Robot path planning method fusing improved black kite algorithm and dynamic window method

Abstract

The application discloses a robot path planning method combining an improved black-winged kite algorithm and a dynamic window method, and comprises the following steps: acquiring environment information and a starting point and a target point of the robot; performing global path planning by using the improved black-winged kite optimization algorithm to generate an initial global path from the starting point to the target point; the improvement comprises optimization of population individual position updating strategies in three stages of initialization, attack behavior and migration behavior of the standard black-winged kite algorithm; path nodes of the global reference path are taken as local navigation sub-targets, and the global path is reconstructed into multiple local paths; firstly, the improved dynamic window method is used to sample a speed space of the robot for the first local path to generate multiple candidate motion trajectories and evaluate an optimal trajectory; local real-time obstacle avoidance planning is performed under the guidance of the local path; and whether the robot reaches the global target point is judged, and the path planning is ended. The application has the characteristics of higher robustness and better smoothness.

Description

Technical Field

[0001] This invention relates to the field of robotics, specifically to a robot path planning method that integrates the improved Blackwing Kite algorithm and the dynamic window method. Background Technology

[0002] With the rapid development of artificial intelligence, mobile robots have been widely applied in fields such as industrial automation, agricultural automation, medical services, and fire rescue. Mobile robot path planning aims to solve problems such as adaptive planning and obstacle avoidance in complex environments. Based on the robot's understanding of its environment, mobile robot path planning is divided into global path planning and local path planning. Global path planning refers to situations where the robot has known environmental information and can reach the predetermined goal by setting relevant parameter points. Based on this attribute, global path planning is also called offline path planning or static path planning. Local path planning refers to situations where the robot has partially or completely unknown environmental information and needs to make adaptive adjustments to address unknown environmental conditions during the planning process. Based on this attribute, local path planning is also called online path planning or dynamic path planning.

[0003] In recent years, swarm intelligence algorithms have gained popularity among scholars and have been widely applied to global path planning. In existing technologies, researchers have addressed the problems of poor path quality and slow convergence speed in the traditional Sparrow Search Algorithm (SSA) by introducing linear path strategies and neighborhood search strategies, proposing an improved Sparrow Search Algorithm (ISSA). The Black-winged Kite Algorithm (BKA) has advantages such as simple structure and ease of implementation; however, the original algorithm is prone to getting trapped in local optima, slow convergence, and lack of adaptability to dynamic environmental changes when dealing with mobile robot path planning problems in complex environments.

[0004] Local path planning algorithms mainly include Dynamic Window (DWA), Artificial Potential Field (APF), and Reinforcement Learning (RL). For example, researchers have proposed a new local algorithm by combining DWA with an improved In-Follow Gap Method (IFGM) to ensure that mobile robots can generate collision-free trajectories. To address the poor robustness and insufficient smoothness of the generated paths in the DWA, researchers have proposed an improved DWA based on fuzzy logic. While the DWA algorithm exhibits good adaptability to real-time obstacle avoidance in dynamic environments, a single DWA algorithm lacks global information guidance, leading to poor path quality and failure to reach the target point in complex dynamic environments. In summary, existing technologies address the problems of low global search efficiency and insufficient obstacle avoidance performance for mobile robots in complex environments. Summary of the Invention

[0005] The purpose of this invention is to overcome the above-mentioned shortcomings and propose a robot path planning method that integrates the improved Blackwing Kite algorithm and the dynamic window method, which has higher robustness, better smoothness, and can plan better paths for mobile robots.

[0006] This invention provides a robot path planning method that integrates the improved Black-winged Kite algorithm and the dynamic window method, comprising the following steps:

[0007] Step 1: Obtain the global map information of the robot's environment, as well as the robot's starting point and target point;

[0008] Step 2: Based on the global map information, the improved Black-winged Kite optimization algorithm is used to perform global path planning and generate an initial global path from the starting point to the target point; the improvement includes optimization of the population individual position update strategy in the three stages of the standard Black-winged Kite algorithm: initialization, attack behavior, and migration behavior.

[0009] Step 3: During the robot's movement along the initial global path, the robot's sensor data is acquired in real time to perceive local dynamic obstacles; the path nodes of the global reference path are used as local navigation sub-targets to reconstruct the global path into multiple local paths; firstly, the robot's velocity space is sampled using an improved dynamic window method for the first local path to generate multiple candidate motion trajectories and evaluate the optimal trajectory.

[0010] Step 4: Based on the sensor data and the current motion state, output the linear velocity and angular velocity control commands corresponding to the optimal trajectory to drive the robot to move and perform local real-time obstacle avoidance planning under the guidance of the local path;

[0011] Step 5: Determine if the robot has reached the global target point. If not, return to step 3 to continue planning the remaining local paths. If the target point has been reached, end the path planning.

[0012] The above-described robot path planning method, which integrates the improved Black-winged Kite algorithm and the dynamic window method, includes the following: In step 2, the initialization phase of the improved Black-winged Kite optimization algorithm introduces a optimal point set strategy to optimize population diversity; the optimal point set... The mathematical model is as follows:

[0013]

[0014] in, Indicates taking the decimal. It is the size of the point set. It's a good spot; , yes The unit cube in 3D space;

[0015] Let the population individual search space be the first The next and upper limits of Dimension are respectively and , Represent the updated position of individual black-winged kites; map the point set to the population search space to construct the population individual initialization and update model:

[0016] .

[0017] The above-mentioned robot path planning method integrates the improved Black-winged Kite algorithm and the dynamic window method, wherein: in step 2, the improved Black-winged Kite optimization algorithm introduces a leader-based Levy flight search strategy in the attack behavior phase to improve the global search efficiency of the population; the Levy flight function Defined as:

[0018]

[0019] in, Indicates step size, and It follows a standard normal distribution;

[0020] Combining the Lévy flight function and the current best individual, a population update model is constructed at this stage:

[0021]

[0022] in, This represents the leader of the current population, i.e., the optimal solution; Represents the current position of the individual; For random coefficients, It is a random number within the interval [0,1]. It is a constant.

[0023] The robot path planning method described above, which integrates the improved Black-winged Kite algorithm and the dynamic window method, includes the following: In step 2, the improved Black-winged Kite optimization algorithm introduces a spiral search strategy in the migration behavior stage to improve the accuracy of population solving; the mathematical model for the spiral search path is as follows:

[0024]

[0025] in, It is the spiral radius, which controls the search range; Let these be variables; construct a population update model at this stage:

[0026]

[0027] in, These are the parameters for the spiral shape; It is a variable; Indicates the first In the next iteration, the black-winged kite... The optimal individual in the dimension, and represent the first Only Black-winged Kite Based on the First The problem of dimensionality in the first Position in the next iteration; Indicates the first In the nth iteration, any black-winged kite in the nth iteration... The fitness value at the current position in the dimension; Indicates the first In the nth iteration, any black-winged kite in the nth iteration... Fitness values ​​at random locations in the dimension; This indicates a Cauchy mutation.

[0028] The above-mentioned robot path planning method, which integrates the improved Black-winged Kite algorithm and the dynamic window method, wherein: in step 3, the improved dynamic window method controls the input linear velocity. and angular velocity Output motion trajectory; when At that time, the output is an arc-shaped trajectory, and the model is as follows:

[0029]

[0030] in, and Indicates the robot's current coordinates. Indicates the robot's heading angle. The radius of the arc representing the trajectory of motion. Represents the predicted time; when When the time comes, the output is a straight line trajectory, and the model is as follows:

[0031]

[0032] The robot path planning method described above, which integrates the improved Black-winged Kite algorithm and the dynamic window method, includes the following: In step 3, when evaluating the optimal trajectory based on the improved dynamic window method, the constructed evaluation function... for:

[0033]

[0034] in, For adaptive heading weights, This indicates the deviation between the predicted trajectory endpoint direction and the target direction after normalization. For adaptive obstacle avoidance weights, This represents the minimum distance between the truncated trajectory and the obstacle after normalization. For adaptive speed weights, This represents the maximum feasible linear velocity after normalization.

[0035] The above-mentioned robot path planning method, which integrates the improved Black-winged Kite algorithm and the dynamic window method, wherein: the adaptive heading weight The model is:

[0036]

[0037] in, Represents the initial fixed heading weight. , The distance between the robot's current position and the target point is Euclidean. The maximum safety threshold distance is set. This represents the heading angle deviation.

[0038] The above-mentioned robot path planning method, which integrates the improved Black-winged Kite algorithm and the dynamic window method, wherein: the adaptive obstacle avoidance weights The model is:

[0039]

[0040] in, Represents the initial fixed value obstacle avoidance weight and the danger penetration rate. This indicates the percentage of obstacles within the robot's safe zone; dynamic safe distance. , Represents the radius of the obstacle. Represents the maximum deceleration. This represents the timeframe for safety forecasting.

[0041] The above-mentioned robot path planning method, which integrates the improved Black-winged Kite algorithm and the dynamic window method, wherein: the adaptive velocity weight The model is as follows: When the robot is in a dangerous state, i.e., when it detects an obstacle, the dangerous state adaptive velocity weight model is:

[0042]

[0043] When the robot is in a safe state, i.e., no obstacle is detected, the safe state adaptive velocity weight model is:

[0044]

[0045] in, This represents the initial fixed value of the velocity weight.

[0046] Compared with existing technologies, this invention has significant advantages. As shown in the above scheme, for global planning, this invention proposes an improved Black-winged Kite Algorithm (MEBKA). In the initialization phase of the algorithm, a best-point set strategy is introduced to enrich the population diversity. In the attack behavior search phase, a leader-based Levy flight search strategy is designed, which enhances the global search capability of the algorithm by leveraging the long-step jump characteristics of Levy flight and avoids the population getting trapped in local optima. In the migration behavior search phase, a spiral search strategy is introduced to enhance the flexibility of position updates, and a local reinforcement search is performed in the search space through a spiral path. Secondly, for local planning, an improved Dynamic Window Method (IDWA) is proposed. This method introduces a dynamic safety mechanism into the robot's safe velocity space in the traditional DWA, ensuring the safety of the robot during operation. A nonlinear parameter adaptive strategy is designed to optimize the heading weight coefficient, obstacle avoidance weight coefficient, and velocity weight coefficient of the traditional DWA, avoiding getting trapped in local optima in the local planning phase. Finally, the global path information generated by MEBKA is dynamically integrated into the velocity sampling space of the local planner IDWA, forming a two-layer planning of global information guidance and local correction.

[0047] In summary, the present invention has the following advantages:

[0048] (1) An improved Blackwing Kite Algorithm (MEBKA) is proposed based on a best-set strategy, a leader-based Levy flight search strategy, and a spiral search strategy. The best-set strategy is introduced in the initialization stage of the traditional BKA to enhance population diversity; the leader-based Levy flight search strategy is introduced in the attack behavior search stage to enhance the global search capability of the algorithm through the long step jump characteristics of Levy flight; and the spiral search strategy is introduced in the migration behavior search stage to enhance the flexibility of position updates and perform local reinforcement search in the search space through the spiral path.

[0049] (2) An improved dynamic window method (IDWA) is proposed by introducing a dynamic safety mechanism and a nonlinear parameter adaptive strategy. The dynamic safety mechanism is introduced into the safe velocity space of the robot in the traditional DWA to ensure the safety of the robot during operation; a nonlinear parameter adaptive strategy is designed to optimize the heading weight coefficient, obstacle avoidance weight coefficient, and velocity weight coefficient of the traditional DWA to improve planning efficiency.

[0050] (3) A path planning method integrating the improved Blackwing Kite algorithm and the dynamic window method is proposed. The critical path nodes generated by MEBKA are used as local target points of IDWA in sequence. The global path information is dynamically integrated into the velocity sampling space of the local planner IDWA, forming a two-layer planning structure of global information guidance and local correction, which improves search efficiency and enhances adaptability to dynamic changes in the environment.

[0051] The following specific embodiments further illustrate the beneficial effects of the present invention. Attached Figure Description

[0052] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0053] The following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, features, and effects of the robot path planning method that integrates the improved Blackwing Kite algorithm and the dynamic window method proposed in this invention.

[0054] See Figure 1 This invention discloses a robot path planning method that integrates the improved Black-winged Kite algorithm and the dynamic window method, wherein the method includes the following steps:

[0055] Step 1: Obtain the global map information of the robot's environment, as well as the robot's starting point and target point;

[0056] Step 2: Based on the global map information, an improved Black-winged Kite optimization algorithm is used for global path planning to generate an initial global path from the starting point to the target point. The improvement includes optimization of the population individual position update strategy in three stages: initialization, attack behavior, and migration behavior in the standard Black-winged Kite algorithm. Specifically, it includes:

[0057] Step 2.1: The improved black-winged kite optimization algorithm introduces a prime location set strategy in the initialization phase to optimize population diversity. Prime location set The mathematical model is as follows:

[0058]

[0059] in, Indicates taking the decimal. It is the size of the point set. That's a good spot. , yes A unit cube in 3D space.

[0060] Let the population individual search space be the first The next and upper limits of Dimension are respectively and , This represents the updated position of a Black-winged Kite individual. The point set is mapped to the population search space to construct the population individual initialization and update model:

[0061]

[0062] Step 2.2: The improved Blackwing Kite optimization algorithm introduces a leader-based Lévy flight search strategy in the attack behavior phase to improve the global search efficiency of the population. Lévy flight function. Defined as:

[0063]

[0064] in, Indicates step size, and It follows a standard normal distribution.

[0065] Combining the Lévy flight function and the current best individual, a population update model is constructed at this stage:

[0066]

[0067] in, This represents the leader of the current population, i.e., the optimal solution; Represents the current position of the individual; These are random coefficients used to adjust the search direction in the space. It is a random number within the interval [0,1]. It is a constant, and we take the empirical value of 0.9.

[0068] Step 2.3: The improved black-winged kite optimization algorithm introduces a spiral search strategy in the migration behavior stage to improve the accuracy of population calculation. The mathematical model of the spiral path is:

[0069]

[0070] in, It is the spiral radius, which controls the search range. Let be a variable, with a value range of [0,1]. A population update model is constructed at this stage:

[0071]

[0072] in, These are the parameters of the spiral shape, which determine the tightness of the spiral; It is a variable with a value range of [-1, 1]. Indicates the first In the next iteration, the black-winged kite... The optimal individual in the dimension, and represent the first Only Black-winged Kite Based on the First The problem of dimensionality in the first Position in the next iteration; Indicates the first In the nth iteration, any black-winged kite in the nth iteration... The fitness value at the current position in the dimension; Indicates the first In the nth iteration, any black-winged kite in the nth iteration... Fitness values ​​at random locations in the dimension; This represents Cauchy variation, which increases the randomness of the solution based on this strategy.

[0073] Step 3: During the robot's movement along the initial global path, acquire the robot's sensor data in real time to perceive local dynamic obstacles; use the path nodes of the global reference path... As a local navigation sub-target, select any two adjacent nodes and mark the former node as the starting point of the local path. The next node is denoted as the target point of the local path. Reconstruct the global path as Segment local path First, regarding the local path... An improved dynamic window method is used to sample the robot's velocity space, generate multiple candidate motion trajectories, and evaluate the optimal trajectory. Specifically, this includes:

[0074] Step 3.1: Initialize the velocity sampling range and navigation angle (Current local path segment robot starting direction and...) (The included angle of the axis) is used to establish the linear velocity that the mobile robot can achieve. and angular velocity The two-dimensional velocity space ;

[0075] The two-dimensional velocity space It consists of the intersection of three velocity constraints:

[0076]

[0077] The reachable velocity space representing the robot's dynamic constraints has the following range:

[0078]

[0079] and These represent the robot's maximum linear velocity and maximum angular velocity, respectively.

[0080] The reachable velocity space representing the robot's kinematic constraints has the following range:

[0081]

[0082] and These are the robot's current speed and acceleration, respectively. and These are the robot's linear acceleration and angular acceleration, respectively. That is the sampling time.

[0083] The safe speed space for the robot ensures a safe balance between the robot and obstacles, and its range is as follows:

[0084]

[0085] This represents the distance from the robot's current position to the nearest obstacle.

[0086] Step 3.2: The improved dynamic window method controls the input... Output two different forms of motion trajectories. When At that time, the output is an arc-shaped trajectory, and the model is as follows:

[0087]

[0088] in, and Indicates the robot's current coordinates. This indicates the robot's heading angle, i.e., its current orientation relative to the direction of travel. The angle between the axes, The radius of the arc representing the trajectory of motion. Represents the prediction time, usually taken as ,and .when When the time comes, the output is a straight line trajectory, and the model is as follows:

[0089]

[0090] Step 3.3: Evaluate the optimal trajectory in the current state based on the improved dynamic window method, wherein the evaluation function constructed by the method is... for:

[0091]

[0092] The adaptive heading weights constructed for the method are as follows:

[0093]

[0094] in, Represents the initial fixed heading weight. , The distance between the robot's current position and the target point is Euclidean. The maximum safety threshold distance is set. For heading angle deviation, This is the sigmoid response function.

[0095] This represents the deviation between the predicted trajectory endpoint direction and the target direction after normalization, and its model is as follows:

[0096]

[0097] in, This represents the azimuth angle, i.e., the ray segment formed by the target point and the robot's current position. The included angle of the axis, This indicates the heading angle at the end of the predicted trajectory.

[0098] The adaptive obstacle avoidance weights constructed for the method are modeled as follows:

[0099]

[0100] in, Represents the initial fixed value obstacle avoidance weight and the danger penetration rate. This indicates the percentage of obstacles within the robot's safe zone, and Dynamic safety distance , Represents the radius of the obstacle. Represents the maximum deceleration. This represents the timeframe for safety forecasting.

[0101] The model represents the minimum distance between the normalized truncated trajectory and the obstacle.

[0102]

[0103] The adaptive velocity weights constructed for the method are as follows: When the robot is in a dangerous state, i.e., when an obstacle is detected, the model is:

[0104]

[0105] When the robot is in a safe state, i.e., when no obstacle is detected, the model is as follows:

[0106]

[0107] The normalized maximum feasible linear velocity is represented by the following model:

[0108]

[0109] in, This represents the initial fixed value of the velocity weight.

[0110] Step 4: Based on the sensor data and current motion state, output the linear velocity and angular velocity control commands corresponding to the optimal trajectory to drive the robot's movement. (Within the local path) Local real-time obstacle avoidance planning is carried out under the guidance of [the relevant authority].

[0111] Step 5: Determine if the robot has reached the global target point. If not, return to Step 3 to continue processing the remaining local path. Perform local path planning sequentially; if the destination has been reached, end the path planning.

[0112] Performance Analysis:

[0113] To preliminarily verify the effectiveness of this invention, it was first compared experimentally with nine swarm intelligence algorithms—RUN, WOA, HHO, PSO, NRBO, GOOSE, CFOA, DBO, and BKA—in three different levels of path planning problems: 10×10 grid environment, 20×20 grid environment, and 30×30 grid environment. Each algorithm was tested 30 times, and the path length experimental results are shown in Tables 1, 2, and 3, respectively. The experimental environment was built on a Windows 11 operating system, with a 12th-generation Intel(R) Core(TM) i5-12400F 2.5GHz processor, 16GB RAM, and MATLAB 2023b as the software platform.

[0114] Table 1 10×10 raster environment results

[0115]

[0116] Table 2 20×20 raster environment results

[0117]

[0118] Table 3 30×30 raster environment results

[0119]

[0120] The experimental results show that this invention performs excellently in path planning environments of three different levels of complexity. In the 10×10 grid environment, the average path length of this invention is 13.0603, the lowest among all algorithms, approximately 6.5% lower than the second-ranked RUN. In the 20×20 grid environment, MEBKA has the lowest average path length, approximately 5.9% lower than the second-ranked RUN. In the 30×30 grid environment, MEBKA's average path length is approximately 7.5% lower than the second-ranked RUN, significantly outperforming the other compared algorithms.

[0121] Meanwhile, to further verify the correctness and effectiveness of the fusion improved Blackwing Kite algorithm and dynamic window method (MEBKA-IDWA) described in this application, MEBKA-IDWA and the global algorithm MEBKA were compared experimentally in a 20×20 grid environment. Each algorithm was tested 30 times, and the experimental results of path length are shown in Table 4.

[0122] Table 4. Comparison of Fusion Algorithm Experiment Results

[0123]

[0124] As shown in Table 4, the standard deviation of MEBKA-IDWA is approximately 3.27% lower than that of MEBKA, indicating that its path planning results have less volatility and higher stability. The maximum value of MEBKA-IDWA is approximately 0.67% lower than that of MEBKA, indicating that outliers are effectively controlled in the given task environment. While the minimum values ​​of MEBKA-IDWA and MEBKA are not significantly different, the mean of MEBKA-IDWA is still lower, demonstrating its superior path planning performance.

[0125] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments without departing from the technical essence of the present invention shall still fall within the scope of the present invention.

Claims

1. A robot path planning method integrating the improved Black-winged Kite algorithm and the dynamic window method, characterized in that: Includes the following steps: Step 1: Obtain the global map information of the robot's environment, as well as the robot's starting point and target point; Step 2: Based on the global map information, the improved Black-winged Kite optimization algorithm is used to perform global path planning and generate an initial global path from the starting point to the target point; the improvement includes optimization of the population individual position update strategy in the three stages of the standard Black-winged Kite algorithm: initialization, attack behavior, and migration behavior. Step 3: During the robot's movement along the initial global path, the robot's sensor data is acquired in real time to perceive local dynamic obstacles; the path nodes of the global reference path are used as local navigation sub-targets to reconstruct the global path into multiple local paths; firstly, the robot's velocity space is sampled using an improved dynamic window method for the first local path to generate multiple candidate motion trajectories and evaluate the optimal trajectory. Step 4: Based on the sensor data and the current motion state, output the linear velocity and angular velocity control commands corresponding to the optimal trajectory to drive the robot to move and perform local real-time obstacle avoidance planning under the guidance of the local path; Step 5: Determine if the robot has reached the global target point. If not, return to step 3 to continue planning the remaining local paths. If the target point has been reached, end the path planning.

2. The robot path planning method integrating the improved Blackwing Kite algorithm and the dynamic window method as described in claim 1, characterized in that: In step 2, the improved black-winged kite optimization algorithm introduces a prime location set strategy during the initialization phase to optimize population diversity; prime location set The mathematical model is as follows: in, Indicates taking the decimal. It is the size of the optimal point set. It's a good spot; , yes The unit cube in 3D space; Let the population individual search space be the first The lower and upper bounds of dimension are respectively and , Represent the updated position of individual black-winged kites; map the point set to the population search space to construct the population individual initialization and update model: 。 3. The robot path planning method integrating the improved Black-winged Kite algorithm and the dynamic window method as described in claim 1, characterized in that: In step 2, the improved Black-winged Kite optimization algorithm introduces a leader-based Levy flight search strategy in the attack behavior phase to improve the global search efficiency of the population; Levy flight function Defined as: Where α represents the step size, and It follows a standard normal distribution; Combining the Lévy flight function and the current best individual, a population update model is constructed at this stage: in, This represents the leader of the current population, i.e., the optimal solution; Represents the current position of the individual; For random coefficients, It is a random number within the interval [0,1]. It is a constant.

4. The robot path planning method integrating the improved Black-winged Kite algorithm and the dynamic window method as described in claim 1, characterized in that: In step 2, the improved black-winged kite optimization algorithm introduces a spiral search strategy in the migration behavior stage to improve the accuracy of population solving; the mathematical model of the spiral search path is: in, It is the spiral radius, which controls the search range; Let [variable] be a variable, with a value range of [0,1]; construct a population update model at this stage: in, These are the parameters for the spiral shape; It is a variable, and its value range is [-1, 1]. Indicates the first In the next iteration, the black-winged kite... The optimal individual in the dimension, and represent the first Only Black-winged Kite Based on the First The problem of dimensionality in the first Position in the next iteration; Indicates the first In the nth iteration, any black-winged kite in the nth iteration... The fitness value at the current position in the dimension; Indicates the first In the nth iteration, any black-winged kite in the nth iteration... Fitness values ​​at random locations in the dimension; This indicates a Cauchy mutation.

5. The robot path planning method integrating the improved Blackwing Kite algorithm and the dynamic window method as described in claim 1, characterized in that: The improved dynamic window method in step 3 involves controlling the input linear velocity. and angular velocity Output motion trajectory; when At that time, the output is an arc-shaped trajectory, and the model is as follows: in, and Indicates the robot's current coordinates. Indicates the robot's heading angle. The radius of the arc representing the trajectory of motion. Represents the predicted time; when When the time comes, the output is a straight line trajectory, and the model is: 。 6. The robot path planning method integrating the improved Black-winged Kite algorithm and the dynamic window method as described in claim 1, characterized in that: The evaluation function constructed in step 3 when evaluating the optimal trajectory for: in, For adaptive heading weights, This indicates the deviation between the predicted trajectory endpoint direction and the target direction after normalization. For adaptive obstacle avoidance weights, This represents the minimum distance between the truncated trajectory and the obstacle after normalization. For adaptive speed weights, This represents the maximum feasible linear velocity after normalization.

7. The robot path planning method integrating the improved Black-winged Kite algorithm and the dynamic window method as described in claim 6, characterized in that: The adaptive heading weight The model is: in, Represents the initial fixed heading weight. , The distance between the robot's current position and the target point is Euclidean. The maximum safe threshold distance is set. This represents the heading angle deviation.

8. The robot path planning method integrating the improved Black-winged Kite algorithm and the dynamic window method as described in claim 6, characterized in that: The adaptive obstacle avoidance weights The model is: in, Represents the initial fixed value obstacle avoidance weight and the danger penetration rate. This indicates the percentage of obstacles within the robot's safe zone; Represents the distance from the robot's current position to the nearest obstacle; dynamic safe distance. , Represents the radius of the obstacle. Represents the maximum deceleration. This represents the timeframe for safety forecasting.

9. The robot path planning method integrating the improved Blackwing Kite algorithm and the dynamic window method as described in claim 6, characterized in that: The adaptive speed weight The model is as follows: When the robot is in a dangerous state, i.e., when it detects an obstacle, the dangerous state adaptive velocity weight model is: When the robot is in a safe state, i.e., no obstacle is detected, the safe state adaptive velocity weight model is: in, This represents the initial fixed value of the velocity weight.

Images