A parking space complexity quantification evaluation method

Abstract

The application relates to a parking space complexity quantification evaluation method. The method firstly normalizes each index value to the [0-1] interval through a corresponding quantification method, and maps the index value to the evaluation set {not complex, slightly complex, complex, relatively complex} through a membership function, so as to obtain a fuzzy evaluation matrix. Secondly, the decisive index is regarded as a whole and the remaining indexes participate in the analytic hierarchy process to obtain a primary weight. Thirdly, the corresponding decisive index weight distribution is determined by using the square sum normalization of each decisive index, so that a main factor prominent type comprehensive weight is obtained. Finally, the evaluation result vector is obtained by solving the fuzzy evaluation matrix combined with the comprehensive weight, and the corresponding evaluation grade is determined according to the maximum membership degree principle. The application can provide more accurate guidance for subsequent researches such as intelligent parking lot parking space allocation, automatic valet parking space selection and automatic parking system testing.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic parking, in particular to a parking space complexity quantitative evaluation method. BACKGROUND

[0002] The parking space environment complexity quantitative evaluation gives the complexity level of the target parking space by analyzing various information of the target parking space and the environment, and can be used for guiding parking space allocation in intelligent parking lot, parking space selection in automatic valet parking, and parking scene evaluation in automatic parking system test.

[0003] In the automatic parking system test, the mainstream method of parking space complexity evaluation is to extract real data, simulation data and experience data to obtain automatic parking test scene elements, and then combine different types of environmental elements to control the environmental complexity of the test scene. This method, which classifies the complexity of environmental elements through fuzzy statements and uses artificial design and prior knowledge to obtain environmental complexity rating, cannot accurately quantify the complexity of the parking space environment, which is not conducive to guiding the improvement of the automatic parking test system.

[0004] Establishing a comprehensive evaluation model needs to go through three steps of index selection, weight assignment and comprehensive evaluation. Most of the current models use objective weight assignment methods such as entropy weight method, coefficient of variation method and CRITIC method, which are mostly based on statistics and have a large degree of mutual influence between samples. Moreover, they cannot be well applied to single sample evaluation, and the vehicle environment complexity evaluation is more oriented towards single parking space evaluation. Therefore, the present application proposes a weight secondary distribution method based on the objective data of the sample itself, which uses index square sum normalization to redistribute the weight, and can assign a larger weight to the index when it approaches the critical value. Finally, a parking space environment complexity quantitative evaluation model is established by combining the fuzzy comprehensive evaluation method. SUMMARY

[0005] The purpose of the present application is to provide a parking space complexity quantitative evaluation method, which can provide more accurate guidance for subsequent researches such as intelligent parking lot parking space allocation, automatic valet parking space selection and automatic parking system test.

[0006] To achieve the above-mentioned purpose, the technical solution of the present application is as follows: a parking space complexity quantitative evaluation method, comprising the following steps:

[0007] S1, comprehensively consider various common scenes such as ground and underground, parking lot and on-street parking, extract evaluation factors that can affect the complexity of the parking space environment, including seven types of parking space, parking space line attribute, parking space occupancy on both sides, parking space boundary curvature, parking lane width, parking space size, and parking space obstacle situation, and divide the parking lane width, parking space size and parking space obstacle situation into decisive indicators, and the other indicators are general indicators;

[0008] S2, to exclude the influence of data dimension, all indicators are normalized to the interval [0-1] according to their relative difficulty; for qualitative factors that are difficult to quantify, they are directly assigned according to their relative difficulty; the min-max method is used to quantify quantitative factors whose complexity changes linearly with the value; for quantitative factors whose complexity changes nonlinearly with the value, select a function with the same change trend to represent. i ' is the original value of the indicator, u i is the normalized value of the indicator, and the specific normalization method is as follows:

[0009] 1) Parking space type

[0010]

[0011] 2) Parking space line attribute

[0012]

[0013] 3) Parking space occupancy on both sides

[0014]

[0015] 4) According to the "Code for Design of Urban Road Engineering" (CJJ 37-2012), the minimum radius limit of road design in China is 20m, so the value range of road curvature [curv min , curv max ] is [-1 / 20, 1 / 20], and the quantification formula is

[0016]

[0017] 5) Let L W be the minimum channel width required for a vehicle to enter the garage once, which can be obtained by analyzing the circular-rectilinear path, and k be the scaling factor determined by the data value range, then

[0018]

[0019] 6) Let S be the minimum parking space size required, then

[0020]

[0021] 7) The obstacle indicator u7' is the distance from the center of the obstacle to the center of the parking space. To simplify the calculation, the obstacle is represented as a circle with the geometric center of the obstacle as the center and the maximum distance from the geometric center to the obstacle contour as the radius R, then

[0022]

[0023] S3, the decisive indicators are regarded as a whole and general indicators participate in analytic hierarchy process to solve initial weight, so that the decisive indicators as a whole obtain more than half weight, and then according to the square sum of the normalized value of each decisive indicator, that is, the relative difficulty between each decisive indicator, the corresponding decisive indicator weight is determined, and the main factor prominent weight is obtained; the weight solving steps include:

[0024] 1) the judgment matrix B=(b ij ) z×z , z=5; wherein: b ij is the relative importance of index i relative to index j, which is valued according to table 1; z is the number of input indexes;

[0025] Table 1 judgment matrix importance grade scale

[0026]

[0027] 2) solve the maximum eigenvalue λ max of the judgment matrix B and the corresponding eigenvector T

[0028] 3) judgment matrix consistency check:

[0029] ① calculate the consistency index CI:

[0030]

[0031] ② calculate the consistency ratio CR, and the average random consistency index RI can be obtained by referring to table 2:

[0032]

[0033] Table 2 average random consistency value

[0034]

[0035] ③ if CR<0.1, it meets the consistency requirement, otherwise the judgment matrix should be rebuilt

[0036] 4) after normalization of T, the primary weight W'=(w1, w2, w3, w4, w d )

[0037]

[0038] 5) let u k represent the normalized value of the kth index of the parking space, then the weight of the kth index is

[0039]

[0040] The main factor prominent weight W=(w1, w2, w3, w4, w5, w6, w7) is obtained

[0041] S4, determining the index evaluation grade division criterion f mn That is, the nth index in the mth evaluation grade is defined as the value, and the membership function is constructed accordingly; each index data is fuzzy mapped to the comment set {not complex F1, slightly complex F2, complex F3, and relatively complex F4} through the membership function, and a sample fuzzy evaluation matrix P is obtained; the membership function of each grade is F 1n , F 2n , F 3n , F 4n , and a semi-ladder and triangle combination form is adopted

[0042]

[0043]

[0044]

[0045] S5, the sample fuzzy evaluation matrix is combined with the obtained main factor prominent weight to obtain an evaluation result vector Q, which contains the membership degree of the evaluation sample in each comment set, and the sample complexity rating is determined according to the maximum membership degree principle.

[0046] Q=W·P.

[0047] Compared with the prior art, the present application has the following beneficial effects: the present application can provide more accurate guidance for subsequent researches such as intelligent parking lot space allocation, automatic guest parking space selection, and automatic parking system testing. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 It is a parking space environment complexity quantification evaluation flowchart.

[0049] Figure 2 It is a simulation experiment scene diagram.

[0050] Figure 3 It is a parking space d index weight diagram.

[0051] Figure 4 It is a simulation evaluation result diagram. DETAILED DESCRIPTION

[0052] The technical solutions of the present application will be specifically described below with reference to the accompanying drawings.

[0053] The present application provides a parking space complexity quantification evaluation method, comprising the following steps:

[0054] S1, extract the evaluation factors that can affect the parking space environment complexity shared by the parking scene, including seven kinds of parking space types, parking space line attributes, parking space both sides occupancy, parking space boundary curvature, parking lane width, parking space size, and parking space obstacle conditions, and divide the indexes into general indexes and decisive indexes;

[0055] S2, to exclude the influence of data dimension, all indexes are normalized to the [0-1] interval according to their relative difficulty; for qualitative factors that are difficult to quantify, directly assign values according to their relative difficulty; the min-max method is used to quantify the quantitative factors whose complexity changes linearly with the value; for quantitative factors whose complexity changes nonlinearly with the value, select a function with the same change trend to represent them;

[0056] S3, the decisive indexes are considered as a whole and participate in the analytic hierarchy process to solve the initial weight, so that the decisive indexes as a whole obtain more than half of the weight, and then the corresponding decisive index weight is determined according to the square sum normalization value of the normalized value of each decisive index, that is, the relative difficulty between each decisive index, to obtain the main factor prominent weight;

[0057] S4, determine the index evaluation grade division criteria, and construct the membership function based on the criteria; the index data is respectively fuzzy mapped to the comment set {not complex F1, slightly complex F2, complex F3, and relatively complex F4} through the membership function, to obtain the sample fuzzy evaluation matrix P;

[0058] S5, the sample fuzzy evaluation matrix combined with the main factor prominent weight can obtain the evaluation result vector Q, which contains the membership degree of the evaluation sample in each comment set, and the sample complexity rating is determined according to the maximum membership degree principle.

[0059] The following is a specific embodiment of the present application.

[0060] Appendix Figure 1 The parking space environment complexity quantification evaluation flowchart: first, normalize the index values to the [0-1] interval through the corresponding quantification method, and fuzzy map to the comment set {not complex, slightly complex, complex, and relatively complex} through the membership function, to obtain the fuzzy evaluation matrix. Secondly, the decisive indexes are considered as a whole and participate in the analytic hierarchy process together with the remaining indexes to solve the primary weight. Thirdly, the corresponding index weight distribution is determined by using the square sum normalization of each decisive index, to obtain the main factor prominent comprehensive weight. Finally, the evaluation result vector is obtained by solving the fuzzy evaluation matrix combined with the comprehensive weight, and the corresponding evaluation grade is determined according to the maximum membership degree principle.

[0061] Appendix Figure 2For simulation experiment scene graph: based on MATLAB software to establish simulation test environment, simulate a group of parking spaces. Among them, parking space a to parking space c are designed to increase in difficulty, to verify whether the invention can correctly reflect the parking environment complexity, to ensure the complexity increases, the more difficult parking space should add complex elements on the basis of not changing the control part of the easier parking space; Parking space d is designed as a parking space with critical decisive factor (parking space size), which is evaluated by using the invention, to verify whether the invention can give greater weight to the decisive index when it approaches the critical value to ensure the accuracy of the evaluation result. Among them: (1) According to “Urban Road Engineering Design Specification” (CJJ37-2012), the minimum motor lane width in China is 3.25m, and the parking channel width is 6.5m. (2) The parking space size with parking line is set to 6m x 2.5m according to “Urban Road In-parking Space Setting Specification” (GA / T 850-2021). (3) Only right-hand driving is considered, and the vehicle to be parked enters the garage in reverse.

[0062] Appendix Figure 3 For parking space d index weight graph: when the parking space size index approaches the critical value, the invention gives it a greater weight. For the obstacle index, there are no obstacles around parking space d, which is an invalid index, and the invention does not give it weight.

[0063] Appendix Figure 4 For simulation evaluation result graph: the evaluation results of each parking space and the corresponding membership are (1) Parking space a: not complex (0.6011) (2) Parking space b: slightly complex (0.5685) (3) Parking space c: complex (0.3811) (4) Parking space d: more complex (0.5 = 4872). The evaluation result conforms to the trend of increasing environmental complexity.

[0064] The above is the preferred embodiment of the invention, any changes made according to the technical solutions of the invention, as long as the function generated does not exceed the scope of the technical solutions of the invention, belongs to the protection scope of the invention.

Claims

1. A method for quantitatively evaluating a parking space complexity, characterized by, Comprising the following steps: S1, extract the evaluation factors that can affect the parking space environment complexity shared by parking scenarios, including seven kinds of parking space type, parking space line attribute, parking space both sides occupancy, parking space boundary curvature, parking lane width, parking space size, parking space obstacle situation, and divide the indexes into general indexes and decisive indexes, and divide the parking lane width, parking space size, parking space obstacle situation into decisive indexes, and the other indexes are general indexes; S2, in order to exclude the influence of data dimension, all indexes are normalized to the [0-1] interval according to their relative difficulty; for qualitative factors that are difficult to quantify, directly assign values according to their relative difficulty; use the min-max method to quantify the quantitative factors whose complexity changes linearly with the value; for quantitative factors whose complexity changes nonlinearly with the value, select a function with the same change trend to represent them; S3, consider the decisive indexes as a whole and the general indexes to participate in the analytic hierarchy process to solve the initial weight, so that the decisive indexes as a whole obtain more than half of the weight, and then determine the corresponding decisive index weight according to the square sum of the normalized values of each decisive index, that is, the relative difficulty between each decisive index, to obtain the main factor prominent type weight; S4, determine the index evaluation grade division criteria, and construct the membership function based on this; Each index data is fuzzy mapped to the evaluation set {not complex F1, slightly complex F2, complex F3, relatively complex F4} respectively through the membership function, and a sample fuzzy evaluation matrix P is obtained; S5, the sample fuzzy evaluation matrix combined with the main factor prominent type weight can obtain an evaluation result vector Q, which contains the membership degree of the evaluation sample in each evaluation set, and the sample complexity rating is determined according to the maximum membership degree principle.

2. The method of claim 1, wherein, In step S2, let u i be the original value of the index, and u i be the normalized value of the index. The specific normalization method is as follows: 1) parking space type 2) parking space line attribute 3) parking space both sides occupancy 4) The range of the curvature value [curv min ,curv max ] of the parking space boundary is [-1 / 20, 1 / 20], and the quantization formula is 5) Parking passage width: set L W The minimum passage width required for one-time parking of a vehicle is obtained by circular-rectilinear path analysis, k is a scaling factor, which is determined by the data value range, then 6) parking space size: let S be the minimum parking space size required for parking, then 7) parking space obstacle situation: the obstacle index u7' is the distance from the center of the obstacle to the center of the parking space. To simplify the calculation, the obstacle is represented as a circle with the geometric center of the obstacle as the center and the maximum distance from the geometric center to the obstacle contour as the radius R, then 3. The method of claim 1, wherein the complexity of the parking space is quantified by a value of 0 to 1. Step S3 is implemented as follows: 1) Take 3 decisive indicators as a whole indicator, and construct judgment matrix B=(b ij ) z×z , z=5; wherein: b ij is the relative importance of indicator i relative to indicator j; z is the number of input indicators; 2) Solving the maximum eigenvalue λ of the judgment matrix B and its corresponding eigenvector T max ; 3) judgment matrix consistency test: ①Calculate the consistency index CI: ②Calculate the consistency ratio CR, and find the average random consistency index RI from the table: ③If CR<0.1, it meets the consistency requirement, otherwise, reconstruct the judgment matrix; 4) After T normalization, the primary weights W' = (w1, w2, w3, w4, w d ) w d i.e. the overall index of the 3 decisive indices combined; 5) Let u k The normalized value of the kth indicator representing the parking space is denoted by uk, and the weight of the kth indicator is The main factor prominent type weight W=(w1, w2, w3, w4, w5, w6, w7) is obtained.

4. The method of claim 3, wherein the complexity of the parking space is quantified by a value of 0 to 1. In step S4, the index evaluation grade division criterion f mn is the defined value of the nth index in the mth evaluation grade; the membership function represents the mapping relationship of each index to each comment set, respectively F 1n , F 2n , F 3n , F 4n , adopts the form of semi-ladder combined with triangle, as follows:

5. The method of claim 1, wherein the complexity of the parking space is quantified by a value of 0 to 1. In step S5, the expression of the evaluation result vector Q is as follows: Q=W·P where W is the main factor prominent type weight.

Images