Table 2 Parameters for training algorithms

Maximum number of episodes \(N_{epi}\)

5000

Replay buffer capacity \(C\)

100,000

Initial exploration \(\varepsilon_{0}\)

0.5

Exploration decay rate \(\alpha\)

0.998

Slide window size \(N_{1}\)

32

Outage penalty weight \(\mu\)

30

Obstacle avoidance weight \(\eta\)

50

Maximum step per episode \(N_{step}\)

300

Reaching destination tolerating distance \(D_{tol}\)

20 m