Table 2 Performance comparison of SoC estimation methods across different drive cycles.

From: Quaternion generative adversarial -driven Soc estimation using Tyrannosaurus optimizer for improving hybrid electric vehicles renewably powered energy management

Drive cycle

Methods

RMSE

MSE

MAPE

MAE

BJDST

SFO-EKF16

0. 74

0.049

6.492

0. 418

HMDNN17

0. 625

0.053

8.318

0. 569

IWHODL-BMS18

0. 596

0.092

6.31

0. 635

WSO-HDLNN20

0. 699

0.055

5.32

0. 486

CNN-BWGRU21

0. 422

0.030

6.730

0.586

Proposed

0.093

0.0086

2.14

0.281

US06

SFO-EKF16

0. 84

0.029

4.592

0. 618

HMDNN17

0. 675

0.053

7.318

0.549

IWHODL-BMS18

0. 596

0.062

5.31

0. 735

WSO-HDLNN20

0. 499

0.075

6.32

0. 986

CNN-BWGRU21

0. 322

0.040

7.830

0.486

Proposed

0.0389

0.002

0.235

0.132

FUDS

SFO-EKF16

0. 64

0.039

5. 592

0. 318

HMDNN17

0. 725

0.043

9.318

0. 49

IWHODL-BMS18

0. 696

0.092

6.31

0. 935

WSO-HDLNN20

0. 799

0.055

4.32

0. 886

CNN-BWGRU21

0. 622

0.030

7. 30

0.386

Proposed

0.101

0.01

3.342

0.289