A vehicle intelligent navigation method and system

Abstract

This invention belongs to the field of vehicle navigation technology and provides a method and system for intelligent vehicle navigation. The method involves acquiring 3D point cloud data to be processed during vehicle navigation; spatially partitioning the 3D point cloud data and solving for the normal vectors of each point; calculating the angles between the normal vectors using the obtained normal vectors; and constructing a clustering algorithm that integrates geometric features, along with a clustering objective function, based on these angles. The method employs the first defense mechanism of the crowned porcupine algorithm to improve the hunting phase of the black-winged kite algorithm, obtaining a local optimization strategy. It combines the aggregation behavior of the artificial fish swarm algorithm with the migration mechanism of the black-winged kite algorithm to obtain a global optimization strategy, resulting in a combined intelligent algorithm. This combined intelligent algorithm is used to improve the clustering algorithm, iteratively optimizing the cluster centers until the point cloud cluster division results are obtained. Vehicle navigation is then performed based on the point cloud cluster division results. This invention improves the accuracy of point classification in intelligent vehicle navigation.

Description

Technical Field

[0001] This invention belongs to the field of vehicle navigation technology, specifically relating to a vehicle intelligent navigation method and system. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] LiDAR is a sensor that can accurately detect the position of objects. In the field of intelligent vehicles, in addition to helping vehicles perceive the environment, LiDAR can also assist vehicles in planning routes and intelligent navigation. When existing LiDAR is used for vehicle route planning and intelligent navigation, it generally relies on point cloud data to complete tasks such as environmental perception, path planning and dynamic obstacle avoidance.

[0004] Point cloud data serves as the core basis for vehicles to perceive and adapt to complex environments, and its processing accuracy and efficiency directly determine the performance ceiling of navigation systems. Existing machine learning methods often employ clustering when directly processing point clouds, with similarity overemphasis on the Euclidean distance of points. The cluster outlines are generally spherical, failing to adequately mine local geometric features, especially the poor clustering effect of point clouds at the outline edges. Furthermore, the clustering strategies employed are still limited to traditional methods, making it difficult to accurately represent model features and often failing to meet the high-precision navigation requirements. Summary of the Invention

[0005] To address the aforementioned problems, this invention proposes a vehicle intelligent navigation method and system. This invention utilizes a combined intelligent algorithm for clustering and introduces a cluster similarity judgment method based on the geometric features of the clustered point cloud data. By calculating the angle between point cloud vectors and combining it with Euclidean distance, point clouds with very similar features can be accurately classified into corresponding point cloud clusters, thereby determining the region or category to which each point belongs and improving the accuracy of point classification in vehicle intelligent navigation.

[0006] According to some embodiments, the present invention adopts the following technical solution:

[0007] A vehicle intelligent navigation method includes the following steps:

[0008] Acquire the 3D point cloud data to be processed during vehicle navigation;

[0009] A binary spatial partitioning tree index structure is used to spatially partition the 3D point cloud data to be processed, and the normal vector of each point is solved.

[0010] Using the obtained normal vectors, the angle between the normal vectors is calculated. Based on the angle, a similarity function that incorporates geometric features and a clustering objective function are constructed in the clustering algorithm.

[0011] The first defense mechanism of the crowned porcupine algorithm is used to improve the hunting phase of the black-winged kite algorithm to obtain a local optimization strategy. The aggregation behavior of the artificial fish swarm algorithm and the migration mechanism of the black-winged kite algorithm are combined to obtain a global optimization strategy. The combination of the local optimization strategy and the global optimization strategy yields a combinatorial intelligent algorithm.

[0012] The clustering algorithm is improved by using combinatorial intelligent algorithms, and the cluster centers are iteratively optimized until the final cluster centers are obtained, thereby obtaining the point cloud cluster division results;

[0013] Vehicle navigation is performed based on the point cloud cluster division results.

[0014] As an alternative implementation, the clustering algorithm is the K-means clustering algorithm.

[0015] As an alternative implementation method, the clustering process depends not only on the Euclidean distance between the point to be determined and the center point, but also on the continuity and correlation between the point to be determined and its nearest other point to be determined. This is necessary to accurately describe the object's outline. Therefore, the similarity score fused with geometric features is as follows:

[0016] ;

[0017] in, It is a point to be determined. With the center point The Euclidean distance is given by the formula:

[0018] ;

[0019] in, and It is the first of these two points m Each coordinate value M It is the dimension of the point cloud;

[0020] The similarity value of the angle between the normal vectors is calculated as follows:

[0021] ;

[0022] ;

[0023] in, Point to be determined With the center point of this cluster The angle between the vectors, To traverse the cloud cluster at this point and closest included angle point and The angle between the vectors.

[0024] As an alternative implementation, the clustering objective function for:

[0025] ;

[0026] in, Point to be determined With the center point of this cluster The similarity between them, where K is the total number of clusters, and the smaller the objective function value, the better.

[0027] As an alternative implementation, the process of improving the hunting phase of the Black-winged Kite algorithm with the first defense mechanism of the Crowned Porcupine algorithm to obtain a local optimization strategy includes: The first defense mechanism of the Crowned Porcupine algorithm is as follows:

[0028] ;

[0029] in, and Let represent the specific positions of the i-th individual in the population after iterations t+1 and t, respectively. These are random numbers based on a normal distribution. It is a random number in the interval [0,1]. This represents the optimal solution at the t-th iteration. This represents the position vector between the current hooded porcupine and a randomly selected hooded porcupine from the current hooded porcupine population, used to represent the predator's position in the t-th iteration. Its mathematical formula is: ,in, r It is 1 and the population size N Random numbers between;

[0030] Based on the optimal values ​​obtained from two iterations of the crowned porcupine test, the conditional judgment criteria in the mathematical model of the black-winged kite's attack behavior are calculated. for:

[0031] ;

[0032] This represents the optimal solution at the t-th iteration. Denotes the optimal solution at the (t-1)th iteration, when When it is greater than 0.9, let It is 0.9; When the attack formula is active, the Black-winged Kite's own attack formula is used; otherwise, the Crowned Porcupine's first defensive mechanism formula is incorporated into the Black-winged Kite's attack formula, and the Black-winged Kite's formula is modified accordingly. The correction factor is used to derive the formula for the improved attack behavior of the black-winged kite:

[0033] ;

[0034] ;

[0035] ;

[0036] in, Let represent the specific positions of the i-th individual in the black-winged kite population in the j-th dimension at iterations t+1 and t, respectively; r is a random number in the interval [0, 1] used to control random behavior; n is a dynamic decay coefficient used to control the search step size; T represents the total number of iterations; and t represents the number of the current iteration.

[0037] The goal is to enable the black-winged kite to explore as closely as possible and find the optimal value at that moment, thus achieving local optimization.

[0038] As an alternative implementation method, the process of obtaining a global optimization strategy by combining the aggregation and tail-chasing behaviors of the artificial fish swarm algorithm with the migration mechanism of the black-winged kite algorithm includes:

[0039] In the global exploration phase, based on the formula of artificial fish swarm vision, the first step is to... Centered on the visual perception distance, the number of black-winged kites within the visual perception distance is: Calculate the center position of each black-winged kite. And calculate the fitness value at that location. This value is also the objective function of clustering. This study borrows the aggregation behavior from the artificial fish swarm algorithm as a conditional judgment criterion for the migration behavior of black-winged kites, and sets the optimal value of the artificial fish swarm as the basis for this judgment. Incorporating this into the formula, the improved mathematical model of black-winged kite migration behavior is as follows: ;

[0040] in, Let represent the specific positions of the i-th individual in the black-winged kite population in the j-th dimension at iterations t+1 and t, respectively. This indicates the leading scorer of the Black-winged Kite in the j-th dimension of the t-th iteration so far; Let represent the fitness value of any random position of a black-winged kite in the t-th iteration. For random numbers that conform to the Cauchy distribution, Represents a random number between (0,1). This represents the crowding level of the artificial fish positions in the artificial fish swarm algorithm. The stride length of the artificial fish;

[0041] If the black-winged kite migration behavior formula is used, then the improved mathematical model of black-winged kite migration behavior, which is combined with the artificial fish swarming behavior formula, is used to obtain the improved global optimization strategy.

[0042] As an alternative implementation, the process of combining local optimization strategies and global optimization strategies to obtain a combined intelligent algorithm includes: when a black-winged kite finds a promising region in the solution space, the remaining black-winged kites will gather towards that black-winged kite, which is... The optimal solution obtained by the improved global optimization strategy is compared with the optimal solution obtained by the local optimization strategy. The best solution is selected from the two to obtain the combinatorial intelligent algorithm, where a black-winged kite represents a solution.

[0043] As an alternative implementation method, the process of improving the clustering algorithm using combinatorial intelligent algorithms and iteratively optimizing the cluster centers includes:

[0044] K cluster centers were calculated using a combinatorial intelligent algorithm. Its objective function value is The local stagnation threshold is defined as ;

[0045] Employing a local optimization strategy for the current center Optimize the function and record the changes in the objective function value during continuous iterations;

[0046] If N1 consecutive iterations satisfy:

[0047] ;

[0048] If the system is found to be stuck in a local stagnation, the number of times it occurs is recorded as NS. A new center is then generated using a global optimization strategy. Compare the objective function values: If Then let And enter the local optimization stage; otherwise, maintain the original center. Continue to seek local optimization. , These are the objective function values ​​for the t-th and t-1-th iterations, respectively;

[0049] In the local optimization process, if it occurs N2 times consecutively If local degradation occurs, a global optimization strategy is used to generate [the appropriate generator]. A new center Choose the optimal center:

[0050] ;

[0051] Compare the objective function values, if Then let And enter the local optimization stage; otherwise, maintain the original center. Continue to seek local optimization;

[0052] When NS reaches the set value, it is considered that the optimal cluster center has been found.

[0053] As an alternative implementation method, the process of vehicle navigation based on point cloud clustering results includes identifying passable areas and obstacle outlines in the current environment based on point cloud clustering results, planning a path within the passable area to avoid obstacle outlines, and achieving navigation.

[0054] A vehicle intelligent navigation system, comprising:

[0055] The data acquisition module is configured to acquire the 3D point cloud data to be processed during vehicle navigation.

[0056] The data indexing module is configured to use a binary spatial partitioning tree index structure to spatially partition the 3D point cloud data to be processed and solve for the normal vector of each point.

[0057] The data preprocessing module is configured to use the obtained normal vectors to calculate the angle between the normal vectors, and based on the angle, construct a similarity function that incorporates geometric features and a clustering objective function in the clustering algorithm;

[0058] The algorithm improvement module is configured to improve the hunting phase of the black-winged kite algorithm by adopting the first defense mechanism of the crowned porcupine algorithm to obtain a local optimization strategy, and combine the aggregation behavior of the artificial fish swarm algorithm with the migration mechanism of the black-winged kite algorithm to obtain a global optimization strategy. The combination of the local optimization strategy and the global optimization strategy yields a combinatorial intelligent algorithm.

[0059] The clustering module is configured to improve the clustering algorithm using combinatorial intelligent algorithms, iteratively optimize the cluster centers until the final cluster centers are obtained, and thus obtain the point cloud cluster division results;

[0060] The navigation execution module is configured to perform vehicle navigation based on the point cloud clustering results.

[0061] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0062] This invention utilizes a combined intelligent algorithm to optimize clustering. It introduces a cluster similarity judgment method based on the geometric features of clustered point cloud data. By calculating the angle between point cloud vectors and combining it with Euclidean distance, point clouds with very similar features can be accurately classified into corresponding point cloud clusters, thereby determining the region or category to which each point belongs. This improves the accuracy of point classification in intelligent navigation and effectively enhances the safety of intelligent driving of vehicles.

[0063] First, this invention improves the black-winged kite hunting phase algorithm by employing the first defense mechanism of the crowned porcupine algorithm, resulting in a local optimization strategy. Simultaneously, it combines the aggregation behavior from the artificial fish swarm algorithm with the black-winged kite migration mechanism to escape local optima and perform global optimization, thus obtaining a combined intelligent algorithm.

[0064] Secondly, the clustering process depends not only on the Euclidean distance between the undetermined point and the center point, but also on the continuity and correlation between the undetermined point and the nearest undetermined point. This is necessary to accurately describe the object's outline. By comprehensively considering the angle between undetermined points and the Euclidean distance between the undetermined point and the center point, a more accurate similarity metric is established. This method can improve the similarity of point cloud data of the same object's outline, so that the point cloud data of the same object's outline can be grouped into the same cluster. By incorporating an improved clustering objective function that integrates geometric features, combined with an adaptive optimization mechanism, i.e., an intelligent optimization strategy, global exploration is triggered when the local search of the center point stagnates, and multiple candidate centers are selected when the optimization degenerates, ensuring that the clustering process is accurate and efficient, thereby providing more accurate environmental feature recognition for robot navigation.

[0065] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0066] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0067] Figure 1 This is a schematic flowchart of a vehicle intelligent navigation method according to one embodiment. Detailed Implementation

[0068] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0069] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0070] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0071] Where there is no conflict, the embodiments and features described in this application may be combined with each other.

[0072] Example 1

[0073] Point clouds, as the core carrier for vehicles to understand three-dimensional space, directly determine the performance of navigation systems in path planning and dynamic obstacle avoidance. However, existing machine learning-based point cloud processing methods have significant limitations: the algorithms focus excessively on the Euclidean distance relationship of points, neglecting in-depth mining of local geometric features; the clustering strategies adopted still adhere to traditional modes, making it difficult to achieve efficient feature representation in complex scenes, resulting in environmental perception accuracy and real-time performance that cannot meet the needs of demanding navigation tasks.

[0074] To address the above problems, this embodiment provides a vehicle intelligent navigation method, including the following steps:

[0075] Step 1: Acquire the 3D point cloud data to be processed during vehicle navigation.

[0076] Step 2: Based on the point cloud data acquired by LiDAR scanning, a binary spatial partitioning tree index structure is used to spatially organize it. Then, the normal vector of each point is calculated through the point cloud data. Using the obtained normal vector information, the angle between point cloud data is calculated, and the similarity and objective function of improved clustering that integrates geometric features are constructed.

[0077] Step 3: The first defense mechanism of the Crowned Porcupine algorithm is used to improve the hunting phase of the Black-winged Kite algorithm, resulting in a local optimization strategy. Simultaneously, the aggregation behavior of the artificial fish swarm algorithm and the migration mechanism of the Black-winged Kite algorithm are combined to obtain a global optimization strategy, thus deriving a combined intelligent algorithm. During point cloud data clustering, the combined intelligent algorithm dynamically selects appropriate strategies based on optimization requirements, ultimately obtaining accurate point cloud cluster division results. Based on the point cloud cluster division results, passable areas and obstacle outlines in the current environment are identified. Paths are planned within passable areas to avoid obstacle outlines, achieving navigation.

[0078] After the vehicle acquires three-dimensional point cloud data of the driving environment through LiDAR scanning technology, a binary spatial partitioning tree index is constructed to efficiently organize the point cloud spatially, and the normal vectors of each point cloud cluster are further calculated.

[0079] In this embodiment, the clustering method used is K-means clustering, and the clustering similarity improvement process is as follows:

[0080] ;

[0081] in, The Euclidean distance between the point to be determined and the center point is given by the following formula:

[0082] ;

[0083] in, and They are two points in a 3D point cloud. and These are the m-th coordinates of these two points, where M is the dimension of the point cloud.

[0084] The similarity value of the angle between the normal vectors is specifically calculated using the following formula:

[0085] ;

[0086] ;

[0087] in, Point to be determined With the center point of this cluster The angle between the vectors, To traverse the cloud cluster at this point and closest included angle point and The vector angle is used to improve the similarity of point cloud data of the same object contour, so that the point cloud data of the same object contour can be grouped into the same cluster.

[0088] Based on the new similarity metric, an improved clustering method is constructed, and the objective function is determined as follows:

[0089] ;

[0090] In this embodiment, the smaller the objective function value, the better the clustering effect. The clustering operation of the point cloud data is completed, and each data point is accurately divided into the most suitable point cloud cluster, so as to realize high-precision clustering of SLAM navigation point cloud data and determine the contours of each object in the driving environment.

[0091] The process of improving the Black-winged Kite algorithm using the Crowned Porcupine optimization algorithm is as follows:

[0092] The first step is to perform an optimization phase, using the first defense mechanism formula of the Crowned Porcupine algorithm to improve the attack behavior of the Blackwing Kite.

[0093] The formula for the first defense mechanism of the Crowned Porcupine algorithm is as follows:

[0094] In the formula, and Let represent the specific positions of the i-th individual in the population after iterations t+1 and t, respectively. These are random numbers based on a normal distribution. It is a random number in the interval [0,1]. This represents the optimal solution at the t-th iteration. This represents the position vector between the current hooded porcupine and a randomly selected hooded porcupine from the current hooded porcupine population, used to represent the predator's position in the t-th iteration. Its mathematical formula is: ,in, r It is 1 and the population size N A random number between [a certain number of points].

[0095] Based on the optimal values ​​obtained from two iterations of the crowned porcupine test, the conditional judgment criteria in the mathematical model of the black-winged kite's attack behavior are calculated. for:

[0096] This represents the optimal solution at the t-th iteration. Denotes the optimal solution at the (t-1)th iteration, when When it is greater than 0.9, let It is 0.9. When the attack formula is active, the Black-winged Kite's own attack formula is used; otherwise, the Crowned Porcupine's first defensive mechanism formula is incorporated into the Black-winged Kite's attack formula, and the Black-winged Kite's formula is modified accordingly. The correction factor is used to derive the formula for the improved attack behavior of the black-winged kite:

[0097] ;

[0098] ;

[0099] ;

[0100] in, Let represent the specific positions of the i-th individual in the black-winged kite population in the j-th dimension at iterations t+1 and t, respectively. r is a random number in the interval [0,1] used to control random behavior. n is a dynamic decay coefficient used to control the search step size. T represents the total number of iterations, and t represents the current iteration number.

[0101] This allows the black-winged kite to explore as closely as possible and find the optimal value at that moment.

[0102] The method combines the aggregation behavior of artificial fish swarming algorithms with the migration mechanism of black-winged kites to escape local optima and perform global optimization. The construction process is as follows:

[0103] Next, we enter the global exploration phase. Based on the formula for the artificial fish swarm's field of vision, we first use... The number of black-winged kites within a circle centered on the perceived distance Visual is: ( (excluding the black-winged kite), find their center positions. And calculate the fitness value at that location. This value is also the objective function for clustering, and it serves as the conditional basis for judging the migration behavior of black-winged kites. The improved mathematical model of black-winged kite migration behavior is as follows: ;

[0104] This indicates the leading scorer of the Black-winged Kite in the j-th dimension of the t-th iteration so far. and Let represent the position of the i-th black-winged kite in the j-th dimension, and its position in the t-th and t+1-th iteration steps, respectively. Let represent the fitness value of any random position of a black-winged kite in the t-th iteration, and n represent the current number of black-winged kites. For random numbers that conform to the Cauchy distribution, This represents a random number between (0, 1).

[0105] In some cases, the black-winged kite migration behavior formula is used; otherwise, a black-winged kite migration formula combined with the artificial fish swarming behavior formula is used, thus obtaining an improved global optimization algorithm.

[0106] When a "black-winged kite" (representing a solution) finds a promising region in the solution space, the remaining "black-winged kites" will gather towards that "black-winged kite". The optimal solution obtained through optimization is compared with the optimal solution obtained from the improved black-winged kite attack behavior, and the optimal solution is selected from the two. This yields the combinatorial intelligent algorithm.

[0107] Improving the clustering process using combinatorial intelligent algorithms, such as Figure 1 As shown, perform the following steps:

[0108] K cluster centers were calculated using a combinatorial intelligent algorithm. Its objective function value is The local stagnation threshold is defined as ,make .

[0109] Employing local optimization strategies from combinatorial intelligent algorithms to optimize the current center Optimize the function and record the changes in the objective function value during continuous iterations.

[0110] If the following conditions are met in 10 consecutive iterations:

[0111] ;

[0112] If the system is found to be stuck in a local stagnation, the number of times it occurs is recorded as NS. A new center is then generated using a global optimization strategy. Compare the objective function values: If Then let And enter the local optimization stage; otherwise, maintain the original center. Continue to seek local optimization.

[0113] In the process of local optimization, if it occurs 3 times in a row If local degradation occurs, a global optimization strategy is used to generate [the appropriate generator]. A new center Choose the optimal center:

[0114] ;

[0115] Compare the objective function values: If Then let And enter the local optimization stage; otherwise, maintain the original center. Continue local optimization. The above process dynamically evaluates the changing trend of the objective function and adaptively switches between local refinement and global exploration to ensure that the clustering process maintains convergence accuracy while also having the ability to escape local optima.

[0116] When NS=1000, it is considered that the optimal cluster center has been found.

[0117] Of course, the above values ​​can be adjusted according to requirements such as navigation accuracy and response speed.

[0118] Example 2

[0119] A vehicle intelligent navigation system, comprising:

[0120] The data acquisition module is configured to acquire the 3D point cloud data to be processed during vehicle navigation.

[0121] The data indexing module is configured to use a binary spatial partitioning tree index structure to spatially partition the 3D point cloud data to be processed and solve for the normal vector of each point.

[0122] The data preprocessing module is configured to use the obtained normal vectors to calculate the angle between the normal vectors, and based on the angle, construct a similarity function that incorporates geometric features and a clustering objective function in the clustering algorithm;

[0123] The algorithm improvement module is configured to improve the hunting phase of the black-winged kite algorithm by adopting the first defense mechanism of the crowned porcupine algorithm to obtain a local optimization strategy, and combine the aggregation behavior of the artificial fish swarm algorithm with the migration mechanism of the black-winged kite algorithm to obtain a global optimization strategy. The combination of the local optimization strategy and the global optimization strategy yields a combinatorial intelligent algorithm.

[0124] The clustering module is configured to improve the clustering algorithm using combinatorial intelligent algorithms, iteratively optimize the cluster centers until the final cluster centers are obtained, and thus obtain the point cloud cluster division results;

[0125] The navigation execution module is configured to perform vehicle navigation based on the point cloud clustering results.

[0126] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of one or more computer-usable storage media (including, but not limited to, disk storage, etc.) containing computer-usable program code. CD - ROM It takes the form of a computer program product implemented on (such as optical memory, etc.).

[0127] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0128] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0129] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0130] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art without creative effort within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent navigation of a vehicle, characterized in that, The method comprises the following steps: acquiring three-dimensional point cloud data to be processed in a vehicle navigation process; adopting a binary spatial partition tree index structure to perform spatial division on the three-dimensional point cloud data to be processed, and solving normal vectors of each point; calculating angles between the normal vectors by using the solved normal vectors, and constructing similarity with geometric features in a clustering algorithm and a clustering objective function based on the angles; the similarity with geometric features is: ； wherein, is the pending point and the center point Euclidean distance, the formula is: ； in, and It is the first of these two points m Each coordinate value M It is the dimension of the point cloud; For the normal vector angle similarity value, the normal vector angle similarity value of the undetermined point is calculated as: ； ； wherein, is the pending point the vector angle of the cluster center point with this point cloud cluster, is the point with the closest angle in this point cloud cluster with the vector angle of the cluster center point The clustering objective function is: ； wherein, is the pending point the similarity between this cluster center point and the Kth cluster center, K is the total number of clustering clusters, the smaller the objective function value is, the better. adopting a first defense mechanism of the Crowned Porcupine Algorithm to improve a hunting stage of the Black Kite Algorithm to obtain a local optimization strategy, combining a gathering behavior and a following behavior of the Artificial Fish Swarm Algorithm with a migration mechanism of the Black Kite Algorithm to obtain a global optimization strategy, and combining the local optimization strategy and the global optimization strategy to obtain a combined intelligent algorithm; the process of adopting the first defense mechanism of the Crowned Porcupine Algorithm to improve the hunting stage of the Black Kite Algorithm to obtain the local optimization strategy comprises: ； where, and denote the specific position of the i-th individual of the population after t+1 and t iterations, respectively, is a random number based on a normal distribution, is a random number in the interval [0, 1], denotes the optimal solution at the t-th iteration, denotes the position vector between the current crown-hyena and a randomly selected crown-hyena from the current population of crown-hyenas, which is used to represent the position of the predator at the t-th iteration, and its mathematical formula is: where, R is a random number between 1 and the population size N ; According to the optimal values of the two iterations of the crown porcupine, the condition judgment basis in the mathematical model of the black-winged kite attack behavior is calculated is: ； This represents the optimal solution at the t-th iteration. Denotes the optimal solution at the (t-1)th iteration, when When it is greater than 0.9, let It is 0.9; When the attack formula is active, the Black-winged Kite's own attack formula is used; otherwise, the Crowned Porcupine's first defensive mechanism formula is incorporated into the Black-winged Kite's attack formula, and the Black-winged Kite's formula is modified accordingly. The correction factor is used to derive the formula for the improved attack behavior of the black-winged kite: ； ； ； where, respectively represent the specific position of the i-th individual in the black-winged kite population in the j-th dimension at t+1 and t iterations, r is a random number in the interval [0, 1] for controlling the random behavior; n is a dynamic attenuation coefficient for controlling the search step, T represents the total number of iterations, and t represents the number of the current iteration; making the Black Kite explore as close as possible to obtain an optimal value at this time, and realizing local optimization; The process of combining the local optimization strategy and the global optimization strategy to obtain the combined intelligent algorithm comprises: when one black-winged kite finds a promising area in the solution space, the rest of the black-winged kites will gather around the black-winged kite, and the black-winged kite is the The optimal solution obtained by the improved global optimization strategy is compared with the optimal solution obtained by the local optimization strategy, and the optimal solution is selected, to obtain the combined intelligent algorithm, wherein one black-winged kite represents one solution. the process of combining the gathering behavior and the following behavior of the Artificial Fish Swarm Algorithm with the migration mechanism of the Black Kite Algorithm to obtain the global optimization strategy comprises: In the global exploration stage, based on the formula of artificial fish swarm vision, first take as the center, the number of black-winged kite in the distance Visual is , the center position of each black-winged kite is calculated , and the fitness value of the position is calculated , which is also the objective function of clustering , the aggregation behavior in artificial fish swarm algorithm is used as the condition judgment basis of black-winged kite migration behavior, and the optimal value of artificial fish swarm is integrated into the formula, and the improved mathematical model of black-winged kite migration behavior is as follows: ; wherein, respectively represent the specific position of the i-th individual in the black kite population in the j-th dimension at the t+1 and t iterations, represents the leading score of the black kite in the j-th dimension at the t iteration so far; represents the fitness value of the random position of any one black kite in the t iteration, is a random number conforming to the Cauchy distribution, represents a random number between (0, 1), is the position crowding degree of the artificial fish in the artificial fish school algorithm, is the moving step length of the artificial fish; When the global optimal solution is not found, the improved migration behavior of black kite is adopted, otherwise, the improved migration behavior of black kite combined with the aggregation behavior of artificial fish school is adopted, and thus the improved global optimization strategy is obtained. improving the clustering algorithm by using the combined intelligent algorithm, iteratively optimizing clustering centers, until final clustering centers are obtained, and then obtaining point cloud cluster division results; performing vehicle navigation according to the point cloud cluster division results; the process of improving the clustering algorithm by using the combined intelligent algorithm, iteratively optimizing the clustering centers comprises: K cluster centers were calculated using a combinatorial intelligent algorithm. Its objective function value is The local stagnation threshold is defined as ; The local optimization strategy is adopted to optimize the current center and record the change of the objective function value of continuous iterations. if continuous N1 iterations satisfy: ； If the condition is satisfied, the local stagnation is detected, the number of local stagnation is recorded as NS, and the global optimization strategy is used to generate a new center . The objective function values are compared: if , let , and enter the local optimization phase; otherwise, keep the original center and continue the local optimization, , are the objective function values of the tth and (t-1)th iterations, respectively. In the local optimization process, if the continuous N2 times appear , it is determined that local degeneration occurs, and a global optimization strategy is adopted to generate new centers , and the optimal center is selected. ； Compare the objective function values, if then let and enter the local optimization phase; otherwise keep the original center Continue local optimization; when the NS reaches a set value, the optimal clustering center is considered to be found.

2. The method of claim 1, wherein, The clustering algorithm is a K-means clustering algorithm.

3. The method of claim 1, wherein the step of determining the optimal route is performed by a navigation system of the vehicle. The process of performing vehicle navigation according to the point cloud cluster division results comprises identifying passable areas and obstacle contours in a current environment according to the point cloud cluster division results, planning a path in the passable areas to avoid the obstacle contours, and realizing navigation.

4. A vehicle intelligent navigation system for implementing the vehicle intelligent navigation method according to any one of claims 1-3, characterized in that, comprise: a data acquisition module configured to acquire three-dimensional point cloud data to be processed in a vehicle navigation process; a data index module configured to adopt a binary spatial partition tree index structure to perform spatial division on the three-dimensional point cloud data to be processed, and solve normal vectors of each point; a data preprocessing module configured to calculate angles between the normal vectors by using the solved normal vectors, and construct similarity with geometric features in a clustering algorithm and a clustering objective function based on the angles; an algorithm improvement module configured to adopt a first defense mechanism of the Crowned Porcupine Algorithm to improve a hunting stage of the Black Kite Algorithm to obtain a local optimization strategy, combine a gathering behavior and a following behavior of the Artificial Fish Swarm Algorithm with a migration mechanism of the Black Kite Algorithm to obtain a global optimization strategy, and combine the local optimization strategy and the global optimization strategy to obtain a combined intelligent algorithm; a clustering module configured to improve the clustering algorithm by using the combined intelligent algorithm, iteratively optimize clustering centers, until final clustering centers are obtained, and then obtain point cloud cluster division results; a navigation execution module configured to perform vehicle navigation according to the point cloud cluster division results.

Images