Method for learning driving style based on self-coded regularization network

Abstract

The invention discloses a method for learning driving styles based on a self-coded regularization network. The method mainly comprises the following steps: performing GPS (Global Positioning System) data conversion, performing regularization network self coding, performing target function sum approximation, establishing a run length encoding frame, and establishing the number of drivers, namely, in a group of unknown driving, inputting GPS data of vehicles establishing a statistic characteristic matrix as network input, introducing a marker of a limited training set as a prior into an unsupervised automatic encoder, reconstructing hidden layer RNN (Recurrent Neural Network) characteristics, extracting a neck layer of a regularization self-coding structure as a final driving style characteristic representation layer, and estimating the number of drivers in the driving process. By adopting the method, the limit that the driving style of an unknown driver is hard to describe can be solved, a self-coded regularization network is designed to directly learn driving habits of the driver from the GPS data, then recognition and classification precision of different drivers can be improved, and a relatively safe and accurate method can be provided for design of assistant and automatic driving systems.

Description

technical field [0001] The invention relates to the field of driving recognition, in particular to a method for learning driving style based on self-encoding regularization network. Background technique [0002] Driving recognition is often used in the auto insurance industry and in the fields of autopilot and assisted driving of cars to identify drivers and estimate the number of drivers. In particular, driving style information from GPS data is of particular interest to many fields, because good recognition can answer questions such as whether the driver was driving properly at the time of the accident and how many people shared the car, so reliable information can be made. Risk assessment directly affects the policy fees that insurance companies should pay. In addition, good driving style representation can help to better model and understand human driving behavior, and help improve the design of assisted and automatic driving systems. Although the existing methods can ...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem that existing methods are difficult to describe the driving style of unknown drivers, the purpose of the present invention is to provide a method for learning driving style based on self-encoding regularization network, using self-encoding regularization deep neural network (AutoReNet) and run-length coding framework, Combining supervised and unsupervised feature learning in one architecture, learning driver's driving behavior directly from GPS records improves the accuracy of different driver recognition classification

Images