Urban new energy vehicle travel characteristic modeling method and system based on entropy principle

Abstract

The invention discloses an urban new energy vehicle travel characteristic modeling method and system based on an entropy principle, and relates to the technical field of urban traffic travel analysis. The method has the advantages that the entropy theory is introduced to deduce the mathematical model by directly counting the characteristics of the daily average travel time in different travel areas, so that the phenomenon that the correctness of a conclusion is influenced by data heterogeneity caused by different travel areas is avoided. Actual geographical environment influence factors are considered, and independent items related to path factors are introduced, so that the established model is more in line with actual conditions. On one hand, the reliability of the model is verifiable, and on the other hand, the model is universally applicable.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban traffic travel analysis, and particularly relates to a modeling method and system for the travel characteristics of new energy vehicles in cities based on the entropy principle. Background Art

[0002] The urban traffic system is a key infrastructure of modern society. With the acceleration of the urban intelligentization process and the growth of the population, new energy electric vehicles have gradually joined the urban traffic system, which poses an increasingly great challenge to the scientific management of the urban traffic system. In order to achieve the scientific planning and layout of charging piles related to new energy electric vehicles in cities, it is necessary to deeply and accurately understand the spatio-temporal distribution characteristics of the mobility of new energy electric vehicles, establish a general model reflecting the mobility law of new energy electric vehicles in different regions, and provide theoretical support for the formulation and implementation of relevant policies. For the research on the overall traffic mobility of cities, new energy electric vehicles play an increasingly important role. Understanding the influencing factors and laws behind the mobility of electric vehicles in different regions is a key issue. Due to the complex road environment, local area restrictions, difficulties in inefficient data collection, and insufficient policy implementation efforts, this problem is very challenging. The current research mainly focuses on traditional multi-modal mixed travel or a specific urban area, lacking the research on the characteristics under a single new energy electric vehicle mode in a large range.

[0003] The existing methods and models for analyzing the mobility of traffic vehicles are all based on the statistical analysis of historical data, real-time traffic flow monitoring, simulation, etc. However, these methods and models usually have the following deficiencies:

[0004] First of all, the current traffic vehicle monitoring system often relies on traffic cameras or sensors to collect traffic vehicle flow data. This collection method has a low update frequency and inaccurate data, which will directly affect the reliability of the analysis results. Secondly, when conducting statistical analysis of historical data, different modes are generally not grouped and classified, which will indirectly lead to various data being messy and unable to reveal some characteristic laws of specific travel groups. In addition, the traditional traffic vehicle mobility analysis methods are limited to the analysis of local areas and are difficult to comprehensively evaluate the overall mobility of traffic vehicle systems in different regions of the city from a global perspective. This limitation makes it difficult to accurately predict the impact of traffic flow changes on the entire urban traffic system. In the existing models, machine learning models apply learning algorithms (such as neural networks, regression analysis) to model traffic vehicle flow data, but they are very dependent on good algorithms, require many environmental parameters, and the results cannot be verified. The network analysis model establishes the mobility characteristics of the traffic vehicle network through graph theory and network analysis techniques, but this method can only be applied to the analysis of the mobility characteristics of traffic vehicles in the short term and on a small scale due to the high calculation cost. Summary of the Invention

[0005] To solve the above-mentioned technical problems such as the deficiencies in the method for analyzing the mobility of traffic vehicles and the limitations in modeling, the present invention provides a modeling method for the travel characteristics of new energy vehicles in cities based on the entropy principle. It not only overcomes the heterogeneity of data caused by the travel regions of electric vehicles but also fully considers the influencing factors of the actual environment on the traffic mobility of electric vehicles, providing a new perspective for the analysis of urban traffic mobility, a new idea for revealing the spatio-temporal distribution law of urban traffic, and can provide a reference for formulating urban traffic development policies.

[0006] In the first aspect of the present invention, there is provided a modeling method for the travel characteristics of new energy vehicles in cities based on the entropy principle, including the following steps:

[0007] Step 1: Based on the travel record data of new energy vehicles, obtain the travel time and travel distance;

[0008] Step 1.1: Obtain a number of travel record data of new energy vehicles in different cities and preprocess the travel record data of new energy vehicles;

[0009] The travel record data of new energy vehicles includes vehicle id, journey start time, longitude of the journey start point, latitude of the journey start point, journey end time, longitude of the journey end point, and latitude of the journey end point;

[0010] The preprocessing is to filter out the travel record data of new energy vehicles where the journey start time and journey end time are not within the range of 0 - 24 hours, and at the same time filter out the travel record data of new energy vehicles where the longitude of the journey start point, latitude of the journey start point, longitude of the journey end point, and latitude of the journey end point are not within the corresponding city range;

[0011] Step 1.2: According to the journey start time, longitude of the journey start point, latitude of the journey start point, journey end time, longitude of the journey end point, and latitude of the journey end point in each preprocessed travel record data of new energy vehicles, extract the travel distance and travel time;

[0012] Step 2: Statistically calculate the daily average travel time of each city, and use the daily average travel time of each city to perform mean value processing on all travel times of that city; the daily average travel time of each city is obtained by summing up the travel times of all new energy vehicles in that city and dividing by the number of new energy vehicles;

[0013] Step 3: Based on the entropy principle, construct a probability distribution model of travel time;

[0014] The probability distribution model of travel time is:

[0015]

[0016] where p(T k ) is the probability of a certain travel time, Tk For a certain travel time, b' is a constant that ensures the continuous integral is 1 for the limiting condition, and b' = e -1-λ , λ and μ are Lagrange multipliers, and β is a value representing the slope relationship between the travel time and the probability distribution function, which determines the exponential nature of the probability distribution function of the travel time, and is the average daily travel time;

[0017] Step 4: Introduce the actual path influence factor into the probability distribution model of the travel time obtained in Step 3 to obtain the probability distribution model of the travel time considering the actual path influence factor;

[0018] Step 4.1: Establish a vehicle flow model;

[0019] It is assumed that within a city, new energy vehicles move between z location points, from the journey starting point i = 1, 2,..., z to the journey ending point i = 1, 2,..., z & i ≠ j, where i and j are the numbers of the location points;

[0020] Within a certain period of time, with as the journey ending point, the number of flowing vehicles is modeled as:

[0021]

[0022] Among them, is the number of flowing vehicles from other journey starting points with as the journey ending point, represents the total number of vehicles with as the journey starting point, represents the moving probability from the journey starting point to the journey ending point ;

[0023] Since the number of flowing vehicles arriving at the journey ending point is N, the total travel time, that is, the vehicle flow model is defined as:

[0024]

[0025] Among them, is the total travel time with as the journey ending point, m is the number of the journey, l is the travel distance, and v m is the average speed of new energy vehicle travel;

[0026] Step 4.2: Establish the relationship expression between the number of journey starting points and the travel distance;

[0027] ​First, grid aggregation processing needs to be performed on each city: For a city, divide the range of its longitude and latitude into several grids, set the side length Δx of each grid according to the total number of grids, and then aggregate all the position points falling within the same grid into one position point. For different new energy vehicle travel record data, this position point is the starting point or the ending point of the journey; then select an ending point of the journey and map all the starting points corresponding to this ending point to a unified coordinate range to represent the relative distance between the starting point and the ending point of the journey: With an ending point of the journey as the center, map all the starting points to different circumferences according to the travel distance;

[0028] When considering the number N of starting points of the journey corresponding to as the ending point of the journey at different travel distances l l , establish the relationship expression between the number of starting points of the journey and the travel distance:

[0029]

[0030] where ρ is the homogeneous density of the starting points of the journey after grid division, and θ is a parameter reflecting the size of the mapping range;

[0031] Step 4.3: According to the relationship expression between the number of starting points of the journey and the travel distance and the vehicle flow model, establish the relationship expression between the total travel time and the travel distance, and then obtain the probability distribution expression of the travel time;

[0032] The relationship expression between the total travel time and the travel distance is:

[0033]

[0034] where a is the linear coefficient, l max is the farthest travel distance, and T is the total travel time under all paths;

[0035] Taking the derivative of both sides of formula (10) with respect to the travel distance, the probability distribution expression of the travel time is:

[0036] p(T k ) = aT k (12)

[0037] Step 4.4: According to the probability distribution expression of the travel time, introduce the actual path influence factor into the travel time probability distribution model obtained in Step 3 to obtain the travel time probability distribution model considering the actual path influence factor for each city;

[0038] The final travel time probability distribution model considering the actual path influence factor is:

[0039]

[0040] where b is a constant for the limiting condition that can ensure the continuous integral to be 1;

[0041] Step 5: Replace the travel time in the probability distribution model of travel time considering the actual path influence factor with the travel time after homogenization processing;

[0042]

[0043] where is the travel time after homogenization processing, is the probability of the travel time after homogenization processing, and α is a parameter;

[0044] Step 6: According to the travel time after homogenization processing, use the maximum likelihood method to estimate the parameters in the probability distribution model of travel time considering the actual path influence factor, obtain the final probability distribution model of travel time, and use it for modeling and analysis of vehicle mobility; the parameters α and β in the probability distribution model of travel time considering the actual path influence factor.

[0045] The second aspect of the invention provides a modeling system for the travel characteristics of new energy vehicles in cities based on the entropy principle, which is used to implement the modeling method for the travel characteristics of new energy vehicles in cities based on the entropy principle, and includes:

[0046] A new energy vehicle travel record data acquisition module, which is used to acquire new energy vehicle travel record data; the new energy vehicle travel record data includes vehicle id, journey start time, longitude of the journey start point, latitude of the journey start point, journey end time, longitude of the journey end point, and latitude of the journey end point;

[0047] A data preprocessing module, which is used to preprocess the new energy vehicle travel record data, and extract the travel time and travel distance from each piece of preprocessed new energy vehicle travel record data;

[0048] A homogenization processing module, which is used to calculate the daily average travel time of the city, and use the daily average travel time of the city to perform homogenization processing on all travel times in the city to obtain the travel time after homogenization processing;

[0049] A model construction module, which is used to construct a probability distribution model of travel time considering the actual path influence factor;

[0050] A parameter estimation module, which is used to estimate the parameters in the probability distribution model of travel time considering the actual path influence factor according to the travel time after homogenization processing by using the maximum likelihood method to obtain the final probability distribution model of travel time;

[0051] The travel characteristic analysis module uses the probability distribution model of the final travel time to obtain the probability distribution of the travel time and analyze vehicle mobility;

[0052] The third aspect of the present invention provides an electronic device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device runs, the processor communicates with the memory through the bus. When the machine-readable instructions are executed by the processor, the steps of the modeling method for the travel characteristics of urban new energy vehicles based on the entropy principle are executed;

[0053] The fourth aspect of the present invention provides a computer-readable storage medium. A computer program is stored in the computer-readable storage medium. When the computer program is run by a processor, the steps of the modeling method for the travel characteristics of urban new energy vehicles based on the entropy principle as described above are executed.

[0054] The beneficial effects of the present invention are:

[0055] 1. The advantage of the present invention is that it directly starts from the characteristics of the daily average travel time in different travel regions, introduces the entropy theory, and derives a mathematical model, avoiding the heterogeneity of data caused by different travel regions, thus affecting the correctness of the conclusion.

[0056] 2. Considering the actual geographical environment influencing factors and introducing an independent term related to the path factor, the established model is more in line with the actual situation.

[0057] 3. Based on the maximum likelihood method to estimate experimental parameters and verify the model with different city tram data. On the one hand, it shows that the reliability of the model is verifiable, and on the other hand, it shows that the model is generally applicable. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 is a schematic flow chart of the modeling method for the travel characteristics of urban new energy vehicles based on the entropy principle in an embodiment of the present invention;

[0059] Figure 2 is a probability distribution diagram of the travel time in an embodiment of the present invention;

[0060] Figure 3 is a schematic diagram of the mapping of the journey starting point in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] The following will further clarify the present invention in conjunction with the drawings and specific embodiments. It should be understood that the following specific embodiments are only for illustration and not for limiting the scope of the present invention.

[0062] Figure 1It is a schematic flowchart of a modeling method for the travel characteristics of new energy vehicles in cities based on the entropy principle. As can be seen from the accompanying drawings, a modeling method for the travel characteristics of new energy vehicles in cities based on the entropy principle according to an embodiment of the present application includes the following steps:

[0063] Step 1: Based on the travel record data of new energy vehicles, obtain the travel time and travel distance;

[0064] Step 1.1: Obtain several travel record data of new energy vehicles in different cities, and preprocess the travel record data of new energy vehicles;

[0065] The travel record data of new energy vehicles includes vehicle id, journey start time, longitude of the journey start point, latitude of the journey start point, journey end time, longitude of the journey end point, and latitude of the journey end point;

[0066] The preprocessing is to filter out the travel record data of new energy vehicles whose journey start time and journey end time are not within the range of 0-24 hours, and at the same time filter out the travel record data of new energy vehicles whose longitude of the journey start point, latitude of the journey start point, longitude of the journey end point, and latitude of the journey end point are not within the corresponding city range;

[0067] In this embodiment, the travel record data of new energy vehicles obtained comes from the GPS positioning collection data officially announced by the government;

[0068] Step 1.2: According to the journey start time, longitude of the journey start point, latitude of the journey start point, journey end time, longitude of the journey end point, and latitude of the journey end point in each preprocessed travel record data of new energy vehicles, extract the travel distance and travel time;

[0069] Step 2: Statistically calculate the daily average travel time of each city, and use the daily average travel time of each city to perform mean processing (normalization) on all travel times of the city; the daily average travel time of each city is obtained by summing up the travel times of all new energy vehicles in the city and dividing by the number of new energy vehicles;

[0070] In this embodiment, in order to verify that the fitting function corresponding to the experimental data conforms to the parameters modeled theoretically later, it is necessary to perform logarithmic binning on the mean travel time, and draw the probability function distribution diagram of the travel time of each city. As Figure 2 shown, the probability function distribution diagram of the travel time shows an approximate exponential distribution in the semi-logarithmic coordinate system, that is, the distribution probability of the travel time of new energy vehicles in the city is in the form of an exponential function;

[0071] Step 3: Construct a probability distribution model of travel time based on the entropy principle;

[0072] Using the entropy method, that is, given a set of random numbers X, when their mean is constant, the distribution with the maximum (most likely) entropy among them is the exponential distribution.

[0073] Therefore, let the entropy of the travel time be:

[0074]

[0075] where ent is the entropy of the travel time, p(T k ) is the probability of a certain travel time, T k is a certain travel time, and p(T k ) satisfies two constraints: is the average daily travel time;

[0076] In order to find the distribution when the entropy of the travel time is maximized under the above two constraints, the Lagrangian maximization function L is obtained according to the Lagrange multiplier method as:

[0077]

[0078] where L is the Lagrangian maximization function, and λ and μ are Lagrange multipliers;

[0079] Then, the Lagrange multipliers λ and μ are determined by differentiation. The differentiation of the Lagrangian maximization function L gives:

[0080] -lnp(T k ) - 1 - λ - μT k =0 (3)

[0081] That is:

[0082]

[0083] In order to make the above result also conform to the distribution of the travel time after mean normalization, the probability distribution model of the travel time is constructed according to formula (4) as:

[0084]

[0085] where b' is a constant that can ensure the continuous integral is 1 under the constraint condition, and b' = e -1-λ , and the Lagrange multiplier β is a value representing the slope relationship between the travel time and the probability distribution function, which determines the exponential property of the probability distribution function of the travel time;

[0086] Step 4: Introduce the actual path influence factor into the probability distribution model of the travel time obtained in Step 3 to obtain the probability distribution model of the travel time considering the actual path influence factor;

[0087] Step 4.1: Establish a vehicle flow model for each city;

[0088] It is assumed that within a city, new energy vehicles move between z location points, from the starting point of a journey (i = 1, 2,..., z) to the end point of the journey (i = 1, 2,..., z & i ≠ j), where i and j are the numbers of the location points;

[0089] Within a period of time, with as the end point of the journey, the number of flowing vehicles is modeled as:

[0090]

[0091] Among them, is the number of flowing vehicles coming from other journey starting points with as the end point of the journey, represents the total number of vehicles with as the starting point of the journey, represents the moving probability from the starting point of the journey to the end point of the journey ;

[0092] Since the number of flowing vehicles arriving at the end point of the journey is N, the total travel time, that is, the vehicle flow model is defined as:

[0093]

[0094] Among them, is the total travel time with as the end point of the journey, m is the number of the journey, l is the travel distance, and v m is the average speed of new energy vehicle travel;

[0095] Step 4.2: Establish a relational expression between the number of journey starting points and the travel distance;

[0096] Because the actually collected new energy vehicle travel record data is longitude and latitude scatter point data, first, grid aggregation processing needs to be carried out for each city: for a city, divide its longitude and latitude range into several grids, set the side length Δx of each grid according to the total number of grids, and then aggregate all the location points falling within the same grid into one location point. For different new energy vehicle travel record data, this location point is the starting point or the end point of the journey; then select an end point of the journey, and map all the journey starting points corresponding to this end point to a unified coordinate range to represent the relative distance between the journey starting point and the end point, as Figure 3 shown: with a journey end point as the center, map all the journey starting points to different circumferences according to the travel distance;

[0097] Set the scale of the starting point of the journey after grid-based aggregation processing

[0098]

[0099] where ρ is the homogeneous density of the starting point of the journey after gridding;

[0100] When considering the number N of starting points of the journey corresponding to as the end point of the journey at different travel distances l l , introduce a parameter θ that can reflect the size of the mapping range, and use the area equivalence method to establish the relationship expression between the number of starting points of the journey and the travel distance:

[0101] N l ρ(Δx / 2) 2 = 2πl(Δx)θ, 0 < θ ≤ 1 (9)

[0102] where θ is the parameter reflecting the size of the mapping range, and the number N of starting points of the journey l increases linearly with the increase of the travel distance l;

[0103] In this embodiment, the size of the grid is 100m * 100m;

[0104] Step 4.3: According to the relationship expression between the number of starting points of the journey and the travel distance and the vehicle flow model, establish the relationship expression between the total travel time and the travel distance, and then obtain the probability distribution expression of the travel time;

[0105] Since the travel time for a travel distance of l is T k = l / v m , and the number of starting points of the journey is N l = a·l, where a is the linear coefficient, then T k ~ N l . Therefore, when considering a travel distance of l, there will be a linear relationship between the travel time T k and the number of starting points of the journey during the journey, that is, the relationship expression between the total travel time and the travel distance is:

[0106]

[0107] where l max is the farthest travel distance, and T is the total travel time under all paths;

[0108] That is:

[0109]

[0110] Taking the derivative on both sides gives the probability distribution expression of the travel time as:

[0111] p(T k ) = aT k (12)

[0112] Step 4.4: According to the probability distribution expression of travel time, introduce the actual path influence factor into the travel time probability distribution model obtained in Step 3 to obtain the travel time probability distribution model considering the actual path influence factor for each city;

[0113] Therefore, when explaining the travel time distribution of the overall vehicle, an independent term T related to the path should be considered to be added. k Therefore, the final travel time probability distribution model considering the actual path influence factor is:

[0114]

[0115] where b is a constant for the constraint condition that can ensure the continuous integral is 1.

[0116] Step 5: Replace the travel time in the travel time probability distribution model considering the actual path influence factor with the mean-processed travel time

[0117]

[0118] where, is the mean-processed travel time, is the probability of the mean-processed travel time, and α is a parameter;

[0119] Step 6: According to the mean-processed travel time, use the maximum likelihood method to estimate the parameters in the travel time probability distribution model considering the actual path influence factor to obtain the final travel time probability distribution model, and use it for the modeling and analysis of vehicle mobility;

[0120] The calculation method for estimating each parameter using the maximum likelihood method is as follows:

[0121]

[0122] where, is the estimated parameter value, L(α, β; x) is the likelihood function containing two parameters, and x is the observed data;

[0123] Based on the travel time after homogenization processing, using Jupyter Notebook software, the parameter estimation of the probability distribution model of the travel time considering the actual path influence factors above is carried out based on the maximum likelihood estimation method, and two parameter estimation values are obtained: α reflects the shape parameter of the model, which determines the fitting degree between the fitting function and the experimental data, while β reflects the scale parameter of the probability distribution model of the travel time considering the actual path influence factors, which determines the property that the fitting function has an exponential distribution;

[0124] The fitting curve of the experiment is as Figure 2 shown. The front section of the curve is dominated by the power term for the rising trend, and the latter section is dominated by the exponential term for the decaying trend. It can be obtained from the graph that when, the overall probability integral proportion is very large, indicating that the travel time of this group of new energy vehicles is fixed within an interval related to the average travel time. Among them, according to statistics, the average travel time is a relatively fixed value, indicating that the travel time of the entire electric vehicle group can be predicted with high probability.

[0125] Verify the general applicability of the model using different city datasets:

[0126] First, collect datasets (Beijing and Shenzhen) of electric vehicle trips from different cities. Then, using the method of data preprocessing in step 1 to extract travel time and distance, and again using Jupyter Notebook software, simulate and fit the collected datasets, compare the two experimental parameters α and β, and test the goodness of fit R 2 and the root mean square error RMSE. Among them, the goodness of fit R 2 is close to 1 and the root mean square error RMSE is also very small, verifying that the model still has good universality under heterogeneous conditions in different geographical regions.

[0127] In this embodiment, based on the high-precision GPS trajectory records of electric vehicle trips in different regions provided free of charge by the government official, with a time span of one month, a dataset containing millions to tens of millions of records is constructed. After data preprocessing, the new energy electric vehicle travel dataset D1 in New York City, the dataset D2 in Shenzhen, and the dataset D3 in Beijing are formed;

[0128] In all the collected traffic datasets, key information such as vehicle id, journey start time, journey start longitude and latitude, journey end time, journey end longitude and latitude, and location name is included. Some information of the New York City dataset is shown in Table 1 specifically;

[0129] Table 1 New Energy Vehicle Travel Record Dataset in New York City

[0130] Vehicle ID Journey start time Journey end time Journey start longitude Journey start latitude Journey end longitude Journey end latitude id1080784 2016 / 2 / 29 16:40 42429.699305556 -73.95391846 40.77887344 -73.96387482 40.77116394 id0889885 2016 / 3 / 11 23:35 42440.995138889 -73.98831177 40.73174286 -73.99475098 40.69493103 id0857912 2016 / 2 / 21 17:59 42421.768055556 -73.99731445 40.72145844 -73.94802856 40.7749176 id3744273 2016 / 1 / 5 9:44 42374.41875 -73.96166992 40.75971985 -73.95677948 40.7806282 id0232939 2016 / 2 / 17 6:42 42417.288888889 -74.01712036 40.70846939 -73.98818207 40.7406311 id1918069 2016 / 2 / 14 18:31 42414.788194444 -73.9936142 40.75188446 -73.99542236 40.72386169 id2429028 2016 / 4 / 20 20:30 42480.858333333 -73.96508026 40.75891495 -73.97680664 40.76410675 id1663798 2016 / 6 / 19 16:48 42540.7125 -73.96389008 40.76543427 -73.87242889 40.77420044 id2436943 2016 / 3 / 28 19:17 42457.825 -73.87288666 40.77428055 -73.97901917 40.76187897 id2933909 2016 / 4 / 10 22:01 42470.934027778 -73.98782349 40.74098206 -73.99915314 40.68645096 id2750279 2016 / 1 / 17 19:40 42386.825694444 -73.99711609 40.7220993 -74.00689697 40.70832062 id1338820 2016 / 4 / 22 23:33 42482.991666667 -74.00379181 40.74182892 -73.99333954 40.72188187 id0670329 2016 / 1 / 13 20:25 42382.854166667 -74.00482178 40.7285881 -73.99702454 40.72712326 id3131886 2016 / 6 / 26 13:39 42547.58125 -73.98880005 40.75274658 -73.99594879 40.76725388 id1054271 2016 / 3 / 17 20:35 42446.865277778 -73.99478912 40.74451065 -73.96574402 40.75407028 id2930166 2016 / 2 / 21 1:56 42422.077777778 -74.00287628 40.73413086 -73.95539856 40.76750183 id3867951 2016 / 4 / 13 22:27 42473.945833333 -73.979599 40.76065826 -73.97838593 40.72956085

[0131] In the present invention, the Jupyter Notebook software is used for implementation. A mathematical model is established based on the entropy theory, and the influencing factors of the actual path are added. Then, the collected experimental data is preprocessed, and the travel time and distance features are extracted. Based on the principle of the maximum likelihood estimation method, the fitting function module of the Jupyter Notebook software is used for parameter estimation. The comparison between the parameter estimation results and the theoretical results is shown in Table 2.

[0132] Table 2 Comparison of parameters α and β and fitting index R of the travel time probability model 2 and RMSE in the comparative experiment

[0133]

[0134] From the results in Table 2 above, it can be found that under the tram travel mode in different urban areas, the experimental fitting parameters α and β are not much different from the theoretically obtained values, and there is consistency in different regions, which proves the unity of the model. In addition, the goodness of fit and the root mean square error also evaluate the superiority of the model.

[0135] The implementation cases of the present invention are described in detail above, but the present invention is not limited to the specific analysis details of the above indicators. Within the overall result range of the present invention, the research index method of the present invention can be extended to the study of the probability distribution of travel distance. The present invention will not list more similar methods for research indicators, and these index analysis methods all belong to the protection scope of the present invention.

[0136] This embodiment provides a modeling system for the travel characteristics of new energy vehicles in cities based on the entropy principle, which is used to implement the modeling method for the travel characteristics of new energy vehicles in cities based on the entropy principle, and includes:

[0137] A new energy vehicle travel record data acquisition module, which is used to acquire new energy vehicle travel record data; the new energy vehicle travel record data includes vehicle id, journey start time, longitude of the journey start point, latitude of the journey start point, journey end time, longitude of the journey end point, and latitude of the journey end point.

[0138] A data preprocessing module, which is used to preprocess the new energy vehicle travel record data and extract the travel time and travel distance from each preprocessed new energy vehicle travel record data.

[0139] An equalization processing module, which is used to calculate the daily average travel time of the city and use the daily average travel time of the city to perform equalization processing on all travel times in the city to obtain the equalized travel time.

[0140] A model construction module, which is used to construct a probability distribution model of travel time considering the influencing factors of the actual path.

[0141] A parameter estimation module, configured to estimate parameters in a probability distribution model of travel time considering actual path influence factors by using the maximum likelihood method based on the travel time after homogenization processing, so as to obtain a final probability distribution model of travel time;

[0142] A travel characteristic analysis module, configured to obtain the probability distribution of travel time by using the final probability distribution model of travel time, and analyze vehicle mobility;

[0143] In this embodiment, an electronic device is provided, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device runs, the processor communicates with the memory through the bus. When the machine-readable instructions are executed by the processor, the steps of the modeling method for urban new energy vehicle travel characteristics based on the entropy principle are executed;

[0144] This embodiment provides a computer-readable storage medium, in which a computer program is stored. When the computer program is run by a processor, the steps of the modeling method for urban new energy vehicle travel characteristics based on the entropy principle are executed.

Claims

1. A modeling method for urban new energy vehicle travel characteristics based on the entropy principle, characterized in that: The following steps are involved: Step 1: Based on the travel record data of new energy vehicles, obtain the travel time and travel distance; Step 2: Count the average daily travel time of each city, and use the average daily travel time of each city to average all the travel times of the city; the average daily travel time of each city is obtained by summing up the travel time of all new energy vehicles in the city and dividing it by the number of new energy vehicles; Step 3: Construct a probability distribution model of travel time based on the entropy principle; Step 4: Introduce the actual path influencing factors into the probability distribution model of travel time obtained in step 3 to obtain the probability distribution model of travel time considering the actual path influencing factors; Step 5: Replace the travel time in the probability distribution model of travel time considering the actual path influencing factors with the travel time after mean processing; Step 6: Based on the averaged travel time, use the maximum likelihood method to estimate the parameters in the probability distribution model of the travel time that takes into account the actual path influencing factors, obtain the final probability distribution model of the travel time, and use it to model and analyze vehicle mobility.

2. The modeling method for urban new energy vehicle travel characteristics based on the entropy principle according to claim 1 is characterized in that: Step 1 specifically includes: Step 1.1: Obtain several pieces of new energy vehicle travel record data in different cities, and pre-process the new energy vehicle travel record data; The new energy vehicle travel record data includes vehicle ID, journey starting time, journey starting longitude, journey starting latitude, journey ending time, journey ending longitude and journey ending latitude; The preprocessing is to filter out the travel record data of new energy vehicles whose starting time and ending time are not within the range of 0-24 hours, and filter out the travel record data of new energy vehicles whose longitude of the starting point, latitude of the starting point, longitude of the ending point and latitude of the ending point are not within the corresponding city range; Step 1.2: Extract the travel distance and travel time according to the starting time of the journey, the longitude of the starting time of the journey, the latitude of the starting time of the journey, the time of the journey end, the longitude of the journey end and the latitude of the journey end in each pre-processed new energy vehicle travel record data.

3. The modeling method of urban new energy vehicle travel characteristics based on the entropy principle according to claim 1 is characterized in that: The probability distribution model of travel time in step 3 is: Among them, p(T k ) is the probability of a certain travel time, T k is a certain travel time, b' is a constant that can ensure the continuous integral is 1, and b' = e -1-λ , λ and μ are Lagrange multipliers, β is the value representing the slope relationship between travel time and probability distribution function, which determines the exponential nature of the probability distribution function of travel time, and is the average daily travel time.

4. The modeling method for urban new energy vehicle travel characteristics based on the entropy principle according to claim 1 is characterized in that: Step 4 specifically includes: Step 4.1: Establish vehicle flow model; Set in a city, the new energy vehicle moves between z locations, starting from the starting point of the journey To the end of the journey Among them, i and j are the numbers of the location points; Over a period of time, For the journey end point, the number of mobile vehicles is modeled as: in, For For the journey end from other journey starting points The number of mobile vehicles coming, Indicates that The total number of vehicles that are the starting point of the trip, From the starting point of the journey To the end of the journey The probability of moving; As we reach the end of our journey The number of mobile vehicles is N, so the total travel time, that is, the vehicle flow model is defined as: in, For is the total travel time to the destination of the journey, m is the number of the journey, l is the travel distance, v m is the average speed of new energy vehicle travel; Step 4.2: Establish the relationship between the number of journey starting points and the travel distance; First, each city needs to be gridded and aggregated: for a city, its longitude and latitude range is divided into several grids, and the side length Δx of each grid is set according to the total number of grids. Then all the location points falling into the same grid are aggregated into one location point. For different new energy vehicle travel record data, this location point is the starting point or the end point of the journey; then a journey end point is selected, and all the journey starting points corresponding to the journey end point are mapped to a unified coordinate range to represent the relative distance between the journey starting point and the journey end point: with a journey end point as the center of the circle, all the journey starting points are mapped to different circumferences according to the travel distance; When considering different travel distances l The number of journey starting points corresponding to the journey end point N l When , the relationship between the number of journey starting points and the travel distance is established: N l p(Δx / 2) 2 =2πl(Δx)θ, 0<θ≤1 (9) Among them, ρ is the homogeneous density of the starting point of the journey after gridding, and θ is a parameter reflecting the size of the mapping range; Step 4.3: Based on the relationship between the number of trip starting points and the travel distance and the vehicle flow model, establish the relationship between the total travel time and the travel distance, and then obtain the probability distribution expression of the travel time; The relationship between total travel time and travel distance is expressed as: Among them, a is the linear coefficient, l max is the longest travel distance, T is the total travel time under all paths; The probability distribution expression of travel time obtained by derivation of travel distance on both sides of formula (10) is: p(T k )=aT k (12) Step 4.4: According to the probability distribution expression of travel time, the actual path influencing factor is introduced into the probability distribution model of travel time obtained in step 3 to obtain the probability distribution model of travel time of each city considering the actual path influencing factor; The final probability distribution model of travel time considering the actual path influencing factors is: Where b is a constant that ensures the continuous integral is 1.

5. The modeling method of urban new energy vehicle travel characteristics based on the entropy principle according to claim 1 is characterized in that: In step 5, the travel time in the probability distribution model of travel time considering the actual path influencing factors is replaced by the travel time after mean processing to obtain: in, is the travel time after mean processing, is the probability of travel time after averaging, and α is a parameter.

6. The modeling method of urban new energy vehicle travel characteristics based on the entropy principle according to claim 1 is characterized in that: Step 6 describes the parameters α and β in the probability distribution model of travel time that takes into account the actual path influencing factors.

7. A modeling system for urban new energy vehicle travel characteristics based on the entropy principle, used to implement the modeling method for urban new energy vehicle travel characteristics based on the entropy principle as described in any one of claims 1 to 6, characterized in that: include: A new energy vehicle travel record data acquisition module is used to acquire new energy vehicle travel record data; The new energy vehicle travel record data includes vehicle ID, journey starting time, journey starting longitude, journey starting latitude, journey ending time, journey ending longitude and journey ending latitude; A data preprocessing module, used to preprocess the travel record data of new energy vehicles, and extract the travel time and travel distance from each piece of preprocessed travel record data of new energy vehicles; The averaging processing module is used to calculate the daily average travel time of a city, and use the daily average travel time of the city to average all the travel times of the city to obtain the averaged travel time; A model building module is used to build a probability distribution model of travel time that takes into account factors affecting the actual path; A parameter estimation module is used to estimate the parameters in the probability distribution model of the travel time taking into account the influencing factors of the actual path according to the travel time after mean processing by using the maximum likelihood method to obtain the final probability distribution model of the travel time; The travel characteristics analysis module uses the final probability distribution model of travel time to obtain the probability distribution of travel time and analyze vehicle mobility.

8. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus. When the machine-readable instructions are executed by the processor, the steps of the method for modeling the travel characteristics of urban new energy vehicles based on the entropy principle as described in any one of claims 1 to 6 are performed.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the method for modeling the travel characteristics of urban new energy vehicles based on the entropy principle as described in any one of claims 1 to 6.

Images