A lane change risk assessment and personalized lane change decision method

Abstract

This invention discloses a lane-changing risk assessment and personalized lane-changing decision-making method, comprising the following steps: Step S1, risk data perception; Step S2, lane-changing risk assessment based on minimum safe distance: including risk assessment based on the vehicle in front in the current lane, risk assessment based on the vehicle behind in the target lane, and risk assessment based on the vehicle in front in the target lane; Step S3, calculating the comprehensive lane-changing risk R based on the lane-changing risk assessment results obtained in Step S2; Step S4, calculating the personalized comprehensive lane-changing risk threshold R based on the driver's driving ability and driving style data. Q Step S5: Combine the overall lane change risk R with the personalized overall lane change risk threshold R. Q By comparing different approaches, personalized lane-change decisions are achieved. This invention makes lane-change risk assessment more accurate, and lane-change decisions can meet the driver assistance needs of drivers with different driving styles and abilities.

Description

Technical Field

[0001] This invention belongs to the field of intelligent driving technology and relates to a method for lane change risk assessment and personalized lane change decision-making. Background Technology

[0002] Currently, intelligent vehicle driver assistance systems based on driver habits can reduce human-machine conflicts and other problems during the assistance process, effectively improving the adaptability of intelligent driver assistance systems, and are one of the current research hotspots in the field of intelligent driving.

[0003] In the current lane-changing decision-making process, there are mostly certain problems in assessing the feasibility of lane changes: First, when assessing risks, the risks of each traffic participant are evaluated and compared one by one, ignoring the role of each traffic participant in the overall risk; Second, most of them only consider the influence of the driver's driving style and ignore the role of the driver's driving ability in the decision-making process.

[0004] Therefore, it is evident that existing risk-assessment-based personalized lane change assist systems and related technologies are insufficient to meet the current personalized needs for lane change assist systems. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a lane change risk assessment and personalized lane change decision-making method, which makes lane change risk assessment more accurate and lane change decisions can meet the assisted driving needs of drivers with different driving styles and abilities.

[0006] The technical solution adopted in this invention is a method for lane-changing risk assessment and personalized lane-changing decision-making, comprising the following steps:

[0007] Step S1: Risk Data Perception;

[0008] Step S2, Lane change risk assessment based on minimum safe distance: including risk assessment based on the vehicle in front in the current lane, risk assessment based on the vehicle behind in the target lane, and risk assessment based on the vehicle in front in the target lane;

[0009] Step S3: Calculate the comprehensive lane-changing risk R based on the lane-changing risk assessment results obtained in Step S2;

[0010] Step S4: Calculate the personalized comprehensive lane-changing risk threshold R based on the driver's driving ability and driving style data. Q ;

[0011] Step S5: Combine the overall lane change risk R with the personalized overall lane change risk threshold R. Q By comparing, personalized lane-changing decisions can be made.

[0012] Further, step S1 specifically refers to the driving data acquired by the sensors mounted on the intelligent connected vehicle for sensing the driving environment around the vehicle; the driving data includes the main vehicle data and surrounding vehicle data, wherein:

[0013] The main vehicle data includes the main vehicle's length, width, lateral coordinates, longitudinal coordinates, speed, and maximum braking deceleration; the surrounding vehicle data includes the length, width, speed, maximum braking deceleration, lateral coordinates, and longitudinal coordinates of each surrounding vehicle.

[0014] Furthermore, the expression for the risk assessment based on the vehicle ahead in the current lane in step S2 is as follows:

[0015]

[0016] Among them, C L0 R is the car-following safety distance factor. L0 ∈[0,1] represents the risk of following the carousel, R L0 The larger the value, the higher the risk. a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant.

[0017] The following safety distance coefficient C L0 The calculation method is as follows:

[0018]

[0019] Among them, S1 is the safe following distance for the main vehicle, S L0_actual The current actual longitudinal distance between the main vehicle and the vehicle in front in the current lane;

[0020] The safe following distance S1 of the main vehicle is calculated as follows:

[0021]

[0022] Among them, v M Main vehicle speed, t h a is the sum of the driver's reaction time and the time from when the driver applies the brakes to when deceleration begins. Mmax The maximum deceleration that the main vehicle can achieve, v L0 a is the initial braking speed of the vehicle in front in the current lane. L0max L represents the maximum braking deceleration of the vehicle in front in the current lane. L0 For safety margin.

[0023] Furthermore, the expression for the risk assessment based on the vehicle following in the target lane in step S2 is as follows:

[0024]

[0025] Among them, C Fd R is the lane-changing safety distance coefficient based on the following vehicle in the target lane. Fd ∈[0,1] represents the lane-changing risk based on the vehicle following in the target lane, R Fd The larger the value, the higher the risk. a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant.

[0026] The lane-changing safety distance coefficient C based on the vehicle behind in the target lane Fd The calculation method is as follows:

[0027]

[0028] Where S2 is the minimum safe lane-changing distance based on the vehicle behind in the target lane, S Fd_actual This represents the actual longitudinal distance between the driver vehicle and the vehicle following in the target lane at the initial moment of lane change.

[0029] The minimum safe lane-changing distance S2 for vehicles following in the target lane is calculated as follows:

[0030]

[0031] Among them, v Fd t represents the speed of the vehicle following in the target lane. lc The time required for the main vehicle to change lanes, v M Main vehicle speed, t b a is the reaction time of the driver of the vehicle following in the target lane. Fdmax a is the maximum braking deceleration that the vehicle behind can produce in the target lane. Mmax The maximum braking deceleration that the main vehicle can achieve, LF d为 Safety margin.

[0032] Furthermore, the expression for the risk assessment based on the vehicle ahead in the target lane in step S2 is as follows:

[0033]

[0034] Among them, C Ld R is the lane-changing safety distance coefficient based on the vehicle ahead in the target lane. Ld ∈[0,1] represents the lane-changing risk based on the vehicle ahead in the target lane, R Ld The larger the value, the higher the risk. a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant.

[0035] The lane-changing safety distance coefficient C based on the vehicle ahead in the target lane Ld The calculation method is as follows:

[0036]

[0037] Where S3 is the minimum safe lane-changing distance based on the vehicle ahead in the target lane, S Ld_actual This represents the actual longitudinal distance between the driver vehicle and the vehicle in front in the target lane at the initial moment of lane change.

[0038] The calculation method for the minimum safe lane-changing distance S3 based on the vehicle in front in the target lane is as follows:

[0039]

[0040] Among them, v M Main vehicle speed, t lc The time required for the main vehicle to change lanes, v Ld t represents the speed of the vehicle ahead in the target lane. h a is the sum of the driver's reaction time and the time from when the driver applies the brakes to when deceleration begins. Mmax The maximum braking deceleration that the main vehicle can achieve, a Ldmax L represents the maximum braking deceleration of the vehicle ahead in the target lane. Ld For safety margin.

[0041] Furthermore, the comprehensive lane-changing risk R in step S3 is:

[0042] R = R max +(1-R max (w1R) min +w2R L0 )

[0043] Among them, R max and R min The risk assessment results R based on the vehicle ahead in the target lane are respectively Ld Compared with the risk assessment results of vehicles following in the target lane R Fd The larger and smaller values ​​between them, R L0 The risk assessment results of the vehicle in front in the current lane, w1 and w2 are risk weighting factors, both belonging to [0,1].

[0044] Further, step S4 specifically includes:

[0045] Step S41: Based on the driving data of the driver's driving style, the driving data of the driver's driving style is reduced in dimensionality by using principal component analysis to reduce the amount of computation. Then, the K-means clustering method is used to divide it into three clusters. The cluster centers of the three clusters correspond to the corresponding comprehensive lane change risk thresholds of aggressive, normal and conservative driving styles, respectively.

[0046] Step S42: Based on the driving data of the driver's driving ability, establish a risk exposure level model using the minimum safe distance model. Based on the risk exposure level model, use quantile regression analysis to evaluate the driver's driving ability and divide it into a comprehensive lane change risk threshold for skilled, normal, and unfamiliar driving abilities.

[0047] Step S43: Weight and fuse the comprehensive lane-change risk threshold of driving style and the comprehensive lane-change risk threshold of driving ability to obtain the personalized comprehensive lane-change risk threshold R. Q .

[0048] Furthermore,

[0049] The risk exposure level model is specifically implemented as follows: When a vehicle is traveling and its actual distance is less than the minimum safe distance, the average ratio of the difference between the minimum safe distance and the actual distance to the minimum safe distance is considered the risk exposure level. The specific expression is as follows:

[0050]

[0051] Among them, P severity d represents the risk exposure level value, which is a constant greater than 0. mini Let d be the minimum safe distance obtained in the i-th calculation. i Let be the distance actually maintained in the i-th iteration, and n be the number of sampling iterations that satisfy the condition.

[0052] Further, step S5 specifically includes:

[0053] When the overall lane change risk R is greater than the personalized overall lane change risk threshold R Q If the overall lane-changing risk R is less than or equal to the personalized overall lane-changing risk threshold R, a lane-changing risk warning will be issued and the lane-changing behavior will be terminated. Q If necessary, continue changing lanes.

[0054] The beneficial effects of this invention are

[0055] (1) Based on the minimum safe distance, the risk normalization process is performed using the Sigmoid function, which makes the risk assessment more intuitive and convenient.

[0056] (2) Based on the driving risks of surrounding vehicles to the main vehicle, a fusion strategy was designed, which comprehensively considers the impact of various traffic participants on the overall lane change risk, avoiding the one-sidedness of individual successive participants' judgments and multiple redundant judgments.

[0057] (3) Driving ability factors were incorporated into the personalized lane-changing comprehensive risk threshold assessment. This improved the accuracy of the comprehensive risk threshold assessment, enhanced the adaptability and reliability of the lane-changing decision system, and reduced the false alarm rate of the system. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a framework diagram of the lane-changing risk assessment and personalized lane-changing decision-making method according to an embodiment of the present invention.

[0060] Figure 2 This is a schematic diagram of the lane-changing process according to an embodiment of the present invention.

[0061] Figure 3 This is a diagram showing the lane change risk assessment results of an embodiment of the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] This invention provides a lane-changing risk assessment method, see [link to relevant documentation]. Figure 1 Intelligent connected vehicles are equipped with sensors to perceive the driving environment around the vehicle, acquiring driving data including data from the main vehicle and surrounding vehicles. A minimum safe distance model is then used to calculate the risk posed by surrounding vehicles to the main vehicle's lane-changing behavior.

[0064] The main vehicle data includes vehicle length, vehicle width, lateral coordinates, longitudinal coordinates, speed, acceleration, and maximum braking deceleration; the surrounding vehicle data includes the length, width, speed, maximum braking deceleration, lateral coordinates, and longitudinal coordinates of each surrounding vehicle.

[0065] First, the minimum safe distances to the vehicle in front in the current lane, the vehicle in front in the target lane, and the vehicle behind in the target lane are calculated using perception information. Then, the ratio of the minimum safe distance to the actual distance to the vehicle is used to perform risk normalization using the Sigmoid function, yielding the risk assessment result.

[0066] The specific implementation process of this embodiment is as follows:

[0067] Risk assessment based on the vehicle ahead in the current lane

[0068] Figure 2 This is a schematic diagram of the lane-changing process of the present invention. In car-following scenarios, as long as the following vehicle's speed decreases to match the speed of the vehicle in front, a rear-end collision can be avoided, ensuring following safety. In extreme cases, the risk of collision is greatest when the vehicle in front abruptly stops. Furthermore, the minimum safe distance calculated under this assumption is relatively conservative. During this process, the distance traveled by the vehicle in front is:

[0069]

[0070] Among them, v L0 a is the initial braking speed of the vehicle in front in the current lane. L0max This represents the maximum braking deceleration of the vehicle ahead in the current lane. For the driver vehicle, the distance it travels from the moment the vehicle ahead begins braking until it comes to a complete stop is:

[0071]

[0072] Where v M Main vehicle speed, t h a is the sum of the driver's reaction time and the time from when the driver applies the brakes to when deceleration begins. Mmax The maximum braking deceleration that the main vehicle can achieve.

[0073] Therefore, the safe following distance for the main vehicle in car-following conditions is:

[0074]

[0075] Where L L0 For safety margin.

[0076] The Sigmoid function is used to normalize the car-following risk assessment:

[0077]

[0078]

[0079] Among them, C L0 R is the car-following safety distance factor. L0 ∈[0,1] represents the risk of following the carousel, R L0 The higher the value of S, the higher the risk. L0_actual is the current actual longitudinal distance between the main vehicle and the vehicle in front in the current lane, a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant.

[0080] Risk assessment based on vehicles following in the target lane

[0081] During the lane change, the distance the main vehicle travels forward is:

[0082] SMbd =v M t lc

[0083] Among them, t lc The time required for the main vehicle to change lanes.

[0084] Meanwhile, the distance that the vehicle behind in the target lane travels forward is:

[0085] S Fd =v Fd t lc

[0086] Where v Fd The speed of the vehicle following in the target lane.

[0087] In addition, to prevent rear-end collisions between vehicles in the target lane and the main vehicle after a lane change, vehicles in the target lane should maintain a safe following distance after the main vehicle successfully completes the lane change. Therefore, at the initial moment of the lane change, the minimum safe lane-changing distance based on the following vehicle in the target lane is:

[0088]

[0089] Among them, t b a is the reaction time of the driver of the vehicle following in the target lane. Fdmax L represents the maximum braking deceleration that a vehicle following in the target lane can produce. Fd For safety margin.

[0090] The Sigmoid function is used to normalize the car-following risk assessment:

[0091]

[0092]

[0093] Among them, C Fd R is the lane-changing safety distance coefficient based on the following vehicle in the target lane. Fd ∈[0,1] represents the lane-changing risk based on the vehicle following in the target lane, R Fd The higher the value of S, the higher the risk. Fd_actual This represents the actual longitudinal distance between the main vehicle and the vehicle following in the target lane at the initial moment of lane change.

[0094] Risk assessment based on the vehicle ahead in the target lane

[0095] During the lane change process, the distance the vehicle in front in the target lane travels forward is:

[0096] S Ld =v Ld t lc

[0097] Where v Ld The speed of the vehicle ahead in the target lane.

[0098] In addition, to prevent the main vehicle from rear-ending the vehicle in the target lane after changing lanes, a safe following distance should be maintained between the main vehicle and the vehicle in the target lane after the main vehicle successfully completes the lane change. Therefore, at the initial moment of the lane change, the minimum safe lane change distance based on the vehicle in the target lane is:

[0099]

[0100] Among them, a Ldmax L represents the maximum braking deceleration that the vehicle ahead in the target lane can produce. Ld For safety margin.

[0101] The Sigmoid function is used to normalize the car-following risk assessment:

[0102]

[0103]

[0104] Among them, C Ld R is the lane-changing safety distance coefficient based on the vehicle ahead in the target lane. Ld ∈[0,1] represents the lane-changing risk based on the vehicle ahead in the target lane, R Ld The higher the value of S, the higher the risk. Ld_actual This represents the actual longitudinal distance between the main vehicle and the vehicle in front in the target lane at the initial moment of lane change.

[0105] Based on the three risk assessment results obtained, this invention designs a comprehensive lane-changing risk fusion strategy that considers the impact of each traffic participant on the overall risk, and assesses the current comprehensive lane-changing risk.

[0106] Integrated Lane Change Risk Integration Strategy

[0107] During lane changing, the majority of the risk stems from collisions between the vehicle in front or behind in the target lane and the magistrate vehicle, while the risk of following the vehicle in the current lane has a relatively small impact. Therefore, when designing lane-changing risk fusion strategies, greater emphasis should be placed on risk assessment from vehicles in front or behind the target lane.

[0108] The comprehensive lane-change risk fusion strategy is implemented as follows:

[0109] R max =max{R Ld R Fd}

[0110] R min =min{R Ld R Fd}

[0111] Among them, R max and R min R respectively Ld With R Fd The larger and smaller values ​​between

[0112] Its comprehensive lane-changing risk assessment is as follows:

[0113] R = R max +(1-R max (w1R) min +w2R L0 )

[0114] w1 and w2 are risk weighting factors, both belonging to [0,1].

[0115] This invention also provides a personalized lane-changing decision-making method, based on the aforementioned comprehensive vehicle lane-changing risk assessment method, including: selecting a driving style and driving ability, and comparing the comprehensive lane-changing risk R experienced by the main vehicle with the personalized comprehensive lane-changing risk threshold R corresponding to the driving style and driving ability. Q The comparison is made to determine whether the main vehicle should continue to change lanes.

[0116] Specifically, the overall lane-changing risk R experienced by the main vehicle is compared with the personalized overall lane-changing risk threshold R corresponding to the driving style and driving ability. Q The method for determining whether the main vehicle should continue lane changing is based on a comparison: when the overall lane-changing risk R is greater than the personalized overall lane-changing risk threshold R... Q If the overall lane-changing risk R is less than or equal to the personalized overall lane-changing risk threshold R, a lane-changing risk warning will be issued and the lane-changing behavior will be terminated. Q If necessary, continue changing lanes.

[0117] Furthermore, the method for determining the lane-changing risk threshold corresponding to driving style and driving ability is as follows: on the one hand, based on the driving data of the driver's driving style, the driving data of the driver's driving style is reduced in dimensionality using principal component analysis to reduce the amount of computation, and then the K-means clustering method is used to divide it into three clusters. The cluster centers of the three clusters correspond to the corresponding comprehensive lane-changing risk thresholds for aggressive, normal and conservative driving styles, respectively.

[0118] On the other hand, based on driving data of the driver's driving ability, a risk exposure level model is established using the minimum safe distance model. The risk exposure level model is specifically implemented as follows: when the actual distance maintained by the vehicle is less than the minimum safe distance during the driving process, the average ratio of the difference between the minimum safe distance and the actual distance to the minimum safe distance is the risk exposure level.

[0119]

[0120] Among them, P severity d represents the risk exposure level value, which is a constant greater than 0. mini Let d be the minimum safe distance obtained in the i-th calculation. i Let be the distance actually maintained in the i-th iteration, and n be the number of sampling iterations that satisfy the condition.

[0121] Then, based on the risk exposure level model, quantile regression analysis is used to assess the driver's driving ability, which is divided into skilled, normal, and unfamiliar driving abilities, and a comprehensive lane-changing risk threshold is set.

[0122] Finally, the comprehensive lane-changing risk thresholds based on driving style and driving ability are weighted and fused to obtain the personalized comprehensive lane-changing risk threshold R. Q .

[0123] Example

[0124] Based on the Ubiquitous Traffic Eyes open-source dataset, suitable natural vehicle driving data and the driving data of vehicles in front of and behind the current lane and the target lane were extracted. Then, based on the aforementioned lane change risk assessment method, Matlab was used to evaluate and analyze the following risk of the main vehicle, the lane change risk of the vehicle in front in the target lane, the lane change risk of the vehicle behind in the target lane, and the overall lane change risk. The simulation results are as follows: Figure 3 As shown.

[0125] like Figure 3 The simulation results show that the simulation results of the following risk of the main vehicle, the lane-changing risk based on the preceding vehicle in the target lane, the lane-changing risk based on the following vehicle in the target lane, and the comprehensive lane-changing risk are all within the [0,1] interval throughout the process. This is more intuitive and concise, and is conducive to determining the comprehensive lane-changing risk threshold in the subsequent lane-changing feasibility assessment.

[0126] Through the Figure 3 The analysis shows that when the risk of the main vehicle following, the risk of the vehicle in front changing lanes in the target lane, and the risk of the vehicle behind changing lanes in the target lane are relatively small, the comprehensive lane-changing risk successfully takes into account the impact of multiple lane-changing risks. The result is consistent with the actual situation of drivers taking conservative driving behavior when faced with multiple sources of low to medium lane-changing risks during driving.

[0127] Additionally Figure 3 It is evident that by integrating three different lane-changing risks to obtain a comprehensive lane-changing risk, simplicity is achieved in subsequent lane-changing feasibility decisions, avoiding redundant judgment structures. Furthermore, this approach better aligns with the actual situation where drivers consider the vehicle's current overall driving status rather than comparing and judging each risk objective individually when making lane-changing decisions.

[0128] Compared with existing lane change risk assessment methods, and especially with the Time-To-Collision (TTC) theory, the assessment method proposed in this invention takes into account driver factors more comprehensively, such as driver reaction time and lane change time. Compared with lane change risk assessment methods based on risk field theory, the assessment method proposed in this invention has the advantages of low computational load and strong real-time performance.

[0129] In this invention, all vehicles in this embodiment refer to intelligent connected vehicles, specifically vehicles equipped with onboard sensors (including cameras inside and outside the vehicle), controllers, actuators, and other devices, capable of real-time data interaction with vehicles in other lanes. In practical applications, for example, if it is necessary to know the driver's reaction time, it is sufficient to record the difference between the braking time of the vehicle in front and the driver's braking time. To ensure data accuracy, multiple recorded data can be averaged, and the data can be updated in real-time as the number of records increases to ensure real-time performance.

[0130] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for lane-changing risk assessment and personalized lane-changing decision-making, characterized in that, Includes the following steps: Step S1: Risk Data Perception; Step S2, Lane change risk assessment based on minimum safe distance: including risk assessment based on the vehicle in front in the current lane, risk assessment based on the vehicle behind in the target lane, and risk assessment based on the vehicle in front in the target lane; Step S3: Calculate the comprehensive lane-changing risk based on the lane-changing risk assessment results obtained in Step S2. ; Step S4: Calculate the personalized comprehensive lane-changing risk threshold based on the driver's driving ability and driving style data. ; Step S5: Integrate lane change risks With personalized comprehensive lane change risk threshold Comparison enables personalized lane-changing decisions; The expression for the risk assessment based on the vehicle ahead in the current lane in step S2 is: ； in, To determine the car-following safety distance coefficient, ∈[0,1] represents the risk of following the lead. The larger the value, the higher the risk. a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant. The following safety distance coefficient The calculation method is as follows: ； in, The safe following distance for the main vehicle The current actual longitudinal distance between the main vehicle and the vehicle in front in the current lane; The safe following distance of the main vehicle The calculation method is as follows: ； in, The speed of the main vehicle. It is the sum of the driver's reaction time and the time from when the driver applies the brakes to when deceleration begins. The maximum deceleration that the main vehicle can achieve. The initial braking speed of the vehicle in front in the current lane. The maximum braking deceleration of the vehicle in front in the current lane. For safety margin; The expression for the risk assessment based on the vehicle following in the target lane in step S2 is: ； in, This is based on the lane-changing safety distance coefficient of the vehicle following in the target lane. ∈[0,1] represents the lane-changing risk based on the following vehicle in the target lane. The larger the value, the higher the risk. a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant. The lane-changing safety distance coefficient based on the vehicle behind in the target lane The calculation method is as follows: ； in, To determine the minimum safe lane-changing distance based on the vehicle following in the target lane. This represents the actual longitudinal distance between the driver vehicle and the vehicle following in the target lane at the initial moment of lane change. The minimum safe lane-changing distance for vehicles following in the target lane. The calculation method is as follows: ； in, The speed of the vehicle behind in the target lane. The time required for the main vehicle to change lanes. The speed of the main vehicle. For the reaction time of the driver of the vehicle following in the target lane, The maximum braking deceleration that a vehicle following in the target lane can produce. The maximum braking deceleration that the main vehicle can achieve. For safety margin; The expression for the risk assessment based on the vehicle ahead in the target lane in step S2 is: ； in, Based on the lane-changing safety distance coefficient of the vehicle ahead in the target lane, This indicates the risk of lane changing based on the vehicle ahead in the target lane. The larger the value, the higher the risk. a and b are the adjustment parameters of the Sigmoid function, is a constant, and e is the natural constant. The lane-changing safety distance coefficient based on the vehicle ahead in the target lane The calculation method is as follows: ； in, To determine the minimum safe lane-changing distance based on the vehicle ahead in the target lane, This represents the actual longitudinal distance between the driver vehicle and the vehicle in front in the target lane at the initial moment of lane change. The minimum safe lane-changing distance based on the vehicle ahead in the target lane The calculation method is as follows: ； in, The speed of the main vehicle. The time required for the main vehicle to change lanes. The speed of the vehicle ahead in the target lane. It is the sum of the driver's reaction time and the time from when the driver applies the brakes to when deceleration begins. The maximum braking deceleration that the main vehicle can achieve. The maximum braking deceleration of the vehicle ahead in the target lane. For safety margin; In step S3, the overall lane-changing risk is considered. for: ； in, and These are risk assessment results based on the vehicle ahead in the target lane. Risk assessment results based on the vehicle behind in the target lane The larger and smaller values ​​between Risk assessment results of the vehicle ahead in the current lane. and These are risk weighting factors, all belonging to [0,1].

2. The lane-changing risk assessment and personalized lane-changing decision-making method according to claim 1, characterized in that, Step S1 specifically refers to the driving data acquired by the sensors on the intelligent connected vehicle used to perceive the driving environment around the vehicle; the driving data includes the main vehicle data and the surrounding vehicle data, wherein: The main vehicle data includes the main vehicle's length, width, lateral coordinates, longitudinal coordinates, speed, and maximum braking deceleration; the surrounding vehicle data includes the length, width, speed, maximum braking deceleration, lateral coordinates, and longitudinal coordinates of each surrounding vehicle.

3. The lane-changing risk assessment and personalized lane-changing decision-making method according to claim 1, characterized in that, Step S4 specifically involves: Step S41: Based on the driving data of the driver's driving style, the driving data of the driver's driving style is reduced in dimensionality by using principal component analysis to reduce the amount of computation. Then, the K-means clustering method is used to divide it into three clusters. The cluster centers of the three clusters correspond to the corresponding comprehensive lane change risk thresholds of aggressive, normal and conservative driving styles, respectively. Step S42: Based on the driving data of the driver's driving ability, establish a risk exposure level model using the minimum safe distance model. Based on the risk exposure level model, use quantile regression analysis to evaluate the driver's driving ability and divide it into a comprehensive lane change risk threshold for skilled, normal, and unfamiliar driving abilities. Step S43: Weight and fuse the comprehensive lane-changing risk threshold based on driving style and comprehensive lane-changing risk threshold based on driving ability to obtain a personalized comprehensive lane-changing risk threshold. .

4. The lane-changing risk assessment and personalized lane-changing decision-making method according to claim 3, characterized in that, The risk exposure level model is specifically implemented as follows: When a vehicle is traveling and its actual distance is less than the minimum safe distance, the average ratio of the difference between the minimum safe distance and the actual distance to the minimum safe distance is considered the risk exposure level. The specific expression is as follows: ； in, The risk exposure value is a constant greater than 0. This represents the minimum safe distance obtained from the i-th calculation. Let be the distance actually maintained in the i-th iteration, and n be the number of sampling iterations that satisfy the condition.

5. The lane-changing risk assessment and personalized lane-changing decision-making method according to claim 1, characterized in that, Step S5 specifically involves: When considering the risks of lane changing Greater than the personalized comprehensive lane change risk threshold If this occurs, a lane-changing risk warning will be issued, and the lane-changing behavior will be terminated. When considering the risks of lane changing Less than or equal to the personalized comprehensive lane change risk threshold If necessary, continue changing lanes.

Images