Driving Score Estimator

Directory and File Structure

.
├─ README.md
├─ config.sh
├─ Estimator.sh
├─ InitData
|   ├─ data
|   |   └─ TripFiles
|   └─ src
|       ├─ service_account.json
|       ├─ unzip.sh
|       └─ vlogger.py
├─ RefineData
|   ├─ data
|   |   ├─ refine_file.csv
|   |   ├─ refine_overspeed.csv
|   |   └─ RefineFiles
|   └─ src
|       ├─ refine_classified.sh
|       ├─ refine_pars.sh
|       └─ refine_overspeed.py
├─ Overspeed
|   ├─ data
|   |   ├─ db.csv
|   |   ├─ gps.csv
|   |   └─ overspeed.csv
|   └─ src
|       ├─ initial_data.py
|       └─ overspeed.py
├─ Groundtruth
|   ├─ data
|   |   ├─ dsi_score.csv
|   |   ├─ hca_score.csv
|   |   └─ kmeans_score.csv
|   └─ src
|       ├─ dsi.py
|       ├─ hca.py
|       └─ kmeans.py
└─ Regression
    ├─ data
    |   ├─ rf_estimate.csv
    |	└─ svm_estimate.csv
    └─ src
        ├─ rf.py
        └─ svm.py

Directory	Explanation
InitData	TripFiles Download and unzip
RefineData	Classify the data using normalized values and create a refined refine_file.csv file
Overspeed	Counts the number of overspeed by comparing the user's driving DB with the road information DB
Groundtruth	Using clustering algorithms, we create groundtruth to be used in the driving score
Regression	Groundtruth-based Driving Score Estimate is performed using regression algorithms

Source code	Explanation	Data
README.md
config.sh	Enables configurations and install at once
Estimator.sh	All programs can be executed
vlogger.py	Download TripFiles from Firebase	TripFiles
unzip.sh	TripFiles: trip file unzip	unzip TripFiles
service_account.json	Google Storage account file
refine_classified.sh	Classify TripFiles according to conditions	RefineFiles
refine_pars.sh	Create as 1 usable file	refine_file.csv
refine_overspeed.py	refine_file data + overspeed_file data	refine_overspeed.csv
initial_data.py	Detects the overspeed of all trips inside Tripfiles and creates initial data
overspeed.py	Detects the overspeed of a one trip and adds data to overspeed.csv	overspeed.csv
dsi.py	DSI code for making Groundtruth	dsi_score.csv
hca.py	HCA code among Clustering Algorithms for making Groundtruth	hca_score.csv
kmeans.py	kmeans code among Clustering Algorithms for making Groundtruth	kmeans_score.csv
rf.py	Random Forest code among Regression Algorithms for Driving Score Estimate	rf_estimate.csv
svm.py	Support Vector Machine code among Regression Algorithms for Driving Score Estimate	svm_estimate.csv

Configuration and Install

Configuration and API Install code.

# sudo source config.sh

Estimator

Usage Help

# ./Estimator.sh -h

or

# ./Estimator.sh -?

Program

# ./Estimator.sh -p [Select]

InitData Mode
Ex)

# ./Estimator.sh -p ID

RefineData Mode
Ex)

# ./Estimator.sh -p RD -n [Distance,Count]

Overspeed Mode
Ex)

# ./Estimator.sh -p O

Merge Mode
Ex)

# ./Estimator.sh -p M

Groundtruth Mode
Ex)

# ./Estimator.sh -p G -n [Distance,Count] -g [Model] (dsi)-w [overspeed,accel,decel,stop,start]

Regression Mode
Ex)

# ./Estimator.sh -p R -g [Model] -r [Model]

Description of additional options by program

Normalization

# ./Estimator.sh -n [Distance,Conunt]

Groundtruth Model

# ./Estimator.sh -g [Model]

When selecting DSI among the groundtruth models, the weight value must be entered.
The corresponding options are:

# ./Estimator.sh -g dsi -w [overspeed,accel,decel,stop,start]

Regression Model

# ./Estimator.sh -r [Model]

Contact

Dankook University

Kwanghee Lee
E-main - [email protected]

Sounghyoun Lee
E-mail - [email protected]