A Heavy Goods Vehicle Traffic Flow Prediction Method Based on Incomplete Positioning Data

Abstract

The invention relates to a heavy-duty truck traffic flow prediction method based on non-full sample positioning data. Aiming at the problem of incomplete acquisition of GPS traffic flow, a scheme of first enlarging samples and then predicting is proposed. In the sampling expansion method, a piecewise constant coefficient method based on the distribution of the traffic itself is proposed. In the prediction method, a long-short-term cyclic neural network LSTM model that can solve the long-term memory problem is proposed, so as to achieve the goal that the traffic flow of heavy trucks is closer to the actual flow. The invention overcomes the defects of low precision of traditional coils and video flow collection during data collection, uses GPS data of a satellite positioning device to predict traffic flow, and has high precision. The invention provides a set of more comprehensive and comprehensive methods for predicting the traffic flow of heavy trucks, and can play an active role in alleviating the problem of urban traffic congestion and improving the operating efficiency of urban traffic. The invention has practical application value in engineering, and can be transferred and applied in related fields.

Description

technical field [0001] The invention relates to the field of intelligent transportation, in particular to a heavy truck traffic flow prediction method based on non-full sample positioning data. Background technique [0002] Traffic flow prediction is an important research content of intelligent transportation system, its accuracy and real-time performance play an important role in alleviating traffic congestion. Traditional traffic flow forecasting methods are often based on passenger cars, and there are few in-depth discussions on the flow forecasting of heavy goods vehicles. In addition, traditional passenger car flow data acquisition is mainly based on coil detection and video detection, which often has the problems of inaccurate vehicle type detection and missing flow counts. At the same time, there is no way to comprehensively detect all lanes. In 2014, the State promulgated the "Measures for the Dynamic Supervision and Management of Road Transport Vehicles", requiring...

AI Technical Summary

Problems solved by technology

[0003] In reality, there are still a small number of heavy goods vehicles that are not equipped with satellite positioning devices, and there are cases where the device is damaged and cannot transmit data in real time

Images