Figure 7.

The ToM-SNN model on three kinds of agents in random environments. (A) The bystander with the ToM-SNN model's prediction of pedestrian 2's policy characteristic based on the partial past trajectory [e.g., A(a)] is shown in A(b). In another environment, the bystander with the ToM-SNN model can predict the risks of pedestrian 2 and help it to avoid risks. (B) The bystander with the ToM-SNN model predicts that pedestrian 2's policy characteristic is experienced based on the partial past trajectory [e.g., B(a)] is shown in Figure. In another environment, the bystander with the ToM-SNN model can predict the risks of pedestrian 2 and help it to avoid risks. (C) The bystander with the ToM-SNN model predicts that pedestrian 2 is a cautious agent shown in C(b) based on the partial past trajectory [e.g., C(a)]. In another environment, the bystander with the ToM-SNN model did not interfere with pedestrian 2 's action because it evaluated pedestrian 2's next state was safe [C(c)].