Research Article
Learning-Based Path Planning Algorithm in Ocean Currents for Multi-Glider
Table 1
Parameters of Doc-CNN (example of 128
128).
| | Layer | Detail | Parameters |
| | Processing layer | Conv-00 | 12 5 kernels with stride 1 | | Pool-00 | 3 3 kernels with stride 1 |
| | Branch 1 | Conv-10 | 20 5 kernels with stride 1 | | Pool-10 | 3 3 kernels with stride 2 | | Res-11 | 20 3 kernels with stride 1 | | Pool-11 | 3 3 kernels with stride 2 | | Res-12 | 20 3 kernels with stride 1 | | Pool-12 | 3 3 kernels with stride 2 | | Res-13 | 20 3 kernels with stride 1 | | Pool-13 | 3 3 kernels with stride 2 | | Fc-1 | 192 nodes | | Fc-2 | 12 nodes |
| | Branch 2 | Conv-20 | 20 5 kernels with stride 1 | | Res-21 | 20 3 kernels with stride 1 | | Res-22 | 20 3 kernels with stride 1 | | Res-23 | 20 3 kernels with stride 1 | | Res-24 | 20 3 kernels with stride 1 | | Conv-25 | 12 3 kernels with stride 1 |
| | Output layer | Fc-3 | 4 nodes | | S-1 | 4 nodes |
|
|
