Multimodal personalized vehicle routing method based on brainstorming optimization

Abstract

A multimodal personalized vehicle routing method based on brainstorming optimization consists of quantifying user goal parameters, building a grid map, quantifying path-variable goal parameters, initializing a population, evaluating individuals, partitioning the set, updating individuals, updating the worst-case solution, and outputting a planning result. This method addresses the problem that existing technologies fail to fully consider the dynamic environment of road conditions and user needs, and that the acquisition of technical model parameters is complex, making it unsuitable for multimodal personalized navigation. Field experiments at six locations demonstrated the method's advantages, including simple parameter calculation, high efficiency, fast computation speed, and accurate predictions. It is suitable for widespread use in vehicles and mobile phones.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle navigation, and specifically relates to multimodal personalized vehicle path navigation. Background Art

[0002] Path planning is at the core of navigation research and a crucial component and topic. Path planning involves searching for an optimal or suboptimal path from a starting state to a target state based on factors such as distance, time, and energy consumption. Path planning primarily involves the following issues: using environmental information to establish a reasonable model and then finding an optimal or near-optimal path from the starting state to the target state; handling uncertainties in the environmental model and yaw during path tracking; and utilizing all known information to guide the robot's movements, resulting in relatively optimal behavioral decisions.

[0003] Existing route planning systems often display limited information during navigation, such as route mileage and the number of traffic lights, with minimal variation between routes. However, user requirements for a particular route vary widely. In addition to these requirements, they also include information such as the need to pass through one or more specific locations, driving time, driving comfort, and past driving habits. Users hope to balance multiple objectives while driving. Most current research focuses on minimizing route length, simultaneously planning multiple shortest routes while satisfying constraints. This is a single-objective optimization problem. However, a single objective often fails to meet practical needs. Road conditions can change during driving, such as congestion and temporary accidents. Therefore, vehicle navigation must also consider the dynamic environment of road conditions. Given the uncertainty of the road network, users desire multiple routes to choose from, with good values ​​for objectives such as mileage and the number of congested sections, and with significant variation. Different users also have preferences when driving. By combining navigation data from the user and others, a personalized navigation plan can be created that suits the user's preferences. Although some scholars have conducted in-depth research on the personalized path planning problem, due to the complexity of problem modeling and the difficulty of solving it, there are currently few research results on the multimodal and multi-objective path planning problem.

[0004] Swarm intelligence optimization methods, as an effective approach for solving complex, large-scale optimization problems, have become a research hotspot across many fields. Traditional swarm intelligence optimization methods, such as particle swarm optimization and differential evolution, are primarily model-driven and based on simple iterative formulas. These methods employ the same parameter settings and structure to solve different optimization problems. While simple to implement and requiring no background knowledge of the problem, they suffer from low accuracy and struggle to guarantee performance for specific problems.

[0005] The Brainstorming Optimization (BSO) method is an optimization method that integrates data analysis. During operation, it iteratively searches the solution space by classifying or clustering the solution set data. Using the current information gained from the search, it guides the solution set toward areas with better fitness values. Given the characteristics of multimodal problems, improving the method's solution accuracy and increasing the diversity of its output are crucial. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a multimodal personalized vehicle path navigation method based on brainstorming optimization with simple model parameter calculation, high efficiency, fast testing speed and accurate prediction.

[0007] The technical solution adopted to solve the above technical problems consists of the following steps:

[0008] (1) Parameters for quantifying user goals

[0009] The parameter of the user goal is G∈{1,2,3}. If the user goal is to minimize the time, G is 1; if the user goal is to minimize the distance, G is 2; if the user goal is to pass the least number of traffic lights, G is 3.

[0010] (2) Modeling grid maps

[0011] According to the departure path point and destination selected by the user terminal, the road features in the actual road network map are modeled into a regular grid map Gra. The set of path points P is {1, 2, ..., n}, where n represents the total number of path points in the grid map Gra, and n is a finite positive integer of at least 2.

[0012] (3) Quantifying path difference target parameters

[0013] The path difference target quantification includes the numerical quantification of distance S, driving speed V, number of traffic lights R, interval time RT, driving time T, congestion time C, and congested road section E as follows:

[0014] The distance S between any two path points is determined as follows:

[0015]

[0016] where s p,q Represents the distance from path point p to path point q, p,q∈[2,n].

[0017] Determine the average speed V between any two path points using the following formula:

[0018]

[0019] where v a,bRepresents the average speed of the vehicle from path point a to path point b, a,b∈[2,n].

[0020] Determine the number of traffic lights R between any two path points according to the following formula:

[0021]

[0022] where r c,d Represents the number of traffic lights from path point c to path point d, c,d∈[2,n].

[0023] Determine the interval time RT between the number of traffic lights between any two path points according to the following formula:

[0024]

[0025] where rt u,g Represents the traffic light interval time from path point u to path point g, u,g∈[2,n].

[0026] The driving time T between any two path points is determined as follows:

[0027]

[0028]

[0029] Among them, t e,k represents the driving time from path point e to path point k, e,k∈[2,n].

[0030] The congestion time C between any two path points is determined as follows:

[0031]

[0032] Among them, c z,j represents the congestion time from path point z to path point j, z,j∈[2,n].

[0033] Determine the congested section E between two route points where the congestion time exceeds the limit value ψ minutes as follows:

[0034]

[0035]

[0036] Where ψ represents the traffic congestion time limit, ψ∈[1,30], e o,l 1 means the congestion time from path point o to path point l exceeds the congestion time limit value ψ minutes, o,l∈[2,n], e o,l 0 means the congestion time c from path point o to path point l o,l The limit value ψ minutes is not exceeded;

[0037] (4) Initialize the population

[0038] Initialize the population X according to formula (10) N :

[0039] X N =(x1,x2,…,x N ) (10)

[0040] Among them, x N is the Nth path from the starting point to the destination.

[0041] (5) Evaluate individuals

[0042] The fitness value of the fitness function f(x) is determined according to formulas (11)-(13).

[0043]

[0044]

[0045]

[0046] Among them, w1 is the weight of minimizing driving time, w1∈[1,100], μ is the weight of road congestion time, μ∈[0.0001,100], t x(i-1),x(i) Indicates the average travel time from the i-1th path point to the i path point on the current route x, rt x(i-1),x(i) Indicates the time between the traffic lights from the i-1th path point to the i-th path point on the current route x, c x(i-1),x(i) It represents the congestion time from the i-1th path point to the ith path point of the current route x, i∈{1,2,...,N}, N is a finite positive integer, Δσ is the determination coefficient, Δσ is 0, indicating that the current route is not selectable, Δσ is 1, indicating that the current route is selectable, and the current route is Make e x(i-1),x(i) If Δσ is 1, the current route is blocked and the current route is not selected. If Δσ is 0, the current route is blocked. Make e x(i-1),x(i) is 0, that is, the current route is not blocked, select the current route, Δσ is 1, S x(i-1),x(i) Indicates the path length from the i-1th path point of the current route x to the i-th path point, r x(i-1),x(i)represents the number of traffic lights from the i-1th path point to the i-th path point on the current route x, φ1 is the weight of minimizing the path length, φ1∈[1,100], φ2 is the weight of minimizing the time, φ2∈[0.0001,0.1], φ3 is the weight of minimizing the number of red lights, φ3∈[0.0001,10], δ1 is the weight of minimizing the number of red lights, δ1∈[1,100], δ2 is the weight of minimizing the number of red lights, δ2∈[0.0001,0.1], δ3 is the weight of minimizing the path length, δ3∈[0.0001,10].

[0047] (6) Partitioning the Set

[0048] In each iteration, all solutions are sorted according to their fitness values ​​and divided into elite solution sets. and the general solution set in represents the Eth elite individual, Represents the Nrth ordinary individual.

[0049] (7) Update individual

[0050] Generate a new sample expression according to formula (14)

[0051]

[0052]

[0053] Among them, P c1 and P c2 represents the probability parameter, P c1 ,P c2 ∈(0,1), N(2,1) is a Gaussian random function with mean 2 and variance 1, and rand(0,1) is a random number between (0,1).

[0054] (8) Update the worst solution

[0055] Update the worst solution according to formula (16)

[0056]

[0057] in, are the worst solution, the best solution, and the random solution in the current solution set, respectively. t represents the number of iterations, t∈[20,2000], and F represents the scaling factor, F∈[0.001,2].

[0058] (9) Output navigation results

[0059] Output all alternative routes to the user and navigate based on the selection results.

[0060] In the formulas (11) to (13) for evaluating individuals in step (5) of the present invention, the optimal value of w1 is 50, the optimal value of μ is 50, the optimal value of φ1 is 50, the optimal value of φ2 is 0.05, the optimal value of φ3 is 5, the optimal value of δ1 is 50, the optimal value of δ2 is 0.05, and the optimal value of δ3 is 5.

[0061] In step (7) of the present invention, the P c1 The optimal value of the probability parameter is 0.5, P c2 Represents the probability parameter, and the best value is 0.5.

[0062] In the step (8) of updating the worst solution of the present invention, t is the number of iterations and its optimal value is 1000, and F is the scaling factor and its optimal value is 1.

[0063] By quantifying user goal parameters and assessing individual predictions, this method addresses the problem of existing technologies failing to fully consider the dynamic environment of road conditions and user needs, making them unsuitable for multimodal navigation planning. Quantifying user goal parameters, modeling grid maps, and quantifying path-specific target parameters address the complex model parameter acquisition and unsuitability of existing technologies for personalized navigation planning. This method offers advantages such as simple parameter calculation, high efficiency, fast computation speed, and accurate predictions, making it suitable for widespread use in automobiles and mobile phones. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a flowchart of Example 1 of the present invention. DETAILED DESCRIPTION

[0065] The present invention will be further described in detail below with reference to the accompanying drawings and examples, but the present invention is not limited to the following embodiments.

[0066] Example 1

[0067] The multimodal personalized vehicle path navigation method based on brainstorming optimization in this embodiment consists of the following steps (see Figure 1 ):

[0068] (1) Parameters for quantifying user goals

[0069] The parameter of the user goal is G∈{1,2,3}. If the user goal is to minimize the time, G is 1; if the user goal is to minimize the distance, G is 2; if the user goal is to pass the least number of traffic lights, G is 3.

[0070] (2) Modeling grid maps

[0071] According to the departure path point and destination selected by the user terminal, the road features in the actual road network map are modeled into a regular grid map Gra. The set of path points P is {1, 2, ..., n}, where n represents the total number of path points in the grid map Gra, and n is a finite positive integer of at least 2.

[0072] (3) Quantifying path difference target parameters

[0073] The path difference target quantification includes the numerical quantification of distance S, driving speed V, number of traffic lights R, interval time RT, driving time T, congestion time C, and congested road section E as follows:

[0074] The distance S between any two path points is determined as follows:

[0075]

[0076] where s p,q Represents the distance from path point p to path point q, p,q∈[2,n].

[0077] Determine the average speed V between any two path points using the following formula:

[0078]

[0079] where v a,b Represents the average speed of the vehicle from path point a to path point b, a,b∈[2,n].

[0080] Determine the number of traffic lights R between any two path points according to the following formula:

[0081]

[0082] where r c,d Represents the number of traffic lights from path point c to path point d, c,d∈[2,n].

[0083] Determine the interval time RT between the number of traffic lights between any two path points according to the following formula:

[0084]

[0085] where rt u,g Represents the traffic light interval time from path point u to path point g, u,g∈[2,n].

[0086] The driving time T between any two path points is determined as follows:

[0087]

[0088]

[0089] Among them, t e,k represents the driving time from path point e to path point k, e,k∈[2,n].

[0090] The congestion time C between any two path points is determined as follows:

[0091]

[0092] Among them, c z,j represents the congestion time from path point z to path point j, z,j∈[2,n].

[0093] Determine the congested section E between two route points where the congestion time exceeds the limit value ψ minutes as follows:

[0094]

[0095]

[0096] Where ψ represents the traffic congestion time limit, ψ∈[1,30]. In this embodiment, the value of ψ is 10, e o,l 1 means the congestion time from path point o to path point l exceeds the congestion time limit value ψ minutes, o,l∈[2,n], e o,l 0 means the congestion time c from path point o to path point l o,l The limit value ψ minutes was not exceeded.

[0097] (4) Initialize the population

[0098] Initialize the population X according to formula (10) N :

[0099] X N =(x1,x2,…,x N ) (10)

[0100] Among them, x N is the Nth path from the starting point to the destination.

[0101] (5) Evaluate individuals

[0102] The fitness value of the fitness function f(x) is determined according to formulas (11)-(13).

[0103]

[0104]

[0105]

[0106] Among them, w1 is the weight of minimizing driving time, w1∈[1,100], and the value of w1 in this embodiment is 50, μ is the weight of road congestion time, μ∈[0.0001,100], and the value of μ in this embodiment is 50, t x(i-1),x(i) Indicates the average travel time from the i-1th path point to the i path point on the current route x, rt x(i-1),x(i) Indicates the time between the traffic lights from the i-1th path point to the i-th path point on the current route x, c x(i-1),x(i) It represents the congestion time from the i-1th path point to the ith path point of the current route x, i∈{1,2,...,N}, N is a finite positive integer, Δσ is the determination coefficient, Δσ is 0, indicating that the current route is not selectable, Δσ is 1, indicating that the current route is selectable, and the current route is Make e x(i-1),x(i) If Δσ is 1, the current route is blocked and the current route is not selected. If Δσ is 0, the current route is blocked. Make e x(i-1),x(i) is 0, that is, the current route is not blocked, select the current route, Δσ is 1, S x(i-1),x(i) Indicates the path length from the i-1th path point of the current route x to the i-th path point, r x(i-1),x(i) represents the number of traffic lights from the i-1th path point to the i-th path point of the current route x, φ1 is the weight of minimizing the path length, φ1∈[1,100], and the value of φ1 in this embodiment is 50, φ2 is the weight of minimizing the time, φ2∈[0.0001,0.1], and the value of φ2 in this embodiment is 0.05, φ3 is the weight of minimizing the number of red lights, φ3∈[0.0001,10], and the value of φ3 in this embodiment is 5, δ1 is the weight of minimizing the number of red lights, δ1∈[1,100], and the value of δ1 in this embodiment is 50, δ2 is the weight of minimizing the number of red lights, δ2∈[0.0001,0.1], and the value of δ2 in this embodiment is 0.05, δ3 is the weight of minimizing the path length, δ3∈[0.0001,10], and the value of δ3 in this embodiment is 5.

[0107] (6) Partitioning the Set

[0108] In each iteration, all solutions are sorted according to their fitness values ​​and divided into elite solution sets. and the general solution set in represents the Eth elite individual, Represents the Nrth ordinary individual.

[0109] (7) Update individual

[0110] Generate a new sample expression according to formula (14)

[0111]

[0112]

[0113] Among them, P c1 and P c2 represents the probability parameter, P c1 ,P c2 ∈(0,1), P in this embodiment c1 The value is 0.5, and the P c2 The value is 0.5, N(2,1) is a Gaussian random function with a mean of 2 and a variance of 1, and rand(0,1) is a random number between (0,1).

[0114] (8) Update the worst solution

[0115] Update the worst solution according to formula (16)

[0116]

[0117] in, are the worst solution, the best solution, and the random solution in the current solution set, respectively. t represents the number of iterations, t∈[20,2000]. In this embodiment, the value of t is 1000. F represents the scaling factor, F∈[0.001,2]. In this embodiment, the value of F is 1.

[0118] (9) Output navigation results

[0119] Output all alternative routes to the user and navigate based on the selection results.

[0120] Complete a multimodal personalized vehicle path navigation method based on brainstorming optimization.

[0121] Example 2

[0122] The multimodal personalized vehicle path navigation method based on brainstorming optimization in this embodiment consists of the following steps:

[0123] (1) Parameters for quantifying user goals

[0124] This step is the same as in Example 1.

[0125] (2) Modeling grid maps

[0126] This step is the same as in Example 1.

[0127] (3) Quantifying path difference target parameters

[0128] In formula (9), ψ represents the traffic congestion time limit, ψ∈[1,30], and the value of ψ in this embodiment is 1, eo,l 1 means the congestion time from path point o to path point l exceeds the congestion time limit value ψ minutes, o,l∈[2,n], e o,l 0 means the congestion time c from path point o to path point l o,l The limit value ψ minutes was not exceeded.

[0129] The other steps of this step are the same as those in Example 1.

[0130] (4) Initialize the population

[0131] This step is the same as in Example 1.

[0132] (5) Evaluate individuals

[0133] The fitness value of the fitness function f(x) is determined according to formulas (11)-(13).

[0134]

[0135]

[0136]

[0137] Wherein, w1 is the weight of minimizing driving time, w1∈[1,100], in this embodiment, w1 is set to 1, μ is the weight of road congestion time, μ∈[0.0001,100], in this embodiment, μ is set to 0.0001, t x(i-1),x(i) Indicates the average travel time from the i-1th path point to the i path point on the current route x, rt x(i-1),x(i) Indicates the time between the traffic lights from the i-1th path point to the i-th path point on the current route x, c x(i-1),x(i) It represents the congestion time from the i-1th path point to the ith path point of the current route x, i∈{1,2,...,N}, N is a finite positive integer, Δσ is the determination coefficient, Δσ is 0, indicating that the current route is not selectable, Δσ is 1, indicating that the current route is selectable, and the current route is Make e x(i-1),x(i) If Δσ is 1, the current route is blocked and the current route is not selected. If Δσ is 0, the current route is blocked. Make e x(i-1),x(i) is 0, that is, the current route is not blocked, select the current route, Δσ is 1, S x(i-1),x(i) Indicates the path length from the i-1th path point of the current route x to the i-th path point, r x(i-1),x(i)represents the number of traffic lights from the i-1th path point to the i-th path point of the current route x, φ1 is the weight of minimizing the path length, φ1∈[1,100], and the value of φ1 in this embodiment is 1, φ2 is the weight of minimizing the time, φ2∈[0.0001,0.1], and the value of φ2 in this embodiment is 0.0001, φ3 is the weight of minimizing the number of red lights, φ3∈[0.0001,10], and the value of φ3 in this embodiment is 0.0001, δ1 is the weight of minimizing the number of red lights, δ1∈[1,100], and the value of δ1 in this embodiment is 1, δ2 is the weight of minimizing the number of red lights, δ2∈[0.0001,0.1], and the value of δ2 in this embodiment is 0.0001, and δ3 is the weight of minimizing the path length, δ3∈[0.0001,10], and the value of δ3 in this embodiment is 0.0001.

[0138] (6) Partitioning the Set

[0139] This step is the same as in Example 1.

[0140] (7) Update individual

[0141] Generate a new sample expression according to formula (14)

[0142]

[0143]

[0144] Among them, P c1 and P c2 represents the probability parameter, P c1 ,P c2 ∈(0,1), P in this embodiment c1 The value is 0.1, and the P c2 The value is 0.1, N(2,1) is a Gaussian random function with a mean of 2 and a variance of 1, and rand(0,1) is a random number between (0,1).

[0145] (8) Update the worst solution

[0146] Update the worst solution according to formula (16)

[0147]

[0148] in, are the worst solution, the best solution, and the random solution in the current solution set, respectively. t represents the number of iterations, t∈[20,2000]. In this embodiment, the value of t is 20. F represents the scaling factor, F∈[0.001,2]. In this embodiment, the value of F is 0.001.

[0149] (9) Output navigation results

[0150] Output all alternative routes to the user and navigate based on the selection results.

[0151] Complete a multimodal personalized vehicle path navigation method based on brainstorming optimization.

[0152] Example 3

[0153] The multimodal personalized vehicle path navigation method based on brainstorming optimization in this embodiment consists of the following steps:

[0154] (1) Parameters for quantifying user goals

[0155] This step is the same as in Example 1.

[0156] (2) Modeling grid maps

[0157] This step is the same as in Example 1.

[0158] (3) Quantifying path difference target parameters

[0159] In formula (9), ψ represents the traffic congestion time limit, ψ∈[1,30], and the value of ψ in this embodiment is 30, e o,l 1 means the congestion time from path point o to path point l exceeds the congestion time limit value ψ minutes, o,l∈[2,n], e o,l 0 means the congestion time c from path point o to path point l o,l The limit value ψ minutes was not exceeded.

[0160] The other steps of this step are the same as those in Example 1.

[0161] (4) Initialize the population

[0162] This step is the same as in Example 1.

[0163] (5) Evaluate individuals

[0164] The fitness value of the fitness function f(x) is determined according to formulas (11)-(13).

[0165]

[0166]

[0167]

[0168] Wherein, w1 is the weight of minimizing driving time, w1∈[1,100], in this embodiment, w1 is 100, μ is the weight of road congestion time, μ∈[0.0001,100], in this embodiment, μ is 100, t x(i-1),x(i) Indicates the average travel time from the i-1th path point to the i path point on the current route x, rt x(i-1),x(i) Indicates the time between the traffic lights from the i-1th path point to the i-th path point on the current route x, c x(i-1),x(i) It represents the congestion time from the i-1th path point to the ith path point of the current route x, i∈{1,2,...,N}, N is a finite positive integer, Δσ is the determination coefficient, Δσ is 0, indicating that the current route is not selectable, Δσ is 1, indicating that the current route is selectable, and the current route is Make e x(i-1),x(i) If Δσ is 1, the current route is blocked and the current route is not selected. If Δσ is 0, the current route is blocked. Make e x(i-1),x(i) is 0, that is, the current route is not blocked, select the current route, Δσ is 1, S x(i-1),x(i) Indicates the path length from the i-1th path point of the current route x to the i-th path point, r x(i-1),x(i) represents the number of traffic lights from the i-1th path point to the i-th path point of the current route x, φ1 is the weight of minimizing the path length, φ1∈[1,100], and the value of φ1 in this embodiment is 100, φ2 is the weight of minimizing the time, φ2∈[0.0001,0.1], and the value of φ2 in this embodiment is 0.1, φ3 is the weight of minimizing the number of red lights, φ3∈[0.0001,10], and the value of φ3 in this embodiment is 10, δ1 is the weight of minimizing the number of red lights, δ1∈[1,100], and the value of δ1 in this embodiment is 100, δ2 is the weight of minimizing the number of red lights, δ2∈[0.0001,0.1], and the value of δ2 in this embodiment is 0.1, δ3 is the weight of minimizing the path length, δ3∈[0.0001,10], and the value of δ3 in this embodiment is 10.

[0169] (6) Partitioning the Set

[0170] This step is the same as in Example 1.

[0171] (7) Update individual

[0172] Generate a new sample expression according to formula (14)

[0173]

[0174]

[0175] Among them, Pc1 and P c2 represents the probability parameter, P c1 ,P c2 ∈(0,1), P in this embodiment c1 The value is 0.9, and the P c2 The value is 0.9, N(2,1) is a Gaussian random function with a mean of 2 and a variance of 1, and rand(0,1) is a random number between (0,1).

[0176] (8) Update the worst solution

[0177] Update the worst solution according to formula (16)

[0178]

[0179] in, are the worst solution, the best solution, and the random solution in the current solution set, respectively. t represents the number of iterations, t∈[20,2000]. In this embodiment, the value of t is 2000. F represents the scaling factor, F∈[0.001,2]. In this embodiment, the value of F is 2.

[0180] (9) Output navigation results

[0181] Output all alternative routes to the user and navigate based on the selection results.

[0182] Multimodal personalized vehicle path navigation method based on brainstorming optimization.

[0183] In order to verify the beneficial effects of the present invention, the inventors used the method of Example 1 of the present invention to conduct experiments on six navigation planning problems starting from Xi'an, Shaanxi Province. The experimental results are as follows.

[0184] Using Python to plan routes on Amap can use the route planning service provided by the Amap Web API. The following are the basic steps for using Python to plan routes on Amap:

[0185] Create an application and obtain the AppKey of the AutoNavi application. Select the editor PyCharm Community Edition 2019.1, use the requests library to send an HTTP request to the AutoNavi Map Web API, request the path planning service, and pass the necessary parameters, including the starting point, end point, and travel mode. Parse the returned JSON data. Use the method of Example 1 of the present invention and the default greedy method to plan the six navigation problems respectively. When the ratio of the planned route time corresponding to the method of Example 1 of the present invention to the planned route time corresponding to the fixed greedy method is less than 1, it means compared with the fixed greedy method.

[0186] The experimental results are shown in Table 1.

[0187] Table 1 Experimental results

[0188] route ratio Xi'an North-Big Wild Goose Pagoda 0.73 Bell Tower-Convention and Exhibition Center 0.93 City Library-Sports Park 0.91 Bell Tower-Xi'an North 0.93 Convention and Exhibition Center-Xi'an North 0.81 Municipal Library-Xi'an North 0.98 average 0.88

[0189] As can be seen from Table 1, for all test videos, the average ratio corresponding to the method of Example 1 of the present invention is 0.88, that is, the method of Example 1 of the present invention has simple parameter calculation, high efficiency, fast calculation speed, and accurate prediction, and can obtain the same route in less time.

Claims

1. A multimodal personalized vehicle path navigation method based on brainstorming optimization, characterized by It consists of the following steps: (1) Parameters for quantifying user goals The parameter of the user goal is G∈{1,2,3}. If the user goal is to minimize the time, G is 1; if the user goal is to minimize the distance, G is 2; if the user goal is to pass the least number of traffic lights, G is 3; (2) Modeling grid maps The road features in the actual road network map are modeled into a regular grid map Gra according to the departure path point and destination selected by the user terminal. The set of path points P is {1, 2, ..., n}, where n represents the total number of path points in the grid map Gra and is a finite positive integer of at least 2. (3) Quantifying path difference target parameters The path difference target quantification includes the numerical quantification of distance S, driving speed V, number of traffic lights R, interval time RT, driving time T, congestion time C, and congested road section E as follows: The distance S between any two path points is determined as follows: where s p,q Represents the distance from path point p to path point q, p,q∈[2,n]; Determine the average speed V between any two path points using the following formula: where v a,b represents the average speed of the vehicle from path point a to path point b, a,b∈[2,n]; Determine the number of traffic lights R between any two path points according to the following formula: where r c,d Indicates the number of traffic lights from path point c to path point d, c,d∈[2,n]; Determine the interval time RT between the number of traffic lights between any two path points according to the following formula: where rt u,g represents the traffic light interval time from path point u to path point g, u,g∈[2,n]; The driving time T between any two path points is determined as follows: Among them, t e,k represents the driving time from path point e to path point k, e,k∈[2,n]; The congestion time C between any two path points is determined as follows: Among them, c z,j represents the congestion time from path point z to path point j, z,j∈[2,n]; Determine the congested section E between two route points where the congestion time exceeds the limit value ψ minutes as follows: Where ψ represents the traffic congestion time limit, ψ∈[1,30], e o,l 1 means the congestion time from path point o to path point l exceeds the congestion time limit value ψ minutes, o,l∈[2,n], e o,l 0 means the congestion time c from path point o to path point l o,l The limit value ψ minutes is not exceeded; (4) Initialize the population Initialize the population X according to formula (10) N : X N =(x1,x2,…,x N ) (10) Among them, x N is the Nth path from the starting point to the destination; (5) Evaluate individuals The fitness value of the fitness function f(x) is determined according to formulas (11)-(13): Among them, w1 is the weight of minimizing driving time, w1∈[1, 100], μ is the weight of road congestion time, μ∈[0.0001, 100], t x(i-1)，x(i) Indicates the average travel time from the i-1th path point to the i path point on the current route x, rt x(i-1)，x(i) Indicates the time between the traffic lights from the i-1th path point to the i-th path point on the current route x, c x(i-1)，x(i) represents the congestion time from the i-1th path point to the ith path point of the current route x, i∈{1, 2, ..., N}, N is a finite positive integer, Δσ is the determination coefficient, Δσ is 0, indicating that the current route is not selectable, Δσ is 1, indicating that the current route is selectable, and the current route is Make e x(i-1)，x(i) If Δσ is 1, the current route is blocked and the current route is not selected. If Δσ is 0, the current route is blocked. Make e x(i-1)，x(i) is 0, that is, the current route is not blocked, select the current route, Δσ is 1, S x(i-1)，x(i) Indicates the path length from the i-1th path point of the current route x to the i-th path point, r x(i-1)，x(i) Represents the number of traffic lights from the i-1th path point to the i-th path point on the current route x, φ1 is the weight of minimizing the path length, φ1∈[1,100], φ2 is the weight of minimizing the time, φ2∈[0.0001,0.1], φ3 is the weight of minimizing the number of red lights, φ3∈[0.0001,10], δ1 is the weight of minimizing the number of red lights, δ1∈[1,100], δ2 is the weight of minimizing the number of red lights, δ2∈[0.0001,0.1], δ3 is the weight of minimizing the path length, δ3∈[0.0001,10]; (6) Partitioning the Set In each iteration, all solutions are sorted according to their fitness values ​​and divided into elite solution sets. and the general solution set in represents the Eth elite individual, represents the Nrth ordinary individual (7) Update individual Generate a new sample expression according to formula (14) Among them, P c1 and P c2 represents the probability parameter, P c1 , P c2 ∈(0,1), N(2,1) is a Gaussian random function with mean 2 and variance 1, and rand(0,1) is a random number between (0,1); (8) Update the worst solution Update the worst solution according to formula (16) in, are the worst solution, the best solution, and the random solution in the current solution set, respectively. t represents the number of iterations, t∈[20, 2000], and F represents the scaling factor, F∈[0.001, 2]. (9) Output navigation results Output all alternative routes to the user and navigate based on the selection results.

2. The multimodal personalized vehicle path navigation method based on brainstorming optimization according to claim 1 is characterized in that: In the equations (11) to (13) for evaluating individuals in step (5), the value of w1 is 50, the value of μ is 50, the value of φ1 is 50, the value of φ2 is 0.05, the value of φ3 is 5, the value of δ1 is 50, the value of δ2 is 0.05, and the value of δ3 is 5.

3. The multimodal personalized vehicle path navigation method based on brainstorming optimization according to claim 1 is characterized in that: In step (7) updating the individual, the P c1 Indicates that the probability parameter takes a value of 0.5, P c2 Indicates that the probability parameter takes a value of 0.

5.

4. The multimodal personalized vehicle path navigation method based on brainstorming optimization according to claim 1 is characterized in that: In step (8) of updating the worst solution, t is the number of iterations, which is 1000, and F is the scaling factor, which is 1.

Images