Skip to content

2021 CPWC : Deep learning-based Personalized Autonomous VehiclE

Notifications You must be signed in to change notification settings

dyeong-00/2021_DeepPAVE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

2021_CPWC Maneuver Prediction Model

차선 변경 가능 여부 판단 딥러닝 모델을 구축하였다.

Model Introduction

LSTM layer와 1D Convolution layer 3개를 합친 LSTM-FCN 모델을 사용하였으며 Pytorch로 구현하였다.
LSTM-FCN은 시계열 데이터 분류 작업에서 단순 LSTM보다 더 좋은 성능을 달성하는 것으로 알려진 모델이다.
모델 학습 전에 MinMaxScaler를 이용하여 최대 최소 정규화를 진행하였다. 또한 LSTM 블록에서 Dimension Shuffle layer를 통해 train 시간을 단축하였으며 overfitting을 방지하였다.
input data는 10Hz로 받았으며 time step은 30이다.
최종 output은 0 (lane keeping), 1 (left lane change), 2 (right lane change)로 세가지이며 이는 input feature와 합쳐져서 Trajectory Prediction 모델로 넘어간다.

Model Input & Output

Input Feature
Ego: Speed, Acceleration, Yaw angle
현재 차선 정보(끝차선에 있는지): Left edge, Right edge
LB(Left Behind car): Position x, Position y, Speed, Acceleration
LF(Left Front car): Position x, Position y, Speed, Acceleration
B(Behind car): Position x, Position y, Speed, Acceleration
F(Front car): Position x, Position y, Speed, Acceleration
RB(Rignt Behind car): Position x, Position y, Speed, Acceleration
RF(Rignt Front car): Position x, Position y, Speed, Acceleration

Output
0: lane keeping
1: left lane change
2: right lane change

How to Train Behavior Model

Open the ubuntu terminal and go to the path of the folder where behavior_train.py is located.

$ cd path/to/folder

Enter the following command to train the model

$ python behavior_train.py

Enter the path of the csv file and batch size, learning rate for train in the terminal.
You can check the progress at the terminal

Model Result

result

About

2021 CPWC : Deep learning-based Personalized Autonomous VehiclE

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages