TWI675347B - Traffic congestion prediction system - Google Patents
Traffic congestion prediction system Download PDFInfo
- Publication number
- TWI675347B TWI675347B TW107123879A TW107123879A TWI675347B TW I675347 B TWI675347 B TW I675347B TW 107123879 A TW107123879 A TW 107123879A TW 107123879 A TW107123879 A TW 107123879A TW I675347 B TWI675347 B TW I675347B
- Authority
- TW
- Taiwan
- Prior art keywords
- traffic
- congestion
- average
- data
- predicted
- Prior art date
Links
Landscapes
- Traffic Control Systems (AREA)
Abstract
一種交通壅塞資訊預測系統，以所發明之壅塞程度計算方法，預估未來一週分時壅塞資訊；透過週期性地收集道路交通資訊，對當量數、平均車速、平均車道佔有率資訊進行分群處理，從中觀察道路特性，自訂量化指標定義塞車標準，更以一多層神經網路架構交通演算模組，結合所研發的壅塞程度計算方式，運用於交通壅塞程度的預測。A traffic congestion information prediction system uses the invented congestion degree calculation method to estimate the time-of-day congestion information for the next week. By periodically collecting road traffic information, the equivalent number, average speed, and average lane occupancy information are grouped. Observe the characteristics of the road, customize the quantitative indicators to define the traffic jam standards, and use a multi-layer neural network to construct a traffic calculation module. Combined with the developed traffic jam calculation method, it can be used to predict the traffic jam degree.
Description
本發明係有關於一種交通資訊系統，特別是一種交通壅塞資訊預測系統。The invention relates to a traffic information system, in particular to a traffic congestion information prediction system.
塞車，就像是日常生活中的時間小偷，偷去人們的時間而讓許多人不自覺，如果將每個人被偷走的時間累計起來，其數量是相當驚人的；美國運輸研究委員會（Transportation Research Board, TRB）的報告指出，美國一年因交通壅塞而耗費了駕駛人超過 42 億小時在道路上，換算金錢約相當於800億美元，同時也因為交通壅塞導致超過30億加倫的汽油的額外損失。Traffic jams are like time thieves in daily life, stealing people's time and making many people unconscious. If you add up the time that everyone has stolen, the amount is quite amazing; the Transportation Research Board (Transportation Research The Board (TRB) report states that the United States spends more than 4.2 billion hours on roads due to traffic congestion in one year, which translates to approximately 80 billion U.S. dollars. At the same time, more than 3 billion gallons of gasoline are caused by traffic congestion. Additional loss.
交通壅塞除了讓駕駛人卡在車陣中動彈不得外，車輛因低速產生的一氧化碳等廢氣污染環境也相當嚴重，根據英國公路局（The United Kingdom Highway Agency）的統計資料，當車速為10公里／小時，一氧化碳排放量大約是車速120公里／小時的5倍至6倍，可見車輛低速排放廢氣污染環境的影響並不容輕忽。In addition to traffic jams that prevent drivers from moving in the car array, vehicles polluting the environment due to low-speed carbon monoxide and other exhaust gases are also very serious. According to statistics from The United Kingdom Highway Agency, when the speed is 10 km / h The carbon monoxide emissions are about five to six times the vehicle speed of 120 km / h. It can be seen that the impact of low-velocity emissions of vehicles on the environment is not negligible.
再者，交通壅塞也會浪費個人時間並延誤工作時間，日積月累則對國家經濟發展力有相當大的損耗。而用路人被塞在車陣中無法工作，不僅降低其生產力，而且因為壅塞造成心情沮喪，甚至惱怒等情緒變化也是影響工作效率、工作品質及生產率降低的重要因素之一。In addition, traffic congestion will waste personal time and delay working hours. Accumulation of time will cause considerable loss of national economic development capacity. And passers-by are stuck in the car array and cannot work, not only reducing their productivity, but also emotional changes such as frustration and even anger caused by congestion are also one of the important factors affecting work efficiency, work quality and productivity.
有鑒於此，本發明一實施例提出一種交通壅塞資訊的預測。係先透過車輛偵測器，定期將取得之路況資訊，其他影響交通情形之各變數資料，納入資料庫中，再將存放於該資料庫之多元資料進行格式的轉換與清洗。In view of this, an embodiment of the present invention provides a prediction of traffic congestion information. Firstly, through the vehicle detector, the obtained road condition information and other variables that affect the traffic situation are included in the database, and then the multiple data stored in the database are converted and cleaned.
清洗後的資料，轉入交通預測模組單元，以分群方式進行特徵萃取，產生一分群資訊，進一步運用歷史資料結合分群結果，觀察數據特性及趨勢而定義出塞車標準，再依多層神經網路架構，取得一交通演算模組的預測資訊。藉此塞車標準計算前述預測資訊之交通壅塞程度，以此發明一套交通壅塞預測系統。The cleaned data is transferred to the traffic prediction module unit, and features are extracted in a clustering manner to generate a cluster of information. The historical data is combined with the clustering results to observe the characteristics and trends of the data to define traffic jam standards, and then based on the multilayer neural network Framework to obtain forecast information for a traffic calculation module. Based on the traffic congestion criterion, the traffic congestion degree of the aforementioned forecast information is calculated, so as to invent a traffic congestion prediction system.
本發明所提出之交通壅塞資訊預測系統，其核心為可有效預測交通壅塞程度，更貼近用路人的直觀感受，不僅大幅降低因交通壅塞所耗費的沈沒成本，更可為整個城市提升生活效能。此外，亦可利用此系統進行預測使生活便捷度提升，發揚智慧城市之精神。The core of the traffic congestion information prediction system proposed by the present invention is that it can effectively predict the degree of traffic congestion, which is closer to the intuitive feeling of passers-by, which not only greatly reduces the sunk cost due to traffic congestion, but also improves the life efficiency of the entire city. In addition, you can use this system to make predictions to improve the convenience of life and carry forward the spirit of smart cities.
以下在實施方式中詳細敘述本發明之特徵以及內容，根據本說明書揭露之內容、申請專利範圍及圖式，可輕易理解本發明之目的及優點。The features and contents of the present invention are described in detail in the following embodiments. The objects and advantages of the present invention can be easily understood based on the contents disclosed in this specification, the scope of patent applications, and the drawings.
請參照圖1，係為本發明交通壅塞資訊預測系統的實施例示意圖。本發明第一實施例之交通壅塞資訊預測系統1，包含：資料儲存及處理單元10、交通預測模組單元11。Please refer to FIG. 1, which is a schematic diagram of an embodiment of a traffic congestion information prediction system according to the present invention. The traffic congestion information prediction system 1 according to the first embodiment of the present invention includes a data storage and processing unit 10 and a traffic prediction module unit 11.
本發明之交通壅塞資訊預測系統1，依資料儲存及處理單元10，週期性地接收車輛偵測器之量測資料，並將量測資料與背景資料、事件資料組成多元資料，並進行格式轉換及資料清洗，所述的資料清洗是指，如清除離群值、插補歷史平均值等。資料儲存及處理單元10將多元資料進行格式轉換並清除離群值、遺漏值同時利用插補法將其以歷史平均值取代。According to the traffic congestion information prediction system 1 of the present invention, according to the data storage and processing unit 10, the measurement data of the vehicle detector is periodically received, and the measurement data, background data, and event data are combined into multiple data, and format conversion is performed. And data cleaning, the data cleaning refers to, for example, clearing outliers and interpolating historical averages. The data storage and processing unit 10 converts the multivariate data into a format, removes outliers and missing values, and uses interpolation to replace them with historical average values.
在此，多元資料由量測資料、背景資料及事件資料所組成，量測資料指車輛偵測器之平均車速資訊、當量數、平均車道佔有率及平均車距等資訊、背景資料指道路方向、日期、星期、時間、道路數、天氣、國定假日等非經由量測方式所得之資料、事件資料包含其他影響交通壅塞的事件(變數)資料，如交通管制措施、道路施工、事故資訊等。惟前述關於多元資料僅為舉例，非已此為限。Here, the multivariate data consists of measurement data, background data, and event data. The measurement data refers to the average speed information, equivalent number, average lane occupancy, and average distance of the vehicle detector. The background data refers to the direction of the road. , Date, week, time, number of roads, weather, national holidays and other non-measured data, event data contains other events (variables) that affect traffic congestion, such as traffic control measures, road construction, accident information, etc. However, the foregoing information on multiple elements is only an example, and it is not limited to this.
交通預測模組單元11經集群分析並依其分群結果，訂定出塞車標準，並結合多層神經網路架構之壅塞預測模型計算得出預測數值，將預測數值透過壅塞度計算方式轉換成反映道路真實壅塞度之量化指標。The traffic prediction module unit 11 sets a traffic jam standard based on the cluster analysis and the results of the clustering. The traffic prediction module unit 11 calculates the predicted value in combination with the congestion prediction model of the multi-layer neural network architecture. The predicted value is converted to reflect the road through the congestion degree calculation method. Quantitative indicator of true congestion.
換言之，資料儲存及處理單元10，週期性地接收車輛偵測器及其他相關交通資料（包括但不限於平均車速資訊、當量數、平均車道佔有率、道路方向、日期、星期、時間、道路數及平均車距等資訊），進行資料格式轉換及清洗後，透過交通預測模組單元11進行預測，再以該當量數、平均速率、平均佔有率之預測值計算出交通壅塞指標，即塞車標準及壅塞度等。根據發明出的交通壅塞指標計算方法，作為此交通壅塞資訊預測系統的核心價值。In other words, the data storage and processing unit 10 periodically receives vehicle detectors and other related traffic data (including, but not limited to, average speed information, equivalent number, average lane occupancy, road direction, date, week, time, and number of roads And average vehicle distance, etc.), after data format conversion and cleaning, prediction is performed through the traffic prediction module unit 11, and traffic congestion indicators are calculated based on the predicted values of the equivalent number, average rate, and average occupancy rate, that is, the standard of traffic jams And congestion. According to the invention of the traffic congestion indicator calculation method, it serves as the core value of this traffic congestion information prediction system.
交通預測模組單元11係先以集群分析對資料進行特徵萃取，找出歷史資料的不同趨勢跟特性，進而判定塞車標準。另將資料格式轉換、清洗（插補值）後的歷史資料區分為70%的訓練資料集（Train data）、30%的測試資料集（Test data），以訓練資料集透過機器學習技術建構預測模型，再使用均方誤差（MSE）對模型進行驗證，以取得最佳演算模組，確保模型準確度及可信度而取得最佳演算模組。The traffic prediction module unit 11 first uses cluster analysis to extract features from the data, find out different trends and characteristics of historical data, and then determine the traffic jam criteria. In addition, the historical data after data format conversion and cleaning (interpolation value) is divided into 70% of training data set (Train data) and 30% of test data set (Test data). The training data set is used to build predictions through machine learning technology The model is then verified using the mean square error (MSE) to obtain the best calculation module, ensuring the accuracy and credibility of the model to obtain the best calculation module.
根據分群後的資訊，以平均速率及當量數、平均佔有率及當量數兩指標之關係，分別建立y=f(x)、Y=f(X)之關係方程式，y、Y為當量數，x為平均速率，X則為平均佔有率，並透過歷史資料之分佈找出於方程式圖形的反曲點，即，(x’,y’)為道路開始出現壅塞時之平均速率與當量數，(X’,Y’)為道路開始出現壅塞時之平均佔有率及當量數。Based on the information after clustering, the relationship equations of y = f (x) and Y = f (X) are established based on the relationship between the average rate and the number of equivalents, the average occupancy and the number of equivalents, and y and Y are the equivalent numbers. x is the average speed, X is the average occupancy, and the inflection point in the equation graph is found through the distribution of historical data, that is, (x ', y') is the average speed and equivalent number when the road starts to congestion, (X ', Y') are the average occupancy and the equivalent number when road congestion begins.
承前述方式定義其平均佔有率及平均速率（X’, x’）為塞車標準(k,v)，若預測平均速率低於塞車標準v，且預測平均佔有率高於塞車標準k，則判定為塞車 ; 反之，不塞車。由此，可得知不同道路其各自的塞車標準點，以利後續應用於交通壅塞度之計算。The average occupancy rate and average rate (X ', x') are defined as the traffic jam standard (k, v) according to the foregoing method. If the predicted average rate is lower than the traffic jam standard v and the predicted average occupancy rate is higher than the traffic jam standard k, then it is determined Traffic jams; conversely, no traffic jams. From this, we can know the respective traffic jam standard points of different roads, so as to facilitate the subsequent application to the calculation of traffic congestion.
意即，交通預測模組單元11根據其計算出之預測數值，以壅塞度計算方式進行壅塞指標的轉換而反映一特定時間、地點之壅塞程度的量化指標。In other words, the traffic prediction module unit 11 converts the congestion index by the congestion degree calculation method according to the predicted value calculated by the traffic prediction module unit 11 to reflect a quantitative index of the congestion degree at a specific time and place.
本實施例之交通壅塞預測方法，係為使預測能較貼近真實的道路壅塞情形，故於本系統創造出一套壅塞程度的計算方式。The traffic congestion prediction method of this embodiment is to make the prediction closer to the real road congestion situation, so a set of calculation methods for the congestion degree is created in the system.
交通預測模組單元11，係先依分群結果找出平均速率及平均佔有率的塞車標準，再將平均速率、平均佔有率之歷史資料分別進行標準化，以兩者關係所構成之平面座標空間作為壅塞度計算之分佈範圍。擬透過此指標反映道路壅塞程度，壅塞度越高代表道路越壅塞，且為了消除數值範圍差異導致之預測偏差，須先針對平均速率及平均佔有率進行標準化，由於平均速率與道路壅塞程度呈負相關，平均速率越高通常反映道路越順暢，因此需針對平均速率進行反標準化動作，再以反標準化之平均速率及正標準化之平均佔有率，進行壅塞度計算。The traffic prediction module unit 11 first finds the average rate and average occupancy rate according to the clustering results, and then standardizes the historical data of the average rate and the average occupancy respectively, and uses the plane coordinate space formed by the relationship as the plane coordinate space. The distribution range of the congestion degree calculation. It is intended to reflect the degree of road congestion through this indicator. The higher the degree of congestion, the more congested the road. In order to eliminate the prediction deviation caused by the difference in the value range, the average rate and the average occupation rate must be standardized first. Relevantly, the higher the average speed usually reflects the smoother road, so it is necessary to perform a denormalization action on the average speed, and then calculate the congestion degree based on the denormalized average speed and the positively normalized average occupancy rate.
前述所得之壅塞分布區域，係先藉由自訂量化指標的方式對應塞車標準(k,v)，以訂定其壅塞門檻值，將壅塞門檻值訂為壅塞度60%較能反映道路真實之壅塞程度。如圖2所示，即將車輛偵測器所測得之最高歷史平均速率與百分之零的平均佔有率視為壅塞度0%；以車輛偵測器測得之最低歷史平均速率與百分之百的平均佔有率視為壅塞度100%，以此上下限所圍區域，作為道路壅塞程度之計算範圍。The congestion distribution area obtained previously corresponds to the traffic congestion standard (k, v) by customizing quantitative indicators to determine its congestion threshold. Setting the congestion threshold to 60% of the congestion degree can better reflect the trueness of the road. Degree of congestion. As shown in Figure 2, the highest historical average rate and the zero-percent average occupancy rate measured by the vehicle detector are regarded as 0% congestion; the lowest historical average rate and the 100% The average occupancy rate is regarded as 100% congestion, and the area enclosed by the upper and lower limits is used as the calculation range for the degree of road congestion.
意即，交通預測模組單元11根據自訂量化指標進行壅塞門檻值之訂定，並將壅塞門檻值訂為壅塞度60%較能反映道路真實之壅塞程度，並將車輛偵測器所測得之最高歷史平均速率與百分之零的平均佔有率視為壅塞度0%，並以車輛偵測器測得之最低歷史平均速率與百分之百的平均佔有率視為壅塞度100%，以此上下限作為壅塞區之範圍。縱軸、橫軸則分別為由大至小平均速率之標準化與由小到大平均佔有率之標準化，根據平均速率、平均佔有率各自的塞車標準落於壅塞度計算範圍之位置，將該道路之歷史資料分佈區分為塞車和不塞車兩種情形，綠色為不塞車區域，紅色為塞車區域。That is to say, the traffic prediction module unit 11 sets the congestion threshold based on a custom quantified index, and sets the congestion threshold to be 60% of the congestion degree, which can better reflect the true degree of congestion on the road. The highest historical average rate obtained and the average occupation rate of zero percent are regarded as the congestion degree of 0%, and the lowest historical average rate measured by the vehicle detector and the average occupation rate of 100 percent are regarded as the congestion degree of 100%. The upper and lower limits are used as the range of the congestion area. The vertical axis and the horizontal axis are normalized from the average rate from large to small and the average occupancy from small to large, respectively. According to the average traffic rate and average occupancy rate, the traffic jam falls in the position of the congestion degree calculation range. The distribution of historical data is divided into two cases of traffic jams and non-traffic situations. Green is the non-traffic area and red is the traffic congestion area.
如發生預測數值（預測平均速率、預測平均佔有率）未落於塞或不塞之區域，即藍色區域，則分別計算該預測點與歷史塞車資料點中心的距離、歷史不塞車資料點中心的距離，取兩者間距離較近者，以此決定該預測點要投影到塞車或不塞車之區域內，並根據投影位置計算得出壅塞度。If the predicted value (predicted average rate, predicted average occupancy) does not fall in the congested or uncongested area, that is, the blue area, the distance between the predicted point and the center of the historical traffic jam data point, and the center of the historical traffic jam data point, respectively The distance between the two is taken as the shorter one between them to determine the predicted point to be projected into the area where there is a traffic jam or no traffic jam, and the congestion degree is calculated based on the projection position.
以圖2之A點為例，可知其預測平均速率小於該車輛偵測器平均速率之塞車標準，且預測佔有率大於該車輛偵測器平均佔有率之塞車標準，故該預測點標準化後座標落在紅色的塞車區域，再依A點至平均速率與平均佔有率塞車標準之範圍，計算其佔整個塞車區域之比例，進一步以此比例與壅塞門檻值做換算，得出該預測點之壅塞度。Taking point A in Fig. 2 as an example, it can be known that the predicted average rate is less than the traffic jam criterion of the average speed of the vehicle detector, and the predicted occupancy rate is greater than the traffic jam criterion of the vehicle detector's average occupancy rate, so the predicted point is standardized after coordinates Fall in the red traffic jam area, and then calculate the proportion of the entire traffic jam area according to the average speed and average occupancy standard range from point A, and further convert this ratio to the traffic jam threshold to obtain the traffic jam at the predicted point. degree.
意即，交通預測模組單元11所取得之預測平均速率、預測平均佔有率落於壅塞區之位置，進行壅塞度之計算，當發生一預測點為落於塞車，即預測平均速率小於該車輛偵測器平均速率之塞車標準，且預測平均佔有率大於該車輛偵測器平均佔有率之塞車標準；或預測點為落於不塞車區域，即預測平均速率大於該車輛偵測器平均速率之塞車標準，且預測平均佔有率小於該車輛偵測器平均佔有率之塞車標準，則依預測點至平均速率與平均佔有率之塞車標準，計算其佔整個塞車或不塞車區域之比例，並以此比例與壅塞門檻值做換算，得出預測點之壅塞度；如發生預測點未落於塞或不塞車之區域，則分別計算預測點與歷史塞車資料點中心的距離、歷史不塞車資料點中心的距離，取兩者間距離較近者，以此決定預測點要投影到塞車或不塞車之區域內，並根據投影位置計算得出壅塞度。That is to say, the predicted average speed and predicted average occupancy rate obtained by the traffic prediction module unit 11 fall in the location of the congestion area, and the congestion degree is calculated. When a predicted point occurs that the traffic is congested, the predicted average speed is less than the vehicle. The average traffic rate of the detector and the average traffic occupancy rate is greater than the average traffic rate of the vehicle occupancy rate; or if the predicted point falls in the non-traffic area, the predicted average rate is greater than the average rate of the vehicle traffic rate The standard of traffic jam and the predicted average occupancy rate is less than the average occupancy rate of the vehicle detector. Based on the predicted point to the average rate and the average occupancy criterion, calculate its proportion to the entire traffic jam or non-traffic area. This ratio is converted to the congestion threshold to obtain the congestion degree of the predicted points. If the predicted point does not fall in the area where traffic is congested or not congested, the distance between the predicted point and the center of the historical traffic congestion data point and the historical traffic congestion data point are calculated separately. The distance of the center, whichever is closer, determines the predicted point to be projected into the traffic jam area or no traffic jam area. Position calculated congestion degree.
本實施案例之一種應用方式，請參閱圖3：(步驟S2-1)透過取得之量測資料、背景資料、事件資料等多元資料，匯入系統資料庫，進行資料清洗與處理，使其成為可用數據；(步驟S2-2) 以多層神經網路架構計算壅塞度，即利用預測模型結合所發明出的道路壅塞程度計算方法進行壅塞度計算，期間透過分散式系統運算巨量資料並輸出結果，(步驟S2-3)輸出結果以供後續分析暨服務應用。For an application method of this implementation case, please refer to Figure 3: (step S2-1) through the acquired measurement data, background data, event data and other multiple data, import the system database, and perform data cleaning and processing to make it into Available data; (Step S2-2) Calculate congestion degree with multi-layer neural network architecture, that is, use a prediction model in conjunction with the invented road congestion degree calculation method to calculate congestion degree. During the period, a large amount of data is calculated through a distributed system and the result is output. (Step S2-3) Output the result for subsequent analysis and service application.
綜上所述，本發明所提出之交通壅塞資訊預測系統，其核心為可有效預測交通壅塞程度，更貼近用路人的直觀感受，不僅大幅降低因交通壅塞所耗費的沈沒成本，更可為整個城市提升生活效能。此外，亦可利用此交通壅塞預測使生活便捷度提升，發揚智慧城市之精神。In summary, the core of the traffic congestion information prediction system proposed by the present invention is that it can effectively predict the degree of traffic congestion, which is closer to the intuitive feeling of passers-by, which not only greatly reduces the sunk cost due to traffic congestion, but also can provide Cities improve life efficiency. In addition, you can also use this traffic jam forecast to improve the convenience of life and promote the spirit of smart cities.
1‧‧‧交通壅塞資訊預測系統1‧‧‧Traffic Congestion Information Forecasting System
10‧‧‧資料儲存及處理單元10‧‧‧Data storage and processing unit
11‧‧‧交通預測模組單元11‧‧‧Traffic Forecast Module Unit
圖1，為本發明交通壅塞資訊預測系統之實施例示意圖； 圖2，為本發明交通壅塞資訊預測系統之壅塞程度計算示意圖；及 圖3，為本發明之應用流程圖。FIG. 1 is a schematic diagram of an embodiment of a traffic congestion information prediction system according to the present invention; FIG. 2 is a schematic diagram of a congestion degree calculation system of the traffic congestion information prediction system according to the present invention; and FIG. 3 is a flow chart of an application of the present invention.
Claims (6)
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW106127512 | 2017-08-14 | ||
| CN106127512 | 2017-08-14 | ||
| ??106127512 | 2017-08-14 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW201911092A TW201911092A (en) | 2019-03-16 |
| TWI675347B true TWI675347B (en) | 2019-10-21 |
Family
ID=66590487
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW107123879A TWI675347B (en) | 2017-08-14 | 2018-07-10 | Traffic congestion prediction system |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TWI675347B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI781483B (en) * | 2020-11-30 | 2022-10-21 | 神通資訊科技股份有限公司 | Application system for open geospatial consortium data conversion standardization and method thereof |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103730006A (en) * | 2014-01-26 | 2014-04-16 | 吉林大学 | Short-time traffic flow combined forecasting method |
| CN104616498A (en) * | 2015-02-02 | 2015-05-13 | 同济大学 | Markov chain and neural network based traffic congestion state combined prediction method |
| TWM538202U (en) * | 2016-10-07 | 2017-03-11 | Institute Of Transportation Motc | Traffic information report system |
| TW201734973A (en) * | 2016-03-21 | 2017-10-01 | Chunghwa Telecom Co Ltd | Dynamic congestion information prediction system and method establish model for linked road sections to perform prediction |
-
2018
- 2018-07-10 TW TW107123879A patent/TWI675347B/en active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103730006A (en) * | 2014-01-26 | 2014-04-16 | 吉林大学 | Short-time traffic flow combined forecasting method |
| CN104616498A (en) * | 2015-02-02 | 2015-05-13 | 同济大学 | Markov chain and neural network based traffic congestion state combined prediction method |
| TW201734973A (en) * | 2016-03-21 | 2017-10-01 | Chunghwa Telecom Co Ltd | Dynamic congestion information prediction system and method establish model for linked road sections to perform prediction |
| TWM538202U (en) * | 2016-10-07 | 2017-03-11 | Institute Of Transportation Motc | Traffic information report system |
Also Published As
| Publication number | Publication date |
|---|---|
| TW201911092A (en) | 2019-03-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9111442B2 (en) | Estimating incident duration | |
| CN108345666B (en) | Vehicle abnormal track detection method based on time-space isolated points | |
| CN104766475B (en) | Urban traffic bottleneck mining method | |
| CN108399468A (en) | It is a kind of based on vehicle when cost optimization operation Time segments division method | |
| CN102819954A (en) | Traffic region dynamic map monitoring and predicating system | |
| JP2019211814A (en) | Congestion prediction device and congestion prediction method | |
| Yuan et al. | Modeling real-time cycle-level crash risk at signalized intersections based on high-resolution event-based data | |
| JP7300700B2 (en) | Pedestrian Accident Risk Evaluation Method by Point and Pedestrian Accident Risk Judgment System by Point | |
| CN104408914A (en) | Signal intersection single vehicle stopping delay time estimating method and system based on GPS data | |
| CN115662113A (en) | Signalized intersection people-vehicle game conflict risk assessment and early warning method | |
| CN107731007A (en) | The crossing accident Forecasting Methodology to be developed based on traffic conflict random process | |
| TWI675347B (en) | Traffic congestion prediction system | |
| CN110766940A (en) | Method for evaluating running condition of road signalized intersection | |
| CN116957345A (en) | Data processing method for unmanned system | |
| CN109740823B (en) | Taxi taking decision method and system oriented to real-time scene calculation | |
| Guan et al. | Traffic incident duration prediction based on artificial neural network | |
| Badhrudeen et al. | Headway analysis using automated sensor data under Indian traffic conditions | |
| CN104331747A (en) | Method for detecting malicious fare evasion | |
| Yang et al. | Multiobjective evaluation of midblock crosswalks on urban streets based on topsis and entropy methods | |
| CN115798212A (en) | Traffic jam detection method based on taxi track | |
| CN110349410B (en) | Traffic congestion index measuring method | |
| Ben-Edigbe et al. | Selective estimations of empirical roadway capacity | |
| CN108182805B (en) | Highway vehicle monitoring and management method based on belief double-base modeling | |
| Ankunda et al. | Pedestrian crash risk analysis using extreme value models: new insights and evidence | |
| Li | High-resolution data-based methods for arterial traffic volume estimation |