Research Article
End-to-End Autonomous Exploration with Deep Reinforcement Learning and Intrinsic Motivation
Table 5
Experiment results of learning exploration with fine-tuning method (no extrinsic reward).
| Environment | Method | Reward | MER (%) | IQRE |
| Maze-1 | ICM + scratch | 584.59 | 100.00 | 7.93 | Ours + scratch | 586.32 | 100.00 | 4.72 | ICM + fine-tuning | 585.16 | 100.00 | 7.58 | Ours + fine-tuning | 585.45 | 100.00 | 5.14 | Maze-2 | ICM + scratch | 567.28 | 100.00 | 8.07 | Ours + scratch | 571.87 | 100.00 | 5.15 | ICM + fine-tuning | 566.34 | 100.00 | 6.49 | Ours + fine-tuning | 568.25 | 100.00 | 4.81 | Maze-3 | ICM + scratch | 532.27 | 91.64 | N/A | Ours + scratch | 579.65 | 100.00 | 6.54 | ICM + fine-tuning | 573.49 | 100.00 | 7.23 | Ours + fine-tuning | 572.86 | 100.00 | 4.73 |
|
|