notesum.ai

Published at October 18

Benchmarking Deep Reinforcement Learning for Navigation in Denied Sensor Environments

cs.LG

cs.AI

Released Date: October 18, 2024

Authors: Mariusz Wisniewski¹, Paraskevas Chatzithanos¹, Weisi Guo¹, Antonios Tsourdos¹

Aff.: ¹Cranfield University

Arxiv: https://arxiv.org/abs/2410.14616v1

Refer to caption

Parameter	Value
Learning Rate	0.03 / 0.003 / 0.0003 / 0.00003 (see LR study)
Tau	0.005
Learning Starts (after, steps)	5000
Batch Size	16384 (Lidar), 256 (Camera)
Discount Factor	0.99
Exploration Noise	0.3