RT Journal Article
T1 Reinforcement and Curriculum Learning for Off-Road Navigation of an UGV with a 3D LiDAR
A1 Sánchez-Montero, Manuel
A1 Morales-Rodríguez, Jesús
A1 Martínez-Rodríguez, Jorge Luis
K1 Robótica
K1 Robots autónomos
K1 Aprendizaje automático (Inteligencia artificial)
AB This paper presents the use of deep Reinforcement Learning (RL) for autonomous navigationof an Unmanned Ground Vehicle (UGV) with an onboard three-dimensional (3D) Light Detectionand Ranging (LiDAR) sensor in off-road environments. For training, both the robotic simulatorGazebo and the Curriculum Learning paradigm are applied. Furthermore, an Actor–Critic NeuralNetwork (NN) scheme is chosen with a suitable state and a custom reward function. To employ the3D LiDAR data as part of the input state of the NNs, a virtual two-dimensional (2D) traversabilityscanner is developed. The resulting Actor NN has been successfully tested in both real and simulatedexperiments and favorably compared with a previous reactive navigation approach on the same UGV.
PB MDPI
YR 2023
FD 2023-03-18
LK https://hdl.handle.net/10630/26864
UL https://hdl.handle.net/10630/26864
LA eng
NO Sánchez M, Morales J, Martínez JL. Reinforcement and Curriculum Learning for Off-Road Navigation of an UGV with a 3D LiDAR. Sensors. 2023; 23(6):3239. https://doi.org/10.3390/s23063239
NO Partial funding for open access charge: Universidad de Málaga
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 20 ene 2026