Vehicle path prediction method based on generalized possibility kripke model

Abstract

A vehicle path prediction method based on generalized possibility Kripke structure model is composed of constructing actual road condition model, constructing path transition distribution matrix, quantifying initial traffic condition, constructing vehicle path prediction atomic proposition, quantifying label value, performing model detection, evaluating advantages and disadvantages of road intersection and predicting optimal path. Compared with possibility Kripke structure, the generalized possibility Kripke model is adopted to model road traffic condition, the road intersection is regarded as state space, the smooth degree of road section is regarded as path transition distribution matrix, the average passing speed of vehicle, average waiting time of red light and traffic violation index are regarded as label value, the optimal path prediction method based on generalized possibility Kripke structure modeling is proposed, the technical problem of difficult calculation is solved, the method has the advantages of easy data acquisition, simple and convenient execution process and the like, and has certain value for traffic management and prediction implementation measures.

Description

Technical Field

[0001] This invention belongs to the field of computer technology, and specifically relates to vehicle route navigation. Background Technology

[0002] Traditional optimal route planning methods are mostly single-objective optimization methods aimed at minimizing path length, based on simple iterative calculations. However, representative indicators for urban roads are often multiple and exhibit uncertainty and variability. Therefore, directly utilizing data-driven approaches to find optimal travel routes has gradually become a research hotspot.

[0003] Vehicle location shifting choices typically describe a driver's actual route selection behavior during a trip. Most studies on vehicle location shifting utilize complex machine learning processes, which are often inefficient for finding optimal routes. In actual driving, it can be assumed that historical travel data implicitly contains the driver's optimal route choices. Therefore, based on historical travel data under specific conditions, corresponding driving information can be obtained, and optimal route planning can be performed.

[0004] Probabilistic model checking primarily addresses the model checking problem of uncertain systems arising from stochastic processes, aiming to determine the accuracy of the probabilistic system against quantitative probability specifications. Multivalued model checking mainly deals with model checking problems of systems containing incomplete or inconsistent information. Fuzzy model checking primarily deals with model checking problems of systems with uncertain data representations, focusing more on the true values ​​of system attributes. While probability computational tree logic is more expressive than computational tree logic, it is too restrictive. Some uncertainties can be described using probability theory but cannot be directly handled using probability computational tree logic model checking, such as those modeled by probability Kripke structures with fuzzy label values. Considering that drivers will mostly choose the optimal route, and assuming all drivers believe they have chosen the optimal route, optimal route planning is performed for trips between a specific origin and destination based on historical travel data. Therefore, the optimal route prediction method based on model checking and a generalized probability Kripke model can take into account the impact of different traffic conditions and use the path with the highest probability assessment value as the predicted optimal path.

[0005] In the field of high-tech road transportation, one urgent technical problem to be solved is to provide a method for predicting routes that ensure smooth traffic flow and safety for vehicles. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a vehicle path prediction method based on the generalized probability Kripke model that is easy to obtain data, simple and convenient to execute, and accurate in prediction.

[0007] The technical solutions adopted to solve the above technical problems are as follows:

[0008] (1) Constructing a real road condition model

[0009] Collect road condition data between the starting point and the ending point, and the set S of all road intersections between the starting point and the ending point:

[0010] S=

[0011] in, Let S represent any element in set S, and n represent the number of road intersections.

[0012] (2) Construct the path transition distribution matrix

[0013] Construct the path transition distribution matrix P using the following formula:

[0014] P=

[0015] Among them, s m and s n Indicates a road intersection. Indicates road intersection s m With s n The path accessibility index between them .

[0016] (3) Quantify the initial traffic conditions

[0017] The road intersection s is calculated using the following formula. n The initial traffic conditions are quantized into a column vector I. n :

[0018] ,

[0019] Among them, i n The first intersection the vehicle passes through is s, with 1 for all other elements and 0 for the rest. n .

[0020] (4) Constructing atomic propositions for vehicle route prediction

[0021] The atomic propositions of vehicle routing planning include the average vehicle speed v, the average waiting time at traffic lights t, and the traffic violation index p.

[0022] (5) Quantify label values

[0023] Quantify the average passing speed label value of vehicles using the following formula :

[0024]

[0025] in, v represents the average speed of the vehicle. max This indicates all road intersections that have passed through within the last six months. n The vehicle's maximum speed, v min This indicates all road intersections that have passed through within the last six months. n The vehicle's minimum speed, quantified by the formula, represents the average waiting time label for red lights. :

[0026]

[0027]

[0028] Where t represents the average waiting time at a red light, t max This indicates all road intersections that have passed through within the last six months. n The longest waiting time for a vehicle, t min This indicates all road intersections that have passed through within the last six months. n The shortest waiting time for vehicles is quantified by the following formula to determine the traffic violation index label value. :

[0029]

[0030] in, This represents the traffic violation index, where 'c' indicates the number of road intersections in the past six months. n The number of vehicles committing traffic violations, d represents the number of intersections s in the past six months. n Total number of vehicles passing through.

[0031] (6) Perform model testing

[0032] According to formula (1), determine the probability that each road intersection will have a longer red light waiting time than the next road intersection. :

[0033] (1)

[0034]

[0035]

[0036]

[0037] in, This indicates that the next intersection to be reached from the current intersection will have a long red light waiting time. This represents the max-min operation of the fuzzy matrix. Indicates that the diagonal elements are The diagonal matrix P + As an intermediate variable, D represents the matrix A column vector consisting of all elements on the main diagonal;

[0038] According to formula (2), determine the probability that no traffic violations will occur at each road intersection. :

[0039] (2)

[0040]

[0041]

[0042] in, This means that all road intersections starting from the current intersection will always satisfy the condition that no traffic violations will occur. express The maximum fixed-point function, Z represents the function The fixed point, Indicates that the diagonal elements are A diagonal matrix.

[0043] According to formula (3), determine the probability that each road intersection will eventually allow vehicles to pass at a relatively high speed. :

[0044] (3)

[0045] in, v indicates that at a certain moment, the vehicle will travel at a relatively high speed. Represents the identity matrix. Indicates that the diagonal elements are A diagonal matrix.

[0046] (7) Evaluate the quality of road intersections

[0047] Evaluate road intersections s according to formula (4) n Advantages and disadvantages :

[0048] (4)

[0049]

[0050]

[0051]

[0052] Where e represents the weight of the average vehicle speed. This indicates the probability that the vehicle will eventually travel at a faster speed, and f represents the weight of the average waiting time at the red light. This indicates the probability of a long wait time at the next intersection's red light, and g represents the weight of the traffic violation. This indicates the likelihood that the vehicle will not commit a traffic violation; E1 and E3 assess the probability of traffic violations at road intersections. n E2 is a positive indicator for evaluating road intersections. n A negative indicator.

[0053] (8) Predict the optimal path

[0054] 1) Determine the set C containing the subsequent road intersections.

[0055] Road intersections n The set C containing the subsequent road intersections is as follows:

[0056] C ,

[0057] Among them, s n+q Indicates road intersection s n The subsequent intersection.

[0058] 2) Determine the evaluation set D for subsequent road intersections.

[0059] For each road intersection in set C Calculate its evaluation value according to formula (4). The evaluation set D of the subsequent road intersections is as follows:

[0060] D

[0061] Where, f(s) n+q ) represents a road intersection s n+q The evaluation value of its merits and demerits.

[0062] 3) Determine the next road intersection the vehicle will pass through.

[0063] Select the evaluation set D with the largest evaluation value. The corresponding road intersection As the next road intersection the vehicle passes through.

[0064] 4) Output the optimal path

[0065] By doing this, the next optimal road intersection is selected step by step to obtain the local optimal path.

[0066] Step (2) of this invention constructs the path transition distribution matrix as follows:

[0067] Construct the path transition distribution matrix P using the following formula:

[0068] P=

[0069] in Indicates road intersection s m With s n The path accessibility index between them The optimal value is 0.5.

[0070] Compared with the probabilistic Kripke structure, this invention uses a generalized probabilistic Kripke model to model road traffic conditions. It takes road intersections as the state space, the smoothness of road segments as the path transition distribution matrix, and the average vehicle speed, average red light waiting time, and traffic violation index as label values. It proposes an optimal path prediction method based on the generalized probabilistic Kripke structure model, which solves the technical problem of computational difficulty. It has the advantages of easy data acquisition and simple and convenient execution process, and has certain value for traffic management and forecasting implementation measures. Attached Figure Description

[0071] Figure 1 This is a flowchart of Embodiment 1 of the present invention. Detailed Implementation

[0072] The present invention will be further described below with reference to the accompanying drawings and embodiments, but the present invention is not limited to the following embodiments.

[0073] Example 1

[0074] The vehicle path prediction method based on the generalized probability Kripke model in this embodiment consists of the following steps (see...). Figure 1 ):

[0075] (1) Constructing a real road condition model

[0076] Collect road condition data between the starting point and the ending point, and the set S of all road intersections between the starting point and the ending point: S =

[0077] in, Let S represent any element in set S, and n represent the number of road intersections.

[0078] (2) Construct the path transition distribution matrix

[0079] Construct the path transition distribution matrix P using the following formula:

[0080] P=

[0081] Among them, s m and s n Indicates a road intersection. Indicates road intersection s m With s n The path accessibility index between them The value is 0.5.

[0082] (3) Quantify the initial traffic conditions

[0083] The road intersection s is calculated using the following formula. n The initial traffic conditions are quantized into a column vector I. n :

[0084] ,

[0085] Among them, i n The first intersection the vehicle passes through is s, with 1 for all other elements and 0 for the rest. n .

[0086] (4) Constructing atomic propositions for vehicle route prediction

[0087] The atomic propositions of vehicle routing planning include the average vehicle speed v, the average waiting time at traffic lights t, and the traffic violation index p.

[0088] (5) Quantify label values

[0089] Quantify the average passing speed label value of vehicles using the following formula :

[0090]

[0091] in, v represents the average speed of the vehicle. max This indicates all road intersections that have passed through within the last six months. n The vehicle's maximum speed, v min This indicates all road intersections that have passed through within the last six months. n The vehicle's minimum speed, quantified by the formula, represents the average waiting time label for red lights. :

[0092]

[0093] Where t represents the average waiting time at a red light, t max This indicates all road intersections that have passed through within the last six months. n The longest waiting time for a vehicle, t min This indicates all road intersections that have passed through within the last six months. n The shortest waiting time for vehicles is quantified by the following formula to determine the traffic violation index label value. :

[0094]

[0095] in, This represents the traffic violation index, where 'c' indicates the number of road intersections in the past six months. n The number of vehicles committing traffic violations, d represents the number of intersections s in the past six months. n Total number of vehicles passing through.

[0096] Compared to the generalized probability Kripke structure, the generalized probability Kripke model in this embodiment does not have normality constraints on its initial and state transition distributions, thus conforming to actual urban road traffic conditions. Furthermore, the label values ​​of the generalized probability Kripke structure are fuzzy, containing fuzzy information. In the optimal path planning problem, road traffic conditions are formalized as a generalized probability Kripke structure.

[0097] (6) Perform model testing

[0098] According to formula (1), determine the probability that each road intersection will have a longer red light waiting time than the next road intersection. :

[0099] (1)

[0100]

[0101]

[0102]

[0103] in, This indicates that the next intersection to be reached from the current intersection will have a long red light waiting time. This represents the max-min operation of the fuzzy matrix. Indicates that the diagonal elements are The diagonal matrix P + As an intermediate variable, D represents the matrix A column vector consisting of all elements on the main diagonal.

[0104] According to formula (2), determine the probability that no traffic violations will occur at each road intersection. :

[0105] (2)

[0106]

[0107]

[0108] in, This means that all road intersections starting from the current intersection will always satisfy the condition that no traffic violations will occur. express The maximum fixed-point function, Z represents the function The fixed point, Indicates that the diagonal elements are A diagonal matrix.

[0109] According to formula (3), determine the probability that each road intersection will eventually allow vehicles to pass at a relatively high speed. :

[0110] (3)

[0111] in, v indicates that at a certain moment, the vehicle will travel at a relatively high speed. Represents the identity matrix. Indicates that the diagonal elements are A diagonal matrix.

[0112] (7) Evaluate the quality of road intersections

[0113] Evaluate road intersections s according to formula (4) n Advantages and disadvantages :

[0114] (4)

[0115]

[0116]

[0117]

[0118] Where e represents the weight of the average vehicle speed. This indicates the probability that the vehicle will eventually travel at a faster speed, and f represents the weight of the average waiting time at the red light. This indicates the probability of a long wait time at the next intersection's red light, and g represents the weight of the traffic violation. This indicates the likelihood that the vehicle will not commit a traffic violation; E1 and E3 assess the probability of traffic violations at road intersections. n E2 is a positive indicator for evaluating road intersections. n A negative indicator.

[0119] (8) Predict the optimal path

[0120] 1) Determine the set C containing the subsequent road intersections.

[0121] Road intersections n The set C containing the subsequent road intersections is as follows:

[0122] C ,

[0123] Among them, s n+q Indicates road intersection s n The subsequent intersection.

[0124] 2) Determine the evaluation set D for subsequent road intersections.

[0125] For each road intersection in set C Calculate its evaluation value according to formula (4). The evaluation set D of the subsequent road intersections is as follows:

[0126] D

[0127] Where, f(s) n+q ) represents a road intersection s n+q The evaluation value of its merits and demerits.

[0128] 3) Determine the next road intersection the vehicle will pass through.

[0129] Select the evaluation set D with the largest evaluation value. The corresponding road intersection As the next road intersection the vehicle passes through.

[0130] 4) Output the optimal path

[0131] Compared with the probabilistic Kripke structure, this invention uses a generalized probabilistic Kripke model to model road traffic conditions. It takes road intersections as the state space, the smoothness of road segments as the path transition distribution matrix, and the average vehicle speed, average red light waiting time, and traffic violation index as label values. It proposes an optimal path prediction method based on the generalized probabilistic Kripke structure model, which solves the technical problem of computational difficulty and has the advantages of easy data acquisition and simple and convenient execution.

[0132] By doing this, the next optimal road intersection is selected step by step to obtain the local optimal path.

[0133] A vehicle route prediction method based on the generalized probability Kripke model was developed.

[0134] Example 2

[0135] The vehicle path prediction method based on the generalized probability Kripke model in this embodiment consists of the following steps:

[0136] (1) Constructing a real road condition model

[0137] The steps are the same as in Example 1.

[0138] (2) Construct the path transition distribution matrix

[0139] Construct the path transition distribution matrix P using the following formula:

[0140] P=

[0141] Among them, s m and s n Indicates a road intersection. Indicates road intersection s m With s n The path accessibility index between them The value is 0.1.

[0142] The other steps are the same as in Example 1.

[0143] A vehicle route prediction method based on the generalized probability Kripke model was developed.

[0144] Example 3

[0145] The vehicle path prediction method based on the generalized probability Kripke model in this embodiment consists of the following steps:

[0146] (1) Constructing a real road condition model

[0147] The steps are the same as in Example 1.

[0148] (2) Construct the path transition distribution matrix

[0149] Construct the path transition distribution matrix P using the following formula:

[0150] P=

[0151] Among them, s m and s n Indicates a road intersection. Indicates road intersection s m With s n The path accessibility index between them The value is 1.

[0152] The other steps are the same as in Example 1.

[0153] A vehicle route prediction method based on the generalized probability Kripke model was developed.

Claims

1. A vehicle route prediction method based on the generalized probability Kripke model, characterized in that... It consists of the following steps: (1) Constructing a real road condition model Collect road condition data between the starting point and the ending point, and the set S of all road intersections between the starting point and the ending point: S= in, Let S represent any element in set S, and n represent the number of road intersections; (2) Construct the path transition distribution matrix Construct the path transition distribution matrix P using the following formula: Among them, s m and s n Indicates a road intersection. Indicates road intersection s m With s n The path accessibility index between them ; (3) Quantify the initial traffic conditions The road intersection s is calculated using the following formula. n The initial traffic conditions are quantized into a column vector I. n : ， Among them, i n The first intersection the vehicle passes through is s, with 1 for all other elements and 0 for the rest. n ; (4) Constructing atomic propositions for vehicle route prediction The atomic propositions of vehicle routing planning include the average vehicle speed v, the average waiting time at traffic lights t, and the traffic violation index p; (5) Quantify label values Quantify the average passing speed label value of vehicles using the following formula : in, v represents the average speed of the vehicle. max This indicates all road intersections that have passed through within the last six months. n The vehicle's maximum speed, v min This indicates all road intersections that have passed through within the last six months. n The vehicle's minimum speed, quantified by the formula, represents the average waiting time label for red lights. : Where t represents the average waiting time at a red light, t max This indicates all road intersections that have passed through within the last six months. n The longest waiting time for a vehicle, t min This indicates all road intersections that have passed through within the last six months. n The shortest waiting time for vehicles is quantified by the following formula to determine the traffic violation index label value. : in, This represents the traffic violation index, where 'c' indicates the number of road intersections in the past six months. n The number of vehicles committing traffic violations, d represents the number of intersections s in the past six months. n Total number of vehicles passing through; (6) Perform model testing According to formula (1), determine the probability that each road intersection will have a longer red light waiting time than the next road intersection. : （1） in, This indicates that the next intersection to be reached from the current intersection will have a long red light waiting time. This represents the max-min operation of the fuzzy matrix. Indicates that the diagonal elements are The diagonal matrix P + As an intermediate variable, D represents the matrix A column vector consisting of all elements on the main diagonal; According to formula (2), determine the probability that no traffic violations will occur at each road intersection. : （2） in, This means that all road intersections starting from the current intersection will always satisfy the condition that no traffic violations will occur. express The maximum fixed-point function, Z represents the function The fixed point, Indicates that the diagonal elements are a diagonal matrix; According to formula (3), determine the probability that each road intersection will eventually allow vehicles to pass at a relatively high speed. : （3） in, v indicates that at a certain moment, the vehicle will travel at a relatively high speed. Represents the identity matrix. Indicates that the diagonal elements are a diagonal matrix; (7) Evaluate the quality of road intersections Evaluate road intersections s according to formula (4) n Advantages and disadvantages : （4） Where e represents the weight of the average vehicle speed. This indicates the probability that the vehicle will eventually travel at a faster speed, and f represents the weight of the average waiting time at the red light. This indicates the probability of a long wait time at the next intersection's red light, and g represents the weight of the traffic violation. This indicates the likelihood that the vehicle will not commit a traffic violation; E1 and E3 assess the probability of traffic violations at road intersections. n E2 is a positive indicator for evaluating road intersections. n Negative indicators; (8) Predict the optimal path 1) Determine the set C containing the subsequent road intersections. Road intersections n The set C containing the subsequent road intersections is as follows: C ， Among them, s n+q Indicates road intersection s n The subsequent intersection; 2) Determine the evaluation set D for subsequent road intersections. For each road intersection in set C Calculate its evaluation value according to formula (4). The evaluation set D of the subsequent road intersections is as follows: D Where, f(s) n+q ) represents a road intersection s n+q The evaluation value of its merits and demerits; 3) Determine the next road intersection the vehicle will pass through. Select the evaluation set D with the largest evaluation value. The corresponding road intersection As the next road intersection the vehicle passes through; 4) Output the optimal path By doing this, the next optimal road intersection is selected step by step to obtain the local optimal path.

2. The vehicle route prediction method based on the generalized probability Kripke model according to claim 1, characterized in that... The path transition distribution matrix constructed in step (2) is as follows: Construct the path transition distribution matrix P using the following formula: P= in Indicates road intersection s m With s n The path accessibility index between them The value is 0.5.

Images