Figures & data
Table 1. Comparison of network structures used in 3 different continuous actions space DRL methods.
Table 2. Environment Settings.
Table 3. Reward function.
Figure 10. Reward comparison of single image model, continuous image model and short-term memory model during training.
![Figure 10. Reward comparison of single image model, continuous image model and short-term memory model during training.](/cms/asset/e091a60a-8db6-44fa-849c-c3b73e477ac6/ccos_a_2183828_f0010_oc.jpg)
Figure 11. Average moving distance comparison of single image model, continuous image model and short-term memory model during training.
![Figure 11. Average moving distance comparison of single image model, continuous image model and short-term memory model during training.](/cms/asset/ed0f8e77-9168-42fb-bd40-99022e188b8c/ccos_a_2183828_f0011_oc.jpg)
Table 4. MobileNet V2 structure.
Table 5. Test results of short-term memory model.
Table 6. Test results of continuous images model.
Table 7. Test results of single image model.
Table 8. Test results of Mobile Net + TD3 model.
Table 9. Test results in rearranged environment.
Table 10. Side collision count in 100 flights of 3 models.