterized in that it comprises the following steps: Step 1: Environmental data collection and integration 1.1 Start the sensor module: The laser radar and millimeter wave radar on the unmanned boat begin to collect obstacle data in the surrounding environment, including location ( , ), distance and speed information; 1.2 Data fusion: Multi-sensor data is fused through Kalman filter algorithm; Step 2: Adaptive adjustment of weight coefficients 2.1 Environmental density analysis: Real-time detection of the density of obstacles around the unmanned boat and calculation of the number of obstacles in a unit area; 2.2 Dynamically adjust weight coefficient: official: ; in, For the distance cost, is the target direction cost; Step 3: Path planning and dynamic window method calculation 3.1 Setting the dynamic window: According to the current position of the unmanned boat ( , ),speed and direction, setting the dynamic window range to limit all feasible paths within the current speed and acceleration range; 3.2 Calculate the cost function: Distance Cost : Evaluate the minimum distance to obstacles in each path to ensure the safety of the path; ； in, Indicates the distance between the current position of the unmanned boat and the obstacle; Target direction cost : Evaluate the deviation of the current path from the target direction and try to keep the unmanned boat facing the target; ； Comprehensive cost calculation: According to the weight coefficient adjusted in real time, the path with the lowest cost is selected as the optimal obstacle avoidance path; Step 4: Optimal Path Execution and Adaptive PID Control 4.1 Path execution: DWA algorithm outputs the target speed of the optimal path and the target location ( , ), the unmanned boat performs obstacle avoidance along the optimal path; 4.2 Adaptive PID control: The PID control module dynamically adjusts the ratio based on the target speed and current position speed planned by DWA. ,integral and differential Gain parameters, real-time adjustment of the speed and direction of the unmanned boat; Error calculation: Calculate the speed error e(t)= ; PID control formula: e(t)+ ; When the path is complex, increase the proportional and differential gains to quickly respond to path changes; when the path is stable, reduce the gain parameters to keep the unmanned boat stable. Step 5: Real-time feedback and path updates 5.1 Real-time monitoring and feedback loop: The system continuously monitors the position of the unmanned boat and the distribution of obstacles during the driving process. When the position or movement of the obstacle is detected to change, the system returns to step 2 through the feedback loop to readjust the weight coefficient and update the optimal path; 5.2 Adaptive window adjustment: Dynamically adjust the DWA window range according to the density of obstacles around the unmanned boat and the driving speed. When obstacles are dense, the window is reduced to improve obstacle avoidance accuracy. Enlarge the window when obstacles are sparse; Step 6: Task completion and system reset 6.1 Arrival at the target: After the unmanned boat completes obstacle avoidance and reaches the target location, the system enters standby mode and is ready to receive new mission instructions; 6.2 System reset and self-check: The unmanned boat resets PID control parameters and cleans up sensor data to prepare for the next operation.

8. The method for operating the real-time intelligent perception and real-time obstacle avoidance system for unmanned marine boats according to claim 7, characterized in that: In step 2 and Adaptive adjustment is made based on obstacle density and direction deviation. The adjustment rules are as follows: When obstacles are dense, increase the distance cost weight , ensuring that the unmanned boat has priority in avoiding obstacles; When obstacles are sparse and deviate from the target direction, increase the target direction cost weight , ensuring that the unmanned boat moves towards the target direction first; The algorithm logic is as follows: ； and Represents the adaptive weight coefficient based on obstacle density and direction deviation.

Images