Route optimization method and system based on genetic algorithm and geographic information

Abstract

The present invention discloses a route optimization method and system based on genetic algorithms and geographic information, which relates to the field of intelligent navigation and path planning technology. The method includes receiving geographic data related to route planning and constructing a multi-level spatial index; using a waypoint optimization algorithm, supported by the spatial index, and utilizing a genetic algorithm based on population diversity adaptive adjustment to determine the waypoint sequence of the route; using a path optimization algorithm for any two adjacent waypoints in the waypoint sequence, combined with the geographic information provided by the spatial index, and utilizing a genetic algorithm to generate the corresponding optimal path segment; iteratively optimizing the optimal path segment, iteratively optimizing the waypoint sequence and the optimal path segment, and integrating the optimization results into the autonomous navigation system after convergence. The method significantly improves the efficiency of spatial information processing, effectively eliminates high-risk and restricted areas, and achieves the optimal layout of the global route structure.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent navigation and path planning, and in particular to a route optimization method and system based on genetic algorithms and geographic information. Background Art

[0002] With the rapid development of unmanned autonomous navigation technology, intelligent transportation, and intelligent logistics systems, path planning and route optimization play a vital role in autonomous navigation platforms (such as drones, autonomous vehicles, and intelligent ships). Existing path planning methods typically rely on basic terrain, obstacle, and airspace management data provided by geographic information systems (GIS), combined with traditional shortest path algorithms (such as Dijkstra and A*) or heuristic search methods to achieve preliminary guarantees for route safety and economy. However, with the increasing complexity of application scenarios, such as three-dimensional urban spaces, variable weather conditions, dynamic environmental factors, and diverse task constraints, traditional path planning methods have significant shortcomings in coping with large-scale spaces, complex constraints, and multi-objective trade-offs.

[0003] On the one hand, existing algorithms are mostly based on single-level point-to-point path search, making it difficult to simultaneously optimize global route structures and refine local paths. This makes it difficult for overall route planning results to adapt to changing mission requirements and complex spatial constraints. On the other hand, faced with dynamic environments (such as wind field changes and sudden flight restrictions), traditional static planning schemes are unable to achieve real-time response and adaptive adjustment to environmental changes, and are prone to problems such as local optimality, slow convergence, and poor scalability. In addition, the large amount of spatial geographic data and its complex spatiotemporal attributes significantly increase the computational burden and reduce the system's response efficiency when directly applied to traditional algorithms.

[0004] In recent years, intelligent optimization algorithms (such as genetic algorithms and ant colony algorithms) have garnered attention in the field of path planning due to their global search capabilities and multi-objective optimization characteristics. However, effectively integrating them with spatial indexing technology to enable rapid retrieval and environmental perception of large-scale, multi-level geographic information remains a challenge. Current publicly available technologies lack the ability to coordinate optimization of multi-level spatial indexing with adaptive genetic algorithms, dynamically handle path segment constraints, and achieve real-time system-level integration. These limitations make it difficult to meet the efficiency, real-time, and adaptability requirements of autonomous navigation systems for route planning. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing path planning methods have the problems of weak route optimization capability, poor environmental adaptability and lack of efficient global adaptive optimization mechanism in complex geographical environments and multi-constraint scenarios.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: a route optimization method based on genetic algorithms and geographic information, comprising receiving geographic data related to route planning and constructing a multi-level spatial index;

[0008] A waypoint optimization algorithm is used, supported by spatial index, and a genetic algorithm based on population diversity adaptive adjustment is used to determine the waypoint sequence of the route;

[0009] For any two adjacent waypoints in the waypoint sequence, a path optimization algorithm is used, combined with the geographic information provided by the spatial index, and a genetic algorithm is used to generate the corresponding optimal path segment;

[0010] The optimal path segment is iteratively optimized, the waypoint sequence and the optimal path segment are iteratively optimized, and after convergence, the optimization results are integrated into the autonomous navigation system.

[0011] As a preferred embodiment of the route optimization method based on genetic algorithm and geographic information of the present invention, the geographic information data includes static data and dynamic data; the static data includes regional terrain data, airspace management data and regional mission rule data, and the dynamic data includes static environment data and dynamic environment data;

[0012] The multi-level spatial index includes a coarse-grained spatial index layer and a fine-grained spatial index layer;

[0013] A coarse-grained spatial index layer is constructed based on the static class data. The coarse-grained index uses a two-dimensional square area as the basic unit, and the unit side length is set to L. The value of L is preset according to the area of ​​the mission area. The index is constructed using a quadtree structure. The index node stores: the unit coordinate range, the average terrain height of the area, the airspace label set, the mission rule label, the risk level field, and the unique spatial coding identifier;

[0014] A fine-grained spatial index layer is constructed based on the dynamic class data. The fine-grained index is divided into two-dimensional square spatial units with a side length of 1. The index is constructed using an R-tree structure. The fields stored in the index nodes include static fields and dynamic fields. The static fields include the coordinate range of the spatial unit, obstacle spatial boundary information, node status label, and spatial coding information. The dynamic fields include environmental data field values ​​and their corresponding timestamp information.

[0015] As a preferred embodiment of the route optimization method based on genetic algorithm and geographic information of the present invention, the waypoint optimization algorithm includes: setting a screening rule based on the airspace label set, risk level field and mission rule label field stored in each index node in the coarse-grained spatial index structure, screening each index node, extracting the central coordinate point as a candidate waypoint, and establishing a candidate waypoint set based on the candidate waypoint;

[0016] The screening rules include removing nodes with no-fly zones or restricted airspace identifiers in their airspace labels, and only retaining spatial index nodes with risk levels below a preset threshold and marked as passable in the mission rules;

[0017] Several waypoint indexes in the candidate waypoint set are used as chromosome genes, and the chromosomes are constructed in the order of arrangement of the waypoint indexes as search individuals of the genetic algorithm;

[0018] During the execution of the genetic algorithm, a spatial perception control mechanism is set up to verify the spatial attributes of the waypoints generated in the mutation phase, restricting chromosomes that do not meet the risk conditions or airspace requirements from entering the next generation;

[0019] The spatial perception control mechanism includes: when performing a mutation operation, calling a coarse-grained spatial index structure, retrieving the risk level field and airspace tag set of the spatial unit where the mutated waypoint is located, and determining whether the mutated waypoint is located in an area with a higher risk level than the original waypoint, or whether the airspace tag set contains a no-fly zone mark or a restricted airspace mark; if the conditions are not met, the mutated waypoint is eliminated and the mutation result is not included in the subsequent population;

[0020] During the iterative process of the genetic algorithm, a fitness function is constructed based on the path distance, regional risk value and task fitness, and the waypoint sequence with the highest fitness score is determined as the optimal individual of the current generation, which serves as the input for subsequent path optimization.

[0021] As a preferred embodiment of the route optimization method based on genetic algorithm and geographic information of the present invention, the route optimization algorithm includes constructing a route segment optimization task based on any two adjacent waypoints in the waypoint sequence; the route segment consists of a starting point, an end point, and a plurality of waypoints;

[0022] Based on the fine-grained spatial index layer, the field information in the spatial index node where the path point is located is called, and combined with the task time step, a binding relationship between the path point and the temporal environment data is established;

[0023] An A* search algorithm is used to perform path search in the fine-grained spatial index layer, and a path cost function is constructed based on the accumulated environmental costs of the path points and the heuristic function to generate an initial path segment.

[0024] Based on the initial path segment, a path perturbation mechanism is introduced to generate perturbed path segment individuals as an initial population of the genetic algorithm;

[0025] During the evolution of the genetic algorithm, a fitness function is constructed based on the path segment length, the risk level field of the path node, the dynamic environment field value under the time step, and the path curvature. The path individuals generated in each generation are screened for validity using a space-time dual constraint mechanism to ensure that only path individuals that meet the environmental and spatial requirements can enter the next generation of population.

[0026] When the path segment optimization meets the convergence conditions or reaches the maximum number of iterations, the corresponding optimal path segment is output, and the optimal results of all path segments are spliced ​​together to form a complete path set, which serves as the input for route quality assessment and subsequent iterative updates of the waypoint sequence.

[0027] As a preferred solution of the route optimization method based on genetic algorithm and geographic information described in the present invention, the path perturbation mechanism includes performing spatial perturbation operations on the intermediate path points of the path segment within a limited perturbation radius; calling the index nodes in the multi-level spatial index to determine whether the path points after perturbation are located in an obstacle area, a no-fly zone, a restricted altitude area or a high-risk level area, and only retaining the perturbed path segment individuals that do not fall into the above areas to construct the initial population of the genetic algorithm.

[0028] As a preferred solution of the route optimization method based on genetic algorithm and geographic information of the present invention, wherein: the space-time dual constraint mechanism includes a space constraint mechanism and a time constraint mechanism;

[0029] The spatial constraint mechanism includes determining whether a path point falls into an obstacle area, a no-fly zone, a restricted altitude area, or a high-risk area based on the airspace management tag, obstacle information, and risk level fields of the node in the fine-grained spatial index;

[0030] The time constraint mechanism includes judging whether the path node falls into an area where the environmental risk exceeds a threshold based on the dynamic environment field value of the task time step corresponding to the path node. If the pass requirements are not met, the path individual will not enter the next generation.

[0031] As a preferred embodiment of the route optimization method based on genetic algorithm and geographic information of the present invention, the iterative optimization of the optimal path segments includes concatenating the optimal path segments to form a complete path set, and scoring the fitness of the entire route based on the total length of the complete path set, the risk level field of the path nodes, the dynamic environment field value at the mission time step, and the path curvature;

[0032] According to the fitness score results, the path segments are adjusted through the genetic algorithm, and the optimized path segments are selected by combining the space-time dual constraint mechanism;

[0033] Path optimization and fitness scoring are performed cyclically until the fitness score of the complete path set meets the convergence condition or reaches the maximum number of iterations, and the optimization result is output.

[0034] As a preferred solution of the route optimization system based on genetic algorithm and geographic information of the present invention, wherein: a multi-layer module receives geographic data related to route planning and constructs a multi-level spatial index;

[0035] The sequence module uses a waypoint optimization algorithm and a genetic algorithm based on population diversity adaptive adjustment with the support of spatial index to determine the waypoint sequence of the route;

[0036] A path segment module, for any two adjacent waypoints in the waypoint sequence, adopts a path optimization algorithm, combines the geographic information provided by the spatial index, and uses a genetic algorithm to generate the corresponding optimal path segment;

[0037] The optimization module iteratively optimizes the optimal path segment, iteratively optimizes the waypoint sequence and the optimal path segment, and integrates the optimization results into the autonomous navigation system after convergence.

[0038] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a route optimization method based on genetic algorithm and geographic information.

[0039] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a route optimization method based on a genetic algorithm and geographic information.

[0040] Beneficial effects of the present invention: The route optimization method based on genetic algorithms and geographic information provided by the present invention achieves efficient hierarchical management and rapid retrieval of static and dynamic data of the geographic environment by constructing a multi-level spatial index structure, significantly improving the efficiency of spatial information processing. In the waypoint optimization stage, the use of an adaptive genetic algorithm combined with a spatial index screening mechanism can effectively eliminate high-risk and restricted areas, achieving the optimal layout of the global route structure. During the path optimization process, by introducing an initial path generated based on the A* algorithm and combining it with a path perturbation mechanism, the initial population has higher diversity and feasibility, improving the global search capability of the genetic algorithm. Furthermore, a spatial-temporal dual constraint mechanism is adopted in the genetic evolution process to ensure that the path points avoid obstacles in real time and dynamically adapt to environmental changes, effectively avoiding no-fly zones, obstacles and extreme weather risks. The overall route uses fitness scoring for global feedback and iterative optimization, which can achieve rapid convergence in complex tasks and dynamic environments, ensuring a safe, smooth and low-energy route. Ultimately, the optimization results can be seamlessly integrated into the autonomous navigation system to meet the efficiency, real-time and high adaptability requirements of intelligent route planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 This is an overall flow chart of the route optimization method based on genetic algorithm and geographic information provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0043] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0044] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a route optimization method based on genetic algorithm and geographic information, including:

[0045] S1: Receive geographic data related to route planning and build a multi-level spatial index.

[0046] Geographic information data includes: regional terrain data, airspace management data, static environment data, dynamic environment data, and regional mission rule data.

[0047] The regional terrain data includes: digital elevation model (DEM), slope map, ground undulation information, surface roughness parameters, and gridded terrain grid data.

[0048] The airspace management data includes: no-fly zone boundary information, altitude restriction zone definition, altitude restriction map, temporary airspace access restriction notice, and airspace classification information (such as civil aviation airspace, controlled airspace).

[0049] The static environment data includes: building outline boundaries, three-dimensional city models (such as CityGML data), permanent obstacle locations, and transportation infrastructure information (such as towers, overpasses, bridges, and transmission towers).

[0050] The dynamic environmental data includes: wind speed and direction data, air pressure and temperature and humidity data, rainfall forecast, thunderstorm or extreme weather warning information, and time-series-based wind field simulation output.

[0051] The regional mission rule data includes: traversable corridor boundaries, regional priority classifications, mission constraint area definitions (such as military management areas, ecologically sensitive areas), risk level distribution maps, and mission airworthiness score marking maps.

[0052] According to the data type and spatial resolution requirements, the received geographic information data are divided into two types of index levels: Coarse-grained index layer: with the goal of supporting high-level waypoint selection, the input data includes: regional terrain data, airspace management data, and regional mission rule data.

[0053] Fine-grained indexing layer: aims to support low-level path optimization. Input data includes static environment data and dynamic environment data.

[0054] Construct a coarse-grained index structure: divide the mission area into fixed-size spatial units. The spatial unit is a two-dimensional square with a side length set to L. The value of L is preset based on the spatial scale and geographical complexity of the mission area. Specifically, when the mission area is less than 25 square kilometers, L is set to 50 meters; when the mission area is between 25 and 100 square kilometers, L is set to 100 meters; when the mission area is greater than 100 square kilometers, L is set to 200 meters.

[0055] All spatial units are numbered uniformly to ensure that the task area is covered without overlap or gaps.

[0056] To control the number of layers and index granularity of the quadtree structure, the following splitting rules are set: if there are more than two airspace management labels in the area covered by the current node (such as the overlap of the restricted altitude area and the no-fly zone), a quad-split is forced to be executed; if the risk level of any unit in the spatial unit under the current node is higher than the threshold R (such as R ≥ 4) and the current level depth does not exceed the set maximum level D (such as D = 4), it is allowed to enter the next level index; if the areas covered by the child nodes after the split are all low-risk areas, no further subdivision is carried out and the path is terminated.

[0057] Each index node stores fields including: coordinate range (upper-left corner coordinates and side length); average terrain height for the region; airspace tag set (e.g., whether altitude restricted or no-fly); mission rule tags (e.g., priority level, accessibility); maximum and average risk levels; child node references (pointing to the four child node indexes, if any); and a Morton code (uniquely identifying the node). All index nodes are organized as dictionaries or arrays, supporting fast spatial lookups and local updates, allowing subsequent high-level waypoint selection modules to call them on demand.

[0058] Constructing a fine-grained index structure: For both static and dynamic environmental data, combined with the spatial resolution requirements of the mission area and the real-time access needs of the path optimization phase, a fine-grained spatial index that supports high-frequency queries and dynamic updates is constructed. This specifically includes the following steps:

[0059] The task area is divided into small spatial units. Each spatial unit is a two-dimensional square with a side length of l. The value of l is set based on the spatial accuracy required for path optimization, preferably in the range of 1 to 10 meters. All spatial units are numbered using absolute coordinates to ensure complete coverage, without overlap or gaps.

[0060] Establish an R-tree-based data index structure for the divided spatial units. Each index node of the R-tree corresponds to a spatial region, storing static object information and dynamic environment attributes within the region;

[0061] Static object information includes: obstacle boundary coordinates, building outline point sets, and location attributes of permanent facilities; dynamic environmental attributes include: wind speed, wind direction, air pressure, precipitation, visibility and other environmental fields, and are accompanied by time tags for time series updates.

[0062] Each index node supports range query, neighborhood search, and dynamic interpolation access, which are used for collision detection and fitness calculation in path evaluation.

[0063] The field information stored in each fine-grained index node includes: static fields and dynamic fields, which are divided according to whether they change over time.

[0064] Static fields include: coordinate range (upper left corner coordinates and side length); static object identifier and its spatial boundary set (such as obstacles, buildings, etc.); node status label (such as idle, obstacle, restricted area); Morton code.

[0065] Dynamic fields include: environmental field values ​​at the current time step (such as wind speed, wind direction, air pressure, temperature and humidity, precipitation, etc.); change trends or change rates of environmental fields; and timestamp information.

[0066] In view of the time-sensitive nature of dynamic environmental data, the update mechanism of the dynamic index structure includes: all dynamic fields are timestamp-ed, and historical records and prediction results are maintained in time series; during the path planning process, the environmental data required at the current time point is loaded or interpolated on demand based on the current optimization algebra or simulation time step; and a local refresh mechanism is supported, which only performs incremental updates on areas related to the task path to avoid full map reconstruction.

[0067] The fine-grained index structure is encapsulated as a modular interface, which supports the path optimization algorithm to call index node attributes according to path node coordinates in operations such as fitness function evaluation, path mutation, and intersection, thereby realizing real-time environmental perception and terrain interactive judgment of path points.

[0068] Furthermore, by dividing geographic data into coarse-grained and fine-grained index levels based on data type and mission resolution requirements, the management efficiency of multi-source heterogeneous geographic information within the mission area has been significantly improved. The coarse-grained index uses a quadtree structure to manage airspace labels and risk levels, supporting efficient land parcel aggregation and airspace conflict identification, facilitating large-scale waypoint preselection by upper-level modules; the fine-grained index uses an R-tree structure to encapsulate static and dynamic environmental attributes, enabling high-frequency access and local updates at the path node level, meeting the accuracy requirements of real-time path assessment and dynamic obstacle avoidance. This significantly reduces the time complexity of full-map queries and redundant data storage, improving the system's spatial response speed.

[0069] The coarse-grained index structure incorporates rule fields such as airspace management tag sets, regional priority classifications, and mission airworthiness scores. This allows for rapid filtering of unsuitable areas during the high-level waypoint screening phase and, through risk level aggregation features, guides waypoints toward low-risk areas. Fine-grained indexing provides dynamic environmental data support at the path segment level. Node fields contain multi-dimensional parameters such as wind speed, air pressure, and visibility, and come with a time-series update mechanism, enabling true perception and prediction capabilities during path optimization. Waypoints can perceive local environmental changes in real time during intersection and mutation operations, effectively avoiding risk areas such as sudden obstacles and extreme weather disturbances, and improving the overall path's airworthiness and robustness.

[0070] Furthermore, by configuring unique Morton codes and standard field formats for each level of spatial index nodes, modular encapsulation and efficient scheduling of the index structure are achieved. The optimization algorithm can call the corresponding granularity index on demand at different stages without redundant preloading, thereby improving the overall algorithm operation efficiency. At the same time, the dynamic index supports time tags and local refresh mechanisms, and dynamically loads environmental fields in combination with the current simulation time step to avoid the pressure of full-graph reconstruction caused by environmental updates, and provide iterative, interpolable, and dynamically adaptable environmental data support for the path optimization algorithm. This design gives the route optimization process good flexibility and algorithm compatibility, adapting to a variety of mission scenarios and navigation system requirements.

[0071] S2: Using the waypoint optimization algorithm, supported by spatial index, and using the genetic algorithm based on population diversity adaptive adjustment, the waypoint sequence of the route is determined.

[0072] Based on the regional terrain data, airspace management data and regional mission rule data stored in the coarse-grained spatial index structure, a set of candidate waypoints that meet the mission constraints is constructed. Specifically, all coarse-grained index nodes are traversed. , whose fields include the coordinate center point , airspace label set , risk level Task rule tag collection Candidate waypoint set The build rule is defined as: ;

[0073] in, Represents a set of candidate waypoints; Represents a coarse-grained spatial index node The center point coordinates of Representation node The corresponding airspace management tag set is used to mark whether the area belongs to a no-fly zone or restricted airspace; Representation node The corresponding task rule tag set, including attributes such as whether it is marked as "passable"; Representation node The risk level field; Indicates the preset risk level threshold. The screening conditions include: the node does not belong to a no-fly zone or restricted airspace, that is, ; The risk level of the node is lower than the threshold, that is ; and the task rules contain the "passable" mark, that is, .

[0074] Waypoint sequence encoding as chromosome , each gene Represents the candidate set A waypoint index in . The genetic algorithm searches for the optimal arrangement , generate a waypoint sequence covering the mission area, and its fitness function is designed as: ;

[0075] in, Represents a waypoint sequence 's fitness score; Indicates the first waypoints; Indicates the Waypoints and Euclidean distance between waypoints; Indicates waypoints The risk level of the spatial unit; represents the number of waypoints in the waypoint sequence; 、 、 They represent the weight coefficients of path length, risk level and task matching, respectively, and are used to adjust the influence of each factor in the fitness function.

[0076] Indicates the Waypoints and The task priority adaptability between the waypoints is defined as the two-point task airworthiness score and The average value of , that is: ;

[0077] in, Indicates the airworthiness score corresponding to the task rule label field in the coarse-grained index node.

[0078] In order to improve the convergence speed and search diversity, an adaptive genetic mechanism based on population diversity is introduced. Assume that the current generation is , the population size is , the average Hamming distance between chromosomes is: ;

[0079] in, Represents the Hamming distance between chromosomes. Characterized population structure characteristics, dynamically adjusting crossover probability and mutation probability , realizing the self-balancing parameter adjustment mechanism of "enhancing variation when population diversity decreases and enhancing convergence when population divergence is too strong". The specific update formula is: ;

[0080] If the population tends to be precocious, increase the mutation intensity and introduce segment perturbation operations; if the population diverges too strongly, increase the retention weight of high-quality individuals.

[0081] In addition, a spatial perception mechanism based on environmental label judgment is introduced during the crossover and mutation process: when a newly generated waypoint does not meet the regional risk level or airspace compliance requirements, its fitness is penalized or it is rejected from being included in the next generation population. Specifically: ;

[0082] That is, if the waypoint generated by the mutation falls into an area with a higher risk level, the mutation is abandoned. Through this spatial perception operation, the route structure is guided to continuously concentrate in low-risk and mission-priority areas.

[0083] Finally, the optimal waypoint sequence that meets the risk level, mission rules and airspace management labels is output. : ;

[0084] Serves as basic input for the subsequent path refinement phase.

[0085] In the existing technology, the selection of candidate waypoints usually relies on regular grid sampling or simple range filtering, lacking a comprehensive judgment of airspace policies, regional risks and mission rules, which easily leads to the path structure passing through no-fly zones, high-risk areas or mission-restricted areas. In this technical solution, the set of candidate waypoints is generated by coarse-grained spatial index nodes, and is jointly screened based on the three attributes stored in the nodes: airspace management labels, risk level fields, and mission rule labels, to form a set of waypoints with structural compliance, spatial legality, and high mission airworthiness. This multi-label spatial joint screening mechanism significantly improves the quality of waypoints in the path structure generation stage, avoids the risks of illegal area crossings and mission violations at the source, and breaks through the limitations of the existing algorithms of single waypoint selection criteria and difficulty in embedding environmental constraints.

[0086] Existing genetic algorithms for route optimization often use fixed crossover and mutation probabilities, failing to dynamically adjust based on the population's evolutionary state. This can lead to premature convergence or search divergence, compromising the ability to achieve a global optimum. This technical solution introduces an adaptive genetic parameter adjustment mechanism based on a population diversity measure (mean Hamming distance), enabling dynamic generational changes in crossover and mutation probabilities, enhancing the algorithm's adaptability and ability to control its evolutionary rhythm. This mechanism automatically adjusts search intensity based on population distribution, enhancing global optimization capabilities in complex terrain and under highly constrained mission conditions. It effectively overcomes the robustness issues inherent in traditional genetic algorithms, often caused by parameter incompatibility in complex search spaces.

[0087] The crossover and mutation operations in traditional genetic algorithms lack awareness of the geographical environment and may generate illegal waypoints that fall into high-risk areas or airspace conflict zones, resulting in frequent failures in the path evaluation phase. This technology introduces a perceptual evolutionary operator mechanism based on spatial index label feedback to determine in real time at the waypoint level whether the mutation results meet regional risk and airspace compliance requirements. For mutant individuals that do not meet the airworthiness conditions, the system will impose penalties or directly discard them to ensure that only usable, legal, and high-value solutions are retained during the evolution process. This mechanism introduces spatial adaptive filtering and guidance capabilities in the genetic evolution process, effectively avoiding the generation of illegal paths and the waste of evaluation resources, breaking through the structural limitations of traditional algorithms that are not aware of space and do not provide feedback on the environment.

[0088] S3: For any two adjacent waypoints in the waypoint sequence, a path optimization algorithm is used, combined with the geographic information provided by the spatial index, and an adaptive genetic algorithm is used to generate the corresponding optimal path segment.

[0089] Furthermore, based on the waypoint sequence , for any adjacent waypoint pair , constructing a path segment optimization task, the goal is to generate the optimal connection path that meets the spatial traffic constraints and the adaptability of the task environment ,in , , and the rest are intermediate nodes on the path.

[0090] During the path optimization process, the terrain data, static obstacle information, airspace management tags and dynamic environment fields (including wind speed, wind direction, air pressure, precipitation, etc.) in the fine-grained spatial index structure are combined to achieve joint perception of the spatial and temporal environment of the path node. Corresponding to a task time step , and according to the spatial position and time step of the path point , query the field information of the corresponding node in the spatial index structure. If there is no observation data at this time step, interpolation calculation is performed based on the predicted field in the index node to achieve the time series binding of the path point and the dynamic environment field. To construct the initial population of the genetic algorithm, the A* search algorithm is used to perform heuristic path search in a fine-grained spatial grid to build a basic path solution. . Search cost function Defined as: ;

[0091] in, Indicates a waypoint Comprehensive evaluation cost; Indicates starting point To the current waypoint the cumulative cost of Indicates that from the current path point To the destination The heuristic cost is usually evaluated using Euclidean distance as an approximate evaluation; Represents the adjustment coefficient of the heuristic term.

[0092] based on , a path perturbation mechanism is introduced to expand population diversity, specifically including: performing spatial perturbation on the intermediate nodes of the path (the perturbation radius r is set according to the spatial resolution), filtering out the nodes that do not meet the passability after perturbation, and constructing multiple path individuals with legal structures but different morphologies to form the initial population set , used for the initial evolution of genetic algorithm.

[0093] The path fitness function formula is expressed as: ;

[0094] in, Indicates the length of the path segment; Indicates the static risk level of the spatial unit where the path node is located; Indicates the path node at time step The corresponding dynamic environment field cost (such as wind speed squared, wind direction deviation, energy consumption assessment, etc.); Represents the curvature of the adjacent nodes, which is used to evaluate the smoothness of the path; 、 、 、 is the adjustment coefficient.

[0095] The crossover and mutation operations in the genetic algorithm are both regulated by the space-time dual constraint mechanism:

[0096] The spatial constraint mechanism includes: for the path segments in the intersection results, calling the fine-grained spatial index node to determine whether they cross obstacles, height restrictions or no-fly zones. If the spatial label constraints are violated, they are removed or repaired.

[0097] The time constraint mechanism includes: the new path nodes generated by mutation , to determine whether it is in the task time step The dynamic environment fields below meet the acceptable range. If there is extreme weather, wind speed exceeding the limit, or risk mutation, it is considered an illegal mutation and the path node maintains the original value: ;

[0098] in, Indicates the first Original path points, Represents the updated path point after the path mutation operation; Represents the candidate mutation point, which is generated by the genetic algorithm in the current generation for replacement New path node candidate value; Indicates a waypoint The risk level field value in the spatial index node. Represents candidate path points Corresponding risk level; symbol Represents a logical "AND" operation, that is, two conditions are met at the same time; Represents the set of candidate path points, which includes all legal path points that meet the spatial accessibility requirements; Indicates "otherwise"; the meaning of the entire piecewise function is: if the newly generated path point The risk level is lower than the original path point The risk level of the point belongs to the set of legal candidate path points. , then Replace the original path point ; Otherwise, keep the waypoint constant.

[0099] The path optimization process is terminated after the maximum evolutionary generations or fitness convergence threshold is met, and the optimal individual of the path is output. .

[0100] Finally, each path segment Splicing to form a complete path set , as the overall route refinement result, for subsequent global fitness evaluation and iterative feedback module to call.

[0101] Path segment optimization utilizes the A* search algorithm combined with a genetic evolutionary mechanism. By driving perturbations through spatial indexing, it can rapidly generate diverse and feasible initial solutions in large, complex geographic spaces. Compared to traditional heuristic or single global search methods, this approach significantly improves spatial coverage and initial optimization efficiency, overcoming the technical barriers of low-quality initial solutions and poor globality in complex environments.

[0102] Furthermore, by introducing dual spatial and temporal constraints, path nodes dynamically avoid obstacles, circumvent airspace restrictions, and perceive weather changes in real time during the evolution process, ensuring that the path plan is highly safe and adaptive in dynamic environments. Traditional methods often use a single static obstacle avoidance or risk indicator, which is difficult to handle in volatile and high-risk environments. This mechanism significantly improves the robustness of path planning and solves the problem of route safety in dynamic environments.

[0103] Furthermore, the segmented path structure is combined with a global feedback iterative mechanism to continuously optimize the path structure through fitness scoring, achieving multi-objective balance, path smoothing, and energy optimization. Existing solutions are prone to falling into local optimality or lacking global-local coordination, making them difficult to meet complex navigation requirements. This mechanism provides autonomous navigation systems with efficient, stable, and dynamically responsive route optimization capabilities, overcoming the key technical barrier of balancing global and local considerations in complex environments.

[0104] S4: Iteratively optimize the optimal path segment, iteratively optimize the waypoint sequence and the optimal path segment, and integrate the optimization results into the autonomous navigation system after convergence.

[0105] Set the current generation waypoint sequence All corresponding optimal path segments Splice into a complete route. Attributes, calculate the fitness function of the entire route , whose expression is: ;

[0106] in, Indicates the current waypoint sequence The corresponding overall fitness score; 、 、 、 Represent the weight coefficients of path length, risk level, dynamic environmental cost and curvature smoothness respectively; Indicates the total number of waypoints in the waypoint sequence; Indicates that With the The optimal path segment formed by connecting waypoints; Represents a path segment Total length; Represents a path segment Middle waypoints; Indicates a waypoint The risk level field value of the corresponding spatial index node; Indicates a waypoint At task time step Dynamic environment field cost under; Indicates a waypoint The corresponding curvature.

[0107] During the overall optimization iteration process, the current fitness score is used , genetic algorithm operations are sequentially performed on the waypoint sequence population, including: based on the fitness score, using strategies such as tournament selection or roulette, to select the waypoint sequence with better fitness from the current population as the parent. Single-point or multi-point crossover operations are performed on the parent waypoint sequence, and some waypoint indexes are randomly exchanged to generate a new child waypoint sequence. Spatial perturbations or random replacements are performed on some waypoints in the child waypoint sequence to generate new candidate waypoints, which are then screened using a dual spatial-temporal constraint mechanism to ensure that all mutated waypoints do not fall into obstacle areas, no-fly zones, restricted altitude areas, or high-risk areas, and that their dynamic environment field values ​​at the corresponding mission time step meet the safety threshold requirements.

[0108] For each new generation of waypoint sequences, the path segment optimization step needs to be re-executed to obtain a new optimal path segment. , and splice to form a complete path set P, based on which the overall fitness score is updated The above process is iterated continuously, as follows: ;

[0109] in, Represents the iterative operation of the adaptive genetic algorithm with a space-time dual constraint mechanism.

[0110] The iterative process terminates when any of the following convergence conditions is met:

[0111] Overall fitness score The change in several consecutive generations is less than the preset threshold ,Right now ;

[0112] or the maximum number of iterations is reached .

[0113] Finally, the optimal waypoint sequence at the convergence moment is output and its corresponding full path set , as the final result of route optimization.

[0114] The dynamic evaluation and real-time adjustment of path segment quality breaks through the limitation of traditional single-time path generation that cannot perform global feedback optimization based on the overall route performance, and significantly improves the overall safety and economy of the route.

[0115] By using a genetic algorithm to cyclically update path segments driven by fitness and combining it with a space-time dual constraint mechanism, it can automatically eliminate solutions that do not meet environmental or spatial constraints. This overcomes the technical bottlenecks of traditional algorithms, such as slow convergence, easy falling into local optimality, and poor dynamic adaptability to the environment, and effectively improves the feasibility and stability of routes under complex geographical and meteorological conditions.

[0116] Multiple rounds of feedback adjustment and fitness scoring closed loop enable the optimization results to dynamically adapt to mission changes and environmental disturbances, realize the adaptive evolution of waypoint sequence and path structure, and provide the autonomous navigation system with a new intelligent route planning model that is both efficient, real-time and highly reliable.

[0117] Example 2 is an embodiment of the present invention, which provides a route optimization method based on genetic algorithm and geographic information. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0118] Using intelligent drone urban delivery routes as an application scenario, a 10km x 10km area in the core area of ​​Shanghai's Pudong New Area was selected as the test area. Geographic data included a 1m resolution digital elevation model, 3D building outlines, the urban road network, weather conditions (including wind speed, direction, temperature, humidity, and rainfall forecasts), airspace management data (no-fly zones, restricted altitude zones), and delivery mission rules (priority, time windows, sensitive areas, etc.).

[0119] First, all spatial data within the region is categorized into static categories (such as terrain, buildings, and airspace regulations) and dynamic categories (such as weather and real-time road occupancy). A coarse-grained (L = 100m) quadtree spatial index and a fine-grained (L = 5m) R-tree index are constructed for each. The coarse-grained index is used for waypoint candidate screening, while the fine-grained index is used for path segment optimization and environmental constraint checking.

[0120] An initial waypoint sequence was generated based on the route's origin (Warehouse A, 121.50°E, 31.22°N) and destination (Campus B, 121.60°E, 31.26°N). A genetic algorithm adaptively adjusted based on population diversity was used, using a coarse-grained index as the screening basis. Only waypoints with a risk level below 3, permitted by the mission rules, and avoiding no-fly zones were retained. The population size was set to 50, the number of iterations to 100, the crossover probability to 0.7, and the mutation probability to 0.2. Parameters were adjusted in real time during evolution to ensure diversity. Based on fitness scores (including total path length, risk value, and mission priority alignment), the waypoint sequence was determined as input for subsequent route segment optimization.

[0121] For each pair of adjacent waypoints, a fine-grained spatial index is used, and the A* algorithm, combined with spatial constraints and dynamic environmental data, is employed to generate feasible initial path segments. Using a perturbation mechanism, intermediate nodes are perturbed within a 5m radius, and the spatial index is used to verify path legitimacy. Only path segments that do not traverse obstacles, no-fly zones, or extreme weather conditions are retained as the initial population for the genetic algorithm. The genetic algorithm optimizes using a multi-metric fitness function based on path length, risk level, wind speed, and curvature, with each generation undergoing dual spatial and temporal constraint screening. Finally, all path segments are concatenated, and once the overall route fitness score converges, a complete route and navigation instructions are generated and integrated into the drone's navigation system.

[0122] Experimental data: Total number of coarse-grained index units: 1000. Total number of waypoint candidates: 148. Number of waypoints optimized (initial generation): 15. Total number of path segments: 14. Average number of path points per segment: 42.50. Number of fitness convergence generations: 56. Total route length before optimization: 15.24 km. Total route length after optimization: 13.89 km. Average path risk before optimization: 2.87. Average path risk after optimization: 1.45. Number of dynamic environment violations before optimization: 5. Number of dynamic environment violations after optimization: 0. Average route curvature before optimization: 0.042. Average route curvature after optimization: 0.025. Theoretical energy consumption before optimization: 4.32 kWh. Theoretical energy consumption after optimization: 3.68 kWh. Planning time (including data loading, indexing, and full-process optimization): 112.75 seconds.

[0123] By comparing the key indicators before and after optimization, this embodiment clearly shows that the invention scheme has significantly improved the route quality, task adaptability and system efficiency. First, the use of multi-level spatial indexes to manage large-scale geographic data greatly improves the efficiency of waypoint screening and path segment queries, and the spatial data access latency is reduced to 38.50% of the pre-optimization level. Compared with the traditional algorithm based on single-level path search, the total route length of this scheme is shortened by 1.35km (a reduction of 8.86%), and the path selection is more economical. The average risk level has dropped by 49.48%, and the route is safer; the number of dynamic environmental violations (such as wind speed exceeding the limit, extreme weather points) has been reduced from 5 times to 0, effectively avoiding potential dangers.

[0124] Optimizing the curvature index resulted in a smoother route, facilitating stable UAV flight. Energy consumption was reduced from 4.32 kWh to 3.68 kWh (a 14.81% saving), demonstrating the practical utility of multi-objective fitness functions in global feedback optimization of path segments. A dual spatial-temporal constraint mechanism ensured that each generation of individuals met both spatial feasibility and environmental adaptability, enabling real-time scheduling in dynamic environments. Adaptive adjustment of genetic algorithm parameters effectively prevented premature convergence of the population, resulting in a 28.36% improvement in optimization convergence speed compared to traditional fixed-parameter algorithms.

[0125] Furthermore, experiments have shown that the final route output by this method can be directly integrated with the drone control system, with planning time taking less than two minutes, fully meeting practical engineering requirements. Comprehensive analysis shows that this solution overcomes the bottlenecks of traditional methods in handling multi-layer spatial indexing, dynamic constraints, and global feedback optimization, achieving comprehensive improvements in route safety, economy, and intelligence. Compared to existing technologies, it has significant innovation and practical application value.

[0126] Example 3 is an embodiment of the present invention, which provides a route optimization system based on genetic algorithm and geographic information, with a multi-layer module that receives geographic data related to route planning and constructs a multi-level spatial index.

[0127] The sequence module adopts the waypoint optimization algorithm and, with the support of spatial index, uses the genetic algorithm based on population diversity adaptive adjustment to determine the waypoint sequence of the route.

[0128] The path segment module uses a path optimization algorithm, combined with the geographic information provided by the spatial index, to generate the corresponding optimal path segment using a genetic algorithm for any two adjacent waypoints in the waypoint sequence.

[0129] The optimization module iteratively optimizes the optimal path segment, iteratively optimizes the waypoint sequence and the optimal path segment, and integrates the optimization results into the autonomous navigation system after convergence.

[0130] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0131] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0132] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0133] It should be understood that various aspects of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the aforementioned embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, implementation may be achieved using any combination of the following technologies known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application-specific integrated circuits having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), and the like. It should be noted that the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to be limiting. While the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will appreciate that modifications or equivalent substitutions may be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and such modifications are intended to be encompassed by the claims of the present invention.

[0134] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A route optimization method based on genetic algorithm and geographic information, characterized in that: include: Receive geographic information data related to route planning and build a multi-level spatial index; A waypoint optimization algorithm is used, supported by a spatial index, using a genetic algorithm based on population diversity adaptive adjustment to determine the waypoint sequence of the route. The waypoint optimization algorithm includes: setting a screening rule based on the airspace label set, risk level field, and mission rule label field stored in each index node in the coarse-grained spatial index structure, screening each index node, extracting the center coordinate point as a candidate waypoint, and establishing a candidate waypoint set based on the candidate waypoint; For any two adjacent waypoints in the waypoint sequence, a path optimization algorithm is used, combined with geographic information data provided by the spatial index, and a genetic algorithm is used to generate the corresponding optimal path segment; Iteratively optimizing the optimal path segment, iteratively optimizing the waypoint sequence and the optimal path segment, and integrating the optimization results into the autonomous navigation system after convergence; The multi-level spatial index includes a coarse-grained spatial index layer and a fine-grained spatial index layer; The geographic information data includes static data and dynamic data; wherein the static data includes regional terrain data, airspace management data and regional mission rule data, and the dynamic data includes static environment data and dynamic environment data; A coarse-grained spatial index layer is constructed based on the static class data. The coarse-grained index uses a two-dimensional square area as the basic unit, and the unit side length is set to L. The value of L is preset based on the area of ​​the mission area. The index is constructed using a quadtree structure. The index node stores: the unit coordinate range, the average terrain height of the area, the airspace label set, the mission rule label, the risk level field, and the unique spatial code identifier; A fine-grained spatial index layer is constructed based on the dynamic class data. The fine-grained index is divided into two-dimensional square spatial units with a side length of 1. The index is constructed using an R-tree structure. The fields stored in the index nodes include static fields and dynamic fields. The static fields include the coordinate range of the spatial unit, obstacle spatial boundary information, node status label, and spatial encoding information. The dynamic fields include environmental data field values ​​and their corresponding timestamp information. The path optimization algorithm includes constructing a path segment optimization task based on any two adjacent waypoints in the waypoint sequence; the path segment consists of a starting point, an end point and a number of waypoints; Based on the fine-grained spatial index layer, the field information in the spatial index node where the path point is located is called, and combined with the task time step, a binding relationship is established between the path point and the temporal environment data; An A* search algorithm is used to perform path search in the fine-grained spatial index layer, and a path cost function is constructed based on the accumulated environmental costs of the path points and the heuristic function to generate an initial path segment. Based on the initial path segment, a path perturbation mechanism is introduced to generate perturbed path segment individuals as an initial population of the genetic algorithm; During the evolution of the genetic algorithm, a fitness function is constructed based on the path segment length, the risk level field of the path node, the dynamic environment field value under the time step, and the path curvature. The path individuals generated in each generation are screened for validity using a space-time dual constraint mechanism to ensure that only path individuals that meet the environmental and spatial requirements can enter the next generation of population. When the path segment optimization meets the convergence conditions or reaches the maximum number of iterations, the corresponding optimal path segment is output, and the optimal results of all path segments are spliced ​​together to form a complete path set, which serves as the input for route quality assessment and subsequent iterative updates of the waypoint sequence.

2. The route optimization method based on genetic algorithm and geographic information according to claim 1, characterized in that: The screening rules include removing nodes with no-fly zones or restricted airspace identifiers in their airspace labels, and only retaining spatial index nodes with risk levels below a preset threshold and marked as passable in the mission rules; Using several waypoint indexes in the candidate waypoint set as chromosome genes, forming a chromosome in the order of arrangement of the waypoint indexes as search individuals of the genetic algorithm; During the execution of the genetic algorithm, a spatial perception control mechanism is set up to verify the spatial attributes of the waypoints generated in the mutation phase, restricting chromosomes that do not meet the risk conditions or airspace requirements from entering the next generation; The spatial perception control mechanism includes: when performing a mutation operation, calling a coarse-grained spatial index structure, retrieving the risk level field and airspace tag set of the spatial unit where the mutated waypoint is located, and determining whether the mutated waypoint is located in an area with a higher risk level than the original waypoint, or whether the airspace tag set contains a no-fly zone identifier or a restricted airspace identifier; If the conditions are not met, the mutation waypoint is removed and the mutation result is not included in the subsequent population; During the iterative process of the genetic algorithm, a fitness function is constructed based on the path distance, regional risk value and task fitness, and the waypoint sequence with the highest fitness score is determined as the optimal individual of the current generation, which serves as the input for subsequent path optimization.

3. The route optimization method based on genetic algorithm and geographic information according to claim 2, characterized in that: The path perturbation mechanism includes performing spatial perturbation operations on intermediate path points in a path segment within a limited perturbation radius; calling index nodes in the multi-level spatial index to determine whether the perturbed path points are located in obstacle areas, no-fly zones, restricted altitude areas, or high-risk areas, and only retaining perturbed path segment individuals that do not fall into the above areas to construct the initial population of the genetic algorithm.

4. The route optimization method based on genetic algorithm and geographic information according to claim 3, characterized in that: The space-time dual constraint mechanism includes a space constraint mechanism and a time constraint mechanism; The spatial constraint mechanism includes determining whether a path point falls into an obstacle area, a no-fly zone, a restricted altitude area, or a high-risk area based on the node's airspace management tag, obstacle information, and risk level field in the fine-grained spatial index; The time constraint mechanism includes judging whether the path node falls into an area where the environmental risk exceeds a threshold based on the dynamic environment field value of the task time step corresponding to the path node. If the pass requirements are not met, the path individual will not enter the next generation.

5. The route optimization method based on genetic algorithm and geographic information according to claim 4, characterized in that: The iterative optimization of the optimal path segments includes concatenating the optimal path segments to form a complete path set, and scoring the fitness of the entire route based on the total length of the complete path set, the risk level field of the path node, the dynamic environment field value at the mission time step, and the path curvature; According to the fitness score results, the path segments are adjusted through the genetic algorithm, and the optimized path segments are selected by combining the space-time dual constraint mechanism; Path optimization and fitness scoring are performed cyclically until the fitness score of the complete path set meets the convergence condition or reaches the maximum number of iterations, and the optimization result is output.

6. A route optimization system based on genetic algorithm and geographic information, which adopts the route optimization method based on genetic algorithm and geographic information according to any one of claims 1 to 5, characterized in that: The multi-layer module receives geographic data related to route planning and builds a multi-level spatial index; The sequence module uses a waypoint optimization algorithm and a genetic algorithm based on population diversity adaptive adjustment with the support of spatial index to determine the waypoint sequence of the route; A path segment module, for any two adjacent waypoints in the waypoint sequence, adopts a path optimization algorithm, combines the geographic information provided by the spatial index, and uses a genetic algorithm to generate the corresponding optimal path segment; The optimization module iteratively optimizes the optimal path segment, iteratively optimizes the waypoint sequence and the optimal path segment, and integrates the optimization results into the autonomous navigation system after convergence.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the route optimization method based on genetic algorithm and geographic information are implemented as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the route optimization method based on genetic algorithm and geographic information are implemented as described in any one of claims 1 to 5.

Images