GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control

Research on UAV (unmanned aerial vehicle) path planning and obstacle avoidance control based on DRL (deep reinforcement learning) still faces limitations, as previous studies primarily utilized current perceptual inputs while neglecting the continuity of flight processes, resulting in low early-stag...

Full description

Saved in:
Bibliographic Details
Main Authors: Jingyi Huang, Yujie Cui, Guipeng Xi, Shuangxia Bai, Bo Li, Geng Wang, Evgeny Neretin
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Drones
Subjects:
Online Access:https://www.mdpi.com/2504-446X/9/4/275
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850144858507837440
author Jingyi Huang
Yujie Cui
Guipeng Xi
Shuangxia Bai
Bo Li
Geng Wang
Evgeny Neretin
author_facet Jingyi Huang
Yujie Cui
Guipeng Xi
Shuangxia Bai
Bo Li
Geng Wang
Evgeny Neretin
author_sort Jingyi Huang
collection DOAJ
description Research on UAV (unmanned aerial vehicle) path planning and obstacle avoidance control based on DRL (deep reinforcement learning) still faces limitations, as previous studies primarily utilized current perceptual inputs while neglecting the continuity of flight processes, resulting in low early-stage learning efficiency. To address these issues, this paper integrates DRL with the Transformer architecture to propose the GTrXL-SAC (gated Transformer-XL soft actor critic) algorithm. The algorithm performs positional embedding on multimodal data combining visual and sensor information. Leveraging the self-attention mechanism of GTrXL, it effectively focuses on different segments of multimodal data for encoding while capturing sequential relationships, significantly improving obstacle recognition accuracy and enhancing both learning efficiency and sample efficiency. Additionally, the algorithm capitalizes on GTrXL’s memory characteristics to generate current drone control decisions through the combined analysis of historical experiences and present states, effectively mitigating long-term dependency issues. Experimental results in the AirSim drone simulation environment demonstrate that compared to PPO and SAC algorithms, GTrXL-SAC achieves more precise policy exploration and optimization, enabling superior control of drone velocity and attitude for stabilized flight while accelerating convergence speed by nearly 20%.
format Article
id doaj-art-2afe2c94a9034fa2aade6de7d4e8c949
institution OA Journals
issn 2504-446X
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Drones
spelling doaj-art-2afe2c94a9034fa2aade6de7d4e8c9492025-08-20T02:28:14ZengMDPI AGDrones2504-446X2025-04-019427510.3390/drones9040275GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous ControlJingyi Huang0Yujie Cui1Guipeng Xi2Shuangxia Bai3Bo Li4Geng Wang5Evgeny Neretin6School of Electronics Information, Northwestern Polytechnical University, 127 Youyi West Road, Xi’an 710072, ChinaSchool of Electronics Information, Northwestern Polytechnical University, 127 Youyi West Road, Xi’an 710072, ChinaSchool of Electronics Information, Northwestern Polytechnical University, 127 Youyi West Road, Xi’an 710072, ChinaSchool of Computing, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon Tong, Hong Kong SAR, ChinaSchool of Electronics Information, Northwestern Polytechnical University, 127 Youyi West Road, Xi’an 710072, ChinaSchool of Electronics Information, Northwestern Polytechnical University, 127 Youyi West Road, Xi’an 710072, ChinaMoscow Aviation Institute, Moscow 125993, RussiaResearch on UAV (unmanned aerial vehicle) path planning and obstacle avoidance control based on DRL (deep reinforcement learning) still faces limitations, as previous studies primarily utilized current perceptual inputs while neglecting the continuity of flight processes, resulting in low early-stage learning efficiency. To address these issues, this paper integrates DRL with the Transformer architecture to propose the GTrXL-SAC (gated Transformer-XL soft actor critic) algorithm. The algorithm performs positional embedding on multimodal data combining visual and sensor information. Leveraging the self-attention mechanism of GTrXL, it effectively focuses on different segments of multimodal data for encoding while capturing sequential relationships, significantly improving obstacle recognition accuracy and enhancing both learning efficiency and sample efficiency. Additionally, the algorithm capitalizes on GTrXL’s memory characteristics to generate current drone control decisions through the combined analysis of historical experiences and present states, effectively mitigating long-term dependency issues. Experimental results in the AirSim drone simulation environment demonstrate that compared to PPO and SAC algorithms, GTrXL-SAC achieves more precise policy exploration and optimization, enabling superior control of drone velocity and attitude for stabilized flight while accelerating convergence speed by nearly 20%.https://www.mdpi.com/2504-446X/9/4/275SACself-attention mechanismTransformerUAV control decision-makingmultimodal data
spellingShingle Jingyi Huang
Yujie Cui
Guipeng Xi
Shuangxia Bai
Bo Li
Geng Wang
Evgeny Neretin
GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control
Drones
SAC
self-attention mechanism
Transformer
UAV control decision-making
multimodal data
title GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control
title_full GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control
title_fullStr GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control
title_full_unstemmed GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control
title_short GTrXL-SAC-Based Path Planning and Obstacle-Aware Control Decision-Making for UAV Autonomous Control
title_sort gtrxl sac based path planning and obstacle aware control decision making for uav autonomous control
topic SAC
self-attention mechanism
Transformer
UAV control decision-making
multimodal data
url https://www.mdpi.com/2504-446X/9/4/275
work_keys_str_mv AT jingyihuang gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol
AT yujiecui gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol
AT guipengxi gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol
AT shuangxiabai gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol
AT boli gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol
AT gengwang gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol
AT evgenyneretin gtrxlsacbasedpathplanningandobstacleawarecontroldecisionmakingforuavautonomouscontrol