Path planning method and device, electronic equipment, storage medium and program product
By performing dimension supplementation on the original solution information and constructing a high-dimensional relation space, and using a chaotic game optimization algorithm to determine the target solution information, the problem of relation space degradation in path planning is solved, and more accurate path planning is achieved.
Patent Information
- Application Number
- CN202410866633.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-28
- Publication Date
- 2025-12-30
AI Technical Summary
In path planning, the constructed relation space is prone to degenerate into one dimension, which means that important features of the target solution information cannot be preserved to the maximum extent, resulting in inaccurate path planning.
By performing dimension supplementation on the original solution information, a high-dimensional relation space is constructed. The target solution information is determined using a chaotic game optimization algorithm, and relation hyperedges are constructed for path planning.
This effectively avoids the degradation of the relation space dimension to one dimension, preserves the important features of the solution information to the maximum extent, and improves the accuracy of path planning.
Smart Images
Figure CN121235233A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of data processing technology, and in particular to a path planning method, apparatus, electronic device, storage medium, and program product. Background Technology
[0002] When performing path planning, a relation space needs to be constructed to determine the target location corresponding to the target solution information, thereby determining the path between the target location and the target location. However, the constructed relation space may degenerate to one dimension, failing to retain the important features of the target solution information to the maximum extent, resulting in inaccurate planned paths. Summary of the Invention
[0003] In view of this, the purpose of this disclosure is to provide a path planning method, apparatus, electronic device, storage medium, and program product to solve or partially solve the above-mentioned technical problems.
[0004] To achieve the above objectives, the first aspect of this disclosure proposes a path planning method, the method comprising:
[0005] Determine the first position;
[0006] Based on the first position, determine the original solution information corresponding to the next position of the first position;
[0007] Determine the target solution information based on the original solution information;
[0008] Path planning is performed based on the target position corresponding to the first position and the target solution information;
[0009] The step of determining the target solution information based on the original solution information includes:
[0010] The dimensions of the original solution information are supplemented to obtain the dimension-supplemented solution information, wherein the dimension of the dimension-supplemented solution information is higher than that of the original solution information.
[0011] A high-dimensional relation space is constructed based on the solution information after dimension supplementation.
[0012] Based on the high-dimensional relation space, construct the relation hyperedges corresponding to the original solution information;
[0013] Based on the aforementioned relational hyperedges, the target solution information is determined.
[0014] Based on the same inventive concept, a second aspect of this disclosure proposes a path planning device, comprising:
[0015] The first position determination module is configured to determine the first position;
[0016] The original solution information determination module is configured to determine the original solution information corresponding to the next position of the first position based on the first position;
[0017] The target solution information determination module is configured to determine the target solution information based on the original solution information;
[0018] The path planning module is configured to perform path planning based on the target position corresponding to the first position and the target solution information;
[0019] The target solution information determination module includes:
[0020] The dimension-complementing processing unit is configured to perform dimension-complementing processing on the original solution information to obtain dimension-complemented solution information, wherein the dimension of the dimension-complemented solution information is higher than the dimension of the original solution information.
[0021] The high-dimensional relation space construction unit is configured to construct a high-dimensional relation space based on the dimension-complemented solution information.
[0022] The relation hyperedge construction unit is configured to construct the relation hyperedge corresponding to the original solution information based on the high-dimensional relation space;
[0023] The target solution information determination unit is configured to determine the target solution information based on the relational hyperedge.
[0024] Based on the same inventive concept, a third aspect of this disclosure proposes an electronic device including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the method described above when executing the computer program.
[0025] Based on the same inventive concept, a fourth aspect of this disclosure provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to perform the methods described above.
[0026] Based on the same inventive concept, a fifth aspect of this disclosure provides a computer program product comprising computer program instructions that, when executed on a computer, cause the computer to perform the method described above.
[0027] As can be seen from the above, the path planning method, apparatus, electronic device, storage medium, and program product provided in this disclosure, by performing dimension supplementation processing on the original solution information, makes the dimension of the obtained dimension-supplemented solution information higher than the high dimension of the original solution information. This ensures that the constructed high-dimensional relation space remains high-dimensional, avoiding the situation where the dimension of the relation space degenerates to one dimension. It can retain the important features of the solution information to the greatest extent, making the planned path more accurate. Attached Figure Description
[0028] To more clearly illustrate the technical solutions in this disclosure or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0029] Figure 1 This is a flowchart of the path planning method according to an embodiment of the present disclosure;
[0030] Figure 2 This is a schematic diagram of the theoretical architecture of the path planning algorithm in an embodiment of this disclosure;
[0031] Figure 3 This is a schematic diagram of the path planning device according to an embodiment of the present disclosure;
[0032] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation
[0033] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.
[0034] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0035] As mentioned above, how to avoid inaccurate path planning has become an important research question.
[0036] Based on the above description, such as Figure 1 As shown, the path planning method proposed in this embodiment includes:
[0037] Step 101: Determine the first position.
[0038] Step 102: Based on the first position, determine the original solution information corresponding to the next position of the first position.
[0039] Step 103: Determine the target solution information based on the original solution information.
[0040] Step 104: Perform path planning based on the target position corresponding to the first position and the target solution information.
[0041] In practice, the original solution information consists of information corresponding to multiple next positions of the first position. Optionally, a Chaotic Game Optimization (CGO) algorithm can be used to determine the target solution information based on the original solution information. The target solution information can be the optimal candidate solution among the candidate solutions corresponding to the original solution information. In this way, the target position corresponding to the target solution information is the optimal position among the next positions of the first position, ensuring that the path between the first position and the target position obtained through path planning is the optimal path.
[0042] For example, the next positions after the first position include the second, third, and fourth positions. The target solution information is determined based on the original solution information corresponding to the next position. If the target position corresponding to the target solution information is the third position, then the path between the first and third positions is considered the optimal path.
[0043] Step 103 includes:
[0044] Step 1031: Perform dimension complementation on the original solution information to obtain dimension-complemented solution information, wherein the dimension of the dimension-complemented solution information is higher than that of the original solution information.
[0045] Step 1032: Construct a high-dimensional relation space based on the solution information after dimension supplementation.
[0046] Step 1033: Based on the high-dimensional relation space, construct the relation hyperedge corresponding to the original solution information.
[0047] Step 1034: Determine the target solution information based on the aforementioned relational hyperedges.
[0048] In practice, the dimensions of the original solution information are supplemented to make the dimensions of the supplemented solution information higher than those of the original solution information. A high-dimensional relation space is constructed based on the supplemented solution information, ensuring that the constructed high-dimensional relation space remains high-dimensional. Based on the constructed high-dimensional relation space, relation hyperedges corresponding to the original solution information are constructed.
[0049] The target solution information is determined based on the relational hyperedges. Specifically, the distance data between each relational hyperedge is determined; based on the distance data, candidate solution information for each relational hyperedge is determined; and the target solution information is determined from the candidate solution information.
[0050] Through the above embodiments, by performing dimension supplementation processing on the original solution information, the dimension of the obtained dimension-supplemented solution information is higher than that of the original solution information. This ensures that the constructed high-dimensional relation space remains high-dimensional, avoiding the situation where the dimension of the relation space degenerates to one dimension. It can retain the important features of the solution information to the greatest extent, making the planned path more accurate.
[0051] In some embodiments, prior to step 1031, the method further includes:
[0052] Step 1031A: Determine the vector dimension and tuple dimension of the original solution information.
[0053] In practice, the original solution information includes: the global optimal solution vector, the mean solution vector, and the original solution vector. The length of any vector in the original solution information is obtained, and this length is used as the vector dimension d of the original solution information. o For example, obtain the length of the original solution vector and use that length as the vector dimension of the original solution information.
[0054] The solution vectors in the original solution information can be represented as tuples, where a tuple refers to a solution tuple, and the decision variables in the tuple collectively define a point in the relation space. Therefore, the tuple dimension d of the original solution information... ot It can be the number of variables in the tuple.
[0055] The vector dimension d of the original solution information o and the tuple dimension d ot Used for data restoration after completing relational calculations and decisions.
[0056] By using the above scheme, and by determining the vector dimension and original group dimension of the original solution information, the data can be restored after the relational calculation decision is completed.
[0057] In some embodiments, step 1031 includes:
[0058] Step 10311: In response to determining that the dimension of the original solution information is greater than or equal to a preset number of dimensions, the statistical relation dimension is determined based on the tuples in the original solution information.
[0059] Step 10312: Determine the first distance from the tuple to the starting point and the second distance from the tuple to the ending point, and determine the distance dimension based on the first distance and the second distance.
[0060] Step 10313: Determine the solution information after dimension complementation based on the statistical relationship dimension and the distance dimension.
[0061] In practice, the dimensions of the original solution information are obtained; it is then determined whether the dimensions of the original solution information are greater than or equal to a preset number of dimensions; if the dimensions of the original solution information are greater than or equal to the preset number of dimensions, dimension supplementation processing is performed on the original solution information. For example, the preset number of dimensions is 2. When the dimensions of the original solution vector are greater than or equal to 2, dimension supplementation processing is performed on the original solution information.
[0062] Specifically, the statistical relation dimensions are determined based on the tuples in the original solution information. These statistical relation dimensions include the statistical relation dimensions between numbers in a tuple, the statistical relation dimensions between numbers and relations, the statistical relation dimensions between relations, and the statistical relation dimensions between tuples.
[0063] Specifically, the distance dimension is determined based on the tuples in the original solution information. This distance dimension reflects the distance from the tuples in the original solution information to both the starting and ending points.
[0064] Furthermore, when the dimension of the original solution information is less than the preset number of dimensions, no dimension supplementation is performed on the original solution information. For example, if the preset number of dimensions is 2, and the dimension of the original solution information is 1, no dimension supplementation is performed on the original solution information.
[0065] The above scheme allows for dimension supplementation when the dimension of the original solution information is greater than or equal to the preset dimension number. This results in a higher dimension for the supplemented solution information, which in turn maintains the high dimension of the high-dimensional relation space constructed based on the supplemented solution information, thus preserving the important features of the solution information to the greatest extent possible.
[0066] In some embodiments, step 10311 includes:
[0067] Step 103111: Obtain the granularity of the relation space.
[0068] Step 103112: In response to determining the fineness of the relation space as a first fineness, determine the first statistical relation dimension between the numbers in the tuple.
[0069] Step 103113: In response to determining the relation space granularity as the second granularity, determine the first statistical relation dimension between numbers in the tuple, the second statistical relation dimension between numbers and relations, the third statistical relation dimension between relations, and the fourth statistical relation dimension between tuples.
[0070] The second level of precision is more precise than the first level of precision.
[0071] In practice, the granularity of the acquired relation space is the level of granularity preset by the user. For example, relation space granularity includes: low granularity (normal mode) and high granularity (deep mode).
[0072] When the relation space granularity is set to the first granularity (low granularity or normal mode), the first statistical relation dimension between numbers in the tuples of the original solution information is determined based on preset statistical relations. The preset statistical relations can be user-defined and include at least one of the following: mean, variance, and expectation.
[0073] For example, the original solution information is a pair (2, 3). The average of the numbers 2 and 3 in the pair, 2.5, is taken as the first statistical relation dimension, and the average of 2.5 is the third dimension other than the numbers 2 and 3 in the pair.
[0074] When the relation space granularity is the second granularity (high granularity or deep mode), the determined statistical relation dimensions, in addition to the first statistical relation dimension between numbers in tuples under the first granularity, also include at least one of the following: the second statistical relation dimension between numbers in tuples and relations, the third statistical relation dimension between relations, and the fourth statistical relation dimension between tuples. Here, "relation" refers to a pre-defined statistical relation.
[0075] The second statistical relation dimension between the numbers in the tuples of the original solution information and the preset statistical relation is determined based on the preset statistical relation. For example, if the tuple of the original solution information is a pair (2, 3), and the first statistical relation dimension between the numbers in the pair is the average of 2 (numbers) and 3 (numbers) in the pair, which is 2.5, then the second statistical relation dimension between the numbers and the relation can be the average of 2 (numbers) and the average of 2.5 (relation) in the pair (i.e., the average value). ).
[0076] The third statistical relation dimension between the preset statistical relations in the tuples of the original solution information is determined based on the preset statistical relations. For example, if the tuple of the original solution information is a binary tuple (2, 3), the third statistical relation dimension between the relations can be the average of the sum of 2 and 3 in the binary tuple (relation) and the average of the mean of 2 and 3 in the binary tuple (relation) (i.e., the average value). ).
[0077] The fourth statistical relation dimension between at least two tuples in the original solution information is determined based on a predefined statistical relation. For example, if the tuples in the original solution information are tuples (2, 3) and tuple (4, 4), the fourth statistical relation dimension between the tuples can be the average of the mean of tuple (2, 3) and the mean of tuple (4, 4) (i.e., the average value). ).
[0078] The above scheme involves performing dimension supplementation on the original solution information to determine the statistical relation dimension of the original solution information, making the dimension of the dimension-supplemented solution information higher than that of the original solution information. In this way, the dimension of the high-dimensional relation space constructed based on the dimension-supplemented solution information remains high-dimensional.
[0079] For example, when the tuples of the original solution information are binary tuples and the dimension of the original solution information is two-dimensional, performing dimension-addition processing on the original solution information at the first level of refinement determines the first statistical relation dimension as the third dimension. In this way, the dimension of the dimension-addition solution information is three-dimensional, and the high-dimensional relation space constructed based on the dimension-addition solution information is also three-dimensional, thus avoiding the situation where the relation space degenerates to one dimension.
[0080] In some embodiments, step 10312 includes:
[0081] Step 103121: Summate the first distance and the second distance to obtain the distance dimension.
[0082] In practice, when performing dimension supplementation processing on the original solution information, regardless of whether the relation space refinement is the first refinement or the second refinement, the distance dimension needs to be determined.
[0083] Specifically, the first distance (including penalty) from the tuples of the original solution information to the starting point is determined, and the second distance (including penalty) from the tuples of the original solution information to the ending point is determined. The distance dimension d is obtained by summing the first and second distances. dist .
[0084] The above scheme involves performing dimension supplementation on the original solution information to determine its distance dimension, which facilitates the subsequent determination of candidate solution information based on the constructed relation space and distance dimension.
[0085] In some embodiments, the original solution information includes: a global optimal solution vector, a mean solution vector, and an original solution vector; step 1032 includes:
[0086] Step 10321: Each tuple in the global optimal solution vector, the mean solution vector, and the original solution vector is augmented according to the augmented solution information to obtain augmented tuples.
[0087] Step 10322: Use the tuples with the added dimensions as points in the high-dimensional relation space, and construct the high-dimensional relation space based on the points in the high-dimensional relation space.
[0088] In practice, for each tuple in each type of solution vector among the global optimal solution vector, the mean solution vector, and the original solution vector, the new tuple in the dimension-added solution information is taken as a point in the high-dimensional relation space. The space where all the points of the dimension-added new tuples are located is the high-dimensional relation space itself.
[0089] The above scheme involves performing dimension supplementation on the original solution information when its dimension is greater than or equal to a preset number. This results in a dimension-supplemented solution information, which is then used to construct a high-dimensional relation space. This ensures that the constructed high-dimensional relation space maintains the same dimension as the dimension-supplemented solution information, preventing it from degenerating to one dimension. This approach maximizes the preservation of key features of the solution information, leading to more accurate path planning.
[0090] For example, if the preset dimension is 2, then when the dimension of the original solution information is greater than or equal to 2, the dimension of the original solution information is supplemented, resulting in a dimension of greater than or equal to 3. The dimension of the high-dimensional relation space constructed based on the supplemented solution information is also greater than or equal to 3. This ensures that the constructed high-dimensional relation space maintains at least three dimensions, preventing the relation space from degenerating to one dimension and preserving the important features of the candidate solution information to the greatest extent.
[0091] In some embodiments, step 1033 includes:
[0092] Step 10331: Determine the category to which the original solution information belongs.
[0093] Step 10332: Construct relational superedges and relational sub-superedges using tuples of the same category according to the category to which the original solution information belongs.
[0094] Wherein, the relational sub-hyperedge is a subset of the relational hyperedges.
[0095] In practice, in the constructed high-dimensional relation space, tuples belonging to the same category are grouped together to form a relation hyperedge according to the category to which the original solution information belongs. The points in the relation hyperedge together constitute a whole and are related to each other.
[0096] For each tuple belonging to the original solution vector, construct relational sub-hyperedges in the relational hyperedges of the overall original solution vector according to the tuples they belong to in the original solution vector.
[0097] The above scheme constructs relational hyperedges and relational sub-hyperedges in a high-dimensional relational space, making the constructed relational hyperedges and relational sub-hyperedges more accurate, thereby enabling a more accurate determination of the candidate solution information for each relational hyperedge.
[0098] In some embodiments, step 1034 includes:
[0099] Step 10341: Determine the distance data between each relational hyperedge.
[0100] Step 10342: Based on the distance data, determine the candidate solution information for each relational hyperedge.
[0101] Step 10343: Determine the target solution information from the candidate solution information.
[0102] In practice, the distance between each relation hyperedge is determined by evaluating its affinity. Based on this distance, candidate solutions for each relation hyperedge are identified. The target solution is then obtained by weighting and evaluating these candidate solutions.
[0103] The target solution information is the optimal candidate solution among the candidate solution information. Therefore, the target position corresponding to the determined target solution information is the optimal position among the next positions after the first position. Thus, path planning is performed based on the first position and the target position to obtain the optimal path.
[0104] The above scheme determines candidate solution information based on the distance data between each relational hyperedge. This ensures that the candidate solution information incorporates the distances from the tuples in the original solution information to the starting and ending points, making the determination of candidate solution information more accurate and enabling a precise determination of the target solution information. The path obtained by path planning based on the target position corresponding to the first position and the target solution information is the optimal path.
[0105] In some embodiments, step 10341 includes:
[0106] Step 103411: Determine the distance data between each relational hyperedge based on the tuples in each relational hyperedge and the solution information after dimension supplementation.
[0107] Step 103412: Determine whether there are relation sub-overedges in the relation superedge.
[0108] Step 103413: In response to determining that there are relational sub-hyperedges in the relational hyperedge, the distance data between each relational sub-hyperedge is determined based on the tuples in each relational sub-hyperedge and the solution information after dimension supplementation.
[0109] In practice, when there are no relational sub-hyperedges in a relational hyperedge, the distance data between tuples in each relational hyperedge is calculated by matching them one by one, and the distance data between each target tuple and other tuples in the relational hyperedge, the tuple with the smallest distance data, and the number of tuples are recorded. The distance data is the result of a weighted polynomial calculation consisting of the distance between elements in the tuples and the statistical relational dimension after each dimension supplementation. The weights can be preset by the user.
[0110] When a relational superedge contains relational sub-superedges, first match each tuple in the relational sub-superedge to calculate the distance between the tuples, thus obtaining the distance between each relational sub-superedge. Then, match each tuple in the relational superedge containing the relational sub-superedges to calculate the distance between the tuples, thus obtaining the distance between each relational superedge.
[0111] The above scheme determines the distance between each relation hyperedge by checking if there are relation sub-hyperedges within the relation hyperedges. If no relation sub-hyperedges exist, the distance between each relation hyperedge is determined directly. If relation sub-hyperedges exist, the distance between each relation hyperedge is determined after the distance between each relation sub-hyperedge is determined. This ensures accurate determination of the distance between each relation hyperedge in both cases.
[0112] In some embodiments, step 10342 includes:
[0113] Step 103421: The average value of the distance data in the relational hyperedge and the average value of the distance dimension of the solution information after dimension supplementation are used as the first ranking index.
[0114] Step 103422: The number of tuples whose distance dimension of the distance data and the solution information after dimension supplementation in the relation hyperedge is less than the first ranking index is used as the second ranking index.
[0115] Step 103423: Sort the tuples in the relational hyperedge according to the first sorting index and the second sorting index to obtain the sorting result, and use the sorting result as the candidate solution information of the relational hyperedge.
[0116] In practice, the affinity value of each relational hyperedge includes: a first ranking index and a second ranking index.
[0117] Specifically, the first ranking metric is the average of the sum of distance data for each tuple in the relation hyperedge and the distance dimension d of each tuple in the relation hyperedge. dist The effect of the average value. The effect refers to meaningful computational processing, and generally does not involve division or division-related operations.
[0118] Specifically, the second ranking metric is the sum of the distance data of each tuple in the relation hyperedge and the distance dimension d of each tuple in the relation hyperedge. dist The result of this is less than the number of tuples in the first ranking index.
[0119] The tuples in the relational hyperedge are sorted according to the first and second sorting indices to obtain the sorting result, which is the candidate solution information.
[0120] Furthermore, when the dimension of the original solution information is less than the preset dimension number, no dimension supplementation is performed on the original solution information, nor is a high-dimensional relation space or relation hyperedge constructed based on the dimension-supplemented solution information. However, a relation space (not a high-dimensional relation space) and relation hyperedges still exist. For cases where the dimension of the original solution information is less than the preset dimension number, the distance dimension d is directly used. dist The relation hyperedges are sorted, and the sorted results are the candidate solution information. For example, if the preset dimension is 2, and the original solution information has a dimension of 1, it is directly sorted according to the distance dimension d. dist The relational hyperedges are sorted, and the sorting result is the candidate solution information.
[0121] The above scheme sorts the tuples in the relational hyperedge according to the first and second sorting indices, making the sorting results more accurate and thus enabling a more accurate determination of candidate solution information.
[0122] In some embodiments, step 10343 includes:
[0123] Step 103431: According to the sorting results, determine a preset number of candidate solutions from the candidate solution information corresponding to the relation superedge as the first target solution information.
[0124] Step 103432: Determine whether there are relation sub-overedges in the relation superedge.
[0125] Step 103433: In response to the determination that there is a relation sub-superedge in the relation superedge, a preset number of candidate solutions are determined from the candidate solution information corresponding to the relation sub-superedge as the second target solution information.
[0126] In practice, for each relation hyperedge and each relation sub-hyperedge, a preset number of target solution information is determined according to the sorting order of the sorting results. This preset number can be pre-defined by the user. Typically, the preset number is 1.
[0127] When a relational superedge contains a relational sub-superedge, a predetermined number of candidate solutions are first determined from the candidate solution information corresponding to the relational superedge as the first target solution information. Then, a predetermined number of candidate solutions are determined from the candidate solution information corresponding to the relational sub-superedge as the second target solution information. This process continues until all target solution information corresponding to the relational sub-superedges in the relational superedge is determined. The CGO algorithm is then used to evaluate relational superedges of different classes, and the relational superedge with the smallest evaluation value is taken as the target solution information.
[0128] Additionally, when the dimension of the original solution information is less than the preset dimension, it is based on the distance dimension d. distThe relational hyperedges are sorted, and a predetermined number of candidate solutions are directly determined as the target solution information. For example, if the predetermined dimension is 2, and the original solution information has a dimension of 1, it is directly sorted according to the distance dimension d. dist Sort the relational hyperedges and use the first tuple in the sorted order as the target solution information.
[0129] By sequentially removing the elements with added dimensions from the elements in the tuple of the target solution information, the dimensionality reduction operation from the high-dimensional relation space is achieved, and the tuple required for the actual application scenario of path planning is obtained.
[0130] By employing the above scheme, a predetermined number of target solutions are determined from the candidate solution information corresponding to the relational hyperedges according to the sorting results. This allows for the identification of the optimal target solution from multiple candidate solutions. Consequently, when performing path planning based on the target position corresponding to the first position and the target solution information, the planned path is the optimal path, improving the accuracy of path planning.
[0131] Through the above embodiments, by performing dimension supplementation processing on the original solution information, the dimension of the obtained dimension-supplemented solution information is higher than the high dimension of the original solution information. This ensures that the constructed high-dimensional relation space remains high-dimensional, avoiding the situation where the dimension of the relation space degenerates to one dimension. It can retain the important features of the solution information to the greatest extent, making the planned path more accurate.
[0132] It should be noted that the embodiments of this disclosure can also be further described in the following ways:
[0133] I. Path Planning Algorithm:
[0134] Figure 2 This is a schematic diagram of the theoretical architecture of the path planning algorithm according to an embodiment of this disclosure. Figure 2 As shown, the path planning algorithm specifically includes:
[0135] Step 1: Input necessary solution information.
[0136] Based on the methods provided in the CGO algorithm, obtain the original solution vector set, the average vector of the original solution vector set, and the optimal vector of the original solution vector set.
[0137] Step 2, determine the dimension of the solution information.
[0138] A. Obtain the length of any vector in the original solution information, which is the vector dimension of the original solution vector, and record the dimension d of the original solution vector. o and the dimension d of each tuple ot It is used to restore the data after completing the relational calculation and decision.
[0139] B. Users can first select the desired relational space granularity based on their expectations. Relational space granularity includes, but is not limited to, low granularity (normal mode) and high granularity (deep mode), or other equivalent descriptions. If users have special requirements, they can pre-set the desired relational space granularity level and the corresponding main processing logic.
[0140] C. Solution Information Dimension Filling. If the dimension of the original solution vector in the original solution information is ≥2, in normal mode, the statistical relation dimension only includes the first statistical relation dimension between the numbers in each N-tuple (N≥2) element of the original solution vector. For example, for the tuple (2, 3), the statistical relation dimension only includes various statistical relations (mean, variance, expectation, etc.) around 2 and 3, and the specific statistical relations can be preset by the user. Each statistical relation corresponds to a dimension. For example, the mean of the numbers 2 and 3 in the above tuple, 2.5, is the third dimension besides the numbers 2 and 3. In deep mode, in addition to the first statistical relation dimension between numbers that will be present in normal mode, the statistical relation dimension also includes, but is not limited to: the second statistical relation dimension between numbers and relations, the third statistical relation dimension between relations, and the fourth statistical relation dimension between tuples. For example, if the mean in the pair (2, 3) is 2.5, then the second statistical relationship dimension between a number and a relation could be the average of 2 (the number) and the mean 2.5 (the relation); the third statistical relationship dimension between relations could be the average of the sum of 2 and 3 in the pair (the relation) and the mean of 2 and 3 in the pair (the relation); the fourth statistical relationship dimension between tuples could be the expected value of the mean of the pair (2, 3) and the mean of the pair (4, 4), and so on. Similarly, specific statistical relationships can be preset by the user.
[0141] It is important to note that this step includes a special case where the original solution vector has a dimension of 1. When this case occurs, no padding operation will be performed, and steps 3 to 5 and step 8 will be skipped (because relation construction is meaningless in this case), and it will be treated as a special case in steps 6 and 7.
[0142] Furthermore, when performing dimension padding, regardless of the granularity of the relation space, the distance dimension d must be the sum of the first distance (including penalty) from the tuples of the original solution information to the starting point and the second distance (including penalty) from the tuples of the original solution information to the ending point. dist .
[0143] Step 3: Construct a high-dimensional relation space.
[0144] The Global Optimal Solution (CGO) contains three types of solution vector information: the global optimal solution vector, the mean solution vector, and the original solution vector. For each type of solution vector in the original solution information, each tuple is augmented with its own dimension, resulting in a new tuple that is considered a point in the high-dimensional relation space. The space containing all the augmented tuples is the high-dimensional relation space itself.
[0145] Step 4: Construct relational hyperedges.
[0146] In the constructed high-dimensional relation space, tuples belonging to the same category are grouped together to form a relation hyperedge according to the category of the original solution information (the points in the relation hyperedge together constitute a whole and are related to each other). For tuples belonging to the original solution vector, relation sub-hyperedges are constructed sequentially in the relation hyperedge of the overall original solution vector according to the tuples they belong to in the original solution vector.
[0147] Step 5: Evaluate the affinity value of each relation superedge.
[0148] Specifically, there are two cases: the relation superedge does not contain a relation sub-superedge, and the relation superedge contains a relation sub-superedge.
[0149] If there are no relational sub-superedges in the relational superedge, first match each tuple in the relational superedge to calculate the distance data between tuples (the distance data is the weighted polynomial calculation result composed of the distance between the elements in the tuples and the statistical relation dimension after each supplementary dimension, and the weights of each item can be preset by the user). Then record the distance data between each target tuple in the corresponding relational superedge and other tuples, the tuple with the smallest distance data, and the number of tuples.
[0150] For relational superedges containing relational sub-superedges, first perform the above calculation for each relational sub-superedge and record the corresponding results. Then, perform the above calculation again for relational superedges containing relational sub-superedges and record the corresponding results.
[0151] Step 6: Calculate candidate solutions for each relation hyperedge.
[0152] The affinity value of each relation hyperedge is calculated by the average of the distance data of each tuple within that relation hyperedge and the distance dimension d of each tuple within that relation hyperedge. dist The result of the average (any meaningful operation is acceptable, but division and related operations should be avoided as much as possible) is used as the first ranking criterion, and the distance data of each tuple and the distance dimension d corresponding to each tuple are used. dist The number of tuples whose result is less than the first ranking criterion is the second ranking criterion. The ranking result is the candidate solution information for each relation hyperedge.
[0153] It's important to note that this step includes a special case where the original solution information vector has a dimension of 1. In this case, the candidate solution information will be directly extracted from the original solution vector according to the distance dimension d.dist Each tuple is sorted, and the sorting result is the candidate solution information.
[0154] Step 7: Weighted decision based on candidate solution information.
[0155] For each relation superedge (including relation sub-superedges) candidate solution information, select a preset number of tuples according to the sorting order (generally select the first tuple in the sorting, or the user can preset the number of tuples to be selected in the order and then merge them).
[0156] When a relational hyperedge contains relational sub-hyperedges, the point pairs selected for each relational sub-hyperedge are further filtered according to the rules in the relational hyperedges containing the relational sub-hyperedges. This process is repeated until no outermost relational hyperedges exist. It is important to note that at this point, relational hyperedges of different categories should no longer contain any relational sub-hyperedges. Finally, the CGO method is used to evaluate the final relational hyperedges of each different category, and the relational hyperedge with the smallest evaluation value is the optimal candidate solution (i.e., the target solution information) for this round.
[0157] It should be noted that this step includes a special case: if the vector dimension of the original solution information is 1, then the first tuple in the sorting in step 6 is directly selected as the optimal candidate solution (i.e., the target solution information) for this round.
[0158] Step 8: Output the optimal candidate solution (i.e., the target solution information).
[0159] Remove the dimension information elements from the tuples of the optimal candidate solutions (i.e., target solution information) obtained in step 7 from the back to the front (refer to d when removing). ot This enables dimensionality reduction operations from a high-dimensional relational space, yielding the tuples required in practical application scenarios.
[0160] Furthermore, the external system global loop is defined by variables in the CGO algorithm. Therefore, users do not need to define variables; they only need to call the optimizer as needed.
[0161] II. Single Robot Path Optimization Scenarios:
[0162] Taking a 2D map as an example. Assuming the starting point is known, starting from that point (recorded in the path trajectory set), the CGO method pre-generates a set of initial solution information, i.e., the next step location information set from the starting point. This includes the average vector and optimal vector calculated from the next step location information set. Then, the obtained initial solution information, the average vector of the initial solution information, and the optimal vector of the initial solution information are input into an optimizer containing the aforementioned path planning algorithm to obtain the optimal candidate solution (i.e., the target solution information). The optimal candidate solution (i.e., the target solution information) is then selected as the target location for the next step and recorded in the path trajectory set. This iteration is repeated until the CGO main loop is completed, resulting in a complete path trajectory set, which contains a movement trajectory consisting of all necessary points traversed from the starting point. With a destination in mind, this movement trajectory is a path from the starting point to the destination.
[0163] III. Multi-robot collaborative path optimization scenarios:
[0164] In this scenario, each robot first plans its path using the method described in Scenario 2. The difference lies in the fact that after each robot obtains its planned path, it compares the resulting paths laterally, selects the best one, and broadcasts it, causing the paths of other robots with non-optimal routes to converge towards the optimal one. Specifically, there are two cases: the first case is that each robot is placed at a random position in the given scene (i.e., different starting points but the same ending point); the second case is that all robots start from the same point. Since, in either case, as long as the ending point is the same, all robots' planned paths will eventually converge to a single route. A slight difference is that in the second case, the robots always travel along almost the same route. This seems contradictory to the previous description, but in reality, due to the computational characteristics of the CGO algorithm, the optimal paths obtained by each robot during its individual path planning are usually not quite the same. However, because they start from the same point, after sharing the optimal path information, there is almost no fundamental deviation from the optimal path during the path convergence process. This is why, when the starting and ending points are the same, the movement trajectories of the robots starting simultaneously from the same point are basically the same.
[0165] There is also a slightly special case where some robots have certain prior information. These robots are designated as pioneers, and optimal path planning is performed using the method described above, guided by their prior knowledge. Once the optimal path is found, it is shared with all robots. The pioneer robots' paths begin to converge towards the optimal path, while the non-pioneer robots completely follow the optimal path (similar to ant colony movement).
[0166] IV. Other scenarios:
[0167] This optimizer, designed for the CGO algorithm, is applicable to all domains where the CGO algorithm can be used. In short, it is suitable for scenarios where the initial state, objective state, and constraints are known, and the optimal objective solution must be selected from multiple possible candidate solutions.
[0168] Through the above embodiments, the game optimizer can efficiently perform the decision-making function of finding the optimal solution in a high-dimensional relation space. This reduces the difficulty of extending the game optimizer to an N-dimensional relation space, simplifies the theoretical framework in high-dimensional relation spaces, improves computational efficiency in high-dimensional relation spaces, and reduces subsequent maintenance costs. It also ensures that the high-dimensional relation space remains at least three-dimensional, preventing the relation space from degenerating to one dimension, and maximizing the preservation of important features of candidate solution information.
[0169] It should be noted that the method of this disclosure embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this disclosure embodiment, and the multiple devices will interact with each other to complete the method described.
[0170] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0171] Based on the same inventive concept, corresponding to any of the above-described embodiments, this disclosure also provides a path planning device.
[0172] refer to Figure 3 The path planning device includes:
[0173] The first position determination module 301 is configured to determine the first position;
[0174] The original solution information determination module 302 is configured to determine the original solution information corresponding to the next position of the first position based on the first position;
[0175] The target solution information determination module 303 is configured to determine the target solution information based on the original solution information.
[0176] The path planning module 304 is configured to perform path planning based on the target position corresponding to the first position and the target solution information;
[0177] The target solution information determination module 303 includes:
[0178] The dimension-complementing processing unit is configured to perform dimension-complementing processing on the original solution information to obtain dimension-complemented solution information, wherein the dimension of the dimension-complemented solution information is higher than the dimension of the original solution information.
[0179] The high-dimensional relation space construction unit is configured to construct a high-dimensional relation space based on the dimension-complemented solution information.
[0180] The relation hyperedge construction unit is configured to construct the relation hyperedge corresponding to the original solution information based on the high-dimensional relation space;
[0181] The target solution information determination unit is configured to determine the target solution information based on the relational hyperedge.
[0182] In some embodiments, the target solution information determination module 303 further includes:
[0183] The dimension determination unit is configured to determine the vector dimension and tuple dimension of the original solution information.
[0184] In some embodiments, the dimension-addition processing unit includes:
[0185] The statistical relation dimension determination subunit is configured to determine the statistical relation dimension based on the tuples in the original solution information in response to determining that the dimension of the original solution information is greater than or equal to a preset number of dimensions.
[0186] The distance dimension determination subunit is configured to determine a first distance from the tuple to the starting point and a second distance from the tuple to the ending point, and to determine the distance dimension based on the first distance and the second distance.
[0187] The dimension-completion processing subunit is configured to determine the solution information after dimension completion based on the statistical relation dimension and the distance dimension.
[0188] In some embodiments, the statistical relation dimension determining sub-unit is specifically configured as follows:
[0189] Obtain the granularity of the relation space;
[0190] In response to determining the fineness of the relation space as a first fineness, a first statistical relation dimension between the numbers in the tuple is determined;
[0191] In response to determining the relation space granularity as the second granularity, the first statistical relation dimension between numbers in the tuple, the second statistical relation dimension between numbers and relations, the third statistical relation dimension between relations and relations, and the fourth statistical relation dimension between tuples are determined.
[0192] The second level of precision is more precise than the first level of precision.
[0193] In some embodiments, the distance dimension determining subunit is specifically configured as follows:
[0194] The distance dimension is obtained by summing the first distance and the second distance.
[0195] In some embodiments, the original solution information includes: the global optimal solution vector, the mean solution vector, and the original solution vector;
[0196] The high-dimensional relation space construction unit includes:
[0197] The dimension-addition processing subunit is configured to add dimension to each tuple in the global optimal solution vector, the mean solution vector, and the original solution vector according to the dimension-addition solution information to obtain dimension-addition tuples.
[0198] The high-dimensional relation space construction subunit is configured to use the dimension-complemented tuples as points in the high-dimensional relation space, and to construct the high-dimensional relation space based on the points in the high-dimensional relation space.
[0199] In some embodiments, the relational hyperedge construction unit includes:
[0200] The category determination subunit is configured to determine the category to which the original solution information belongs;
[0201] The relational hyperedge construction sub-unit is configured to construct relational hyperedges and relational sub-hyperedges using tuples of the same category according to the category to which the original solution information belongs;
[0202] Wherein, the relational sub-hyperedge is a subset of the relational hyperedges.
[0203] In some embodiments, the target solution information determining unit includes:
[0204] The distance data determination sub-unit is configured to determine the distance data between each relational hyperedge;
[0205] The candidate solution information determination subunit is configured to determine the candidate solution information for each relational hyperedge based on the distance data;
[0206] The target solution information determination subunit is configured to determine the target solution information from the candidate solution information.
[0207] In some embodiments, the distance data determination subunit is specifically configured as follows:
[0208] Based on the tuples in each relation hyperedge and the solution information after dimension supplementation, determine the distance data between each relation hyperedge;
[0209] Determine if a relational superedge contains a relational sub-superedge;
[0210] In response to the determination that there are relational sub-hyperedges in relational hyperedges, the distance data between each relational sub-hyperedge is determined based on the tuples in each relational sub-hyperedge and the solution information after dimension supplementation.
[0211] In some embodiments, the candidate solution information determining subunit is specifically configured as follows:
[0212] The average value of the distance data in the relational hyperedge and the average value of the distance dimension of the solution information after dimension supplementation are used as the first ranking index;
[0213] The number of tuples whose distance dimension of the distance data and the solution information after dimension supplementation in the relation hyperedge is less than the first ranking index is used as the second ranking index.
[0214] The tuples in the relational hyperedge are sorted according to the first sorting index and the second sorting index to obtain the sorting result, and the sorting result is used as the candidate solution information of the relational hyperedge.
[0215] In some embodiments, the target solution information determining subunit is specifically configured as follows:
[0216] Based on the sorting results, a preset number of candidate solutions are determined from the candidate solution information corresponding to the relation hyperedge as the first target solution information;
[0217] Determine if a relational superedge contains a relational sub-superedge;
[0218] In response to the existence of relation sub-superedges in the determined relation superedge, a preset number of candidate solutions are determined from the candidate solution information corresponding to the relation sub-superedges as the second target solution information.
[0219] For ease of description, the above apparatus is described in terms of its functions, divided into various modules. Of course, in implementing this disclosure, the functions of each module can be implemented in one or more software and / or hardware.
[0220] The apparatus of the above embodiments is used to implement the corresponding path planning method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0221] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the path planning method described in any of the above embodiments.
[0222] Figure 4 This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.
[0223] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.
[0224] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.
[0225] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.
[0226] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB (Universal Serial Bus), network cable, etc.) or wireless means (such as mobile network, WIFI (Wireless Fidelity), Bluetooth, etc.).
[0227] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.
[0228] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.
[0229] The electronic devices described above are used to implement the corresponding path planning methods in any of the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0230] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this disclosure also provides a non-transitory computer-readable storage medium that stores computer instructions for causing the computer to execute the path planning method as described in any of the above embodiments.
[0231] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.
[0232] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the path planning method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0233] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this disclosure also provides a computer program product, including computer program instructions, which, when run on a computer, cause the computer to execute the path planning method as described in any of the above embodiments.
[0234] The computer program instructions of the computer program product of the above embodiments are used to cause the computer to execute the path planning method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0235] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this disclosure (including the claims) is limited to these examples; within the framework of this disclosure, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this disclosure as described above, which are not provided in detail for the sake of brevity.
[0236] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this disclosure, the provided drawings may or may not show well-known power / ground connections to integrated circuit (IC) chips and other components. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this disclosure, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this disclosure will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuitry) have been set forth to describe exemplary embodiments of this disclosure, it will be apparent to those skilled in the art that the embodiments of this disclosure may be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.
[0237] Although this disclosure has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.
[0238] This disclosure is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.
Claims
1. A path planning method characterized by, The method comprises: determining a first position; determining, according to the first position, original solution information corresponding to a next position of the first position; determining target solution information according to the original solution information; performing path planning according to a target position corresponding to the first position and the target solution information; wherein the determining target solution information according to the original solution information comprises: dimension supplementing the original solution information to obtain supplemented solution information, the dimension of the supplemented solution information being higher than that of the original solution information; constructing a high-dimensional relationship space based on the supplemented solution information; constructing a relationship hyperedge corresponding to the original solution information based on the high-dimensional relationship space; determining target solution information according to the relationship hyperedge.
2. The method of claim 1, wherein, Before the dimension supplementing the original solution information to obtain supplemented solution information, the method further comprises: determining a vector dimension and a tuple dimension of the original solution information.
3. The method of claim 1, wherein, The dimension supplementing the original solution information to obtain supplemented solution information comprises: in response to determining that the dimension of the original solution information is greater than or equal to a preset dimension number, determining a statistical relationship dimension based on a tuple in the original solution information; determining a first distance from the tuple to a starting point position and a second distance from the tuple to an ending point position, and determining a distance dimension based on the first distance and the second distance; determining the supplemented solution information according to the statistical relationship dimension and the distance dimension.
4. The method of claim 3, wherein, The determining a statistical relationship dimension based on a tuple in the original solution information comprises: obtaining relationship space fineness; in response to determining that the relationship space fineness is a first fineness, determining a first statistical relationship dimension between numbers in the tuple; in response to determining that the relationship space fineness is a second fineness, determining a first statistical relationship dimension between numbers in the tuple, a second statistical relationship dimension between numbers and relationships, a third statistical relationship dimension between relationships, and a fourth statistical relationship dimension between tuples; wherein the second fineness is finer than the first fineness.
5. The method of claim 3, wherein, The determining a distance dimension based on the first distance and the second distance comprises: summing the first distance and the second distance to obtain the distance dimension.
6. The method of claim 1, wherein, The original solution information comprises a global optimal solution vector, a mean solution vector, and an original solution vector; The constructing a high-dimensional relationship space based on the supplemented solution information comprises: dimension supplementing each tuple in the global optimal solution vector, the mean solution vector, and the original solution vector according to the supplemented solution information to obtain supplemented tuples; treating the supplemented tuples as points in a high-dimensional relationship space, and constructing the high-dimensional relationship space based on the points in the high-dimensional relationship space.
7. The method of claim 1, wherein, The constructing a relationship hyperedge corresponding to the original solution information based on the high-dimensional relationship space comprises: determining a category to which the original solution information belongs; constructing a relationship hyperedge and a relationship sub-hyperedge using tuples of the same category according to the category to which the original solution information belongs; wherein the relationship sub-hyperedge is a subset of the relationship hyperedge.
8. The method of claim 1, wherein, The determining target solution information according to the relationship hyperedge comprises: determine distance data between each relationship hyperedge; determine candidate solution information of each relationship hyperedge according to the distance data; determine the target solution information from the candidate solution information.
9. The method of claim 8, wherein, The determining of the distance data between each relationship hyperedge comprises: determine distance data between each relationship hyperedge according to the tuples in each relationship hyperedge and the complemented solution information; determine whether there is a relationship sub-hyperedge in the relationship hyperedge; in response to determining that there is a relationship sub-hyperedge in the relationship hyperedge, determine distance data between each relationship sub-hyperedge according to the tuples in each relationship sub-hyperedge and the complemented solution information.
10. The method of claim 8, wherein, The determining of the candidate solution information of each relationship hyperedge according to the distance data comprises: take the average of the distance data in the relationship hyperedge and the average of the distance dimension of the complemented solution information as a first sorting index; take the number of tuples in the relationship hyperedge whose distance data and distance dimension of the complemented solution information are less than the first sorting index as a second sorting index; sort the tuples in the relationship hyperedge according to the first sorting index and the second sorting index to obtain a sorting result, and take the sorting result as the candidate solution information of the relationship hyperedge.
11. The method of claim 8, wherein, The determining of the target solution information from the candidate solution information comprises: determine a preset number of candidate solutions from the candidate solution information corresponding to the relationship hyperedge as first target solution information according to the sorting result; determine whether there is a relationship sub-hyperedge in the relationship hyperedge; in response to determining that there is a relationship sub-hyperedge in the relationship hyperedge, determine a preset number of candidate solutions from the candidate solution information corresponding to the relationship sub-hyperedge as second target solution information.
12. A path planning device characterized by comprising: comprise: a first position determining module configured to determine a first position; an original solution information determining module configured to determine original solution information corresponding to a next position of the first position according to the first position; a target solution information determining module configured to determine target solution information according to the original solution information; a path planning module configured to perform path planning according to the first position and a target position corresponding to the target solution information; wherein the target solution information determining module comprises: a complement dimension processing unit configured to perform complement dimension processing on a dimension of the original solution information to obtain complemented solution information, the dimension of the complemented solution information being higher than the dimension of the original solution information; a high-dimensional relationship space constructing unit configured to construct a high-dimensional relationship space based on the complemented solution information; a relationship hyperedge constructing unit configured to construct a relationship hyperedge corresponding to the original solution information based on the high-dimensional relationship space; a target solution information determining unit configured to determine target solution information according to the relationship hyperedge.
13. An electronic device, comprising: comprise a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method of any one of claims 1 to 11 when executing the program.
14. A non-transitory computer-readable storage medium, comprising: The non-transitory computer readable storage medium stores computer instructions for causing a computer to execute the method of any one of claims 1 to 11. The non-transitory computer readable storage medium stores computer instructions for causing a computer to execute the method of any one of claims 1 to 11.
15. A computer program product comprising computer program instructions, characterised in that, The computer program product comprises computer program instructions which, when run on a computer, cause the computer to perform the method according to any one of claims 1 to 11.