Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios
The autonomous navigation of aerial robots in unknown and complex outdoor environments is a challenging problem that typically requires planners to generate collision-free trajectories based on human expert rules for fast navigation. Presently, aerial robots suffer from high latency in acquiring env...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-09-01
|
| Series: | Drones |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2504-446X/7/10/609 |
| _version_ | 1797574153164816384 |
|---|---|
| author | Kaiyu Hu Huanlin Li Jiafan Zhuang Zhifeng Hao Zhun Fan |
| author_facet | Kaiyu Hu Huanlin Li Jiafan Zhuang Zhifeng Hao Zhun Fan |
| author_sort | Kaiyu Hu |
| collection | DOAJ |
| description | The autonomous navigation of aerial robots in unknown and complex outdoor environments is a challenging problem that typically requires planners to generate collision-free trajectories based on human expert rules for fast navigation. Presently, aerial robots suffer from high latency in acquiring environmental information, which limits the control strategies that the vehicle can implement. In this study, we proposed the SAC_FAE algorithm for high-speed navigation in complex environments using deep reinforcement learning (DRL) policies. Our approach consisted of a soft actor–critic (SAC) algorithm and a focus autoencoder (FAE). Our end-to-end DRL navigation policy enabled a flying robot to efficiently accomplish navigation tasks without prior map information by relying solely on the front-end depth frames and its own pose information. The proposed algorithm outperformed existing trajectory-based optimization approaches at flight speeds exceeding 3 m/s in multiple testing environments, which demonstrates its effectiveness and efficiency. |
| first_indexed | 2024-03-10T21:18:49Z |
| format | Article |
| id | doaj.art-9e83f258ee644e2b947d292181ec955f |
| institution | Directory Open Access Journal |
| issn | 2504-446X |
| language | English |
| last_indexed | 2024-03-10T21:18:49Z |
| publishDate | 2023-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Drones |
| spelling | doaj.art-9e83f258ee644e2b947d292181ec955f2023-11-19T16:15:20ZengMDPI AGDrones2504-446X2023-09-0171060910.3390/drones7100609Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild ScenariosKaiyu Hu0Huanlin Li1Jiafan Zhuang2Zhifeng Hao3Zhun Fan4College of Engineering, Shantou University, Shantou 515063, ChinaCollege of Engineering, Shantou University, Shantou 515063, ChinaCollege of Engineering, Shantou University, Shantou 515063, ChinaCollege of Engineering, Shantou University, Shantou 515063, ChinaCollege of Engineering, Shantou University, Shantou 515063, ChinaThe autonomous navigation of aerial robots in unknown and complex outdoor environments is a challenging problem that typically requires planners to generate collision-free trajectories based on human expert rules for fast navigation. Presently, aerial robots suffer from high latency in acquiring environmental information, which limits the control strategies that the vehicle can implement. In this study, we proposed the SAC_FAE algorithm for high-speed navigation in complex environments using deep reinforcement learning (DRL) policies. Our approach consisted of a soft actor–critic (SAC) algorithm and a focus autoencoder (FAE). Our end-to-end DRL navigation policy enabled a flying robot to efficiently accomplish navigation tasks without prior map information by relying solely on the front-end depth frames and its own pose information. The proposed algorithm outperformed existing trajectory-based optimization approaches at flight speeds exceeding 3 m/s in multiple testing environments, which demonstrates its effectiveness and efficiency.https://www.mdpi.com/2504-446X/7/10/609deep reinforcement learningsoft actor–criticfocus autoencoderunmanned aerial vehicleautonomous navigation |
| spellingShingle | Kaiyu Hu Huanlin Li Jiafan Zhuang Zhifeng Hao Zhun Fan Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios Drones deep reinforcement learning soft actor–critic focus autoencoder unmanned aerial vehicle autonomous navigation |
| title | Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios |
| title_full | Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios |
| title_fullStr | Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios |
| title_full_unstemmed | Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios |
| title_short | Efficient Focus Autoencoders for Fast Autonomous Flight in Intricate Wild Scenarios |
| title_sort | efficient focus autoencoders for fast autonomous flight in intricate wild scenarios |
| topic | deep reinforcement learning soft actor–critic focus autoencoder unmanned aerial vehicle autonomous navigation |
| url | https://www.mdpi.com/2504-446X/7/10/609 |
| work_keys_str_mv | AT kaiyuhu efficientfocusautoencodersforfastautonomousflightinintricatewildscenarios AT huanlinli efficientfocusautoencodersforfastautonomousflightinintricatewildscenarios AT jiafanzhuang efficientfocusautoencodersforfastautonomousflightinintricatewildscenarios AT zhifenghao efficientfocusautoencodersforfastautonomousflightinintricatewildscenarios AT zhunfan efficientfocusautoencodersforfastautonomousflightinintricatewildscenarios |