A method for calculating recommended values of tunnel and interchange exit net distance length

Abstract

This invention provides a method for calculating the recommended clearance length between tunnels and interchange exits, comprising the following steps: constructing a VISSIM simulation model; obtaining the headway, speed, and instantaneous traffic flow density at different cross-sections in the model, and fitting the models to obtain headway models and speed distribution models for different road segments, constructing speed and density models and critical intercalation gap models; constructing a safe lane-changing probability model using the differential method; solving the safe lane-changing probability model based on the Monte Carlo simulation method according to various uncertainties, obtaining the target lane-changing success rate under different traffic volumes, large vehicle ratios, and clearance lengths, and providing a recommended clearance length value. This invention uses a deterministic approach to calculate the problem of large clearance values ​​required under traffic capacity, and utilizes a reliability design method to integrate traffic characteristics, driver characteristics, and traffic safety requirements of small clearance road segments, proposing a method for calculating the recommended clearance length value.

Description

Technical Field

[0001] This invention relates to the field of road traffic safety technology, specifically to a method for calculating the recommended value of the net distance between tunnels and interchange exits. Background Technology

[0002] Given my country's topography, which slopes from west to east, the mountainous terrain and geological conditions in the western region are often complex, with significant elevation differences between mountains. This makes highway construction challenging in some sections, leading to denser tunnel distribution due to limited resources. Sometimes, for reasons of alignment, cost, and regional connectivity, interchanges are placed near tunnels, resulting in excessively close distances between tunnels and interchanges. The need for adapting to changes in light and darkness at tunnel entrances and facilitating lane changes severely disrupts traffic flow in these short-clearance sections. However, the lack of clear requirements for the combined design of clearance indicators and traffic flow conditions presents a challenge for designers in the near and medium term, particularly when traffic volume is below capacity. This leads to difficulties in ensuring safety when dealing with short-clearance sections at exits, and a lack of adequate traffic safety measures to address the specific driving characteristics of these sections. Therefore, it is necessary to conduct a more in-depth study of the traffic characteristics of these short-clearance sections between tunnels and interchange exits, further refining the selection of clearance lengths from the perspective of traffic flow conditions, and providing a new method for calculating recommended clearance lengths.

[0003] The clearance length in the "Detailed Rules for the Design of Highway Grade Separation Intersections" is calculated in segments, including the adaptation process, identification process, lane-changing process, and safety confirmation process. The values ​​used in each part are all "deterministic". Therefore, it is generally believed that a lane change can only be safely completed when the clearance length of the exit section meets the standard value. However, in reality, VISSIM's research on the traffic characteristics of small clearance at exits shows that when the clearance length is less than the standard value, most diverting vehicles can still safely and smoothly exit the main line. Therefore, the clearance length obtained by using the "deterministic" approach cannot reflect the probability of vehicles exiting under different traffic flow conditions on small clearance sections. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides a method for calculating recommended clearance lengths between tunnels and interchange exits based on a reliability algorithm. This method integrates traffic characteristics, driver characteristics, and traffic safety requirements for road sections with small clearances, aiming to provide recommended clearance lengths under different traffic volumes and large vehicle ratios.

[0005] To achieve the above objectives, the present invention provides the following technical solution.

[0006] A method for calculating the recommended value of the net distance between a tunnel and an interchange exit includes the following steps:

[0007] A VISSIM simulation model of the tunnel and interchange exits was constructed and calibrated using measured data.

[0008] The headway, speed, and instantaneous traffic flow density of vehicles at different cross sections in the VISSIM simulation model were obtained, and statistical fitting was performed to obtain headway models and speed distribution models for different road segments.

[0009] Using the K-means clustering algorithm, speed and density models are constructed under different traffic densities; a critical insertable gap model is constructed based on the instantaneous traffic flow density and vehicle position of the target lane.

[0010] Determine the limit state mathematical model between the safe lane change probability and the target lane change probability, and construct the safe lane change probability model using the differential method;

[0011] Using the headway, speed distribution, speed and density, and critical gap as uncertainties, the safe lane-changing probability model is solved based on the Monte Carlo simulation method to obtain the target lane-changing success rate under different traffic volumes, large vehicle ratios, and clearance lengths.

[0012] Based on the target lane change success rate, recommended clearance length values ​​are given for different traffic volumes and the proportion of large vehicles.

[0013] Preferably, the calibration using measured data includes the following steps:

[0014] Without considering weather factors, the measured speeds of road sections with small exit clearances were collected for different vehicle types.

[0015] By comparing the measured speed with the speed at the corresponding location output by the VISSIM simulation model, the driving behavior parameters and expected speed of the VISSIM simulation model are calibrated.

[0016] Preferably, obtaining the headway, speed, and instantaneous traffic flow density at different cross-sections in the VISSIM simulation model includes the following steps:

[0017] Based on the VISSIM simulation model, vehicle speed data acquisition devices are deployed at different cross-sections, namely the tunnel exit, the midpoint of the clearance section, the start and end points of the deceleration lane transition section, the midpoint of the deceleration lane, and the end point of the deceleration lane, corresponding to the main line cross-sections.

[0018] Each data acquisition device outputs ".mer" and ".fzp" format files for the corresponding locations. The headway is calculated using the ".mer" file. Using Python software, the speed and the positions of surrounding vehicles at the corresponding time and section are obtained from the ".fzp" file based on the time and section in the ".mer" file. The instantaneous traffic flow density at a certain section is calculated based on the vehicle positions.

[0019] Preferably, the fitting of the headway model and speed distribution model for different road segments includes the following steps:

[0020] An improved G-type statistic was used to remove outliers, resulting in an effective headway.

[0021] By increasing the correction factor The Weibull distribution was used to fit the probability distribution of headway distance for small exit clearance sections under different working conditions and road sections.

[0022] A one-step fitting was performed on the model parameters under different working conditions to obtain the headway model for different road sections. The speed distribution model was obtained by fitting the Weibull distribution.

[0023] Preferably, the construction of the velocity and density model includes the following steps:

[0024] Speed ​​intervals are divided, and the K-means clustering algorithm is used to perform iterative cluster analysis on the two-dimensional data of speed and instantaneous traffic flow density in the same segment to obtain the corresponding cluster centers and determine the relationship between speed and density.

[0025] Wherein, P(k) a ) represents the density k at a certain velocity. a The probability, density k b The probability of its occurrence is 1-P(k) a ), C i and D i All represent the coefficients of the fitted polynomial; n depends on the sum of squares of the fitted correlation coefficients and R. 2 .

[0026] Preferably, the construction of the critical insertable gap model includes the following steps:

[0027] The K-means clustering algorithm was used to perform iterative clustering analysis on two-dimensional data of headway and instantaneous traffic flow density at different cross sections. This identified a scenario where a convoy with a small headway was formed when vehicle speeds were similar and traffic flow density was high, which is known as the barrier effect.

[0028] The impact of the barrier effect is quantified using a corresponding formula; drivers will select different critical insertion gaps based on the degree of barrier effect, and the formula is as follows:

[0029]

[0030] Among them, t c-k The critical gap that can be inserted by the driver in the original lane when the traffic flow density of the target lane is k; t c-max and t c-minrespectively represent the maximum and minimum values of the critical insertable gap generally acceptable to the driver;

[0031] Regarding the formula of k1 < k < k2, the change trend of the critical insertable gap under different position sections is quantified to obtain the corresponding model;

[0032] For the critical insertable gap of the same lane, the relationship between its model coefficients and positions is further quantified to obtain the critical insertable gap model based on traffic flow density and vehicle position.

[0033] Preferably, the limit state mathematical model between the safe lane change probability and the target lane change probability is shown as follows:

[0034] g(x1, x2,..., x i ) = g[P(z), P T = P(z) - P T

[0035] where P(z) is the safe lane change probability, P T is the target lane change probability, and g is the reliability function for the vehicle to safely and successfully change lanes after waiting for a certain length distance z; g = 0 represents the failure surface; x i represents the variable that satisfies the safe or target lane change probability.

[0036] Preferably, the construction of the safe lane change probability model includes the following steps:

[0037] For a one-way two-lane road, models are respectively built according to two scenarios of the initial vehicle position in the inner lane and the outer lane. Using the differential method, the probability model for the safety of the inner lane is:

[0038]

[0039] The safety lane change probability model for the diverging vehicle in the outer lane is:

[0040]

[0041] where S represents the distance traveled by the diverging vehicle in the inner lane from the tunnel entrance to the first successful lane change, L is the clear distance length, is the critical insertable gap, is the vehicle speed, v m is the speed of the target lane, l is the length of the first lane change transition section of the vehicle, l' is the length of the second lane change transition section of the vehicle, γ, β, α are the parameters of the headway distribution model.

[0042] Preferably, it further includes:

[0043] Based on the concept of "uncertainty" in Monte Carlo simulation, an algorithm for solving the probabilistic model was built using MATLAB. Considering the uncertainties of factors such as vehicle speed, traffic density, critical gap for insertion in the target lane, and headway distribution, the success rate of lane change and exit under different traffic volumes, proportions of large vehicles, and clearance lengths was solved.

[0044] The beneficial effects of this invention are:

[0045] (1) The method of this invention is based on the reliability design method with the fundamental goal of achieving “safe and controllable” route design. It can measure the safety problems caused by the deviation between the alignment design value and the standard value, which is something that the traditional fixed value method cannot achieve. The recommended value obtained is more in line with the actual road conditions.

[0046] (2) The reliability algorithm of this invention combines the traffic characteristics of small clearance road sections and uses the concept of "uncertainty" to study the clearance length under different traffic volumes and large vehicle ratios. The proposed clearance recommendation value can better make up for the deficiencies of relevant specifications.

[0047] (3) The present invention uses VISSIM, a traffic flow simulation software that has been widely used in recent years, to expand the working conditions, so that the measured speed, headway and other data are more universal for exit small clearance road sections under other working conditions. Attached Figure Description

[0048] Figure 1 This is a flowchart of the calculation method according to an embodiment of the present invention;

[0049] Figure 2 This is a road model diagram constructed in VISSIM according to an embodiment of the present invention;

[0050] Figure 3 This is a data collector layout diagram according to an embodiment of the present invention;

[0051] Figure 4 This is a schematic diagram of the exit trajectory of the diversion vehicle in the inner lane according to an embodiment of the present invention;

[0052] Figure 5 This is a schematic diagram of the exit trajectory of the diversion vehicle in the outer lane according to an embodiment of the present invention;

[0053] Figure 6 This is a flowchart of the VISSIM simulation process according to an embodiment of the present invention;

[0054] Figure 7 This is a simulation flowchart of the lane-changing exit probability of the diverting vehicle with its initial position in the inner lane according to an embodiment of the present invention;

[0055] Figure 8 This is a simulation flowchart of the probability of a diverting vehicle initially positioned in the outer lane in an embodiment of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0057] Example 1

[0058] The present invention provides a method for calculating the recommended value of the net distance between tunnel and interchange exit, such as... Figure 1-8 As shown, it includes the following steps:

[0059] 1) Collect the measured speed of the exit-close road section by vehicle type and compare it with the speed of the corresponding position output by VISSIM software to calibrate the driving behavior parameters and expected speed of the VISSIM simulation model;

[0060] 2) In the VISSIM road simulation model, data acquisition devices are deployed at the mainline sections corresponding to the tunnel exit, the midpoint of the clearance section, the start and end points of the deceleration lane transition section, the midpoint of the deceleration lane, and the end point of the deceleration lane. The corresponding ".mer" and ".fzp" format files are output. The headway is calculated using the ".mer" file. Then, using Python software, the speed and the position of surrounding vehicles at the corresponding time and section are obtained from the ".fzp" file based on the time and section in the ".mer" file. The instantaneous traffic flow density of a certain section is calculated based on the vehicle position.

[0061] 3) Using the valid headway data from step 2), outliers are removed using the improved G-statistic, and a correction coefficient is added. The Weibull distribution is used to fit the headway probability distribution of the small clearance section at the exit under different working conditions and road sections. Finally, the model parameters under different working conditions are fitted to obtain the headway model for different road sections.

[0062] 4) Process the velocity data in step 2) using normal distribution and Weibull distribution to construct a velocity distribution model.

[0063] 5) Using Origin software and the K-means clustering algorithm, speed and density models were constructed under different traffic densities.

[0064] 6) Considering the impact of the barrier effect of the target lane on lane-changing vehicles, a critical insertable gap model based on instantaneous traffic flow density and vehicle position was constructed.

[0065] 7) Determine the limit state mathematical model between the safe lane-changing probability and the target lane-changing probability, and use the differential method to construct the safe lane-changing probability model;

[0066] 8) Considering the uncertainties of factors such as vehicle speed, traffic density, critical gap for insertion in the target lane, and headway distribution, an algorithm for solving the probability model was built using the "uncertainty" concept of Monte Carlo simulation. The model was solved using MATLAB, and the lane change success rate under different traffic volumes, large vehicle ratios, and clearance lengths was obtained.

[0067] 9) Based on the target lane change success rate (i.e. target reliability), recommended values ​​for clearance length under different traffic volumes and large vehicle ratios are given.

[0068] This method for calculating the recommended net distance between tunnels and interchange exits is mainly applicable to calculating the recommended net distance for road sections with small exit net distances.

[0069] This calculation method adopts a reliability algorithm process and combines the traffic characteristics of small clearance sections. It uses the concept of "uncertainty" to study the clearance length under different traffic volumes and the proportion of large vehicles, which can effectively avoid the problem that the required clearance value under the traffic capacity calculated by the "deterministic" concept in the "Specification" is too large.

[0070] This invention utilizes VISSIM software to study the traffic characteristics of small clearances at exits and proposes a method for calculating recommended values ​​for the clearance length between tunnel and interchange exits based on a reliability algorithm.

[0071] The following example illustrates this method using a small clearance section at the exit of a one-way two-lane mountain expressway. The specific process is as follows:

[0072] According to the specifications, the lane widths of the mainline and ramps at the corresponding design speeds were determined. The Wiedemann 99 car-following model was selected as the driving behavior model for the corresponding highway, and small cars (Cars) and heavy-duty trucks (HGVs) were chosen as the large / small vehicle types for the test. Measured speeds were collected for each vehicle type at the exit's short clearance section, and compared with the speeds output by the VISSIM software at the corresponding locations. This calibrated the driving behavior parameters and desired speeds of the VISSIM simulation model. To reduce the number of tests and improve efficiency, L... 16 (4 5 An orthogonal array was used for the experiment, considering four factors: net distance length L, traffic volume Q, exit ratio P, and large vehicle ratio H. The optimal solution was selected to calibrate the VISSIM simulation model. Figure 6 ).

[0073] 2) Install 6 cross-section data acquisition devices at the mainline locations corresponding to the tunnel exit, the midpoint of the clearance section, the start and end points of the deceleration lane transition section, the midpoint of the deceleration lane, and the end point of the deceleration lane (e.g., Figure 2 , Figure 3The system measures the average speed of various vehicle types. Based on the ".mer" and ".fzp" format files output by VISSIM, the accuracy of vehicle timestamps in the ".mer" file is adjusted to 0.10 seconds, resulting in a corrected ".mer" file. Using the "segment number" and "lane index," all speed-changing lanes on all segments are designated as lane 1, with the outer and inner lanes designated as lanes 2 and 3 respectively, resulting in a corrected ".fzp" file. Then, based on the vehicle number and time from the corrected ".mer" file, the row of the vehicle at that time can be located from the corrected ".fzp" file, thus obtaining the corrected lane. The number of rows in the same lane is then filtered out, and the row numbers before and after represent the actual vehicle sequence in the simulation. Python software is then used to filter out the density and speed at the desired time and position, thereby obtaining the headway and speed of the vehicle and multiple vehicles in front and behind it. Finally, using the above data and formulas... Calculate density using the formula Calculation speed.

[0074] Where k refers to the traffic flow density calculated based on n vehicles, veh / (km·ln); x i,j The distance between the front ends of the i-th car and the (i+1)-th car is m; v is the average speed of these n cars, km / h; v i The speed of the i-th vehicle is expressed in km / h.

[0075] 3) Using the effective headway data from step 2), the improved G-type statistics by Fei Helang et al. are applied. Remove outliers.

[0076] In this experiment, it is assumed that when 0 < X (k) When ≤10, even if G r >G(α, n), X (r) Still within the normal range. Specifically, the headway distances obtained from each road segment are sorted from smallest to largest to obtain the order of headway observations, which is 0 < X. (1) ≤X (2) ≤…≤X (n) , 1≤r≤n; G r Refers to statistics, if G r If X > G(α, n), then X (r) This is an outlier.

[0077] Using an increased correction factor Weibull distribution To fit the headway probability distribution of the exit small clearance section under different working conditions and road sections, γ≤t<∞, α>0, γ≥0, β>γ, where P(h tP(h ≥ t) is the probability that the headway between vehicles is greater than t; f is the probability density distribution; the value of α determines the shape of the probability distribution curve. When α = 1, it is the negative exponential distribution. When α = 3, it is similar to the normal distribution. The larger α is, the narrower the distribution range; β is the scale parameter of the probability distribution; γ is the starting parameter of the probability distribution; is the correction coefficient;

[0078] Finally, use the "Levenberg - Marquarqt (LM) + General Global Optimization Algorithm (UGO)" of the non - linear analysis professional software 1stOpt to perform one - step fitting and iteration on the model parameters under different working conditions, and obtain the headway model for different lane sections.

[0079] 4) The vehicle speeds in the acceleration lane are fitted according to the normal distribution; while the speed distribution in the short - clear - distance exit section follows the Weibull distribution. Therefore, in this study, the modified Weibull distribution is used to fit the outer lanes of the clear - distance section, the outer lanes of the acceleration section, and the inner and outer lanes of the tunnel entrance. First, the probability distribution statistics of the average vehicle speeds at each position are carried out in sections of 1 km / h, and then the non - linear analysis professional software 1stOpt is used for Weibull distribution fitting, thus obtaining the corresponding speed distribution models for each position.

[0080] 5) Using the origin software, the K - value clustering algorithm is used to perform polynomial fitting on the clustering centers 1 (v a , k a ) and clustering center 2 (v b , k b ), and obtain P(k b ) = 1 - P(k a ); where P(k a ) is the probability of the density k a appearing at a certain speed, and the probability of the density k b appearing is 1 - P(k a ), C i and D i both represent the coefficients of the fitting polynomial; n depends on the sum of squares of the fitting correlation coefficient R 2 , and take the polynomial degree when R 2 is better.

[0081] 6) Use the K - means clustering algorithm to perform iterative clustering analysis on the two - dimensional data of headway and instantaneous traffic flow density under different cross - sections to identify the barrier effect; use the corresponding formula to quantify the influence of the barrier effect intensity, and further construct the critical insertable gap model. The formula is Under the critical insertable gap considering the change of traffic flow density and vehicle position, the formula for k1 < k < k2 is changed to t c-(k,x)=2+e f(x)+g(x)k Then, the data is normalized according to the location of each data acquisition unit, and then a polynomial is used. By fitting the coefficients A1 and A2 to the normalized vehicle position respectively, the critical insertable gap considering the vehicle position can be obtained. Where t c-k The critical gap that a driver in the original lane can insert into when the target lane traffic flow density is k, s; t c-max and t c-min represents the maximum and minimum critical gaps that are generally acceptable to drivers; x represents the longitudinal distance of the vehicle from the tunnel entrance, in meters; s1 and s2 represent the normalized positions of the outer lane and the speed change lane, respectively; B i The coefficients of the fitted polynomial are represented by n; n depends on the sum of squares of the fitted correlation coefficients and R. 2 Take R 2 The degree of the polynomial in a better case.

[0082] 7) The mathematical model for the limiting state between the safe lane-changing probability and the target lane-changing probability is determined as g(x1,x2,...,x...). i )=g[P(z),P T ]=P(z)-P T Where P(z) is the safe lane-changing probability, PT is the target lane-changing probability, and g is the reliability function function for a vehicle to safely and successfully change lanes after waiting a certain distance z; g = 0 indicates the failure surface; x i Variables representing the probability of satisfying a safe or target lane change. A safe lane change probability model is constructed using the differential method; models are built for two scenarios: the initial vehicle position is in the inner lane and the outer lane. The probability model for a safe lane change in the inner lane is as follows:

[0083]

[0084] The safe lane-changing probability model for vehicles diverting from the outer lane is as follows:

[0085]

[0086] Where S represents the distance traveled by a vehicle diverting from the inner lane from the tunnel entrance to the point where it successfully changes lanes for the first time (i.e., moves to the outer lane), and L is the clearance length. For critical insertable gaps, v is the vehicle speed. m Let l be the target lane speed, l be the length of the vehicle's first lane change transition segment, and l' be the length of the vehicle's second lane change transition segment. γ, β, and α are parameters of the headway distribution model (e.g., ...). Figure 4 , 5 ).

[0087] 8) Considering the uncertainties of factors such as vehicle speed, traffic density, critical gap for insertion in the target lane, and headway distribution, MATLAB software can be used to simulate and calculate the probability of a vehicle successfully exiting the main lane under different traffic volumes Q, large vehicle proportions H, and clearance lengths L. For diverting vehicles initially positioned in the inner lane, the simulation process is as follows: Figure 7 For the initial position of the diverting vehicle in the outer lane, the simulation process is as follows: Figure 8 The model was solved by applying the concept of "uncertainty" in Monte Carlo simulation, and the lane change success rate was obtained under different traffic volumes, large vehicle ratios, and clearance lengths.

[0088] 9) Based on the target lane change success rate (i.e., target reliability), recommended clearance length values ​​are given for different traffic volumes and the proportion of large vehicles. For a single-lane highway with a design speed of 80 km / h, the clearance lengths are given in Tables 1 to 3 below, based on target reliability of 85%, 90%, and 95%.

[0089] Table 1. Recommended net distance length (m) for a target reliability of 85%.

[0090]

[0091] Table 2 Recommended net distance lengths (m) for a target reliability of 90%.

[0092]

[0093] Table 3 Recommended net distance lengths (m) for a target reliability of 95%.

[0094]

[0095] Note 1: "—" indicates that a clearance length of 300m is unlikely to meet the target lane change success rate (i.e., target reliability) under this working condition.

[0096] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for calculating the recommended value of the net distance between a tunnel and an interchange exit, characterized in that, Includes the following steps: A VISSIM simulation model of the tunnel and interchange exits was constructed and calibrated using measured data. The headway, speed, and instantaneous traffic flow density of vehicles at different cross sections in the VISSIM simulation model were obtained, and statistical fitting was performed to obtain headway models and speed distribution models for different road segments. Using the K-means clustering algorithm, speed and density models are constructed under different traffic densities. A critical insertable gap model is constructed based on the instantaneous traffic flow density and vehicle position of the target lane. Different critical insertable gaps are selected according to the degree of barrier effect, and the formula is as follows: in, The target lane traffic flow density is k At that time, the critical gap that the driver in the original lane selected; and These represent the maximum and minimum critical intercalation gaps that are generally acceptable to the driver; Determine the limit state mathematical model between the safe lane change probability and the target lane change probability, and construct the safe lane change probability model using the differential method; The construction of the safe lane-changing probability model includes the following steps: For a one-way two-lane road system, models are built for two scenarios: the initial vehicle position is in the inner lane and the outer lane. The probability model for the safety of the inner lane using the differential method is as follows: The safe lane-changing probability model for vehicles diverting from the outer lane is as follows: Where S represents the distance traveled by the diverted vehicle in the inner lane from the tunnel entrance to the point where the first lane change is successful. L It is the net distance length. For critical insertable gaps, For vehicle speed, For the target lane speed, l This refers to the length of the initial lane change phase for the vehicle. l’ This refers to the length of the vehicle's second lane change transition phase. , , , These are the parameters for the headway distribution model. Using the headway, speed distribution, speed and density, and critical gap as uncertainties, the safe lane-changing probability model is solved based on the Monte Carlo simulation method to obtain the target lane-changing success rate under different traffic volumes, large vehicle ratios, and clearance lengths. Based on the target lane change success rate, recommended clearance length values ​​are given for different traffic volumes and the proportion of large vehicles.

2. The method for calculating the recommended value of the net distance between a tunnel and an interchange exit according to claim 1, characterized in that, The calibration using measured data includes the following steps: Without considering weather factors, the measured speeds of road sections with small exit clearances were collected for different vehicle types. By comparing the measured speed with the speed at the corresponding location output by the VISSIM simulation model, the driving behavior parameters and expected speed of the VISSIM simulation model are calibrated.

3. The method for calculating the recommended value of the net distance between a tunnel and an interchange exit according to claim 1, characterized in that, The process of obtaining the headway, speed, and instantaneous traffic flow density at different cross-sections in the VISSIM simulation model includes the following steps: Based on the VISSIM simulation model, vehicle speed data acquisition devices are deployed at different cross-sections, namely the tunnel exit, the midpoint of the clearance section, the start and end points of the deceleration lane transition section, the midpoint of the deceleration lane, and the end point of the deceleration lane, corresponding to the main line cross-sections. Each data acquisition unit outputs ".mer" and ".fzp" format files for the corresponding locations. The headway is calculated using the ".mer" file. Using Python software, the speed and positions of surrounding vehicles at the corresponding time and section are obtained from the ".fzp" file based on the time and section in the ".mer" file. The instantaneous traffic flow density at a certain section is calculated based on the vehicle positions.

4. The method for calculating the recommended value of the net distance between a tunnel and an interchange exit according to claim 1, characterized in that, The fitting of the headway model and speed distribution model for different road segments includes the following steps: An improved G-type statistic was used to remove outliers, resulting in an effective headway. By increasing the correction factor The Weibull distribution was used to fit the probability distribution of headway distance for small exit clearance sections under different working conditions and road sections. A one-step fitting was performed on the model parameters under different working conditions to obtain the headway model for different road sections. The speed distribution model was obtained by fitting the Weibull distribution.

5. The method for calculating the recommended value of the net distance between a tunnel and an interchange exit according to claim 1, characterized in that, The construction of the velocity and density model includes the following steps: Speed ​​intervals are divided, and the K-means clustering algorithm is used to perform iterative cluster analysis on the two-dimensional data of speed and instantaneous traffic flow density in the same segment to obtain the corresponding cluster centers and determine the relationship between speed and density. , , , ; Wherein, P( k a ) as density at a certain velocity k a probability, density k b The probability of its occurrence is 1-P( k a ), C i and D i All represent the coefficients of the fitted polynomial; n Depends on the sum of squares of the fitted correlation coefficients R 2 .

6. The method for calculating the recommended value of the net distance between a tunnel and an interchange exit according to claim 1, characterized in that, The limiting state mathematical model between the safe lane-changing probability and the target lane-changing probability is shown in the following equation: Wherein, P( z P represents the probability of a safe lane change. T The probability of changing lanes for the target g Waiting for a vehicle for a certain distance z A reliable function that enables safe and successful lane changes; g =0 indicates a failure surface; x i A variable representing the probability of satisfying a safe or target lane change.

Images