Research on the Agricultural Machinery Path Tracking Method Based on Deep Reinforcement Learning

Table 3

DQN algorithm path tracking results.

	Target end	Path length (m)	Moving time (s)
1	(−1.13, 0.67)	1.49	8.19
2	(0.84, 1.04)	1.49	8.21
3	(−0.27, 0.34)	1.61	8.78
4	(−1.17, −0.96)	1.73	9.16
5	(0.86, 1.36)	1.85	9.43
6	(0.05, 1.66)	1.89	9.90
7	(−0.91, 1.41)	1.95	10.60
8	(−1.18, 1.33)	1.86	10.83
9	(−1.29, 1.50)	2.17	11.56
10	(1.44, 1.98)	—	—