Human-robot handover on HandoverSim (S1: Unseen Subjects)
75Success Rate (%)Ours
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| OursExecution Mode=Sequential, Evaluation with ground-truth states=No, Fine-tuning=Yes2023.03 | 75 | 7.108 | 0.159 | 7.267 | 5 | 12.59 | 7.41 | 25 | |
| OursExecution Mode=Simultaneous, Evaluation with ground-truth states=No, Fine-tuning=Yes2023.03 | 73.33 | 5.633 | 0.158 | 5.791 | 5.56 | 15.37 | 5.74 | 26.67 | |
| Ours w/o ft.Execution Mode=Sequential, Evaluation with ground-truth states=No, Fine-tuning=No2023.03 | 68.15 | 7.151 | 0.164 | 7.314 | 6.85 | 12.96 | 12.04 | 31.85 | |
| OMG PlannerExecution Mode=Sequential, Evaluation with ground-truth states=Yes2023.03 | 62.78 | 8.012 | 1.355 | 9.366 | 33.33 | 2.22 | 1.67 | 37.22 | |
| Yang et al.Execution Mode=Sequential, Evaluation with ground-truth states=Yes2023.03 | 62.78 | 4.719 | 0.039 | 4.758 | 14.44 | 7.78 | 15 | 37.22 | |
| Ours w/o ft.Execution Mode=Simultaneous, Evaluation with ground-truth states=No, Fine-tuning=No2023.03 | 62.78 | 5.695 | 0.164 | 5.859 | 5.93 | 17.59 | 13.7 | 37.22 | |
| GA-DDPGExecution Mode=Sequential, Evaluation with ground-truth states=No2023.03 | 55 | 6.791 | 0.136 | 6.927 | 8.89 | 15 | 21.11 | 45 | |
| GA-DDPGExecution Mode=Simultaneous, Evaluation with ground-truth states=No2023.03 | 33.33 | 4.261 | 0.132 | 4.393 | 15.56 | 21.67 | 29.44 | 66.67 |