A collision prediction method for intelligent connected hybrid vehicle fleet based on transfer function stability identification

Abstract

The present invention proposes a collision prediction method for an intelligent networked hybrid fleet based on transfer function stability identification, including constructing a traditional human-driven vehicle following model and a networked automatic driving vehicle following model; setting intelligent networked hybrid fleet scene parameters, and loading the traditional human-driven vehicle following model and the networked automatic driving vehicle following model into a microscopic traffic system; traversing the key parameters of the intelligent networked hybrid fleet, obtaining the fleet trajectory data under each set of parameters, and calculating the fleet's transfer function and its corresponding infinite norm at the same time, dividing the fleet's theoretical state according to the infinite norm, establishing a corresponding relationship between the fleet's theoretical state and actual state, and constructing an intelligent networked hybrid fleet collision prediction model; inputting the obtained actual state of the intelligent networked hybrid fleet into the intelligent networked hybrid fleet collision prediction model to predict its operating state. The present invention realizes the monitoring and early warning of the state of the intelligent networked hybrid fleet, and improves the safety of the fleet operation.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent transportation technology, and in particular relates to a collision prediction method for an intelligent networked hybrid vehicle fleet based on transfer function stability identification. Background Art

[0002] With the development of intelligent automobile technology, the main components of traffic flow on urban roads in the future will gradually transition from HV (Human Vehicles) to CAV (Connected and Automated Vehicles). During the transition process, road traffic flow will mainly exist in the form of mixed traffic flow composed of HV and CAV. The mixed traffic of connected autonomous vehicles and traditional human-driven vehicles has brought new safety issues: how to accurately predict the risk of rear-end collisions in mixed traffic flow scenarios. With the development of the Internet of Things and intelligent automobile technology, roadside RSU equipment can obtain and process the operating data of intelligent connected mixed fleets within the observation range. Therefore, it is an important task to improve road traffic safety to realize the prediction of mixed fleet collision risks by utilizing the dynamic information of mixed fleets.

[0003] At present, the research on traffic flow collision risk prediction mainly extracts macro traffic flow parameters from the aggregate data of the road section before the accident, and uses descriptive statistics of traffic flow parameters to establish the relationship between different traffic flow states and collision risks, but the research cannot understand the exact traffic changes near the accident site; or uses the micro trajectory data of two adjacent vehicles to construct safety alternative indicators for collision risk prediction, but the research cannot reflect the impact of the dynamic evolution of traffic flow on collision risk. Therefore, the results of current research cannot achieve the early prediction of collision risk in traffic flow. How to comprehensively consider the impact of the dynamic evolution of traffic flow on collision risk and the exact traffic changes at the collision location is crucial for the accurate prediction of collision risk of intelligent connected hybrid fleets. Summary of the invention

[0004] The technical problem to be solved by the present invention is: to propose a collision prediction method for an intelligent connected hybrid fleet based on transfer function stability identification, by dividing the road traffic flow into fleets, identifying the fleet status based on the transfer function of the intelligent connected hybrid fleet, and realizing the collision risk prediction of the fleet, so as to provide a theoretical basis for road traffic safety management and control.

[0005] The present invention adopts the following technical solutions to solve the above technical problems:

[0006] A collision prediction method for an intelligent connected hybrid vehicle fleet based on transfer function stability identification includes the following steps:

[0007] S1. In the intelligent connected hybrid vehicle fleet, build a traditional human-driven vehicle following model and a connected autonomous driving vehicle following model, classify the following behavior and determine the parameter range.

[0008] S2. Set the intelligent connected hybrid fleet scenario parameters, and load the traditional human-driven vehicle following model and the connected autonomous driving vehicle following model into the micro-transportation system.

[0009] S3. Traverse the key parameters of the intelligent connected hybrid fleet, obtain the fleet trajectory data under each set of parameters, and calculate the transfer function of the intelligent connected hybrid fleet and its corresponding infinite norm.

[0010] S4. According to the infinite norm of the intelligent connected hybrid fleet, the theoretical state of the fleet is divided, the corresponding relationship between the theoretical state and the actual state of the fleet is established, and a collision prediction model for the intelligent connected hybrid fleet is constructed.

[0011] S5. Input the obtained actual state of the intelligent connected hybrid fleet into the collision prediction model of the intelligent connected hybrid fleet to predict its operating state.

[0012] Furthermore, in step S1, constructing the model includes the following:

[0013] IDM (Intelligent Driver Model) is used to construct a traditional human-driven vehicle following model. The specific expression is:

[0014]

[0015]

[0016] Among them, a k (t+τ hv ) indicates that the kth vehicle is at t+τ hv The acceleration at the moment; τ hv Indicates driver reaction delay; a max Indicates the maximum acceleration; v k (t) represents the speed of the kth vehicle at time t; v f represents the free flow speed; s k (t) represents the inter-vehicle distance of the kth vehicle at time t; represents the expected inter-vehicle distance of the kth vehicle at time t; s0 represents the static safety distance; T represents the safety time interval; b represents the absolute value of the comfortable deceleration; v k-1 (t) represents the speed of the k-1th vehicle at time t, and vehicle k-1 is located in front of vehicle k.

[0017] The CACC (Cooperative Adaptive Cruise Control) model is used to build a car-following model for connected autonomous driving vehicles. The specific expression is:

[0018]

[0019] Among them, a k (t+τ cav ) indicates that the kth vehicle is at t+τ cav Acceleration at the moment, g d and g v is the feedback gain, t c represents the expected headway, τ cav represents the sum of communication delay and execution delay, and η represents the data sampling interval.

[0020] There are four types of car-following behaviors: CAV following HV, CAV following CAV, HV following HV, and HV following CAV. The parameter ranges of the four types of car-following behaviors are:

[0021] Furthermore, in step S2, determining the scene parameters includes the following:

[0022] Set the intelligent network hybrid fleet scenario parameters, including the road length L road , total running time t total , state update time step Δt; vehicle basic parameters include speed range [0,v max ], acceleration range Vehicle length car ; The size of the intelligent connected hybrid fleet is N platoon ; The disturbance duration of the leading vehicle is t disturbance ; The disturbance intensity of the leading vehicle is d disturbance ; The proportion of HVs in the intelligent connected hybrid fleet is p HV , the proportion of CAV is p CAV . Detection devices are deployed on the roadside to obtain vehicle trajectory information in real time.

[0023] Furthermore, in step S3, the key parameters of the intelligent connected hybrid fleet are traversed, including the following:

[0024] Smart connected hybrid fleet size Disturbance duration of the lead vehicle Disturbance intensity of the lead vehicle The proportion of CAV p CAV ∈[0,1] and the parameters of the four following behaviors.

[0025] For each set of key parameter combinations of the intelligent connected hybrid fleet, the actual state of the fleet is obtained by analyzing the trajectory data of each vehicle in the fleet. sim , respectively: stable state, unstable state, and collision state.

[0026] Furthermore, determine the actual state of the intelligent connected hybrid fleet sim Includes the following:

[0027] (1) Calculate the maximum speed fluctuation of each vehicle in the intelligent connected hybrid fleet If the maximum fluctuation amplitude gradually decreases along the convoy, it means that the convoy is in a stable state, then state sim =0:

[0028] ε1>ε2>…>ε k >…>ε N ;

[0029] Where k∈{1,2,,N} represents the number of the intelligent network hybrid fleet; N represents the number of the last vehicle in the intelligent network hybrid fleet; v t represents the speed of the intelligent connected hybrid fleet during operation; v eq Represents the equilibrium speed of the intelligent connected hybrid fleet.

[0030] (2) When the distance between vehicles is less than 0, it means that the convoy is in a collision state, then state sim =2.

[0031] (3) The rest of the cases are unstable. sim =1.

[0032] Furthermore, in step S3, calculating the transfer function and its corresponding infinite norm includes the following contents:

[0033] According to the parameters of the car-following behavior, the transfer function between the kth vehicle in the intelligent connected hybrid fleet and its front adjacent vehicle is expressed as:

[0034]

[0035] Among them, G k (λ) represents the transfer function of the kth vehicle; λ represents the Laplace transform operator; f v represents the partial derivative of the following behavior with respect to speed; f Δv represents the partial derivative of the following behavior with respect to the speed difference; f s represents the partial derivative of the following behavior with respect to the vehicle spacing; τ∈{τ cav ,τ hv} indicates the delay of CAV or HV.

[0036] According to different types of car-following behaviors, f in the transfer function v 、f Δv and f s There are two ways to calculate:

[0037] (1) The partial derivatives of the IDM describing the traditional human driving following behavior with respect to speed, speed difference, and vehicle spacing are expressed as follows:

[0038]

[0039] in, Indicates the equilibrium speed of the vehicle driven by a person; Indicates the balance distance between human-driven vehicles.

[0040] when Indicates HV following HV; when Indicates HV following CAV.

[0041] (2) The partial derivatives of the CACC model describing the car-following behavior of connected autonomous vehicles with respect to speed, speed difference, and vehicle spacing are expressed as follows:

[0042]

[0043] when Indicates that CAV follows HV; when Indicates CAV following CAV.

[0044] The head-to-tail transfer function of the intelligent connected hybrid fleet is the product of the transfer functions of all vehicles in the fleet except the head vehicle, and its expression is:

[0045]

[0046] Where T(λ) represents the head-to-tail transfer function.

[0047] Calculate the infinite norm corresponding to the head-tail transfer function of the intelligent connected hybrid fleet. The specific formula is:

[0048]

[0049] Where j represents the imaginary unit, ω represents the angular frequency, and T(λ) ∞ Represents the infinite norm corresponding to the head-to-tail transfer function.

[0050] Furthermore, in step S4, constructing a collision prediction model for an intelligent connected hybrid fleet includes the following sub-steps:

[0051] S401. According to the infinite norm of the intelligent networked hybrid fleet, the theoretical state of the fleet is theory Divided into three types:

[0052] (1) If T(λ) ∞ ≤1, then state theory =0, indicating theoretical stable state;

[0053] (2) If 1<T(λ) ∞ <μ, then state theory =1, indicating theoretical instability;

[0054] (3) If T(λ) ∞ ≥μ, then state theory =2, indicating the theoretical collision state;

[0055] Among them, μ represents the threshold between the theoretical instability state and the theoretical collision state.

[0056] S402: Using the threshold μ between the theoretical instability state and the theoretical collision state as the threshold for theoretical collision detection of the intelligent networked hybrid fleet, and setting the warning threshold μ warning It is 2 / 3 of the theoretical collision detection threshold and is used for early warning of collision risks.

[0057] S403. The final collision prediction model of the intelligent connected hybrid fleet is:

[0058]

[0059] Among them, 1 means there is a collision risk, and 0 means there is no collision risk.

[0060] Furthermore, determining the theoretical collision detection threshold includes the following:

[0061] Step 1: Use the grid search method to determine each group {N platoon ,t disturbance ,d disturbance Theoretical instability state and theoretical collision threshold under the condition

[0062] Step 2: Use the least squares method to fit the threshold and the size of the intelligent connected hybrid fleet N platoon , disturbance duration t disturbance , disturbance intensity d disturbance , the specific expression is:

[0063] μ=F(N platoon ,t disturbance ,d disturbance ,c1,c2,…,c n );

[0064] Among them, c i,i∈{1,2,…,n} represents the unknown parameters.

[0065] Furthermore, in step S5, the size of the intelligent networked hybrid fleet, the disturbance duration of the leading vehicle, and the disturbance intensity are input into the intelligent networked hybrid fleet collision prediction model to calculate the thresholds μ and μ warning , and the infinite norm of the head-tail transfer function of the intelligent connected hybrid fleet ||T(λ)|| ∞ In the input model, compare ||T(λ)|| ∞ and μ warning size, thereby predicting the operating status of the fleet.

[0066] The application process in actual traffic is as follows: the trajectory data of the intelligent connected hybrid fleet on the actual road is obtained through roadside detection equipment, the following model of each vehicle in the fleet is calibrated using a genetic algorithm, and the infinite norm ||T(λ)|| corresponding to the head-to-tail transfer function of the hybrid fleet is calculated. ∞ ; Wavelet analysis is used to obtain the disturbance duration and disturbance intensity of the head vehicle of the mixed convoy, and combined with the convoy size, the current collision threshold μ and collision risk threshold μ are determined warning ; If ||T(λ)|| ∞ ≥μ warning , it is determined that the mixed fleet has a collision risk.

[0067] Furthermore, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the above-mentioned method for predicting collision of an intelligent connected hybrid vehicle fleet based on transfer function stability identification are implemented.

[0068] Furthermore, the present invention also proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned intelligent connected hybrid vehicle fleet collision prediction method based on transfer function stability identification is executed.

[0069] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0070] The intelligent connected hybrid fleet collision prediction method based on transfer function stability identification proposed in the present invention divides the traffic flow into several fleets, determines the collision risk threshold of the fleet through the infinite norm of the transfer function of the intelligent connected hybrid fleet, and realizes the prediction of the collision risk of the fleet, which overcomes the shortcomings of traditional collision risk prediction methods that cannot simultaneously take into account the impact of the dynamic evolution of traffic flow on the collision risk and the exact traffic changes at the collision location. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is an overall implementation flow chart of the present invention.

[0072] Figure 2 Schematic diagram of four types of car-following behaviors in an embodiment of the present invention.

[0073] Figure 3 It is a schematic diagram of an intelligent connected hybrid vehicle fleet experimental platform in an embodiment of the present invention. DETAILED DESCRIPTION

[0074] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings.

[0075] The present invention proposes a collision prediction method for intelligent networked hybrid fleet based on transfer function stability identification. Figure 1 As shown, the following steps are included:

[0076] S1. Construct the traditional human-driven vehicle following model and the connected autonomous driving vehicle following model, classify the following behaviors and determine the parameter range. The specific contents are as follows:

[0077] IDM is used to construct a traditional human-driven vehicle following model. The specific expression is:

[0078]

[0079]

[0080] Among them, a k (t+τ hv ) indicates that the kth vehicle is at t+τ hv The acceleration at the moment; τ hv Indicates driver reaction delay; a max Indicates the maximum acceleration; v k (t) represents the speed of the kth vehicle at time t; v f represents the free flow speed; s k (t) represents the inter-vehicle distance of the kth vehicle at time t; represents the expected inter-vehicle distance of the kth vehicle at time t; s0 represents the static safety distance; T represents the safety time interval; b represents the absolute value of the comfortable deceleration; v k-1 (t) represents the speed of the k-1th vehicle at time t, and vehicle k-1 is located in front of vehicle k.

[0081] The CACC model is used to build a car-following model for connected autonomous driving vehicles. The specific expression is:

[0082]

[0083] Among them, a k (t+τ cav) indicates that the kth vehicle is at t+τ cav Acceleration at the moment, g d and g v represents the feedback gain, tc represents the expected headway, τ cav represents the sum of communication delay and execution delay, and η represents the data sampling interval.

[0084] In this embodiment, η is 0.1 seconds.

[0085] like Figure 2 As shown in the figure, car-following behaviors are divided into four categories: CAV following HV, CAV following CAV, HV following HV, and HV following CAV. The parameter ranges of the four types of car-following behaviors are:

[0086] S2. Set the intelligent networked hybrid fleet scenario parameters, and load the traditional human-driven vehicle following model and the networked autonomous driving vehicle following model into the micro-traffic system, such as Figure 3 The specific contents are as follows:

[0087] Set the intelligent network hybrid fleet scenario parameters, including the road length L road = 1000 meters, total running time t total = 2000 seconds, state update time step Δt = 0.1 seconds; basic vehicle parameters include the speed interval [0,33.3] m / s, that is, v max =33.3 m / s, acceleration interval [-5,5] m / s square, vehicle length l car =5 meters.

[0088] Figure 3 In the figure, the red vehicles are intelligent connected vehicles, the white vehicles are human-driven vehicles, and the yellow squares are roadside detection units, which draw maps of urban expressways and deploy detection devices on the roadside.

[0089] In the intelligent networked hybrid fleet simulation experiment platform, human-driven vehicles are controlled by IDM, and networked autonomous driving vehicles are controlled by CACC. The platform updates the vehicle's trajectory information (position, speed, acceleration, etc.) in real time. If the running time reaches the total running time t total Or if the convoy collides, it stops.

[0090] S3, traverse the key parameters of the intelligent networked hybrid fleet, obtain the fleet trajectory data under each set of parameters, and calculate the transfer function of the intelligent networked hybrid fleet and its corresponding infinite norm. The specific contents are as follows:

[0091] The size of the intelligent connected hybrid fleet is N platoon ∈[5,100], the traversal step length is 5; the disturbance duration of the leading vehicle is t disturbanc∈ e [1,10] seconds, the traversal step is 1; the disturbance intensity of the leading vehicle is d disturbance ∈[-0.5,-0.1] m / s2, the traversal step is 0.1; the ratio of CAV is p CAV ∈[0,1], the traversal step size is 0.1; the model parameters of the four types of car-following behaviors, the traversal step size of each parameter is 0.1.

[0092] For each set of key parameter combinations of the intelligent connected hybrid fleet, the actual state of the fleet is obtained by analyzing the trajectory data of each vehicle in the fleet. sim , respectively stable state, unstable state, collision state, the specific contents are:

[0093] (1) Calculate the maximum speed fluctuation of each vehicle in the intelligent connected hybrid fleet If the maximum fluctuation amplitude gradually decreases along the convoy, it means that the convoy is in a stable state, then state sim =0:

[0094] ε1>ε2>…>ε k >…>ε N ;

[0095] Where k∈{1,2,…,N} represents the number of the intelligent networked hybrid fleet; N represents the number of the last vehicle in the intelligent networked hybrid fleet; v t represents the speed of the intelligent connected hybrid fleet during operation; v eq Represents the equilibrium speed of the intelligent connected hybrid fleet.

[0096] (2) When the distance between vehicles is less than 0, it means that the convoy is in a collision state, then state sim =2.

[0097] (3) The rest of the cases are unstable. sim =1.

[0098] According to the parameters of the car-following behavior, the transfer function between the kth vehicle in the intelligent connected hybrid fleet and its front adjacent vehicle is expressed as:

[0099]

[0100] Among them, G k (λ) represents the transfer function of the kth vehicle; λ represents the Laplace transform operator; f v represents the partial derivative of the following behavior with respect to speed; f Δv represents the partial derivative of the following behavior with respect to the speed difference; f s represents the partial derivative of the following behavior with respect to the vehicle spacing; τ∈{τ cav ,τ hv} indicates the delay of CAV or HV.

[0101] According to different types of car-following behaviors, f in the transfer function v 、f Δv and f s There are two ways to calculate:

[0102] (1) The partial derivatives of IDM with respect to speed, speed difference, and workshop spacing are expressed as follows:

[0103]

[0104] in, Indicates the equilibrium speed of the vehicle driven by a person; Indicates the balance distance between human-driven vehicles.

[0105] when Indicates HV following HV; when Indicates HV following CAV.

[0106] (2) The partial derivatives of the CACC model with respect to speed, speed difference, and workshop spacing are expressed as follows:

[0107]

[0108] when Indicates that CAV follows HV; when Indicates CAV following CAV.

[0109] The head-to-tail transfer function of the intelligent connected hybrid fleet is the product of the transfer functions of all vehicles in the fleet except the head vehicle, and its expression is:

[0110]

[0111] Where T(λ) represents the head-to-tail transfer function.

[0112] Calculate the infinite norm corresponding to the head-tail transfer function of the intelligent connected hybrid fleet. The specific formula is:

[0113]

[0114] Where j represents the imaginary unit, ω represents the angular frequency, ||T(λ)|| ∞ Represents the infinite norm corresponding to the head-to-tail transfer function.

[0115] S4. According to the infinite norm of the intelligent networked hybrid fleet, the theoretical state of the fleet is divided, the corresponding relationship between the theoretical state and the actual state of the fleet is established, and the collision prediction model of the intelligent networked hybrid fleet is constructed. The specific contents are as follows:

[0116] S401. According to the infinite norm of the intelligent networked hybrid fleet, the theoretical state of the fleet is theory Divided into three types:

[0117] (1) If ||T(λ)|| ∞ ≤1, then state theory =0, indicating theoretical stable state;

[0118] (2) If 1<||T(λ)|| ∞ <μ, then state theory =1, indicating theoretical instability;

[0119] (3) If ||T(λ)|| ∞ ≥μ, then state theory =2, indicating the theoretical collision state;

[0120] Among them, μ represents the threshold between the theoretical instability state and the theoretical collision state.

[0121] Determining the theoretical collision detection threshold involves the following:

[0122] Step 1: Use the grid search method to determine each group {N platoon ,t disturbance ,d disturbance Theoretical instability state and theoretical collision threshold under the condition

[0123] Step 2: Use the least squares method to fit the threshold and the size of the intelligent connected hybrid fleet N platoon , disturbance duration t disturbance , disturbance intensity d disturbance , the specific expression is:

[0124] μ=F(N platoon ,t disturbance ,d disturbance ,c1,c2,…,c n );

[0125] Among them, c i ,i∈{1,2,…,n} represents the unknown parameters.

[0126] S402: Using the threshold μ between the theoretical instability state and the theoretical collision state as the threshold for theoretical collision detection of the intelligent networked hybrid fleet, and setting the warning threshold μ warning It is 2 / 3 of the theoretical collision detection threshold.

[0127] S403. The final collision prediction model of the intelligent connected hybrid fleet is:

[0128]

[0129] Among them, 1 means there is a collision risk, and 0 means there is no collision risk.

[0130] S5. Input the obtained actual state of the intelligent connected hybrid fleet into the collision prediction model of the intelligent connected hybrid fleet to predict its operating state.

[0131] The intelligent connected mixed fleet collision prediction method based on transfer function stability identification proposed in the present invention divides the traffic flow into several fleets, determines the collision risk threshold of the mixed fleet through the infinite norm of the mixed fleet transfer function, and realizes the prediction of the collision risk of the mixed fleet. It overcomes the deficiency that the traditional collision risk prediction method cannot simultaneously take into account the influence of the dynamic evolution of the traffic flow on the collision risk and the exact traffic changes at the collision location, so that the collision risk prediction result is more accurate and reliable, and it has practical engineering application value in improving the road accident risk warning capability.

[0132] The embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. It should be noted that when the processor executes the computer program, the specific steps of the method provided in the embodiment of the present invention are corresponding to the specific steps of the method provided in the embodiment of the present invention, and the processor has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided in the embodiment of the present invention.

[0133] The embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. It should be noted that when the computer program is executed by the processor, it corresponds to the specific steps of the method provided in the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided in the embodiment of the present invention.

[0134] The above description is only an exemplary embodiment of the present invention, and does not limit the scope of patent protection of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the scope of patent protection of the present invention.

Claims

1. A collision prediction method for intelligent connected hybrid vehicle fleet based on transfer function stability identification, characterized in that: The following steps are involved: S1. Construct the traditional human-driven vehicle following model and the connected autonomous driving vehicle following model, classify the following behaviors and determine the parameter range; the specific contents are as follows: The intelligent driver model is used to construct the traditional human-driven vehicle following model. The specific expression is: Among them, a k (t+τ hv ) indicates that the kth vehicle is at t+τ hv The acceleration at the moment; τ hv Indicates driver reaction delay; a max Indicates the maximum acceleration; v k (t) represents the speed of the kth vehicle at time t; v f represents the free flow speed; s k (t) represents the inter-vehicle distance of the kth vehicle at time t; represents the expected inter-vehicle distance of the kth vehicle at time t; s0 represents the static safety distance; T represents the safety time interval; b represents the absolute value of the comfortable deceleration; v k-1 (t) represents the speed of vehicle k-1 at time t, where vehicle k-1 is in front of vehicle k; The cooperative adaptive cruise control model is used to build a connected autonomous driving vehicle following model. The specific expression is: Among them, a k (t+τ cav ) indicates that the kth vehicle is at t+τ cav Acceleration at the moment, g d and g v is the feedback gain, t c represents the expected headway, τ cav represents the sum of communication delay and execution delay, η represents the data sampling interval; There are four types of car-following behaviors: CAV following HV, CAV following CAV, HV following HV, and HV following CAV. The parameter ranges of the four types of car-following behaviors are: Among them, CAV stands for connected autonomous vehicle, and HV stands for human-driven vehicle; S2. Set the scenario parameters of the intelligent networked hybrid fleet, and load the traditional human-driven vehicle following model and the networked autonomous driving vehicle following model into the micro-traffic system; the specific contents are as follows: Set the intelligent network hybrid fleet scenario parameters, including the road length L road , total running time t total , state update time step Δt; vehicle basic parameters include speed range [0,v max ], acceleration range Vehicle length car ; The size of the intelligent connected hybrid fleet is N platoon ; The disturbance duration of the leading vehicle is t disturbance ; The disturbance intensity of the leading vehicle is d disturbance ; The proportion of HVs in the intelligent connected hybrid fleet is pHV , the proportion of CAV is p CAV ; S3, traverse the key parameters of the intelligent networked hybrid fleet, obtain the fleet trajectory data under each set of parameters, and calculate the transfer function of the intelligent networked hybrid fleet and its corresponding infinite norm; the specific content is: Smart connected hybrid fleet size Disturbance duration of the lead vehicle Disturbance intensity of the lead vehicle The proportion of CAV p CAV ∈[0,1] and the parameters of the four following behaviors; For each set of key parameter combinations of the intelligent connected hybrid fleet, the actual state of the fleet is obtained by analyzing the trajectory data of each vehicle in the fleet. sim , respectively: stable state, unstable state, collision state; the specific contents are: (1) Calculate the maximum speed fluctuation of each vehicle in the intelligent connected hybrid fleet If the maximum fluctuation amplitude gradually decreases along the convoy, it means that the convoy is in a stable state, then state sim =0: e1>e2>…>e k >…>e N ; Where k∈{1,2,…,N} represents the number of the intelligent networked hybrid fleet; N represents the number of the last vehicle in the intelligent networked hybrid fleet; v t represents the speed of the intelligent connected hybrid fleet during operation; v eq represents the equilibrium speed of the intelligent connected hybrid fleet; (2) When the distance between vehicles is less than 0, it means that the convoy is in a collision state, then state sim =2; (3) The rest of the cases are unstable. sim =1; Computing the transfer function and its corresponding infinity norm involves the following: According to the parameters of the car-following behavior, the transfer function between the kth vehicle in the intelligent connected hybrid fleet and its front adjacent vehicle is expressed as: Among them, G k (λ) represents the transfer function of the kth vehicle; λ represents the Laplace transform operator; f v represents the partial derivative of the following behavior with respect to speed; f Δv represents the partial derivative of the following behavior with respect to the speed difference; f s represents the partial derivative of the following behavior with respect to the vehicle spacing; τ∈{τ cav ,τ hv } indicates the delay of CAV or HV; According to different types of car-following behaviors, f in the transfer function v 、f Δv and f s There are two ways to calculate: (1) The partial derivatives of the intelligent driver model with respect to speed, speed difference, and vehicle spacing are expressed as follows: in, Indicates the equilibrium speed of the vehicle driven by a person; Indicates the balance distance between human-driven vehicles; when Indicates HV following HV; when Indicates HV following CAV; (2) The partial derivatives of the cooperative adaptive cruise control model with respect to speed, speed difference, and vehicle spacing are expressed as follows: when Indicates that CAV follows HV; when Indicates CAV following CAV; The head-to-tail transfer function of the intelligent connected hybrid fleet is the product of the transfer functions of all vehicles in the fleet except the head vehicle, and its expression is: Where T(λ) represents the head-to-tail transfer function; Calculate the infinite norm corresponding to the head-tail transfer function of the intelligent connected hybrid fleet. The specific formula is: Where j represents the imaginary unit, ω represents the angular frequency, ||T(λ)|| ∞ represents the infinite norm corresponding to the head-to-tail transfer function; S4. According to the infinite norm of the intelligent networked hybrid fleet, the theoretical state of the fleet is divided, the corresponding relationship between the theoretical state and the actual state of the fleet is established, and a collision prediction model for the intelligent networked hybrid fleet is constructed; the specific contents are as follows: S401. According to the infinite norm of the intelligent networked hybrid fleet, the theoretical state of the fleet is theory Divided into three types: (1) If ||T(λ)|| ∞ ≤1, then state theory =0, indicating theoretical stable state; (2) If 1<||T(λ)|| ∞ <μ, then state theory =1, indicating theoretical instability; (3) If ||T(λ)|| ∞ ≥μ, then state theory =2, indicating the theoretical collision state; Among them, μ represents the threshold between the theoretical instability state and the theoretical collision state; S402: Using the threshold μ between the theoretical instability state and the theoretical collision state as the threshold for theoretical collision detection of the intelligent networked hybrid fleet, and setting the warning threshold μ warning It is 2 / 3 of the theoretical collision detection threshold; S403. The final collision prediction model of the intelligent connected hybrid fleet is: Among them, 1 indicates that there is a collision risk, and 0 indicates that there is no collision risk; S5. Input the obtained actual state of the intelligent connected hybrid fleet into the collision prediction model of the intelligent connected hybrid fleet to predict its operating state.

2. The intelligent networked hybrid vehicle fleet collision prediction method based on transfer function stability identification according to claim 1 is characterized in that: Determining the theoretical collision detection threshold involves the following: Step 1: Use the grid search method to determine each group {N platoon ,t disturbance ,d disturbance Theoretical instability state and theoretical collision threshold under the condition Step 2: Use the least squares method to fit the threshold and the size of the intelligent connected hybrid fleet N platoon , disturbance duration t disturbance , disturbance intensity d disturbance , the specific expression is: μ=F(N platoon ,t disturbance ,d disturbance ,c1,c2,…,c n ); Among them, c i ,i∈{1,2,…,n} represents the unknown parameters.

3. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 2 are implemented.

4. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 2 is performed.

Images