Research Article
Investigations on Driver Unique Identification from Smartphone’s GPS Data Alone
Table 12
Comparison of the proposed method with other existing methods.
| Method | Sensors | Data | Obtained accuracy |
| Our method | GPS only | 38 drivers, 2 months | 82.3% |
| Enev et al. [6] | Brake pedal | 15 drivers | 87% |
| Zhang et al. [7] | Multisensory | 20 drivers, simulated driving | 85% |
| Hallac et al. [8] | 12 sensor readings, other than GPS | 10 cars, 64 drivers | 76% (for 2 drivers) 50% (for 5 drivers) |
| Burton et al. [21] | Multisensor, brake pedal, gas pedal, speed, etc. | 10 drivers, simulated driving | 95% confidence interval |
| Fung et al. [22] | Location and speed | 14-old-age drivers | 30–61% |
| Moreira-Matias and Farah [23] | Multisensor data | Dataset of 217 families collected over one year | 88% |
| Martínez et al. [24] | Multisensor data | 11 different drivers 25 Km (40 min.) data | Above 80% |
|
|