论文部分内容阅读
Summary
Nowadays，with the development of science and technology and social life，more and more people choose to go out by car，and traffic jams occur from time to time. Therefore，the paper builds a congestion length model based on vehicle GPS data，and adopts the time series prediction algorithm to predict the time of traffic congestion.
In the first stage，a fuzzy comprehensive evaluation method is used to establish an evaluation system for factors affecting traffic congestion. In the second stage，time series prediction algorithm is adopted based on the traffic jam length model，and historical data is used to predict the future traffic jam length. In the third stage，the cellular automata traffic simulation method is adopted.
To sum up，the model in this paper is a feasible and reasonable model that can adapt to various situations，hoping to provide a new solution for predicting the problem of traffic jam time.
Keywords：fuzzy comprehensive evaluation，evaluation system，traffic congestion length model，time series prediction algorithm，cellular automata traffic simulation
Establishment and solution of the model
1. Selection of evaluation factors for traffic congestion
The average speed of traffic flow is the average speed of all vehicles on a road. This index can directly reflect the permanent condition of the road. Generally speaking，the higher the average speed of traffic flow is，the more unobstructed the road is，and the less the traffic jam is. The smaller the average speed of traffic flow，the more blocked the road and the more serious the traffic jam. The calculation formula is as follows：
2. Prediction model of road congestion length
According to the above estimation of the road jam length and the average speed of traffic flow，the road congestion situation can be obtained. In the time period ，two error signals are defined
Let the smoothing parameter be
When the model is accurate， should be close to 0，while the smoothing error is always more than zero，so the value range of smoothing parameter is between 0 and 1. The formula of the exponential smoothing prediction model is obtained as follows
By putting the predicted traffic flow data into the estimation model of traffic jam length，the prediction of traffic jam length can be realized
3. Traffic congestion time prediction model
The prediction model of traffic jam time in this paper is based on the time series model. in the first step. Previously，the length of traffic jam has been estimated. Now select a certain period of traffic jam，and we can get the calculation formula of the time of historical traffic jam
Now considering the weighted average of the historical data in chronological order as average，which USES time sequence in the case of exponential smoothing model to forecast the traffic time：
The formula is based on the moving average formula，and the recurrence formula of the known moving average is
Now as to be the best estimate，then have
References
[1] analysis and research on GPS coordinate time series of Ming Feng [J]，journal of surveying and mapping，2019，48（10）：1340.
[2] Pei Yujie，Zhang qianwen，Dong jiali. Application research of MATLAB simulation technology in evaluating the impact of community opening on road traffic [J]. Jiangsu science and technology information，2017（07）：61-62.
Nowadays，with the development of science and technology and social life，more and more people choose to go out by car，and traffic jams occur from time to time. Therefore，the paper builds a congestion length model based on vehicle GPS data，and adopts the time series prediction algorithm to predict the time of traffic congestion.
In the first stage，a fuzzy comprehensive evaluation method is used to establish an evaluation system for factors affecting traffic congestion. In the second stage，time series prediction algorithm is adopted based on the traffic jam length model，and historical data is used to predict the future traffic jam length. In the third stage，the cellular automata traffic simulation method is adopted.
To sum up，the model in this paper is a feasible and reasonable model that can adapt to various situations，hoping to provide a new solution for predicting the problem of traffic jam time.
Keywords：fuzzy comprehensive evaluation，evaluation system，traffic congestion length model，time series prediction algorithm，cellular automata traffic simulation
Establishment and solution of the model
1. Selection of evaluation factors for traffic congestion
The average speed of traffic flow is the average speed of all vehicles on a road. This index can directly reflect the permanent condition of the road. Generally speaking，the higher the average speed of traffic flow is，the more unobstructed the road is，and the less the traffic jam is. The smaller the average speed of traffic flow，the more blocked the road and the more serious the traffic jam. The calculation formula is as follows：
2. Prediction model of road congestion length
According to the above estimation of the road jam length and the average speed of traffic flow，the road congestion situation can be obtained. In the time period ，two error signals are defined
Let the smoothing parameter be
When the model is accurate， should be close to 0，while the smoothing error is always more than zero，so the value range of smoothing parameter is between 0 and 1. The formula of the exponential smoothing prediction model is obtained as follows
By putting the predicted traffic flow data into the estimation model of traffic jam length，the prediction of traffic jam length can be realized
3. Traffic congestion time prediction model
The prediction model of traffic jam time in this paper is based on the time series model. in the first step. Previously，the length of traffic jam has been estimated. Now select a certain period of traffic jam，and we can get the calculation formula of the time of historical traffic jam
Now considering the weighted average of the historical data in chronological order as average，which USES time sequence in the case of exponential smoothing model to forecast the traffic time：
The formula is based on the moving average formula，and the recurrence formula of the known moving average is
Now as to be the best estimate，then have
References
[1] analysis and research on GPS coordinate time series of Ming Feng [J]，journal of surveying and mapping，2019，48（10）：1340.
[2] Pei Yujie，Zhang qianwen，Dong jiali. Application research of MATLAB simulation technology in evaluating the impact of community opening on road traffic [J]. Jiangsu science and technology information，2017（07）：61-62.