Research Article
Gated Recurrent Unit with RSSIs from Heterogeneous Network for Mobile Positioning
Table 2
Positioning error comparison of deep learning techniques with different memory lengths and RSSIs from different networks.
| Memory length (s) | RNN | LSTM | GRU | Cell | Wi-Fi | Cell + Wi-Fi | Cell | Wi-Fi | Cell + Wi-Fi | Cell | Wi-Fi | Cell + Wi-Fi |
| 1 | 12.43 | 8.00 | 7.63 | 11.42 | 7.23 | 9.06 | 11.18 | 7.10 | 7.16 | 2 | 14.16 | 7.16 | 9.43 | 15.00 | 7.85 | 10.22 | 13.59 | 6.28 | 5.86 | 3 | 24.66 | 9.75 | 10.89 | 24.61 | 9.69 | 12.98 | 20.00 | 9.14 | 7.12 | 5 | 13.13 | 7.16 | 7.36 | 11.55 | 7.14 | 7.30 | 11.24 | 7.18 | 7.08 | 10 | 11.76 | 7.48 | 7.76 | 6.88 | 7.18 | 7.28 | 10.01 | 6.78 | 7.06 | 15 | 11.68 | 6.77 | 7.73 | 6.74 | 6.76 | 6.80 | 6.86 | 6.73 | 6.74 |
|
|