Robot Navigation in Crowd Based on Dual Social Attention Deep Reinforcement Learning

Table 2

Results in the visible setting.

Methods	Success	Collison	Time (s)	Reward
ORCA	1.00	0.00	12.19	0.2552
CADRL	0.92	0.08	10.76	0.2734
LSTM-RL	0.93	0.07	10.42	0.2854
SARL	0.99	0.01	10.34	0.3493
DSARL (our)	1.00	0.00	10.24	0.3505