SLAM dynamic feature optimization method based on robot self-adaptive navigation

Abstract

This invention relates to the field of robot navigation technology and discloses a SLAM dynamic feature optimization method based on robot adaptive navigation. The method includes the following steps: S1, collecting robot sensor data and environmental dynamic feature data; S2, performing dynamic feature extraction processing; and S3, performing SLAM parameter adaptive adjustment processing. During the dynamic feature optimization processing in the robot navigation process, parameters of static obstacles and dynamically moving objects in the environment are collected by multimodal sensors, and the motion features of moving objects are extracted by combining feature recognition algorithms to ensure the accuracy of dynamic feature recognition under different environmental conditions. At the same time, the SLAM positioning accuracy parameters are adjusted in real time according to environmental change indicators, which can detect and correct feature distortion caused by high-speed movement of dynamic objects in real time, ensuring the accuracy of robot positioning and reducing positioning errors in high-speed dynamic scenes.

Description

Technical Field

[0001] This invention relates to the field of robot navigation technology, specifically to a SLAM dynamic feature optimization method based on robot adaptive navigation. Background Technology

[0002] A robot is an intelligent machine capable of semi-autonomous or fully autonomous operation. Robots can perform tasks such as manual labor or movement through programming and automatic control. With the development of society, economy, and science and technology, intelligent robots are becoming increasingly popular, and their applications are expanding in all aspects of life. Robots possess basic characteristics such as perception, decision-making, and execution, and can assist or even replace humans in completing dangerous, arduous, and complex tasks, improving work efficiency and quality, and serving human life. In recent years, the development of computer, sensor, and network technologies has made it possible for robots to enter homes. The focus has also shifted from fixed robotic arms and hands in structured environments to intelligent robots that autonomously move in unstructured, unknown environments.

[0003] Currently, due to the uncertainty and complexity of the motion state of dynamic objects in the robot navigation environment, it is impossible to accurately distinguish the motion trajectory of high-speed moving objects from static environmental features in real time when performing SLAM dynamic feature extraction. When the motion speed of dynamic objects exceeds the sensor sampling frequency threshold, it will cause distortion of feature extraction data, resulting in a decrease in positioning accuracy.

[0004] Therefore, a dynamic feature optimization method for SLAM based on robot adaptive navigation is proposed to solve the above problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a dynamic feature optimization method for SLAM based on robot adaptive navigation. The SLAM dynamic feature optimization method provided by this invention solves the problems of feature extraction data distortion and the resulting decrease in positioning accuracy mentioned in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a SLAM dynamic feature optimization method based on robot adaptive navigation, the method comprising the following steps:

[0007] S1. Collect robot sensor data and environmental dynamic feature data, use multimodal sensors to obtain real-time environmental information, including static obstacle parameters and dynamic moving object parameters, and generate relevant data;

[0008] S2. Perform dynamic feature extraction processing. Based on sensor data and environmental dynamic feature data, use feature recognition algorithms to extract the speed, direction and size parameters of moving objects and generate dynamic feature extraction data.

[0009] S3. Perform adaptive adjustment of SLAM parameters. Based on dynamic feature extraction data and environmental change indicators, adjust the positioning accuracy parameters and map update frequency parameters of the SLAM algorithm in real time to generate adaptive adjustment data.

[0010] S4. Construct SLAM optimized combined data, aligning and combining dynamic feature extraction data and adaptive adjustment data according to time series.

[0011] S5. Perform dynamic feature optimization algorithm matching processing. Based on the SLAM optimization combination data, perform similarity matching with the pre-stored standard optimization algorithm data, search for the optimal dynamic feature optimization algorithm type, and generate target optimization algorithm type data.

[0012] S6. Perform SLAM dynamic feature optimization processing, call the optimization program corresponding to the target optimization algorithm type data to process the data, filter out dynamic feature noise and enhance the accuracy of static features, and generate optimized SLAM map and robot position data.

[0013] Preferably, step S1 includes the following steps:

[0014] S11. The robot uses a lidar sensor to scan the environment, acquire environmental point cloud data, including obstacle distance and shape parameters, and generate static environmental feature data.

[0015] S12. Video stream is captured by a robot equipped with a camera sensor. The trajectory of moving objects is identified based on optical flow analysis to generate dynamic environmental feature data.

[0016] S13. The static environmental feature data and dynamic environmental feature data are fused together and processed synchronously according to timestamps to generate robot sensor data and dynamic environmental feature data.

[0017] Preferably, step S2 includes the following steps:

[0018] S21. Based on the robot sensor data and environmental dynamic feature data, a convolutional neural network model is used to perform feature classification processing, distinguish between static features and dynamic features, and generate feature classification data.

[0019] S22. Perform cluster analysis on the dynamic features in the feature classification data, calculate the centroid position and motion vector of the moving object, and generate dynamic feature extraction data.

[0020] Preferably, step S3 includes the following steps:

[0021] S31. Monitor the rate of change of dynamic environmental characteristics. When the density of dynamic characteristics exceeds the threshold, trigger the adaptive adjustment mechanism.

[0022] S32. Adjust the number of particle filters in the SLAM algorithm according to the rate of change of dynamic characteristics to reduce the computational load;

[0023] S33. Synchronously optimize map update frequency, dynamically set update interval based on environmental stability index, and generate adaptive adjustment data.

[0024] Preferably, step S4 includes the following steps:

[0025] S41. Align the dynamic feature extraction data and the adaptive adjustment data according to the mileage parameter to construct a time-space matrix;

[0026] S42. Mark the dynamic feature hotspot regions in the matrix to generate SLAM optimized combination data.

[0027] Preferably, step S5 includes the following steps:

[0028] S51. Establish a pre-stored standard optimization algorithm database, including standard optimization algorithm data corresponding to genetic algorithm, particle swarm optimization algorithm and simulated annealing algorithm;

[0029] S52. Perform pattern matching between the SLAM optimized combination data and the standard optimization algorithm data, and calculate the similarity score, wherein the formula for calculating the similarity score is:

[0030] ;

[0031] in To score the similarity, Let i be the i-th feature parameter in the SLAM optimization combined data, where i is the feature parameter index. For the i-th standard parameter corresponding to the h-th algorithm in the standard optimization algorithm data, For algorithm type index, This represents the total number of feature parameters.

[0032] Preferably, step S52 includes the following steps:

[0033] S521. Initialize the biomimetic optimization search population and randomly set the search agent positions for the algorithm; the position initialization formula is:

[0034] ;

[0035] in To search for the initial location of the agent, The lower boundary of the search space, Let r be the upper boundary of the search space, and let r represent a random number in the interval [0,1].

[0036] S522. During the exploration phase, a simulated search agent randomly detects similar targets, updates their locations, and evaluates their fitness values; the location update formula is:

[0037] ;

[0038] in For the updated position, Current position For the target similarity position, To adjust the coefficients, if the updated fitness value is better, the original position is replaced;

[0039] S523. During the development phase, optimize the search agent location to lock in the optimal matching algorithm and output the target optimization algorithm type data.

[0040] Preferably, step S6 includes the following steps:

[0041] S61. Call the optimization program according to the target optimization algorithm type data to process the dynamic feature noise in the SLAM optimization combination data;

[0042] S62. Enhance the accuracy of static features and reconstruct the environment map;

[0043] S63. Integrate robot position data to generate an optimized SLAM map and robot position data.

[0044] Preferably, step S62 includes the following steps:

[0045] S621. The iterative nearest-point algorithm is used to correct static feature alignment.

[0046] S622: Apply Kalman filtering to smooth the robot trajectory and output an optimized map.

[0047] Preferably, the method further includes step S7:

[0048] S71. Real-time output of optimized SLAM map and robot position data to the navigation system;

[0049] S72. Adjust the robot path planning based on the output data to achieve adaptive navigation.

[0050] (III) Beneficial Effects

[0051] Compared with existing technologies, this invention provides a dynamic feature optimization method for SLAM based on robot adaptive navigation, which has the following beneficial effects:

[0052] 1. In this invention, during the dynamic feature optimization process of robot navigation, parameters of static obstacles and dynamic moving objects in the environment are collected by multimodal sensors, and the motion features of moving objects are extracted by feature recognition algorithms to ensure the accuracy of dynamic feature recognition under different environmental conditions. At the same time, the SLAM positioning accuracy parameters are adjusted in real time according to environmental change indicators, which can detect and correct feature distortion caused by high-speed movement of dynamic objects in real time, ensuring the accuracy of robot positioning and reducing positioning errors in high-speed dynamic scenes.

[0053] 2. In this invention, when performing dynamic feature optimization algorithm matching, a pre-stored standard optimization algorithm database is constructed and a similarity score is calculated. Combined with a biomimetic optimization search mechanism, the optimal algorithm type is locked, enabling the system to respond to algorithm matching requirements caused by extreme environmental changes. Furthermore, when algorithm matching is delayed or mismatched, the search direction is corrected in real time through the position optimization formula during the development stage, ensuring the timeliness and accuracy of algorithm switching and avoiding path planning deviations caused by navigation decision delays.

[0054] 3. In this invention, during the SLAM map reconstruction process, the static feature alignment is corrected by the iterative nearest point algorithm, and the robot's motion trajectory is smoothed by Kalman filtering. This achieves the filtering out of dynamic feature noise and the reconstruction of static environmental features, enabling the system to separate and process complex feature interference caused by dynamic object occlusion, solve the problem of multi-dimensional feature coupling, and improve the reliability of the environmental map and the robustness of the robot's adaptive navigation. Attached Figure Description

[0055] Figure 1 This is a flowchart of the SLAM dynamic feature optimization method based on robot adaptive navigation according to the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] For specific implementation examples, please refer to: Figure 1 A dynamic feature optimization method for SLAM based on robot adaptive navigation, comprising the following steps:

[0058] S1. Collect robot sensor data and environmental dynamic feature data, use multimodal sensors to obtain real-time environmental information, including static obstacle parameters and dynamic moving object parameters, and generate relevant data;

[0059] S2. Perform dynamic feature extraction processing. Based on sensor data and environmental dynamic feature data, use feature recognition algorithms to extract the speed, direction and size parameters of moving objects and generate dynamic feature extraction data.

[0060] S3. Perform adaptive adjustment of SLAM parameters. Based on dynamic feature extraction data and environmental change indicators, adjust the positioning accuracy parameters and map update frequency parameters of the SLAM algorithm in real time to generate adaptive adjustment data.

[0061] S4. Construct SLAM optimized combined data, aligning and combining dynamic feature extraction data and adaptive adjustment data according to time series.

[0062] S5. Perform dynamic feature optimization algorithm matching processing. Based on the SLAM optimization combination data, perform similarity matching with the pre-stored standard optimization algorithm data, search for the optimal dynamic feature optimization algorithm type, and generate target optimization algorithm type data.

[0063] S6. Perform SLAM dynamic feature optimization processing, call the optimization program corresponding to the target optimization algorithm type data to process the data, filter out dynamic feature noise and enhance the accuracy of static features, and generate optimized SLAM map and robot position data.

[0064] S1 includes the following steps:

[0065] S11. The robot uses a lidar sensor to scan the environment, acquire environmental point cloud data, including obstacle distance and shape parameters, and generate static environmental feature data.

[0066] S12. Video stream is captured by a robot equipped with a camera sensor. The trajectory of moving objects is identified based on optical flow analysis to generate dynamic environmental feature data.

[0067] S13. Integrate static environmental feature data and dynamic environmental feature data, process them synchronously according to timestamps, and generate robot sensor data and dynamic environmental feature data.

[0068] S2 includes the following steps:

[0069] S21. Based on robot sensor data and environmental dynamic feature data, a convolutional neural network model is used to perform feature classification processing, distinguish between static features and dynamic features, and generate feature classification data.

[0070] S22. Perform cluster analysis on the dynamic features in the feature classification data, calculate the centroid position and motion vector of the moving object, and generate dynamic feature extraction data. The formula for calculating the centroid position is:

[0071] ;

[0072] in For the centroid coordinates of dynamic object clusters, Let j be the spatial coordinates of the j-th dynamic feature point. This represents the number of dynamic feature points.

[0073] S3 includes the following steps:

[0074] S31. Monitor the rate of change of dynamic environmental characteristics. When the density of dynamic characteristics exceeds the threshold, trigger the adaptive adjustment mechanism.

[0075] S32. Adjust the number of particle filters in the SLAM algorithm according to the rate of change of dynamic characteristics to reduce the computational load. The formula for adjusting the number of particles is:

[0076] ;

[0077] in To adjust the number of particles, The number of basic particles, For the rate of change of dynamic characteristics, The maximum permissible rate of change threshold;

[0078] S33. Synchronously optimize map update frequency, dynamically set update interval based on environmental stability index, and generate adaptive adjustment data.

[0079] S4 includes the following steps:

[0080] S41. Align the dynamic feature extraction data and adaptive adjustment data according to the mileage parameter, and construct a time-space matrix. The matrix construction formula is as follows:

[0081] ;

[0082] in For time-space matrix, Let k be the mileage parameter at time k. Let k be the spatial coordinates at time k. For dynamic feature data;

[0083] S42. Mark the dynamic feature hotspot regions in the matrix to generate SLAM optimized combination data.

[0084] S5 includes the following steps:

[0085] S51. Establish a pre-stored standard optimization algorithm database, including standard optimization algorithm data corresponding to genetic algorithm, particle swarm optimization algorithm and simulated annealing algorithm;

[0086] S52. Perform pattern matching between the SLAM optimized combination data and the standard optimization algorithm data, and calculate the similarity score. The formula for calculating the similarity score is as follows:

[0087] ;

[0088] in To score the similarity, Let i be the i-th feature parameter in the SLAM optimization combined data, where i is the feature parameter index. For the i-th standard parameter corresponding to the h-th algorithm in the standard optimization algorithm data, For algorithm type index, This represents the total number of feature parameters.

[0089] S52 includes the following steps:

[0090] S521. Initialize the biomimetic optimization search population, randomly set the search agent positions using the algorithm, and the position initialization formula is:

[0091] ;

[0092] in To search for the initial location of the agent, The lower boundary of the search space, Let r be the upper boundary of the search space, and let r represent a random number in the interval [0,1].

[0093] S522. During the exploration phase, the simulated search agent randomly detects similar targets, updates the location, and evaluates the fitness value. The location update formula is:

[0094] ;

[0095] in For the updated position, Current position For the target similarity position, To adjust the coefficients, if the updated fitness value is better, the original position is replaced;

[0096] S523. During the development phase, optimize the search agent position to lock in the optimal matching algorithm, and output the target optimization algorithm type data. The position optimization formula is:

[0097] ;

[0098] in For the optimal position, This is the current optimal position. For learning rate, As a global attraction factor, This is the local oscillation suppression coefficient.

[0099] S6 includes the following steps:

[0100] S61. Call the optimization program according to the target optimization algorithm type data to process the dynamic feature noise in the SLAM optimization combination data;

[0101] S62. Enhance the accuracy of static features and reconstruct the environment map;

[0102] S63. Integrate robot position data to generate an optimized SLAM map and robot position data.

[0103] S62 includes the following steps:

[0104] S621. The iterative nearest-point algorithm is used to correct static feature alignment. The formula for minimizing the error is:

[0105] ;

[0106] in For the nearest point error in iteration, For the i-th point in the target point cloud, Let i be the i-th point in the source point cloud. For rotation matrix, It is a translation vector. For the number of matching point pairs;

[0107] S622: Apply Kalman filtering to smooth the robot trajectory and output an optimized map.

[0108] The method also includes step S7:

[0109] S71. Real-time output of optimized SLAM map and robot position data to the navigation system;

[0110] S72. Adjust the robot path planning based on the output data to achieve adaptive navigation.

[0111] The steps of this SLAM dynamic feature optimization method based on robot adaptive navigation are as follows:

[0112] Step 1: Multimodal environment perception and deep data fusion

[0113] The robotic system uses a high-precision LiDAR to perform omnidirectional environmental scanning, capturing the 3D spatial coordinates and geometric contour features of static obstacles. Simultaneously, it utilizes a high-definition camera to continuously capture dynamic video stream data, combining this with optical flow analysis algorithms to analyze the displacement trajectory, velocity vector, and acceleration trends of moving objects in real time. A timestamp synchronization fusion technique is employed to register static point cloud data and dynamic trajectory data in a unified spatiotemporal coordinate system, constructing a comprehensive environmental feature dataset that includes environmental topology, dynamic object behavior patterns, and real-time changing characteristics. This provides a complete perceptual data foundation for subsequent dynamic feature optimization processing.

[0114] Step 2: Intelligent Extraction and Refined Classification of Dynamic Features

[0115] This study utilizes a deep convolutional neural network architecture to perform multi-level feature mining on fused data. Adaptive convolutional kernels are used to identify static structural features in the environment, including continuous wall contours and fixed obstacle boundaries, as well as dynamic behavioral features, including the instantaneous velocity direction and trajectory curvature of moving objects. Density clustering optimization is applied to the classified dynamic features to calculate the centroid displacement trajectory, rate of change of motion direction, and group interaction behavior patterns of moving objects. This generates a standardized set of dynamic object behavior feature vectors, addressing data distortion issues caused by motion blur and feature overlap.

[0116] Step 3: Dynamic Adjustment and Resource Optimization Mechanism for SLAM Parameters

[0117] The system monitors the gradient of dynamic feature density changes in the environment in real time. When the number of moving objects exceeds the environmental carrying capacity threshold or the movement speed reaches the sensor sampling limit, it autonomously triggers a multi-level parameter adjustment strategy. It dynamically adjusts the particle filter size to balance computational accuracy and real-time requirements, and nonlinearly adjusts the map update frequency based on the environmental stability index. This establishes a dynamic balance model between resource consumption and positioning accuracy, ensuring system response efficiency during sudden dynamic events.

[0118] Step 4: Construction of Spatiotemporal Correlation Matrix and Feature Hotspot Labeling

[0119] Tensor fusion technology is employed to align the dynamic object behavior feature vectors with SLAM parameter adjustment data in both time and space, constructing a multi-dimensional feature matrix that includes time series, spatial coordinates, and motion parameters. A hotspot detection algorithm is used to identify intersections of moving object trajectories, points of abrupt velocity changes, and regions of abnormal behavior, forming a dynamic feature topology with spatiotemporal evolution characteristics, providing a structured input framework for intelligent matching algorithms.

[0120] Step 5: Optimize the algorithm for intelligent matching and dynamic noise filtering

[0121] Multi-dimensional similarity matching is performed between the spatiotemporal feature matrix and a pre-stored algorithm database. A biomimetic optimization algorithm is used to search for high-matching regions in a wide area during the exploration phase, and a gradient descent strategy is employed to lock in the optimal algorithm type during the development phase. Frequency domain filtering of dynamic feature noise is performed to enhance the reconstruction accuracy of static environmental features. An iterative nearest-point algorithm is applied to correct sensor pose offset errors, and adaptive Kalman filtering is combined to achieve smooth optimization of motion trajectories, generating a highly consistent environmental map.

[0122] Step Six: Closed-Loop Navigation Decision and Dynamic Obstacle Avoidance Control

[0123] The optimized SLAM map and robot localization data are input into the navigation decision engine in real time. Based on the dynamic obstacle motion prediction model and static environment topology analysis, the optimal obstacle avoidance path and speed control strategy are generated. The path execution deviation is calibrated in real time through a feedback mechanism, establishing a closed-loop control system of perception-decision-execution to achieve response to sudden dynamic events and continuous safe navigation.

[0124] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0125] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic feature optimization method for SLAM based on robot adaptive navigation, characterized by: The method includes the following steps: S1. Collect robot sensor data and environmental dynamic feature data, use multimodal sensors to obtain real-time environmental information, including static obstacle parameters and dynamic moving object parameters, and generate relevant data, including the following steps: S11. The robot uses a lidar sensor to scan the environment, acquire environmental point cloud data, including obstacle distance and shape parameters, and generate static environmental feature data. S12. Video stream is captured by a robot equipped with a camera sensor. The trajectory of moving objects is identified based on optical flow analysis to generate dynamic environmental feature data. S13. The static environmental feature data and dynamic environmental feature data are fused together and processed synchronously according to timestamps to generate robot sensor data and dynamic environmental feature data. S2. Perform dynamic feature extraction processing. Based on sensor data and environmental dynamic feature data, use feature recognition algorithms to extract the speed, direction and size parameters of moving objects and generate dynamic feature extraction data. S3. Perform adaptive adjustment of SLAM parameters. Based on dynamic feature extraction data and environmental change indicators, adjust the positioning accuracy parameters and map update frequency parameters of the SLAM algorithm in real time to generate adaptive adjustment data. S4. Construct SLAM optimized combined data, aligning and combining dynamically extracted data with adaptively adjusted data according to time series, including the following steps: S41. Align the dynamic feature extraction data and the adaptive adjustment data according to the mileage parameter to construct a time-space matrix; S42. Mark the dynamic feature hotspot regions in the matrix to generate SLAM optimized combination data; S5. Perform dynamic feature optimization algorithm matching processing. Based on the SLAM optimization combination data, perform similarity matching with the pre-stored standard optimization algorithm data, search for the optimal dynamic feature optimization algorithm type, and generate target optimization algorithm type data. S6. Perform SLAM dynamic feature optimization processing, call the optimization program corresponding to the target optimization algorithm type data to process the data, filter out dynamic feature noise and enhance the accuracy of static features, and generate optimized SLAM map and robot position data.

2. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 1, characterized in that: S2 includes the following steps: S21. Based on the robot sensor data and environmental dynamic feature data, a convolutional neural network model is used to perform feature classification processing, distinguish between static features and dynamic features, and generate feature classification data. S22. Perform cluster analysis on the dynamic features in the feature classification data, calculate the centroid position and motion vector of the moving object, and generate dynamic feature extraction data.

3. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 1, characterized in that: S3 includes the following steps: S31. Monitor the rate of change of dynamic environmental characteristics. When the density of dynamic characteristics exceeds the threshold, trigger the adaptive adjustment mechanism. S32. Adjust the number of particle filters in the SLAM algorithm according to the rate of change of dynamic characteristics to reduce the computational load; S33. Synchronously optimize map update frequency, dynamically set update interval based on environmental stability index, and generate adaptive adjustment data.

4. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 1, characterized in that: S5 includes the following steps: S51. Establish a pre-stored standard optimization algorithm database, including standard optimization algorithm data corresponding to genetic algorithm, particle swarm optimization algorithm and simulated annealing algorithm; S52. Perform pattern matching between the SLAM optimized combination data and the standard optimization algorithm data, and calculate the similarity score, wherein the formula for calculating the similarity score is: ； in To score the similarity, Let i be the i-th feature parameter in the SLAM optimization combined data, where i is the feature parameter index. For the i-th standard parameter corresponding to the h-th algorithm in the standard optimization algorithm data, For algorithm type indexing, This represents the total number of feature parameters.

5. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 4, characterized in that: S52 includes the following steps: S521. Initialize the biomimetic optimization search population and randomly set the search agent positions for the algorithm; the position initialization formula is: ； in To search for the initial location of the agent, The lower boundary of the search space, Let r be the upper boundary of the search space, and let r represent a random number in the interval [0,1]. S522. During the exploration phase, a simulated search agent randomly detects similar targets, updates their locations, and evaluates their fitness values; the location update formula is: ； in For the updated position, Current position For the target similarity position, To adjust the coefficients, if the updated fitness value is better, the original position is replaced; S523. During the development phase, optimize the search agent location to lock in the optimal matching algorithm and output the target optimization algorithm type data.

6. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 1, characterized in that: S6 includes the following steps: S61. Call the optimization program according to the target optimization algorithm type data to process the dynamic feature noise in the SLAM optimization combination data; S62. Enhance the accuracy of static features and reconstruct the environment map; S63. Integrate robot position data to generate an optimized SLAM map and robot position data.

7. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 6, characterized in that: S62 includes the following steps: S621. The iterative nearest-point algorithm is used to correct static feature alignment. S622: Apply Kalman filtering to smooth the robot trajectory and output an optimized map.

8. The SLAM dynamic feature optimization method based on robot adaptive navigation according to claim 1, characterized in that: The method further includes step S7: S71. Real-time output of optimized SLAM map and robot position data to the navigation system; S72. Adjust the robot path planning based on the output data to achieve adaptive navigation.

Images