Automatic driving decision-making method and system based on switch-type deep learning network model

Abstract

The invention discloses an automatic driving decision-making method based on a switch-type deep learning network model. The method comprises the following steps: collecting driving environment data invehicle driving under each navigation instruction through a plurality of cameras; and S11, importing the acquired driving environment data and a real-time navigation instruction into a pre-trained and optimized automatic driving decision-making module which adopts a switch type deep learning network model, wherein the switch-type deep learning network model comprises a multi-stage convolutional neural network (CNN) layer, a feature selection layer, a long-term and short-term memory (LSTM) neural network layer and an output layer; and S12, outputting a steering wheel angle and an expected driving speed of the vehicle by the automatic driving decision module according to the input driving environment data and the type of the real-time navigation instruction so as to control the vehicle to realize automatic driving. The invention further discloses a corresponding system. Different feature switches can be selected and activated according to different driving instructions, and automatic driving under complex road conditions can be achieved.

Description

technical field [0001] The invention belongs to the field of automobile automatic driving, and relates to an automatic driving decision-making method and system based on a switch type deep learning network model. Background technique [0002] The method of using deep learning to realize automatic driving of vehicles is the cutting-edge automatic driving algorithm model in the industry. The general idea is to design a deep learning network, then use the original image collected by the sensor as the input of the deep learning network, and then output operations such as braking, acceleration, and steering through the network as outputs, and then train the deep learning network. The advantage is that the model can respond directly to the sensory input without the intervention of humans writing rules. This deep learning autonomous driving technology means that as long as people provide enough training data, the system can automatically learn driving skills. [0003] But in some...

AI Technical Summary

Problems solved by technology

However, the existing deep learning automatic driving system uses the original image as input and outputs the steering wheel angle. It can only drive in a single lane, and cannot complete driving tasks such as intersections, ramps, and lane changes.

Moreover, the existing deep learning automatic driving system only uses the steering wheel angle as the output, without controlling the brake and accelerator, and the achieved automatic driving ability is relatively weak.

Images