A method, device and storage medium for path planning of a robot dog

Abstract

This application relates to a path planning method, device, and storage medium for robotic dogs. The method includes: proposing a multi-strategy improved dung beetle optimization algorithm to address the high-dimensional feature selection problem in complex environments; generating an initial population through a refraction-based back-learning mechanism; combining an elite retention strategy to select the optimal feature subset and obtain the globally optimal position; constructing a collaborative optimization model integrating an osprey optimization position update module, an adaptive t-distribution perturbation module, and a Jacobi curve smoothing module; and iteratively updating the globally optimal solution using dynamic termination conditions, ultimately outputting a low-dimensional, high-precision feature subset. This application solves the problems of low decision-making efficiency and path planning failure in robotic dogs caused by local optima and slow convergence in traditional feature selection algorithms in complex environments. It effectively reduces data dimensionality, shortens computation time, and enables rapid decision-making and high-success-rate path planning for robotic dogs in dynamic and complex environments, while also exhibiting strong robustness under electromagnetic interference.

Description

Technical Field

[0001] This application relates to the field of machine learning technology, and in particular to a method, apparatus and storage medium for path planning of a robot dog. Background Technology

[0002] With the deepening application of artificial intelligence, intelligent robot dogs designed for complex urban scenarios are gradually undertaking key tasks such as life detection in disaster environments and intelligent inspection of urban infrastructure. These complex urban environments are characterized by three main features: nonlinear spatial structure (narrow passages, unknown obstacles), multi-source heterogeneous perception information (data from multiple modal sensors such as vision, lidar, and infrared), and urgent real-time decision-making requirements. In this scenario, robot dogs need to process high-dimensional environmental data in real time to support target recognition, dynamic path planning, and obstacle avoidance. Feature selection technology, as a key step in data preprocessing, directly affects decision-making efficiency and success rate.

[0003] Existing feature selection methods generally suffer from low computational efficiency, susceptibility to local optima, and poor stability when processing high-dimensional data (such as 12 types of sensor features and 1000+ dimensions), leading to decision delays or path planning failures in robot dogs. To improve feature selection efficiency, swarm intelligence optimization algorithms (such as particle swarm optimization and whale optimization) have been introduced due to their global search capabilities. Among them, the Dung Beetle Optimization (DBO) algorithm achieves efficient search by simulating the rolling ball and reproduction behaviors of dung beetles, demonstrating outstanding performance in engineering optimization. However, in complex alleyway scenarios with dynamic targets, DBO still faces the following key problems: the solution space is complex in complex environments, and the rolling ball mechanism of DBO is easily attracted by local optima, resulting in unstable feature subset quality (e.g., a path planning success rate of only 78%); the linear update strategy of traditional DBO cannot adapt to dynamic threats, and feature selection is time-consuming, failing to meet the requirements of real-time tasks; under noise and electromagnetic interference, the feature selection results oscillate significantly, and the classification accuracy fluctuates greatly, affecting the reliability of robot dog decisions.

[0004] Currently, no effective solution has been proposed for the problems of low decision-making efficiency and path planning failure in robot dogs caused by local optima and slow convergence of traditional feature selection algorithms in complex environments in related technologies. Summary of the Invention

[0005] This application provides a robot dog path planning method, apparatus, and storage medium to at least solve the problems of low decision-making efficiency and path planning failure in robot dogs caused by local optima and slow convergence in traditional feature selection algorithms under complex environments in related technologies.

[0006] In a first aspect, embodiments of this application provide a robot dog path planning method, including:

[0007] Based on the acquired feature information, an initial feature population is generated through a refraction-based reverse learning mechanism; then, through an elite retention strategy and a fitness function, the feature subset with the best fitness value is selected from the initial feature population to generate the current elite population, and the current global optimal position corresponding to the current elite population is obtained.

[0008] The optimization algorithm model is used to input the current global optimal position and the first population position corresponding to the initial feature population into the position update module to update the first population position to the second population position; the second population position is input into the perturbation module to generate a perturbed position; the perturbed position is input into the path smoothing module to generate a smooth position; wherein, the optimization algorithm model includes the position update module, the perturbation module and the smoothing module;

[0009] Determine whether the smoothed position meets a preset termination condition. If the termination condition is not met, update the smoothed position to the first population position and input the updated first population position into the optimization algorithm model for iterative calculation. If the termination condition is met, output the optimal path based on the smoothed position.

[0010] In some embodiments, the feature information includes a feature space dimension and an initial search space; the generation of an initial feature population based on the acquired feature information through a refraction-based reverse learning mechanism includes:

[0011] The refraction inverse learning mechanism expands the initial search space into a refraction search space by adjusting the refractive index parameter, and generates the initial feature population based on the refraction search space and the feature space dimension.

[0012] In some embodiments, the step of inputting the current global optimal position and the first population position corresponding to the initial feature population into the position update module, and updating the first population position to the second population position, includes:

[0013] Based on the obtained current weight factor and the current global optimal position, the first population position corresponding to the initial feature population is updated to the second population position through the global search mechanism of the Osprey optimization algorithm.

[0014] In some embodiments, the step of inputting the second population position into the perturbation module to generate the perturbation position includes:

[0015] Based on the obtained current iteration number and maximum iteration number, generate the degree of freedom parameters; based on the degree of freedom parameters, determine the current distribution function; based on the current distribution function, update the position of the second population to the perturbation position.

[0016] In some embodiments, the step of inputting the disturbance location into the path smoothing module to generate a smoothed location includes:

[0017] Based on the obtained random number and the current global optimal position, the perturbation position is updated to a smooth position through a periodic function.

[0018] In some embodiments, determining whether the smoothed position meets a preset termination condition includes:

[0019] The fitness function is called on all smooth feature subsets corresponding to the smooth position to calculate the fitness value of each smooth feature subset, and the feature subset with the best fitness value is selected from each smooth feature subset to generate the optimal smooth feature subset;

[0020] If the fitness value of the smoothed optimal feature subset is lower than the fitness value of the current global optimal position, then the feature position corresponding to the smoothed optimal feature subset is updated to the current global optimal position, and it is determined that the smoothed position does not meet the termination condition.

[0021] If the fitness value of the smoothed optimal feature subset is not lower than the fitness value of the current global optimal position, then the current iteration number is obtained. If the current iteration number is greater than a preset fitness value constant number threshold, then the smoothed position is determined to meet the termination condition.

[0022] In some embodiments, the step of calling the fitness function on all smooth feature subsets corresponding to the smooth position and calculating the fitness value of each smooth feature subset includes:

[0023] Input all smooth feature subsets corresponding to the smooth position into the classifier and calculate the classification error; based on the classification error and the number of all smooth feature subsets, calculate the fitness value of each smooth feature subset using the fitness function.

[0024] In some embodiments, the termination condition further includes the current iteration number reaching a preset maximum iteration number.

[0025] Secondly, embodiments of this application provide a robot dog path planning device, comprising:

[0026] The information acquisition module is used to generate an initial feature population based on the acquired feature information through a refraction-backward learning mechanism; then, through an elite retention strategy and a fitness function, it selects the feature subset with the best fitness value from the initial feature population to generate the current elite population, and obtains the current global optimal position corresponding to the current elite population.

[0027] An optimization and update module is used to, via an optimization algorithm model, input the current global optimal position and the first population position corresponding to the initial feature population into a position update module, update the first population position to a second population position; input the second population position into a perturbation module to generate a perturbed position; and input the perturbed position into a path smoothing module to generate a smoothed position; wherein, the optimization algorithm model includes the position update module, the perturbation module, and the smoothing module;

[0028] The path generation module is used to determine whether the smoothed position meets a preset termination condition. If the termination condition is not met, the smoothed position is updated to the first population position, and the updated first population position is input into the optimization algorithm model for iterative calculation. If the termination condition is met, the optimal path is output based on the smoothed position.

[0029] Thirdly, embodiments of this application provide a storage medium storing a computer program that, when executed by a processor, implements the robot dog path planning method as described in the first aspect above.

[0030] Compared to related technologies, the robot dog path planning method provided in this application generates an initial population based on refraction-based back learning, selects the optimal feature subset and obtains the globally optimal position by combining an elite retention strategy, iteratively optimizes the population position by integrating an Osprey-optimized position update module, an adaptive t-distribution perturbation module and a Jacobi curve smoothing module, dynamically judges the convergence state and updates the optimal solution, and finally outputs a low-dimensional, high-precision feature subset. This solves the problems of low decision-making efficiency and path planning failure in robot dogs caused by local optima and slow convergence in traditional feature selection algorithms in complex environments. It effectively reduces data dimensionality, shortens computation time, and enables robot dogs to make fast decisions and achieve high-success-rate path planning in dynamic and complex environments, while also possessing strong robustness under electromagnetic interference.

[0031] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0032] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0033] Figure 1 This is a hardware structure block diagram of the terminal of the robot dog path planning method according to an embodiment of the present invention;

[0034] Figure 2 This is a flowchart of a robot dog path planning method according to an embodiment of this application;

[0035] Figure 3 This is a preferred flowchart of the robot dog path planning method according to an embodiment of this application;

[0036] Figure 4 This is a schematic diagram of the UCI dataset used in the feature selection experiment of the robot dog path planning method according to the embodiments of this application;

[0037] Figure 5 This is a bar chart showing the classification accuracy before and after feature selection according to the embodiments of this application;

[0038] Figure 6 This is a schematic diagram showing the average fitness values ​​of various algorithms according to embodiments of this application under UCI;

[0039] Figure 7 This is a structural block diagram of a robot dog path planning device according to an embodiment of this application. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application. Furthermore, it is understood that although the efforts made in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, modifications to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0041] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0042] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application means two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The terms “first,” “second,” “third,” etc., used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0043] The method embodiments provided in this example can be executed on a terminal, computer, or similar computing device. Taking running on a terminal as an example, Figure 1 This is a hardware structure block diagram of the terminal of the robot dog path planning method according to an embodiment of the present invention. For example... Figure 1 As shown, a terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. Optionally, the terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the terminal described above. For example, the terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0044] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the robot dog path planning method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0045] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0046] This embodiment provides a path planning method for a robot dog. Figure 2 This is a flowchart of a robot dog path planning method according to an embodiment of this application, such as... Figure 2 As shown, the process includes the following steps:

[0047] Step S201: Based on the acquired feature information, an initial feature population is generated through a refraction-based reverse learning mechanism; then, through an elite retention strategy and a fitness function, the feature subset with the best fitness value is selected from the initial feature population to generate the current elite population, and the current global optimal position corresponding to the current elite population is obtained.

[0048] The initial feature subset population is generated by simulating the principle of light refraction. The search space can be expanded by dynamically adjusting the refractive index parameter to avoid local optima, thus enhancing the diversity of the initial population and improving global exploration capabilities. The elite retention strategy calculates the fitness value of each individual (feature subset) based on the fitness function (considering both classification error and feature quantity), selecting the N individuals with the best fitness values ​​to form an elite population. The initial global optimum is set as the best individual in the elite population. This step improves the quality of the initial population: refraction-based back-learning allows the initial population to cover more high-quality solution regions, reducing invalid searches; the elite population provides a high-quality starting point, shortening the algorithm's convergence time (experiments show a 37.5% improvement in convergence speed); and a diverse initial population reduces the probability of getting trapped in local optima (e.g., achieving 100% classification accuracy on the Exactly dataset).

[0049] Step S202: Through the optimization algorithm model, the current global optimal position and the first population position corresponding to the initial feature population are input into the position update module to update the first population position to the second population position; the second population position is input into the perturbation module to generate a perturbed position; the perturbed position is input into the path smoothing module to generate a smooth position; wherein, the optimization algorithm model includes a position update module, a perturbation module and a smoothing module;

[0050] The optimization algorithm model employs multi-module collaborative processing. The position update module (integrating Osprey optimization and adaptive weights) takes the current global optimum and the first swarm position as input and outputs the updated second swarm position. The perturbation module (using an adaptive t-distribution or other optimization algorithms, such as the perturbation strategy of particle swarm optimization) takes the updated second swarm position as input and outputs the perturbation position. Initially, a Cauchy distribution with large step sizes is used for perturbation to enhance global exploration; later, it gradually converges to a Gaussian distribution to improve local accuracy. The path smoothing module (using Jacobi curves) takes the perturbation position as input and outputs the smoothed position, utilizing the periodicity of the Jacobi curve to suppress oscillations and ensure path continuity. This step balances global search and local development: Osprey optimization (OOA) enhances the ability to explore high-dimensional space; adaptive weight dynamic adjustment strategy (ω decreasing) avoids premature convergence (e.g., the number of features in the PenglungEW dataset is reduced to 3.4667); adaptive t-distribution perturbation effectively escapes local optima (e.g., classification accuracy is improved by 4.14% in the Hillvalley dataset); Jacobi curve smooths the path and reduces result oscillations (e.g., the standard deviation of fitness values ​​in the HeartEW dataset is 0).

[0051] Step S203: Determine whether the smoothed position meets the preset termination condition. If the termination condition is not met, update the smoothed position to the first population position and input the updated first population position into the optimization algorithm model for iterative calculation. If the termination condition is met, output the optimal path based on the smoothed position.

[0052] The termination condition can include reaching the maximum number of iterations or the fitness value changing less than a threshold for multiple consecutive rounds. If the fitness value of the current smoothed position is better than the historical fitness value of the current global best position, then the current global best position is updated; the smoothed position is used as the first population position for the next round of input, and iteration continues. If the termination condition is met, the optimal feature subset, i.e., the optimal path, is output. This step takes an average of 17.8 seconds for feature selection, meeting the real-time requirements of complex heterogeneous environments such as alleyways (e.g., 96% success rate for path planning on a 20×20 map); in a complex 40×40 map, only 4.1 features are needed to maintain an accuracy of 90.5%; the classification accuracy remains ≥87% under electromagnetic interference, ensuring the reliability of path planning.

[0053] Through the above steps, traditional algorithms, which use random initialization, are prone to insufficient population diversity (e.g., the DBO algorithm has an initial fitness value standard deviation of 0.05 on the Exactly dataset). In contrast, the refraction-based inverse learning mechanism in this application expands the search range to 1.5 times the original space by dynamically adjusting the refractive index parameter. Combined with an elite retention strategy, the N individuals with the best fitness values ​​are selected from the initial population to form an elite population, improving the classification accuracy of the initial global optimum by 3.2% compared to traditional methods. Experiments show that this strategy reduces the initial population fitness value of the Wisconsin Breast Cancer Extended Dataset (BreastEW) by 28.6%, effectively avoiding ineffective searches. The position update module introduces the global search capability of the Osprey Optimization (OOA) algorithm, exploring and developing through adaptive weight ω dynamic balancing. In tests on the PenglungEW high-dimensional dataset, this strategy improves the algorithm's fitness value decrease speed by 2.1 times in the first 50 iterations, significantly outperforming the fixed weight mechanism of DBO. The perturbation module employs an adaptive t-distribution strategy. Initially, a Cauchy distribution (degrees of freedom ν=1) with large-step perturbations enhances global exploration, while later convergence to a Gaussian distribution (ν→∞) improves local accuracy. The path smoothing module introduces the periodicity of the Jacobian curve to suppress feature selection oscillations, reducing the convergence curve fluctuation amplitude by 63%. Dual termination conditions are set: the maximum number of iterations (ensuring convergence across most datasets) and a fitness value change less than a threshold for five consecutive iterations. By dynamically updating the global optimal solution (comparing the current best individual with the historical best individual in each round), feature selection time in a 20×20 lane simulation map is reduced from 21.3 seconds (DBO) to 17.8 seconds, meeting the real-time requirements of the robot dog's path planning. In a 40×40 complex map, this mechanism increases the path planning success rate from 78% to 96%, and reduces the average path length by 15.3%. Through a multi-objective fitness function and an elite retention strategy, classification accuracy remains above 87% even under electromagnetic interference. Comparative experiments show that after adding 20% ​​Gaussian noise to the IonosphereEW dataset, the classification accuracy of the Improved Dung Beetle Optimizer for Feature Selection (IDBO-FS) based on the improved dung beetle optimization algorithm only decreased by 2.1% (DBO decreased by 7.8%), and the increase in the number of features was controlled within 5%. Simulation verification in complex environments shows that the robot dog's path planning success rate improved to 96%, and the stability under electromagnetic interference reached 87%, comprehensively solving the decision delay and path failure problems caused by local optima. It is evident that this method can also be applied to other autonomous robots or non-complex environments.

[0054] In some embodiments, the feature information includes the feature space dimension and the initial search space; based on the acquired feature information, an initial feature population is generated through a refraction-based reverse learning mechanism, including:

[0055] The refraction inverse learning mechanism expands the initial search space into a refraction search space by adjusting the refractive index parameter, and generates an initial feature population based on the dimensions of the refraction search space and the feature space.

[0056] The refraction-backward learning mechanism is a technique that expands the search range by adjusting the refractive index parameter. In feature selection algorithms, it is used to generate an initial feature population with higher diversity and quality. Specifically, let the feature space dimension be D (the number of features in the dataset), and the initial search space be [lb, ub] (the search range of the feature subset in the algorithm). Each feature corresponds to a dimension, and lb and ub represent the lower and upper bounds of that dimension (usually binary values ​​0 or 1). The formula for calculating the refraction-backward solution X' is as follows: ;

[0057] In the above formula, X is the original candidate solution (a randomly generated binary vector); X' is the candidate solution generated through refraction back learning; n is the refractive index parameter (dynamically adjusted, usually taking values ​​n∈(0,1]), controlling the expansion of the search space; lb and ub are the lower and upper bounds of the initial search space (e.g., binary encoding lb=0, ub=1). The operation process is as follows: first, the original candidate solution X is randomly generated, then the refraction back solution X' is calculated, and subsequently X and X' are merged to form the initial feature population. When n→0, X' tends to a random value, enhancing the global exploration capability; when n→1, X' tends to the symmetric back solution of the original solution, enhancing the local development capability.

[0058] Refraction-based inverse learning simulates the refraction of light, expanding the search space by adjusting the refractive index n, thereby generating a broader initial feature subset population. This embodiment expands the search space through refraction-based inverse learning, allowing the algorithm to explore a wider region in the early iterations, avoiding getting trapped in local optima. In complex environments such as dynamically changing alleyways, this mechanism can quickly generate diverse feature subsets, supporting real-time decision-making by the robot dog.

[0059] In some embodiments, the current global optimal position and the first population position corresponding to the initial feature population are input into the position update module, and the first population position is updated to the second population position, including:

[0060] Based on the obtained current weight factor and the current global optimal position, the position of the first population corresponding to the initial feature population is updated to the position of the second population through the global search mechanism of the Osprey optimization algorithm.

[0061] During the algorithm's iteration process, the current global optimal position (Best_pos) is recorded, representing the position of the optimal feature subset found in the current iteration. Simultaneously, the current weight factor (ω) is calculated to balance the algorithm's exploration and exploitation capabilities. The current population position (first population position) and the global optimal position are input into the Osprey Optimization Algorithm (OOA) to generate candidate optimal positions (OOA_pos). The Osprey Optimization Algorithm (OOA) is an optimization algorithm that simulates the hunting behavior of ospreys and possesses strong global search capabilities. The global search mechanism of the Osprey Optimization Algorithm (OOA) simulates the three stages of osprey hunting: Exploration, randomly searching a wide area at high altitude (enhancing global exploration capabilities); Exploitation, rapidly diving to lock onto the target area (strengthening local exploitation accuracy); and Perturbation, randomly adjusting the position when close to the water surface (avoiding local optima). Through the three stages of the Osprey Optimization Algorithm, combined with directional search and random perturbation, candidate optimal positions (OOA_pos) are generated.

[0062] Next, based on the position update formula, and combining the current weight factor, the global optimal position, and candidate optimization positions, the new position (i.e., the second population position) is calculated. The formula considers factors such as the current global optimal solution (Best_pos), the position generated by the Osprey algorithm (OOA_pos), and the coupling coefficient. The position update formula is as follows:

[0063] ;

[0064] In the above formula, This is the updated population position, i.e., the second population position; The first population position corresponds to the initial feature population; Best_pos is the global optimum position (the feature subset with the highest fitness); OOA_pos is the candidate optimization position, which is a new candidate solution generated by combining directional search and random perturbation in the three stages of the Osprey optimization algorithm; α and β are coupling coefficients (usually α+β=1), which control the weight ratio between the global optimum position (Best_pos) and the candidate optimization position (OOA_pos); the current weight factor ω is an adaptive weight factor that decreases linearly with the number of iterations, for example, ω=ω max -(ω max -ω min )×t / T, where t is the current iteration number and T is the total number of iterations.

[0065] The position update formula merges the global optimal position and candidate optimization positions, updating the first population position corresponding to the initial feature population to the second population position. By adaptively adjusting the current weight factor ω, the algorithm can find a balance between exploration (global search) and development (local search), ensuring that the algorithm can find the global optimal solution while also performing a refined search after finding a near-optimal solution, thus improving the quality of the solution. In early iterations, ω is relatively large (close to ω0). max This enhances the global exploration capability of OOA by broadly searching the solution space; however, in later iterations, ω is smaller (close to ω). min This method relies on the global optimal location for local fine-grained search, thereby accelerating convergence.

[0066] This embodiment utilizes the hovering mechanism of the Osprey Optimization Algorithm (OOA) to extensively explore the feature space in the early stages of iteration, avoiding the local optima trapping traditional Direct Optimization (DBO) due to its linear rolling ball behavior. In dynamic obstacle scenarios in alleyways, OOA helps the algorithm quickly escape invalid regions, increasing the path planning success rate to 96%, compared to only 82% for the original DBO. By adaptively adjusting the current weight factor ω, it dynamically balances exploration and development, exploring new regions early and focusing on the neighborhood of the optimal solution later. The Osprey Optimization Algorithm has strong global search capabilities, guiding the algorithm to explore more extensively in the feature space, finding better feature subsets. Furthermore, the Osprey Optimization Algorithm can accelerate the convergence process of the overall algorithm, enabling it to find better feature subsets in fewer iterations, thereby improving algorithm efficiency.

[0067] In some embodiments, the second population location is input into the perturbation module to generate a perturbation location, including:

[0068] Based on the current iteration number and the maximum iteration number, generate the degree of freedom parameters; based on the degree of freedom parameters, determine the current distribution function; based on the current distribution function, update the position of the second population to the perturbation position.

[0069] Among them, based on the current iteration number (t) and the maximum iteration number (T), a degree of freedom parameter (df) is generated. A common generation method is to linearly decrease the degree of freedom parameter. For example, df = T - t. In this way, the degree of freedom is larger in the initial stage of the algorithm, which helps with global search; in the later stage of the algorithm, the degree of freedom decreases, which helps with local development. According to the generated degree of freedom parameter, the currently used distribution function is determined. The t-distribution is adopted, and its degree of freedom is determined by the degree of freedom parameter generated in the previous step. The t-distribution is a commonly used distribution in statistics. Especially in the case of small samples, it has thicker tails than the normal distribution, which helps the algorithm to jump out of local optima during the search process. Based on the current t-distribution function, the positions of the second population are perturbed to generate perturbed positions. The perturbation method is usually to add a random number generated by the t-distribution to the current position. The size and direction of this random number are determined by the properties of the t-distribution, which helps the algorithm to conduct a more extensive exploration in the search space. The specific formula is as follows:

[0070] ;

[0071] ;

[0072] In the above formula, and are dynamically selected probability parameters; r is a constant between [0, 1], which controls the selection of the "exploration" or "development" strategy in each iteration; p is a dynamically adjusted probability parameter (decreasing with the iteration number). In the initial stage, "exploration" is made more frequent, and in the later stage, "development" dominates, balancing the global exploration and local development capabilities of the algorithm. When r < p, a global-optimal-guided exploration strategy is applied, is the position after adaptive t-distribution mutation perturbation; the j-th component of the current global optimal position (representing whether the j-th feature of the current optimal feature subset is selected); represents the number of iterations of the adaptive t-distribution (i.e., the degree of freedom parameter). At this time, a "co-directional expansion" search is conducted based on the global optimal solution, and the step size is dynamically amplified using the iteration number t_iter to enhance the exploration ability of the global optimal neighborhood (to avoid the algorithm falling into local optima in the early stage). When r ≥ p, a development strategy is applied, is the reference benchmark point, which is obtained by fitting the historical optimal positions and reflects the central tendency of the "smooth path"; the current position of the i-th individual at the t-th iteration, Lb and Ub are the lower and upper bounds of the search space; b1 and b2 are adaptive weight coefficients, which are dynamically adjusted with the iteration to balance the "local development intensity" and the "global constraint".

[0073] In this embodiment, the algorithm has a large degree of freedom and a thick tail of the t-distribution in the early stages, which helps the algorithm to explore the search space more extensively and find potential global optima. As the number of iterations increases, the degree of freedom gradually decreases, the tail of the t-distribution becomes thinner, and the algorithm gradually shifts to local development, improving the quality of the solution. The perturbation of the t-distribution helps the algorithm escape local optima during the search process, increasing the probability of finding the global optimum. By dynamically switching between "global exploration" (first branch) and "local development" (second branch) through the segmented strategy r and p, the algorithm avoids getting trapped in local optima in the early stages and can accurately optimize feature subsets in the later stages.

[0074] In some embodiments, the perturbation location is input into the path smoothing module to generate a smoothed location, including:

[0075] Based on the obtained random number and the current global optimal position, the perturbed position is updated to a smooth position through a periodic function.

[0076] A random number r2 within the interval (0, 1) is generated to introduce randomness and prevent the algorithm from getting trapped in local optima. Combining the random number and the global optimal position, a periodic function is used to generate a smooth position. The periodic function can be a Jacobi curve, a periodic mathematical function (such as a sine or cosine curve), whose periodicity is used to smooth the position update path and reduce random fluctuations during the search process.

[0077] The position update formula is:

[0078] r2 <P,(1);

[0079] , r2≥P, (2);

[0080] In equation (1) above, r2 is a random number in the range (0, 1); x i (t) represents the current position of the i-th feature subset at time t; x i (t+1) The new position of the i-th feature subset at time t+1 in the next iteration; x best The current iteration represents the globally optimal position; θ is the angle parameter, θ∈(0, 2π); e θ is the exponential term, the step scaling factor; sinθ and cosθ are trigonometric function terms used for direction control. In equation (2) above, x i (t) represents the current position of the i-th feature subset at time t; x i (t+1) represents the new position of the i-th feature subset at time t+1 in the next iteration; C1 and C2 are adaptive weight coefficients (dynamically adjusted with iteration); Lb b Ub b These are the lower and upper bounds of the Jacobian curve.

[0081] Different from the original algorithm, the individual update strategy in this embodiment is determined according to a random number. p is a dynamically adjusted probability parameter (decreasing with the number of iterations). If r2 < P, update is performed using Equation (1); otherwise, position update is performed using Equation (2). In this embodiment, the upper and lower bounds of the Jacobi curve are used to enforce "path smoothing" of the feature subset update, avoiding meaningless fluctuations in the number or combination of features. The adaptive weight guides the population to focus on the global optimal neighborhood, finely optimizing the feature subset. During the foraging stage, through "biological inspiration (Jacobi curve) + mathematical constraint (upper and lower bounds) + dynamic regulation (weight + probability)", the smoothness of the Jacobi curve is injected, enabling the algorithm to achieve the goals of "stable feature selection, accurate classification accuracy, and fast dynamic scenario" in the high-dimensional feature space of a complex environment.

[0082] In some of these embodiments, determining whether the smoothed position meets a preset termination condition includes:

[0083] Calling a fitness function for all smoothed feature subsets corresponding to the smoothed position, calculating the fitness values of each smoothed feature subset, and screening out the feature subset with the optimal fitness value from each smoothed feature subset to generate a smoothed optimal feature subset;

[0084] If the fitness value of the smoothed optimal feature subset is lower than the fitness value of the current global optimal position, update the feature position corresponding to the smoothed optimal feature subset to the current global optimal position, and determine that the smoothed position does not meet the termination condition;

[0085] If the fitness value of the smoothed optimal feature subset is not lower than the fitness value of the current global optimal position, obtain the current number of iterations. If the current number of iterations is greater than the preset constant number of fitness value threshold, determine that the smoothed position meets the termination condition.

[0086] In this process, the fitness function is called on the smoothed position (i.e., the set of feature subsets updated by the Jacobi curve during the foraging phase) to calculate the fitness value of each subset. The subset with the best fitness is selected and denoted as the smoothed optimal feature subset. The fitness value of the smoothed optimal feature subset is compared with the fitness value of the current global optimal position. If the fitness value of the smoothed optimal feature subset is better, that is, the fitness value is lower (because the fitness function takes into account "classification accuracy + number of features", the lower the value, the better the balance of "fewer features + high classification"), then the global optimal position is updated to the feature position of this subset, and it is determined that "the termination condition is not met". The algorithm continues to iterate to explore a better solution. When the fitness value of the smoothed optimal feature subset is not lower than the fitness value of the current global optimum, it is checked whether the current iteration count exceeds a preset fitness value constant number threshold (e.g., the fitness value constant number threshold is 50 iterations; if there is no improvement after exceeding the constant threshold, the iteration ends). The fitness value constant number threshold is the maximum number of iterations allowed for no improvement in the fitness value of the global optimum. If it exceeds the threshold, the termination condition is determined to be "met"; if it does not exceed the threshold, the termination condition is determined to be "not met," and the iteration continues. This embodiment uses a "global optimum dynamic update" mechanism to ensure that the algorithm continues to iterate when a better feature subset is found, avoiding premature termination due to early local optima; the "iteration count circuit breaker" mechanism prevents the algorithm from oscillating meaninglessly near the optimal solution, ensuring the real-time performance of decision-making in complex environments.

[0087] In some embodiments, a fitness function is called on all smooth feature subsets corresponding to the smoothed position to calculate the fitness value of each smooth feature subset, including:

[0088] Input all smooth feature subsets corresponding to the smooth position into the classifier and calculate the classification error; based on the classification error and the number of all smooth feature subsets, calculate the fitness value of each smooth feature subset through the fitness function.

[0089] The smoothed feature subset corresponding to each smoothed position is input into a K-nearest neighbor classifier (K=5, Euclidean distance metric). The classification error is calculated on the test set to predict the proportion of incorrect samples. The classification error reflects the ability of the feature subset to support practical tasks such as target recognition, threat source detection, and obstacle avoidance. The lower the error, the higher the practical application value of the feature subset.

[0090] The quality of feature subsets is quantified by using a fitness function that balances classification accuracy and feature parsimony.

[0091] ;

[0092] In the above formula, α is the classification accuracy weight (default 0.7, balancing "accuracy priority" and "dimensionality reduction priority"). is the proportion of selected features to the total features. The fewer the number of features, the lower the computational complexity and the faster the decision-making of the robot dog. Error is the classification error; Fitness is the output result, which is the fitness value of each smoothed feature subset.

[0093] In this embodiment, by comparing the fitness values of different smoothed feature subsets, a higher-quality feature subset can be selected, thereby improving the performance of the classifier; the fitness function can quantify the quality of each smoothed feature subset, providing a basis for subsequent screening and optimization. By continuously iterating and optimizing the fitness value, the overall performance of the algorithm can be improved, and a feature subset closer to the global optimal solution can be found.

[0094] In some embodiments, the termination condition further includes that the current iteration number reaches the preset maximum iteration number.

[0095] Among them, the preset maximum iteration number T represents the maximum number of iteration rounds allowed by the algorithm. The current iteration number t is monitored in real time. After each iteration, it is checked whether t is equal to T. If t = T, it is determined that the "termination condition is satisfied", and the algorithm stops iterating and outputs the current global optimal feature subset; if t < T, the "fitness stagnation fusing" logic is continued (that is, it is checked whether there is no fitness improvement for several consecutive iterations).

[0096] In a complex environment, "electromagnetic interference and obstacle mutation" may cause the algorithm to temporarily fall into a local optimum. In this embodiment, the "hard constraint on the iteration number" ensures that even if the fitness stagnates, the algorithm will terminate after T times and output the current optimal solution instead of no result. On the Sonarall dataset with strong noise interference, IDBO reduces 23.6% of invalid iterations due to this constraint, and the standard deviation of the classification accuracy is significantly reduced, ensuring the reliability of decision-making.

[0097] The embodiments of the present application will be described and illustrated below through preferred embodiments.

[0098] Figure 3 is a preferred flowchart of the robot dog path planning method according to the embodiment of the present application. As Figure 3 shown, the robot dog path planning method includes the following steps:

[0099] The first step: Improve the multi-strategy design of the dung beetle optimization algorithm.

[0100] S31-1, initialization stage: refraction reverse learning and elite selection strategy.

[0101] First, initialize the dung beetle population size N, the iteration number Tmax and related parameters.

[0102] The population after refraction back learning is calculated and then subjected to elite selection. To improve the diversity and quality of the initial population, a refraction back learning mechanism is used to generate an initial feature subset population. Let the feature space dimension be D, and the upper and lower bounds of the search space be [lb, ub]. The formula for calculating the refraction back solution X' is:

[0103] (31)

[0104] Where n is the refractive index parameter (the value of n ranges from (0, 1]), and the search range is expanded by dynamically adjusting n. The formula for calculating n is as follows: ;

[0105] After generating the initial population, an elite selection strategy is used to select the N individuals with the best fitness values ​​to form an elite population, thereby improving the global search capability.

[0106] S31-2, Rolling Ball Stage: Integrating Osprey Optimization (OOA) Algorithm with Dynamic Weights.

[0107] The process involves calculating the individual fitness and the global optimal position, and determining whether t ≤ Tmax. If yes, the position of the rolling dung beetle is updated; otherwise, the global optimal position and corresponding fitness value are output, and the process ends.

[0108] Next, determine if R < 0.9. If yes, enter the barrier-free mode and update the position of the rolling dung beetle; otherwise, enter the obstacle-prone mode and update the position of the rolling dung beetle.

[0109] More specifically, a global search mechanism from the osprey optimization algorithm is introduced into the dung beetle rolling ball behavior, with an adaptive weight factor ω added to balance exploration and exploitation capabilities. The position update formula is:

[0110] ,

[0111] (32)

[0112] Where ω decreases linearly with the number of iterations, Best_pos is the current global optimal solution, OOA_pos is the position generated by the Osprey algorithm, and α and β are coupling coefficients.

[0113] S31-3, Reproduction Stage: Adaptive t-distribution perturbation and dynamic selection probability.

[0114] This includes updating the location of breeding dung beetles.

[0115] To avoid the algorithm getting trapped in local optima, an adaptive t-distribution perturbation strategy is introduced during the breeding phase. The individual update formula is:

[0116] (33)

[0117] (34)

[0118] In the formula, and are dynamic selection probability parameters, represents the adaptive t-distribution parameter, r is a constant between [0, 1], is the position after being perturbed by adaptive t-distribution mutation; P is the dynamic probability, ω1 = 0.7; ω2 = 0.1.

[0119] S31-4, Foraging stage: Local exploitation guided by the Jacobi curve.

[0120] Among them, it is judged whether rand < 0.5; if so, the position of the foraging dung beetle is updated according to the corresponding formula of the Jacobi curve; if not, the position of the foraging dung beetle is updated according to the original formula.

[0121] In the foraging stage, the Jacobi curve is introduced to guide the population to search along a smooth path and reduce ineffective jumps. The position update formula is:[[]]

[0122] , r2 < P (35)

[0123] , r2 ≥ P (36)

[0124] Among them, r2 is a random number in (0, 1), θ is the angle parameter, θ ∈ (0, 2π). Different from the original algorithm, the individual update strategy is determined according to the random number. If r2 < P (usually set as r2 < 0.5, that is, set P = 0.5), update using formula (35), otherwise update the position using formula (36).

[0125] After that, update the position of the stealing dung beetle, and update the global optimal position and fitness value according to the greedy selection measurement. Let t = t + 1, and return to the above step of judging whether t ≤ Tmax.

[0126] Second step: Feature selection modeling and fitness function design.

[0127] S32-1, Feature subset encoding.

[0128] Encode the individuals in the dung beetle population as binary vectors , where indicates whether the j-th feature is selected (1 for selected, 0 for not selected).

[0129] S32-2, Multi-objective fitness function.

[0130] (37)

[0131] The weights are used as coefficients, and the classification error is calculated on the test set using KNN (K=5).

[0132] Step 3: Experimental design for selecting tunnel environmental characteristics.

[0133] S33-1, Dataset and Simulation Environment.

[0134] A multi-source dataset of the tunnel environment is used, including 12 types of features (D=12) such as LiDAR point clouds, infrared thermal imaging, and sonar signals. The labels are dynamic target location, obstacle distribution, and safety path level. In an optional embodiment, multiple dynamic target locations also carry identity attributes with adversarial identity capabilities. A 20×20 simulated grid map is constructed to simulate the tunnel scene, with the starting point being the initial position of the mobile unit and the ending point being the task target point. Dynamic obstacles are randomly distributed.

[0135] Figure 4 This is a schematic diagram of the UCI dataset used in the feature selection experiment of the robot dog path planning method according to the embodiments of this application, as shown below. Figure 4 As shown, 15 UCI benchmark datasets were selected for experiments. These datasets cover different sizes, feature dimensions, and number of categories, and are representative. They include Exactly (exact matching dataset), BreastEW (Wisconsin breast cancer extended dataset), CongressEW (Congress voting extended weighted dataset), HeartEW (heart disease extended weighted dataset), Hillvalley (terrain classification dataset), IonosphereEW (ionospheric extended weighted dataset), Breastcancer (breast cancer dataset), Lymphography (lymphography dataset), PenglungEW (lung cancer extended dataset), Sonarall (whole sonar signal classification dataset), ParkinsonsEW (Parkinson's disease extended weighted dataset), Spectf (Parkinson's disease extended weighted dataset), WineEW (spectral feature dataset), Zoo (zoo animal classification dataset), and M-of-N (M-of-N pattern classification dataset), etc. Specific information such as the number of features, sample size, and number of categories can be obtained from the UCI official website. Feature dimensions range from 30 to 1000, and sample sizes range from 100 to 10,000. Standardization process: Z-score standardization is applied to continuous features, and one-hot encoding is performed on discrete features.

[0136] S33-2, Comparison of Algorithm and Parameter Settings.

[0137] Comparison Algorithms: Dung Beetle Algorithm (DBO), Particle Swarm Optimization (PSO), African Vulture Optimization Algorithm (AVOA), Chimpanzee Optimization Algorithm (Chimp), Pelican Optimization Algorithm (POA), Harris Eagle Optimization Algorithm (HHO), Sparrow Search Algorithm (SSA), Whale Optimization Algorithm (WOA).

[0138] Parameter settings: Population size N=50, maximum number of iterations T=500, ω=0.3.

[0139] Step 4: Algorithm comparison and performance evaluation.

[0140] S34-1 calculates performance metrics including fitness value, classification accuracy, number of features, convergence speed, and computation time. The specific method is as follows:

[0141] Fitness Value: The average fitness values ​​of each algorithm were compared on 15 UCI benchmark datasets. The lower the fitness value, the better the algorithm balances classification performance and feature parsimony in the feature selection task, and the higher the quality of the selected feature subset. The average fitness value of each algorithm after multiple runs was calculated using a formula and then compared and analyzed.

[0142] Classification accuracy: The K-Nearest Neighbors (KNN, K = 5) classifier was used, with Euclidean distance as the metric. The classification accuracy of each algorithm on the test set was calculated. This is one of the key indicators for evaluating the effectiveness of feature selection, and it can intuitively reflect the support effect of the selected feature subset on the classification task. The average classification accuracy of each algorithm in 30 independent runs was calculated to compare the performance of different algorithms.

[0143] Feature Count: This section calculates the number of features selected by each algorithm on different datasets to evaluate their feature compression capabilities. Fewer features indicate better performance in removing redundant and irrelevant features, helping to reduce model complexity and computational cost. The average feature count for each algorithm on each dataset is recorded and compared.

[0144] Convergence speed: The average convergence curves of each algorithm on 15 datasets are plotted to show the trend of the average fitness value as a function of the number of iterations, thus intuitively reflecting the convergence speed of the algorithm. Algorithms with fast convergence speed can achieve better fitness values ​​in fewer iterations, improving feature selection efficiency.

[0145] Computation time: Record the average running time of each algorithm on different datasets to examine the computational efficiency of the algorithms. Under the premise of ensuring feature selection effectiveness, the shorter the running time, the more practical the algorithm is and the better it can meet the real-time requirements of practical applications.

[0146] S34-2, Analysis of Experimental Results.

[0147] Figure 5 This is a bar chart showing the classification accuracy before and after feature selection according to the embodiments of this application, such as... Figure 5 As shown, experiments on 15 UCI benchmark datasets demonstrate that the IDBO-FS algorithm performs exceptionally well in feature selection tasks. On the Exactly dataset, IDBO reduces the average fitness value by 38.67% compared to traditional DBO, achieving a classification accuracy of 100% (standard deviation of 0), validating its advantage in balancing feature parsimony and classification accuracy. For biomedical datasets (such as Breastcancer and Lymphography), IDBO improves classification accuracy by 0.31% and 4.14%, respectively, indicating its strong adaptability to heterogeneous data. IDBO effectively reduces feature dimensionality while maintaining high classification accuracy. Figure 6 This is a schematic diagram illustrating the average fitness values ​​of various algorithms according to embodiments of this application under UCI, as shown below. Figure 6 As shown, in the BreastEW dataset, IDBO's Avg is 3.266, a 19% reduction compared to the traditional DBO. In the PenglungEW dataset, IDBO's Avg is 3.4667, significantly lower than DBO and SSA. The convergence curves reveal that IDBO's fitness value drops rapidly in the early stages of iteration. Compared to other algorithms, DBO and AVOA are prone to getting trapped in local optima and have slow convergence speeds. IDBO's median box plot value is significantly better than other algorithms in most datasets, and its IQR is generally narrower than the comparison algorithms. For example, in the HeartEW dataset, IDBO's IQR is 0 (Std = 0), indicating highly concentrated results. In the noisy Sonarall and Spectf datasets, IDBO's IQR is significantly lower than AVOA and WOA.

[0148] In a lane simulation environment, the feature subset selected by IDBO-FS (average 3.2 features) improved the robot dog's path planning success rate to 96% in a 20×20 grid map, reduced the path length by 15.3%, and increased the convergence speed by 37.5% compared to DBO, meeting the requirements of real-time tasks. Furthermore, in a complex 40×40 map, IDBO-FS maintained an accuracy of 90.5% with 4.1 features and maintained stability of over 87% under electromagnetic interference, fully demonstrating its robustness in dynamic and complex environments.

[0149] This embodiment also provides a robot dog path planning device, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the terms "module," "unit," "subunit," etc., can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0150] Figure 7 This is a structural block diagram of a robot dog path planning device according to an embodiment of this application, such as... Figure 7 As shown, the device includes:

[0151] The information acquisition module 10 is used to generate an initial feature population based on the acquired feature information through a refraction reverse learning mechanism; then, through an elite retention strategy and a fitness function, it selects the feature subset with the best fitness value from the initial feature population to generate the current elite population, and obtains the current global optimal position corresponding to the current elite population.

[0152] The optimization and update module 20 is used to input the current global optimal position and the first population position corresponding to the initial feature population into the position update module via the optimization algorithm model, update the first population position to the second population position; input the second population position into the perturbation module to generate a perturbed position; input the perturbed position into the path smoothing module to generate a smoothed position; wherein, the optimization algorithm model includes the position update module, the perturbation module and the smoothing module;

[0153] The path generation module 30 is used to determine whether the smoothed position meets the preset termination condition. If the termination condition is not met, the smoothed position is updated to the first population position, and the updated first population position is input into the optimization algorithm model for iterative calculation. If the termination condition is met, the optimal path is output based on the smoothed position.

[0154] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.

[0155] Furthermore, in conjunction with the robot dog path planning method in the above embodiments, this application embodiment can provide a storage medium for implementation. The storage medium stores a computer program; when the computer program is executed by a processor, it implements any of the robot dog path planning methods in the above embodiments.

[0156] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0157] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A path planning method for a robot dog, characterized in that, include: Based on the obtained feature space dimension and the initial search space, an initial feature population is generated through a refraction-based reverse learning mechanism, including: The refractive index parameter is adjusted by the refraction inverse learning mechanism, the initial search space is expanded into a refraction search space, and the initial feature population is generated based on the refraction search space and the feature space dimension. Then, by using the elite retention strategy and fitness function, the feature subset with the best fitness value is selected from the initial feature population to generate the current elite population, and the current global optimal position corresponding to the current elite population is obtained. The current global optimal position and the first population position corresponding to the initial feature population are input into the position update module in the optimization algorithm model, and the first population position is updated to the second population position, including: Based on the obtained current weight factor and the current global optimal position, the position of the first population corresponding to the initial feature population is updated to the position of the second population through the global search mechanism of the Osprey optimization algorithm in the position update module. The second group position is input into the perturbation module to generate a perturbation position; the perturbation position is input into the path smoothing module to generate a smooth position; wherein, the optimization algorithm model includes the position update module, the perturbation module, and the smoothing module; Determine whether the smoothed position meets a preset termination condition. If the termination condition is not met, update the smoothed position to the first population position and input the updated first population position into the optimization algorithm model for iterative calculation. If the termination condition is met, output the optimal path based on the smoothed position.

2. The robot dog path planning method according to claim 1, characterized in that, The step of inputting the second population position into the perturbation module to generate the perturbation position includes: Based on the obtained current iteration number and maximum iteration number, generate the degree of freedom parameters; based on the degree of freedom parameters, determine the current distribution function; based on the current distribution function, update the position of the second population to the perturbation position.

3. The robot dog path planning method according to claim 1, characterized in that, The step of inputting the disturbance position into the path smoothing module to generate a smoothed position includes: Based on the obtained random number and the current global optimal position, the perturbation position is updated to a smooth position through a periodic function.

4. The robot dog path planning method according to claim 1, characterized in that, The step of determining whether the smoothed position meets the preset termination condition includes: The fitness function is called on all smooth feature subsets corresponding to the smooth position to calculate the fitness value of each smooth feature subset, and the feature subset with the best fitness value is selected from each smooth feature subset to generate the optimal smooth feature subset; If the fitness value of the smoothed optimal feature subset is lower than the fitness value of the current global optimal position, then the feature position corresponding to the smoothed optimal feature subset is updated to the current global optimal position, and it is determined that the smoothed position does not meet the termination condition. If the fitness value of the smoothed optimal feature subset is not lower than the fitness value of the current global optimal position, then the current iteration number is obtained. If the current iteration number is greater than a preset fitness value constant number threshold, then the smoothed position is determined to meet the termination condition.

5. The robot dog path planning method according to claim 4, characterized in that, The step of calling the fitness function on all smooth feature subsets corresponding to the smooth position and calculating the fitness value of each smooth feature subset includes: Input all smooth feature subsets corresponding to the smooth position into the classifier and calculate the classification error; based on the classification error and the number of all smooth feature subsets, calculate the fitness value of each smooth feature subset using the fitness function.

6. The robot dog path planning method according to claim 4, characterized in that, The termination condition also includes the current iteration number reaching a preset maximum iteration number.

7. A path planning device for a robotic dog, characterized in that, include: The information acquisition module is used to generate an initial feature population based on the acquired feature space dimension and the initial search space through a refraction-backward learning mechanism. Then, by using the elite retention strategy and fitness function, the feature subset with the best fitness value is selected from the initial feature population to generate the current elite population, and the current global optimal position corresponding to the current elite population is obtained. The information acquisition module is also used to adjust the refractive index parameter through the refraction inverse learning mechanism, expand the initial search space into a refraction search space, and generate the initial feature population based on the refraction search space and the feature space dimension. The optimization and update module is used to input the current global optimal position and the first population position corresponding to the initial feature population into the position update module in the optimization algorithm model, and update the first population position to the second population position. The second group position is input into the perturbation module to generate a perturbation position; the perturbation position is input into the path smoothing module to generate a smooth position; wherein, the optimization algorithm model includes the position update module, the perturbation module, and the smoothing module; The optimization and update module is further configured to update the first population position corresponding to the initial feature population to the second population position based on the obtained current weight factor and the current global optimal position, through the global search mechanism of the Osprey optimization algorithm in the position update module. The path generation module is used to determine whether the smoothed position meets a preset termination condition. If the termination condition is not met, the smoothed position is updated to the first population position, and the updated first population position is input into the optimization algorithm model for iterative calculation. If the termination condition is met, the optimal path is output based on the smoothed position.

8. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the robot dog path planning method according to any one of claims 1 to 6 when running.

Images