UAV Path Planning Based on Multicritic-Delayed Deep Deterministic Policy Gradient

Table 1

The result of algorithms.

	Learning stage			Exploiting stage
	Success	Collision	Loss	Success	Collision	Loss
DDPG	73.6%	19.3%	7.1%	80.5%	10.1%	9.4%
TD3	78.5%	17.1%	4.4%	88.4%	5.6%	6.0%
MCDDPG	81.9%	15.8%	2.3%	92.1%	3.4%	4.5%
MCD	89.8%	10.1%	0.1%	94.3%	1.9%	3.8%