Parameter calibration method of vehicle delay model at entrance and exit of large parking lot based on simulation
Through simulation experimental data calibration method, combined with field surveys and Vissim simulation software, a vehicle delay model for entrances and exits of large parking lots was constructed, solving the problems of complex vehicle delay calculation and difficult model calibration, and achieving efficient and accurate parameter calibration and traffic flow manage.
Patent Information
- Application Number
- CN202111443611.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-30
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2041-11-30
AI Technical Summary
The delay in the entry and exit of large parking lots is complicated to calculate and the model calibration is difficult. The existing technology fails to effectively consider the comprehensive evaluation of multiple objects, resulting in the complex model being difficult to understand and poor practical application.
Using simulation experimental data calibration method, through a combination of field survey and traffic simulation, a vehicle average delay model based on Vissim simulation software is constructed, vehicle delay data is obtained, model parameters are calibrated, and the practicality and accuracy of the model are improved.
It realizes efficient and accurate parameter calibration of vehicle delay model at entrances and exits of large parking lots, simplifies the parameter calibration process, improves calculation accuracy and scope of application, and supports refined control and management of dynamic and static traffic flows.
Smart Images

Figure CN114239371B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a simulation-based large parking lot entrance and exit vehicle delay model parameter calibration method, belonging to the field of urban road traffic control. Background Art
[0002] With the increase in urban development density, the trend of land intensification is obvious, giving rise to a large number of large parking lots. Vehicles arrive and leave more concentratedly, and the traffic congestion problem at the entrance and exit junctions needs to be solved urgently. At present, the research on the traffic impact of parking lots is mostly limited to a single object, such as connecting roads, entrance and exit passages, or the junction between entrances and exits and connecting roads, lacking a comprehensive evaluation that considers multiple objects. In addition, it fails to directly reveal the relationship between traffic volume and evaluation indicators, resulting in a complex and difficult-to-understand model and poor practical applicability.
[0003] There are many evaluation indicators that describe traffic impact, such as queue length, delay, traffic capacity, etc. Among them, vehicle delay can not only reflect the severity of congestion, fuel consumption and travel time loss of dynamic traffic flow, but also measure the rationality and reliability of traffic channelization design. At the same time, drivers are most sensitive to time changes when they are in congested queues. In addition, vehicle delay is widely used in traffic congestion evaluation, intersection signal control and path guidance. Considering that delay has a certain evaluation significance for both dynamic and static traffic, the present invention selects the average vehicle delay as the evaluation indicator for dynamic and static traffic impact analysis of parking lots.
[0004] Vehicle delay refers to the difference between the actual travel time of a vehicle and the travel time under ideal conditions, which generally includes fixed delay, operation delay, parking delay, queuing delay, etc. The parameters in the average vehicle delay model are all parameters that cannot be calibrated by direct observation and measurement, such as critical gap and follow-up time parameters, which are estimated using the maximum likelihood estimation method, which requires the collection of a large amount of data for calculation.
[0005] Theoretically, the calibration method is feasible by measuring the dynamic and static traffic volume and vehicle delay data. However, the calibration requires a huge amount of data, the traffic survey task is arduous, time-consuming and labor-intensive, and the practical operability is poor. Therefore, this paper proposes to use the simulation experiment data calibration method. The traffic volume is input through the simulation software, and the vehicle delay data is obtained to calibrate the parameters of the vehicle average delay model.
[0006] Advantages of simulation experiment data calibration method: (1) Through simulation software, not only can a large amount of required data be obtained at low cost, less time and high accuracy, but also the traffic volume can be combined arbitrarily and the data range obtained is wide. (2) With the rapid increase in the operating speed of modern computers and the improvement in numerical computing capabilities, the feasibility of data calibration method through simulation software has been greatly improved.
[0007] Therefore, the present invention proposes an average delay calculation model for the delay characteristics of multiple types of vehicles at the entrances and exits of large parking lots, and establishes an efficient and accurate vehicle delay model parameter calibration method by combining field investigation with traffic simulation, thereby improving the practicability of the model. This is a key technology that needs to be studied urgently in the refined control and management of dynamic and static traffic flows at the entrances and exits of large parking lots. Summary of the invention
[0008] The purpose of the invention is to provide a method for calibrating parameters of a vehicle delay model at the entrance and exit of a large parking lot based on simulation, so as to solve the problems of complex calculation of vehicle delay at the entrance and exit of a large parking lot and difficulty in model calibration.
[0009] The technical solution to achieve the purpose of the present invention is: a method for calibrating parameters of a vehicle delay model at an entrance and exit of a large parking lot based on simulation, characterized in that it includes the following steps:
[0010] Step 1: According to the location and size of the parking lot, select parking lots with a total number of parking spaces greater than 300 for field investigation, select the lane closest to the opposite driving lane (the leftmost lane) on the road connected to the parking lot entrance and exit as the relevant road for information collection, and obtain static facility data and dynamic traffic data of the relevant roads and parking lot entrances and exits through field measurement and traffic video shooting.
[0011] Step 2: Based on the relevant road static facility data and dynamic traffic flow data obtained in the field, a road vehicle free driving simulation environment is constructed to perform parameter correction on the car-following model of road traffic flow;
[0012] Construct a simulation environment for free-driving vehicles on roads: Based on Vissim simulation software, conventional simulation process; establish Vissim road network engineering files; model the road network according to the static facility attributes of the survey site, and input the measured road traffic flow; set up data collectors at the starting and ending points of the road section.
[0013] Step 3: Analyze the dynamic traffic flow data of the parking lot entrance and exit obtained in the field, extract and count the data of the video shot in step 1, and obtain relevant modeling parameters. The modeling parameters include the average speed of vehicles arriving at the entrance without interference from road vehicles, the average speed of vehicles leaving the exit, the ratio of vehicles arriving at the parking lot to road traffic, the ratio of traffic volume in each lane leaving the parking lot from different lanes, and the ratio of vehicles leaving the parking lot and entering different lanes. The two parameters of observation path and congestion avoidance are calibrated through intuitive video statistics.
[0014] Among them, the calibrated road section refers to the exit passage and road from the parking lot. The observation path and congestion avoidance parameters are both set parameters in the Vissim simulation software. The conflict behavior observation path and congestion avoidance parameters are calibrated using an intuitive video statistical method. Through second-by-second observation, the total conflict traffic volume of the exit passage, the traffic volume of the main road that gives way, and the traffic volume of the main road that does not give way are counted, and the total conflict traffic volume of the main road, the traffic volume of the exit passage that gives way, and the traffic volume of the exit passage that does not give way are counted; vehicles on the main road or exit passage without conflict behavior are not counted.
[0015] The parameters are explained as follows:
[0016] Among the conflict area attributes, Anticipate routes and Avoid blocking have the greatest impact on conflict behavior. These two attribute parameters affect the critical gap, following time and traffic capacity correction coefficient in the vehicle average delay model.
[0017] The observation path indicates how many proportions of vehicles in the entry channel will follow the road priority rule, consider the vehicles on the road and calculate the gaps to decide whether to accelerate, decelerate or even stop. This is because the drivers of exit vehicles over-optimistically estimate the gaps or directly fail to follow the road priority rule, resulting in entering the conflict area when they should not have, affecting the driving of vehicles on the road.
[0018] Avoiding congestion refers to the proportion of vehicles with priority on the road that may choose not to enter the conflict area and let the vehicles in the exit channel go first because they conservatively estimate that they cannot pass through the conflict area immediately.
[0019] Step 4: Based on the static facility data of the parking lot entrances and exits obtained in the field, a complete simulation scene of the parking lot entrances and exits is constructed, and the road section parameter attributes, entrance and exit attributes, and deceleration zone attributes are set, and the driving time detectors for road vehicles, vehicles arriving at the parking lot, and vehicles leaving the parking lot are set. That is, in VISSIM, the simulation scene can be set according to the actual situation of the parking lot. Enter the road section parameters, entrance and exit attributes, and deceleration zone attributes in the simulation scene constructed in the simulation software, and set the time detector at the location where the driving time needs to be detected in the simulation scene in accordance with the actual parking lot scene, that is, at the starting and ending points of the road and parking lot entrance and exit channels)
[0020] Build a complete simulation scene of the parking lot entrance and exit: The conventional simulation software modeling method is as follows: 1. Build a road network engineering file, for example, you can import the CAD base map of the scene or the Amap Baidu map into Vissim; 2. Build a road network model through road sections and connectors; 3. Based on the static facility attributes obtained in step 1, complete the setting of road section parameter attributes, entrance and exit attributes, and deceleration zone attributes, and set the driving time detectors for road vehicles, parking lot arrival vehicles, and leaving vehicles; 4. According to the modeling parameters related to the dynamic traffic data obtained in step 3, set the input vehicle attributes.
[0021] Step 5. Through the secondary development technology of VISSIM software, input different traffic volume combinations of road vehicles and parking lot entrances and exits to vehicles leaving the vehicle in the established simulation scenario (traffic volume combination: different combinations of road traffic volume, parking lot arrival and departure traffic volume, for example, road traffic volume is 500pcu / h, parking lot arrival traffic volume is 100pcu / h, parking lot departure traffic volume is 200pcu / h), run the simulation experiment, detect the delay time of each vehicle, and obtain the average delay time of the vehicles in the simulation experiment
[0022] Simulation experiment: Based on the complete simulation scene of the parking lot entrance and exit constructed in step 4, through secondary development, call the Simulation, Evaluation, PcuicleInputs and other interfaces in the Vissim COM interface, input the traffic volume combination information, perform simulation, obtain the passing time and delay duration of each vehicle based on the driving time detector, and calculate the average delay duration of the three types of vehicles: vehicles arriving at the parking lot, vehicles leaving the parking lot, and vehicles on the road;
[0023] Step 6: Construct an average delay model for multiple types of vehicles. Calibrate the parameters of the average delay model for multiple types of vehicles according to the input traffic volume and the average delay time of the output vehicles in the simulation experiment, and use the evaluation indicators to determine whether to accept the calibration results. If the calibration results are acceptable, it means that the average vehicle delay model is applicable to the current large parking lot entrances and exits; if the calibration results are unacceptable, it means that the average vehicle delay model is not applicable to the current large parking lot entrances and exits.
[0024] Preferably, the static facility data and dynamic traffic flow data of the relevant roads and parking lot entrances and exits in step 1 refer to:
[0025] The static facility data of relevant roads include: number of lanes, lane width, lane design speed, road section length, intersection location, signal light location, etc.;
[0026] The dynamic traffic flow data of relevant roads include: traffic flow information such as headway, traffic volume, etc.;
[0027] The static facility data of parking lot entrances and exits include: entrance and exit connecting road data, including the geometric dimensions of the connecting roads, number of lanes, lane width, lane design speed, and section length; entrance and exit channel data, including the length, width, slope and other geometric dimension data of the entrance and exit channels, the distance between the exit and the entrance, the angle between the entrance and the exit and the connecting road, the position of the gate in the entrance and exit channels, etc.
[0028] The dynamic traffic flow data of the parking lot entrances and exits include: the traffic flow arriving and leaving the parking lot entrances and exits captured by video, the traffic flow on the connected roads and the traffic flow entering the parking lot; the parking time of vehicles leaving and vehicles arriving through the gates counted by manual timing.
[0029] Preferably, the following model of the road traffic flow in step 2 adopts the Wiedemann 74 model, and its formula is:
[0030]
[0031] d=ax+bx
[0032] Where: d is the distance between the two vehicles; ax is the average parking distance (the average parking distance between two stopped vehicles, with a variation range of ±1m); bx is the safety distance; bx_add is the additional part of the safety distance; bx_mult is the multiple part of the safety distance; v is the vehicle speed in m / s; z is a value between [0,1], which is a standard normal distribution with a mean of 0.5 and a standard deviation of 0.15.
[0033] Preferably, the parameter correction method of the car-following model in step 2 is:
[0034] (1) According to the road traffic volume and headway obtained in the field, the measured headway distribution is fitted and judged, for example, by using a shifted negative exponential distribution function;
[0035] (2) Establish Vissim road network engineering files. Model the road network based on the relevant road static facility data from field surveys, input the measured road traffic volume, and set up data collectors at the cross-section locations;
[0036] (3) Modify the parameters of the car-following model. Generate a particle swarm by programming the particle swarm algorithm in Matlab, then call the I Driving Behavior Par Sets (driving behavior) interface of Vissim to modify the three parameters of the car-following model and run the simulation;
[0037] (4) Obtain the simulated headway. The data collector collects the basic information of each vehicle passing through the cross section, including vehicle type, speed, time, etc., and obtains the headway by taking the difference between the passing time of the front and rear vehicles; calls the I DataCollectionEvaluation interface to obtain basic data, calculate the headway, and finally calculate the distribution parameter estimate, and the single simulation ends;
[0038] (5) Set 10 random seeds for simulation to eliminate accidental errors;
[0039] (6) Seeking the optimal solution. Calculate the distribution parameter estimate of a particle under 10 random seeds, take its average value, and make the difference between it and the distribution parameter estimate of the measured headway, which is the objective function value of the particle; similarly, calculate the objective function value of each particle in a particle population, and determine the minimum objective function value of the particle population; then iterate until the convergence condition is met or the maximum number of iterations is reached, and output the optimal solution.
[0040] Preferably, the calculation formula of the average delay model of multiple types of vehicles in step 6 is as follows:
[0041] (1) Vehicles arriving at the parking lot entrance and exit:
[0042]
[0043]
[0044]
[0045] in, is the average delay of vehicles arriving at the i-th entrance of the parking lot; the unit is s; is the fixed delay of vehicles arriving at the i-th entrance of the parking lot, in seconds; is the average delay of vehicles arriving at the i-th entrance of the parking lot to pass through the entrance gate, in seconds; is the average delay of vehicles entering the i-th entrance of the parking lot from the second lane, in seconds; is the traffic volume reaching the i-th entrance of the parking lot, in pcu / h; is the traffic volume of the first lane entering the i-th entrance of the parking lot, in pcu / h; is the traffic volume of the second lane entering the i-th entrance of the parking lot, in pcu / h; is the traffic volume in the first lane of the road near the i-th entrance and exit of the parking lot, in pcu / h; is the service rate of the gate at the i-th entrance of the parking lot, in pcu / h; is the critical gap of vehicles from the second lane to the i-th entrance of the parking lot, in s; is the follow-up time of the vehicle from the second lane to the i-th entrance of the parking lot, in seconds; is the capacity correction factor for vehicles in the second lane to cross the first lane and enter the i-th entrance of the parking lot; n i is the number of lanes of the road near the i-th entrance and exit of the parking lot; i is the entrance and exit number.
[0046] (2) Leaving the car park entrance or exit:
[0047]
[0048]
[0049]
[0050] in, is the average delay of vehicles leaving the i-th exit of the parking lot, in seconds; is the fixed delay of vehicles leaving the parking lot at the i-th exit, in seconds; is the average delay of vehicles leaving the parking lot at the i-th exit through the exit gate, in seconds; is the average delay for vehicles leaving the i-th exit of the parking lot to enter the first lane of the road, in seconds; is the service rate of the exit gate at the i-th exit of the parking lot, in pcu / h; is the traffic volume leaving the i-th exit of the parking lot, in pcu / h; is the traffic volume of the first lane of the i-th exit of the parking lot, in pcu / h; is the critical gap for vehicles leaving the i-th exit of the parking lot to enter the first lane of the road, in seconds; is the following time of the vehicle leaving the i-th exit of the parking lot, in seconds; It is the capacity correction coefficient for vehicles leaving the i-th exit of the parking lot and entering the first lane; i is the entrance and exit number.
[0051] (3) Road vehicles:
[0052]
[0053]
[0054]
[0055]
[0056] in, is the average delay of vehicles on the roads near the i-th entrance and exit of the parking lot, in seconds; is the average delay caused by vehicles leaving the parking lot in the first lane of the road near the i-th entrance and exit of the parking lot, in seconds; is the average delay caused by vehicles in the first lane of the road near the i-th entrance and exit of the parking lot due to the passing of the second lane, in seconds; The time that vehicles in the second lane of the road near the i-th entrance and exit of the parking lot wait for the vehicle ahead to find an acceptable gap to enter the parking lot, including the queue delay, in seconds; is the total traffic volume of the roads near the i-th entrance and exit of the parking lot, in pcu / h; is the traffic volume leaving the i-th exit of the parking lot, in pcu / h; is the traffic volume of the first lane of the road near the i-th entrance and exit of the parking lot, in pcu / h; is the traffic volume in the first lane affected by vehicles leaving the i-th exit of the parking lot, in pcu / h; is the traffic volume in the first lane of the road near the i-th entrance and exit of the parking lot affected by vehicles in the second lane, in pcu / h; is the traffic volume of the second lane of the road near the i-th entrance and exit of the parking lot, in pcu / h; is the traffic volume from the first lane into the i-th entrance of the parking lot, in pcu / h; is the traffic volume from the second lane into the i-th entrance of the parking lot, in pcu / h; is the critical gap between the vehicle in the first lane and the conflict zone of the i-th exit of the parking lot, in seconds; is the following time of the vehicle in the first lane entering the conflict zone with the i-th exit of the parking lot, in seconds; is the critical gap between the first lane vehicle entering the conflict zone and the second lane vehicle passing through the road near the i-th entrance and exit of the parking lot, in seconds; is the following time of the first lane vehicle entering the conflict zone with the second lane vehicle passing through the road near the i-th entrance and exit of the parking lot, in seconds; is the following time of vehicles in the first lane of the road near the i-th entrance and exit of the parking lot, in seconds; is the follow-up time of the vehicle from the second lane to the i-th entrance of the parking lot, in seconds; is the critical gap of vehicles from the second lane to the i-th entrance of the parking lot, in s; is the capacity correction factor for vehicles in the second lane entering the i-th entrance of the parking lot crossing the first lane; is the traffic capacity correction coefficient of the first lane entering the conflict zone with the i-th exit of the parking lot; is the traffic capacity correction coefficient of the conflict zone between vehicles entering the first lane and passing the second lane of the road near the i-th entrance and exit of the parking lot; n i is the number of lanes near the i-th entrance and exit of the parking lot; i is the entrance and exit number.
[0057] Preferably, the evaluation index for calibration of the average delay model parameters of multiple types of vehicles in step 6 is R-square (coefficient of determination), and its value range is [0,1]. It is generally believed that R-square (coefficient of determination) is greater than 0.9 and the calibration value is accepted. The calculation formula of R-square (coefficient of determination) is as follows:
[0058]
[0059] in: is the model prediction value under the i-th group of traffic volume combination; is the mean of the simulation data; y i is the simulation data value under the i-th traffic volume combination, i∈[1,m], and m is the total number of traffic volume combinations.
[0060] Compared with the prior art, the present invention has the following significant advantages: the present invention aims to comprehensively consider the mutual influence relationship between the entrances and exits of large parking lots and the connecting roads of various types of vehicles, construct a calculation model for vehicle delays, and propose a simulation-based parameter calibration method. Compared with the traditional vehicle delay calculation and traffic flow parameter calibration methods, the present invention can significantly reduce costs, improve efficiency, expand the scope of application, and enhance calculation accuracy;
[0061] Therefore, the present invention can provide new ways and methods for accurately calculating delays of various types of vehicles and analyzing dynamic and static traffic conditions at the entrances and exits of large parking lots, and provide a theoretical basis for solving the bottleneck problems of entrances and exits of large parking lots and connecting roads, thereby supporting the refined management and control of traffic flows at the entrances and exits of large parking lots. BRIEF DESCRIPTION OF THE DRAWINGS
[0062] Figure 1 It is a flow chart of the present invention.
[0063] Figure 2 Parameter correction iteration process and error value diagram of the car-following model.
[0064] Figure 3 Speed and distance between two vehicles.
[0065] Figure 4 Simulation environment modeling of the west (1) and east (2) entrances and exits of a large parking lot.
[0066] Figure 5 Fitting surface of the delay model for vehicles leaving the west exit of the large parking lot. DETAILED DESCRIPTION
[0067] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.
[0068] like Figure 1 As shown, a method for calibrating parameters of a vehicle delay model at an entrance and exit of a large parking lot based on simulation includes the following steps:
[0069] Step 1: According to the location and size of the parking lot, a large parking lot with a total number of parking spaces greater than 300 is selected for field investigation. The lanes around it that are least affected by lane-changing behavior are selected as relevant roads for information collection. The geometric data of the parking lot entrances and exits and relevant roads are measured, and the traffic flow videos of the parking lot entrances and exits and relevant roads are taken to obtain the static facility data and dynamic traffic flow data of the relevant roads and parking lot entrances and exits.
[0070] The parking lot selected in the embodiment of the present invention is a parking lot of a large commercial complex in Nanjing, which has three floors, namely B1, B2, and B3, with a total of 1,800 parking spaces. Its main entrances and exits are located on the east and west sides of the parking lot, respectively. The east entrance and exit is connected to the city's main road and its two-lane auxiliary road; the west entrance and exit is connected to the three-lane city branch road. The leftmost lane of the southbound city main road at the east entrance and exit is the least affected by the lane change behavior and is used as a related road for information collection.
[0071] Through field investigation, the static facility data of relevant roads were obtained, including: number of lanes, lane width, lane design speed, section length, intersection location, signal light location, etc.; as well as the static facility data of the entrances and exits of the two parking lots, including: entrance and exit connecting road data, including the geometric dimensions of the connecting roads, number of lanes, lane width, lane design speed, and section length; entrance and exit channel data, including the length, width, slope and other geometric dimension data of the entrance and exit channels, the distance between the exit and the entrance, the angle between the entrance and the exit and the connecting road, the location of the gate in the entrance and exit channels, etc.
[0072] In order to obtain dynamic traffic flow data, the embodiment of the present invention conducts a field survey, which is divided into a parking lot entrance and exit survey and a related road survey.
[0073] For the entrances and exits of parking lots, the video survey uses five cameras to collect data, and the best angles for shooting vehicles entering and exiting the entrances and exits and road traffic are determined in advance. Four of the cameras are set up at the entrances and exits of the two entrances and exits, and one is set up on the central dividing strip connecting the main road and the auxiliary road. The video captures the traffic volume arriving and leaving the entrances and exits of the parking lot, the traffic volume on the connected roads, and the traffic volume entering the parking lot. Manual timing is used to count the parking time of vehicles leaving the gate within half an hour and the parking time of vehicles arriving at the gate within half an hour. It can be obtained that the parking time of vehicles arriving at the gate is 2 seconds, and the parking time of vehicles leaving the gate conforms to the normal distribution with a mean of 4 seconds. Considering the parking characteristics of commercial parking lots, the video survey time is selected as 15:00-22:00 on weekends, and the manual timing is selected as 19:00-19:30 on weekends.
[0074] For the relevant roads, the survey used video shooting methods, and a tripod was set up on the central dividing strip next to the observation section to shoot vehicles passing through the observation section. The time for collecting traffic flow was 16:00-19:00 on weekends. The data during the peak traffic period was more in line with the purpose of this study.
[0075] Step 2: Based on the relevant road static facility data and dynamic traffic flow data obtained in the field, a road vehicle free driving simulation environment is constructed to perform parameter correction on the car-following model of road traffic flow;
[0076] The headway times within the three hours of the survey were statistically analyzed and fitted in Matlab based on the shifted negative exponential distribution function. The fitting results are shown in Table 1 below.
[0077] Table 1 Fitting results corresponding to each period
[0078]
[0079] Based on Vissim, a free-driving simulation environment for road vehicles was constructed. The particle swarm algorithm was used to generate particle swarms. The Vissim interface was called to modify the parameters of the following model, obtain the simulated headway and its distribution parameters, and calculate the average headway obtained by simulation, so as to obtain the objective function value, that is, the absolute value of the difference between the measured average headway and the average headway obtained by simulation. Through repeated iterations, the parameter combination with the smallest objective function value was selected and determined as the optimal solution based on the particle swarm algorithm.
[0080] The parameter correction uses the measured data at 18-19 o'clock, sets 20 iterations, 20 particle populations, and 10 random seeds, and a total of 20*20*10=4000 simulation experiments are required. After 20 iterations, the iterative process and error value are as follows Figure 2As shown. After 11 iterations, the objective function tends to be flat, and the objective function value is 0.1467 seconds. The average headway of the negative exponential distribution of the measured data shift is 5.4119 seconds, and the average headway obtained by the simulation experiment is 5.2652 seconds. The relative error is 2.71%, and the optimal solution is determined. The optimal solution is that the average parking distance is 1.8654 meters, the additional part of the safety distance is 1.5643, and the multiple part of the safety distance is 2.5795.
[0081] In order to verify the accuracy of parameter correction and avoid a large number of simulation experiments, the measured headway data at 16-17 o'clock and the measured headway data at 17-18 o'clock were used to conduct experiments in a simulation environment modified to the correction parameters. The objective function value at 16-17 o'clock was 0.1752, with a relative error of 2.49%, and the objective function value at 17-18 o'clock was 0.1563, with a relative error of 2.64%, both of which were within an acceptable range, verifying the accuracy of the correction parameters.
[0082] The default values of the three parameters of the following model in Vissim are: the average parking distance is 2 meters, the additional part of the safety distance is 2, and the multiple of the safety distance is 3. After the parameters are calibrated, they are 1.8654, 1.5643, and 2.5795. The curve of the distance between the two cars changing with speed (z value is 0.5) is as follows Figure 3 Observation Figure 3 It can be seen that the distance between the two vehicles is smaller after the correction parameters, which is consistent with the traffic characteristics of heavy traffic volume during peak hours and reduced vehicle distance.
[0083] Step 3: Analyze the dynamic traffic flow data of the parking lot entrance and exit obtained in the field to obtain relevant modeling parameters, including the speed of vehicles arriving at the entrance without interference from road vehicles, the speed of vehicles leaving the exit, the ratio of vehicles arriving at the parking lot to road traffic, the ratio of vehicles leaving the parking lot from different lanes, the ratio of vehicles leaving the parking lot and entering different lanes, etc., and calibrate the two parameters of observation path and congestion avoidance through intuitive video statistics;
[0084] The speed of vehicles arriving at the entrance and leaving the exit without interference from road vehicles can be obtained by statistics:
[0085] Table 2 Vehicle speed distribution at the parking lot entrance and exit
[0086]
[0087]
[0088] As for the ratio of vehicles arriving at the parking lot to road traffic, the vehicle flow entering the parking lot from the east entrance and the vehicle flow on the east entrance and exit auxiliary road are counted at intervals of 15 minutes, and the ratio between the two is stable at around 40%; the vehicle flow entering the parking lot from the west entrance and the vehicle flow on the west entrance and exit branch road are counted at intervals of 15 minutes, and the ratio between the two fluctuates around 20%.
[0089] The ratio of vehicles entering the parking lot from different lanes is statistically analyzed at 15-minute intervals, and it can be found that the ratio of vehicles entering the parking lot from the second lane and from the first lane at the east and west entrances is 1:1.
[0090] As for the ratio of vehicles leaving the parking lot and entering different lanes, the ratio of vehicles leaving the parking lot at the east exit and entering the auxiliary road to vehicles entering the main road was counted at intervals of 15 minutes, and the ratio fluctuated around 2:8. Therefore, the ratio of vehicles leaving the east exit and entering the main road to vehicles leaving the parking lot was set at 2:8, and all vehicles leaving the west exit entered the leftmost lane of the branch road.
[0091] At the same time, through video statistics, conflict behavior traffic data can be obtained, and some data are shown in the following table:
[0092] Table 3 Statistics of vehicle conflict behaviors at parking lot entrances and exits
[0093]
[0094] Get the conflict behavior parameters: the observation path value is 0.8 and the avoidance of congestion value is 0.2.
[0095] Step 4: Based on the static facility data of the parking lot entrances and exits obtained in the field, a complete simulation scene of the parking lot entrances and exits is constructed, and attributes such as road section parameters, entrances and exits, and deceleration zones are set. Also, driving time detectors for road vehicles, vehicles arriving at the parking lot, and vehicles leaving the parking lot are set.
[0096] Construct a simulation environment for large parking lot entrances and exits. Figure 4 shown.
[0097] Step 5: Run the simulation experiment by inputting different combinations of road vehicles and parking lot entrances and exits to vehicles leaving the vehicle, detect the delay time of each vehicle, and obtain the average delay time of the vehicles in the simulation experiment;
[0098] According to the traffic volume data collected from the survey, the combined parameters of the facility simulation traffic volume are as follows:
[0099] In the east entrance and exit scenario, the auxiliary road inputs a traffic volume of 100 to 1050 pcu / h, and increases the traffic volume with an interval of 50 pcu / h, for a total of 20 traffic volumes; the exit channel inputs a traffic volume of 30 to 315 pcu / h, and increases the traffic volume with an interval of 15 pcu / h, for a total of 20 traffic volumes; therefore, there are a total of 20*20=400 input combinations, the random seed is set to 10, and a total of 4000 simulation experiments are conducted.
[0100] In the west entrance and exit scenario, the branch road inputs a traffic volume of 200 to 2100 pcu / h, and increases the traffic volume with an interval of 100 pcu / h, for a total of 20 traffic volumes; the exit channel inputs a traffic volume of 30 to 315 pcu / h, and increases the traffic volume with an interval of 15 pcu / h, for a total of 20 traffic volumes; therefore, there are a total of 20*20=400 input combinations, the random seed is set to 10, and a total of 4000 simulation experiments are required.
[0101] Run the program to perform batch simulation.
[0102] Step 6: Construct an average delay model for multiple types of vehicles. Calibrate the parameters of the average delay model for multiple types of vehicles based on the input traffic volume and the average delay time of the output vehicles in the simulation experiment, and determine whether to accept the calibration results through evaluation indicators.
[0103] The large parking lot has two entrances and exits, the east and west. Each entrance and exit has delay models for three types of vehicles: vehicles arriving at the parking lot entrance and exit, vehicles arriving at the parking lot entrance and exit, and road vehicles, totaling six vehicle average delay models.
[0104] In this invention, the vehicle delay model leaving the west entrance is taken as an example, and the parking lot entrance number i = 1. The input variable is the total road traffic volume and leaving the parking lot The unit is pcu / h. The number of lanes is 3.
[0105] It is known that the road departure rate is 20%, the traffic volume entering the parking lot from the second lane accounts for half of the total arriving traffic volume, the total traffic volume of the three lanes is evenly distributed, and the traffic volume of the first and second lanes is the difference between the road traffic volume and the parking lot leaving traffic volume, so the road traffic volume of the first lane is
[0106] The departure vehicle delay model is constructed as follows:
[0107]
[0108]
[0109]
[0110] After fitting, the parameter calibration results are shown in Table 4.
[0111] Table 4 Calibration results of the delay model parameters for vehicles leaving the west exit
[0112]
[0113] Overall evaluation of model fitting: SSE (the sum of squared errors between the model fitting value and the simulation data point) is 192.7, and R-square (coefficient of determination) is 0.9478. Because the coefficient of determination is greater than 0.9, the fitting effect is good, so the parameter calibration result is accepted. Substituting the parameter calibration result into the formula, the average delay model of vehicles leaving the west exit is obtained.
[0114]
[0115] The fitting surface of the delay model for vehicles leaving the west exit of a large parking lot is as follows: Figure 5 shown.
[0116] Through the analysis of the delay model, simulation data, and fitting surface, it can be seen that when other parameters in the model remain unchanged, the average delay of vehicles leaving the parking lot increases slowly first and then sharply as the vehicles leave. When there are fewer vehicles leaving, the increase in the average delay of vehicles leaving the parking lot by the total road traffic volume is not obvious. As the number of vehicles leaving the parking lot increases, the vehicles on the road will cause a lot of delays to the vehicles leaving the parking lot. The average delay of vehicles leaving the parking lot increases sharply as the leaving traffic volume and the total road traffic volume increase.
[0117] In summary, the present invention selects a suitable large parking lot for field investigation, obtains static facility data and dynamic traffic data of relevant roads and parking lot entrances and exits; based on the static facility data and dynamic traffic data of relevant roads obtained in the field, constructs a road vehicle free driving simulation environment, and performs parameter correction on the following model of road driving traffic; analyzes the dynamic traffic data of parking lot entrances and exits obtained in the field, and obtains relevant modeling parameters; constructs a complete simulation scene of parking lot entrances and exits based on the static facility data of parking lot entrances and exits obtained in the field; runs simulation experiments by inputting different traffic volume combinations, detects the delay time of each vehicle, and obtains the average delay time of vehicles in the simulation experiment; constructs an average delay model for multiple types of vehicles, and calibrates the parameters of the average delay model for multiple types of vehicles according to the input traffic volume and output average delay time of vehicles in the simulation experiment. The present invention can simplify the parameter calibration process, improve the efficiency of practical data use, and provide support for the efficient establishment of a refined management strategy for dynamic and static traffic applicable to the entrances and exits of large parking lots. It is a new means of calculating the average delay of multiple types of vehicles at the entrances and exits in the context of the continuous increase of large urban parking lots.
[0118] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for calibrating parameters of a vehicle delay model at the entrance and exit of a large parking lot based on simulation, characterized in that: The following steps are involved: Step 1: According to the location and size of the parking lot, select the parking lot with a total number of parking spaces greater than 300 for field survey, select the lane closest to the opposite driving lane on the road connected to the parking lot entrance and exit, that is, the leftmost lane as the relevant road for information collection, and obtain the static facility data and dynamic traffic data of the relevant roads and parking lot entrances and exits through field measurement and traffic video shooting; Step 2: Based on the relevant road static facility data and dynamic traffic flow data obtained in the field, a road vehicle free driving simulation environment is constructed to perform parameter correction on the car-following model of road traffic flow; Step 3: Analyze the dynamic traffic flow data of the parking lot entrance and exit obtained in the field, extract and count the data of the video shot in step 1, and obtain relevant modeling parameters. The modeling parameters include the average speed of vehicles arriving at the entrance without interference from road vehicles, the average speed of vehicles leaving the exit, the ratio of vehicles arriving at the parking lot to road traffic, the ratio of traffic volume in each lane leaving the parking lot from different lanes, and the ratio of vehicles leaving the parking lot and entering different lanes. The two parameters of observation path and congestion avoidance are calibrated through intuitive video statistics. Step 4: Based on the static facility data of the parking lot entrance and exit obtained in the field, a complete simulation scene of the parking lot entrance and exit is constructed, and the road section parameter attributes, entrance and exit attributes, and deceleration zone attributes are set, and the driving time detectors of road vehicles, parking lot arriving vehicles, and leaving vehicles are set; Step 5: Through the secondary development technology of VISSIM software, input different road vehicles and traffic volume combinations of vehicles from parking lot entrances and exits into the established simulation scenario, run the simulation experiment, detect the delay time of each vehicle, and obtain the average delay time of vehicles in the simulation experiment; Step 6: Construct an average delay model for multiple types of vehicles, calibrate the parameters of the average delay model for multiple types of vehicles according to the input traffic volume and the average delay time of the output vehicles in the simulation experiment, and determine whether to accept the calibration results through evaluation indicators; The calculation formula of the average delay model for multiple types of vehicles is as follows: (1) Vehicles arriving at the parking lot entrance and exit: in, is the average delay of vehicles arriving at the i-th entrance of the parking lot; the unit is s; is the fixed delay of vehicles arriving at the i-th entrance of the parking lot, in seconds; is the average delay of vehicles arriving at the i-th entrance of the parking lot to pass through the entrance gate, in seconds; is the average delay of vehicles entering the i-th entrance of the parking lot from the second lane, in seconds; is the traffic volume reaching the i-th entrance of the parking lot, in pcu / h; is the traffic volume of the first lane entering the i-th entrance of the parking lot, in pcu / h; is the traffic volume of the second lane entering the i-th entrance of the parking lot, in pcu / h; is the traffic volume in the first lane of the road near the i-th entrance and exit of the parking lot, in pcu / h; is the service rate of the gate at the i-th entrance of the parking lot, in pcu / h; is the critical gap of vehicles from the second lane to the i-th entrance of the parking lot, in s; is the follow-up time of the vehicle from the second lane to the i-th entrance of the parking lot, in seconds; is the capacity correction factor for vehicles in the second lane to cross the first lane and enter the i-th entrance of the parking lot; n i is the number of lanes of the road near the i-th entrance and exit of the parking lot; i is the entrance and exit number; (2) Leaving the car park entrance or exit: in, is the average delay of vehicles leaving the i-th exit of the parking lot, in seconds; is the fixed delay of vehicles leaving the parking lot at the i-th exit, in seconds; is the average delay of vehicles leaving the parking lot at the i-th exit through the exit gate, in seconds; is the average delay for vehicles leaving the i-th exit of the parking lot to enter the first lane of the road, in seconds; is the service rate of the exit gate at the i-th exit of the parking lot, in pcu / h; is the traffic volume leaving the i-th exit of the parking lot, in pcu / h; is the traffic volume of the first lane of the i-th exit of the parking lot, in pcu / h; is the critical gap for vehicles leaving the i-th exit of the parking lot to enter the first lane of the road, in seconds; is the following time of the vehicle leaving the i-th exit of the parking lot, in seconds; is the capacity correction coefficient for vehicles leaving the i-th exit of the parking lot and entering the first lane; i is the exit number; (3) Road vehicles: in, is the average delay of vehicles on the roads near the i-th entrance and exit of the parking lot, in seconds; is the average delay caused by vehicles leaving the parking lot in the first lane of the road near the i-th entrance and exit of the parking lot, in seconds; is the average delay caused by vehicles in the first lane of the road near the i-th entrance and exit of the parking lot due to the passing of the second lane, in seconds; The time that vehicles in the second lane of the road near the i-th entrance and exit of the parking lot wait for the vehicle ahead to find an acceptable gap to enter the parking lot, including the queue delay, in seconds; is the total traffic volume of the roads near the i-th entrance and exit of the parking lot, in pcu / h; is the traffic volume leaving the i-th exit of the parking lot, in pcu / h; is the traffic volume of the first lane of the road near the i-th entrance and exit of the parking lot, in pcu / h; is the traffic volume of the first lane affected by vehicles leaving the i-th exit of the parking lot, in pcu / h; is the traffic volume in the first lane of the road near the i-th entrance and exit of the parking lot affected by vehicles in the second lane, in pcu / h; is the traffic volume of the second lane of the road near the i-th entrance and exit of the parking lot, in pcu / h; is the traffic volume from the first lane into the i-th entrance of the parking lot, in pcu / h; is the traffic volume from the second lane into the i-th entrance of the parking lot, in pcu / h; is the critical gap between the vehicle in the first lane and the conflict zone of the i-th exit of the parking lot, in seconds; is the following time of the vehicle in the first lane entering the conflict zone with the i-th exit of the parking lot, in seconds; is the critical gap between the first lane vehicle entering the conflict zone and the second lane vehicle passing through the road near the i-th entrance and exit of the parking lot, in seconds; is the following time of the first lane vehicle entering the conflict zone with the second lane vehicle passing through the road near the i-th entrance and exit of the parking lot, in seconds; is the following time of the vehicle in the first lane of the road near the i-th entrance and exit of the parking lot, in seconds; is the follow-up time of the vehicle from the second lane to the i-th entrance of the parking lot, in seconds; is the critical gap of vehicles from the second lane to the i-th entrance of the parking lot, in s; is the capacity correction factor for vehicles in the second lane entering the i-th entrance of the parking lot crossing the first lane; is the traffic capacity correction coefficient of the first lane entering the conflict zone with the i-th exit of the parking lot; is the traffic capacity correction coefficient of the conflict zone between vehicles entering the first lane and passing the second lane of the road near the i-th entrance and exit of the parking lot; n i is the number of lanes near the i-th entrance and exit of the parking lot; i is the entrance and exit number.
2. A method for calibrating parameters of a vehicle delay model at an entrance and exit of a large parking lot based on simulation according to claim 1, characterized in that: The static facility data and dynamic traffic flow data of the relevant roads and parking lot entrances and exits in step 1 refer to: The static facility data of relevant roads include: number of lanes, lane width, lane design speed, road section length, intersection location, and signal light location; The dynamic traffic data of relevant roads include: headway and traffic volume; The static facility data of the parking lot entrances and exits include: entrance and exit connecting road data, including the geometric dimensions of the connecting road, the number of lanes, lane width, lane design speed, and section length; entrance and exit channel data, including the length, width, slope geometric dimension data of the entrance and exit channel, the distance between the exit and the entrance, the angle between the entrance and the exit and the connecting road, and the position of the gate in the entrance and exit channel; The dynamic traffic flow data of the parking lot entrances and exits include: the traffic flow arriving and leaving the parking lot entrances and exits captured by video, the traffic flow on the connected roads and the traffic flow entering the parking lot; the parking time of vehicles leaving and vehicles arriving through the gates counted by manual timing.
3. The method for calibrating parameters of a vehicle delay model at an entrance and exit of a large parking lot based on simulation according to claim 1, characterized in that: The following model of road traffic in step 2 adopts the Wiedemann 74 model, and its formula is: d=ax+bx Where: d is the distance between the two vehicles; ax is the average parking distance, the average stopping distance between two stopped vehicles, with a variation range of ±1m; bx is the safety distance; bx_add is the additional part of the safety distance; bx_mult is the multiple part of the safety distance; v is the vehicle speed, in m / s; z is a value between [0,1], which is a standard normal distribution with a mean of 0.5 and a standard deviation of 0.
15.
4. The method for calibrating parameters of a vehicle delay model at an entrance and exit of a large parking lot based on simulation according to claim 1, characterized in that: The parameter correction method of the car-following model in step 2 is: S41, fitting and judging the measured headway distribution according to the road traffic volume and headway obtained in the field; S42. Establish Vissim road network engineering files; model the road network based on the relevant road static facility data from field surveys, input the measured road traffic volume, and set up data collectors at the cross-section locations; S43, modify the parameters of the car-following model; generate a particle swarm by programming a particle swarm algorithm through Matlab, and then call the IDriving Behavior Par Sets interface of Vissim to modify the three parameters of the car-following model, and run the simulation; S44, obtaining the simulated headway time; The data collector collects the basic information of each vehicle passing through the cross section, including the vehicle type, speed, and time, and obtains the headway time by subtracting the passing time of the front and rear vehicles; Call the I DataCollectionEvaluation interface to obtain basic data, calculate the headway, and finally calculate the distribution parameter estimate, and the single simulation ends; S45, setting 10 random seeds for simulation to eliminate accidental errors; S46, seek the optimal solution; calculate the estimated distribution parameters of a particle under 10 random seeds, take its average value, and make the difference between it and the estimated distribution parameters of the measured headway, which is the objective function value of the particle; similarly, calculate the objective function value of each particle in a particle population, and determine the minimum objective function value of the particle population; then iterate until the convergence condition is met or the maximum number of iterations is reached, and output the optimal solution.
5. The method for calibrating parameters of a vehicle delay model at an entrance and exit of a large parking lot based on simulation according to claim 1, characterized in that: The evaluation index for the calibration of the average delay model parameters of multiple types of vehicles in step 6 is R-square, whose value range is [0,1]. It is generally believed that R-square is greater than 0.9 and the calibration value is accepted; the calculation formula of R-square is as follows: in: is the model prediction value under the i-th group of traffic volume combination; is the mean of the simulation data; y i is the simulation data value under the i-th traffic volume combination, i∈[1,m], and m is the total number of traffic volume combinations.
Citation Information
Patent Citations
VISSIM simulation-based method for determining length of accelerating lane in left-in ramp confluence area
CN111695241A
Intersection traffic state parameter estimation method and system based on multi-source data
CN113380027A
Method and system for intersection group-based traffic control
WO2017166474A1
