Table 10 Comparison between the proposed prediction approach and previous works.

From: Multistage prediction approach of EVs charging performance in smart transportation systems by deep learning technique

Source

Prediction

Model

Features

Results

Year

First Proposed Prediction Approach

Initial SoC

Required SoC

FDNN

Distance, Road Characteristics, and Weather Data

Index

SoC-In

SoC-Req

2025

SMAPE

0.0007

0.00015

MAE

0.00015

0.00013

MSE

6.8*10−7

1.8*10−7

RMSE

0.00083

0.00043

R2

1

1

Second

Proposed

Prediction

Approach

Initial SOC

Required SOC

Arrival Time

Departure Time

FDNN

Distance, Road Characteristics, and

Weather Data

Index

SoC-In

SoC-Req

2025

SMAPE

0.00044

0.00018

MAE

0.00017

0.00015

MSE

5*10−7

4.2*10−7

RMSE

0.00071

0.00065

R2

1

1

Ref.31

SoC

ANN

SVM

Linear GPR

Ensemble

Boosting

Ensemble

Bagging

Battery Capacity

Battery Voltage

Battery Current

and Battery

Temperature

MSE

0.00054

2021

MAE

0.00027

RMSE

0.02329

R2

0.999

Ref.32

SL, Energy

Consumption

RF, SVM,

XGBoost, &

ANN

HCD, Weather,

Traffic, Events

Data

SMAPE

9.92%

2021

MAE

66.5 min

SMAPE

11.6% Consumption

R2

0.7

Ref.33

Departure Time

XGBoost

HCD, Vehicle Type, Charging Location

MAE

82 min

2020

Ref.34

Energy

Requirements

XGBoost

HCD

MAE

4.6 kWh

2020

R2

0.52

Ref.35

SL, Energy

Consumption

Ensemble

Model of

SVM, RF, &

DKDE

HCD

SMAPE

10.4% Duration

2019

7.5% Consumption