Autonomous Navigation of the UAV through Deep Reinforcement Learning with Sensor Perception Enhancement

Algorithm 1

ANOAU algorithm for autonomous navigation of the UAV.

Input: Actor network parameters
Critic network parameters
Enhance network parameters
Output: UAV’s navigation policy
Initialize , , , episode experience pool
forepisode = 1, 2, …Mdo
Initialize UAV status and get sensor information
forstep = 1, …Tdo
Enhance by use sensor perception enhancement, generate
Select a UAV action  =
Obtain reward and observation by run UAV action
Store episode experience in where handle by use sensor perception enhancement
end
fori = 1, …Ndo
Sample minibatch of experience from D
Gradient descent step on Equation 26 with respect to the weights ,
Gradient aescent step on Equation 25 with respect to the weights
end
	end