Whole vehicle load prediction method based on LSTM neural network

Abstract

The invention provides a whole vehicle load prediction method based on an LSTM neural network. The method comprises the following steps of collecting vehicle data through a vehicle-mounted terminal; integrating vehicle data, and carrying out data cleaning and standardized processing to obtain training data; training through a long-term and short-term memory neural network model to obtain a load model; and inputting predicted data into the trained load model to predict and acquiring load data. By using the whole vehicle load prediction method based on the LSTM neural network, a big data calculating platform is used, based on an application of the neural network model in a prediction field, a longitudinal dynamic vehicle mass calculation equation is combined, and real-time and accurate prediction of the whole vehicle load can be realized.

Description

technical field [0001] The invention relates to a vehicle load prediction method based on an LSTM neural network, and belongs to the technical fields of Internet of Vehicles and logistics transportation. Background technique [0002] At present, there are two main methods of traditional load forecasting: [0003] Method 1: Based on the longitudinal dynamics, the weight of the vehicle is estimated by the recursive least squares method with the forgetting factor. [0004] Disadvantages: The shifting operation is not considered, and the estimation during the shifting process will lead to low estimation accuracy and slow convergence speed of the algorithm. Moreover, in reality, due to real-time changes in braking force, air resistance, rolling friction and other factors, it is difficult to obtain it accurately, which also affects the prediction accuracy to a certain extent. It is difficult to control the error range between the predicted result and the real value within 15%. ...

AI Technical Summary

Problems solved by technology

[0004] Disadvantages: The shifting operation is not considered, and the estimation during the shifting process will lead to low estimation accuracy and slow convergence speed of the algorithm

[0006] Disadvantages: Additional sensors need to be installed on each vehicle. The price of an ordinary vehicle load sensor is more than RMB 10,000, which is expensive, and the sensor based on deformation is easy to age, which will increase the maintenance cost in the later period, and the calculation of load accuracy also have an impact

Images