Energy-Efficient UAV Movement Control for Fair Communication Coverage: A Deep Reinforcement Learning Approach

. 2022 Mar 1;22(5):1919. doi: 10.3390/s22051919

Algorithm 1: Pseudocode for the state-based game part.

1.2

1:
Initialize the UAV network.
2:
for each timeslot t in T do
3:
for each UAV i in N do
4:
for each action j in $a c t i o n s$ do
5:
Evaluate $I P$ and $F I$ values using action j.
6:
Find $E_{s e n s}$ and $E_{m o v}$ using action j.
7:
Store $I P, F I, E_{s e n s}, E_{m o v}$ .
8:
end for
9:
Select the action with the maximum reward value using Equation (16).
10:
Update the UAV location based on the action.
11:
while The new UAV location is outside the region or the UAV loses its connectivity do
12:
Cancel the new movement.
13:
Select the next maximum action.
14:
Update the UAV location based on the new action.
15:
end while
16:
Update the last $I P, F I, E_{s e n s}, E_{m o v}$ values for UAV i.
17:
end for
18:
end for