A Path-Planning Approach Based on Potential and Dynamic Q-Learning for Mobile Robots in Unknown Environment

Table 2

A comparison between PDQL, CQL, RRT, APF, and fuzzy for path planning. Both the lowest (best) mean and the values of path length bigger than the level 0.05 of significance are highlighted.

Env	Statistics	PDQL	CQL	RRT	APF	Fuzzy
M01	Path length	26.93	28.06	32.02	27.36	28.12
	t-test	—	4.94e − 2	2.16e − 8	—	—
	Angle	0.14	0.59	0.74	0.02	0.02
	Time	5.88	6.17	7.41	2.89	3.43
M02	Path length	29.45	30.79	32.98	31.75	30.83
	t-test	—	8.06e − 6	3.57e − 7	—	—
	Angle	1.21	1.57	0.95	0.93	0.98
	Time	7.11	7.33	7.71	4.12	4.54
M03	Path length	30.17	31.91	32.26	31.42	32.68
	t-test	—	2.84e − 4	3.56e − 5	8.29e − 4	1.69e − 6
	Angle	1.14	1.63	1.28	1.49	1.30
	Time	6.50	7.03	5.03	3.88	4.51
M04	Path length	24.16	24.85	24.91	24.30	26.40
	t-test	—	3.47e − 2	1.32e − 4	—	—
	Angle	0.59	0.66	0.71	0.65	0.72
	Time	6.73	7.12	7.21	3.69	4.32
M05	Path length	22.63	25.23	24.53	26.28	26.99
	t-test	—	0.27e − 2	2.77e − 2	—	—
	Angle	1.62	1.74	1.80	1.68	1.65
	Time	6.85	7.38	6.03	3.46	3.78
M06	Path length	24.83	27.26	25.99	27.02	29.02
	t-test	—	2.94e − 4	5.71e − 2	—	—
	Angle	1.57	1.86	1.81	1.82	1.65
	Time	6.08	6.56	7.02	4.29	4.37
M07	Path length	31.60	32.81	33.57	34.55	34.56
	t-test	—	1.94e − 3	3.01e − 5	—	—
	Angle	2.37	2.50	2.03	1.88	1.82
	Time	6.64	7.26	7.26	3.35	3.67
M08	Path length	30.44	31.41	32.81	32.61	32.31
	t-test	—	6.61e − 3	9.84e − 6	—	—
	Angle	2.73	2.77	2.28	2.45	2.58
	Time	6.51	7.14	7.84	4.10	4.50
M09	Path length	29.99	31.18	30.67	31.83	32.15
	t-test	—	8.54e − 3	2.57e − 2	—	—
	Angle	1.89	1.96	1.76	2.05	2.10
	Time	6.56	6.79	6.22	4.04	4.48