End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration
Autonomous navigation of drones using computer vision has achieved promising performance. Nano-sized drones based on edge computing platforms are lightweight, flexible, and cheap; thus, they are suitable for exploring narrow spaces. However, due to their extremely limited computing power and storage...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-01-01
|
| Series: | Drones |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2504-446X/8/2/33 |
| _version_ | 1797298443001004032 |
|---|---|
| author | Ning Zhang Francesco Nex George Vosselman Norman Kerle |
| author_facet | Ning Zhang Francesco Nex George Vosselman Norman Kerle |
| author_sort | Ning Zhang |
| collection | DOAJ |
| description | Autonomous navigation of drones using computer vision has achieved promising performance. Nano-sized drones based on edge computing platforms are lightweight, flexible, and cheap; thus, they are suitable for exploring narrow spaces. However, due to their extremely limited computing power and storage, vision algorithms designed for high-performance GPU platforms cannot be used for nano-drones. To address this issue, this paper presents a lightweight CNN depth estimation network deployed on nano-drones for obstacle avoidance. Inspired by knowledge distillation (KD), a Channel-Aware Distillation Transformer (CADiT) is proposed to facilitate the small network to learn knowledge from a larger network. The proposed method is validated on the KITTI dataset and tested on a Crazyflie nano-drone with an ultra-low power microprocessor GAP8. This paper also implements a communication pipe so that the collected images can be streamed to a laptop through the on-board Wi-Fi module in real-time, enabling an offline reconstruction of the environment. |
| first_indexed | 2024-03-07T22:36:00Z |
| format | Article |
| id | doaj.art-d62acadd55dd4cc59a4253fdbcac55fb |
| institution | Directory Open Access Journal |
| issn | 2504-446X |
| language | English |
| last_indexed | 2024-03-07T22:36:00Z |
| publishDate | 2024-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Drones |
| spelling | doaj.art-d62acadd55dd4cc59a4253fdbcac55fb2024-02-23T15:14:09ZengMDPI AGDrones2504-446X2024-01-01823310.3390/drones8020033End-to-End Nano-Drone Obstacle Avoidance for Indoor ExplorationNing Zhang0Francesco Nex1George Vosselman2Norman Kerle3ITC Faculty Geo-Information Science and Earth Observation, University of Twente, 7522 NH Enschede, The NetherlandsITC Faculty Geo-Information Science and Earth Observation, University of Twente, 7522 NH Enschede, The NetherlandsITC Faculty Geo-Information Science and Earth Observation, University of Twente, 7522 NH Enschede, The NetherlandsITC Faculty Geo-Information Science and Earth Observation, University of Twente, 7522 NH Enschede, The NetherlandsAutonomous navigation of drones using computer vision has achieved promising performance. Nano-sized drones based on edge computing platforms are lightweight, flexible, and cheap; thus, they are suitable for exploring narrow spaces. However, due to their extremely limited computing power and storage, vision algorithms designed for high-performance GPU platforms cannot be used for nano-drones. To address this issue, this paper presents a lightweight CNN depth estimation network deployed on nano-drones for obstacle avoidance. Inspired by knowledge distillation (KD), a Channel-Aware Distillation Transformer (CADiT) is proposed to facilitate the small network to learn knowledge from a larger network. The proposed method is validated on the KITTI dataset and tested on a Crazyflie nano-drone with an ultra-low power microprocessor GAP8. This paper also implements a communication pipe so that the collected images can be streamed to a laptop through the on-board Wi-Fi module in real-time, enabling an offline reconstruction of the environment.https://www.mdpi.com/2504-446X/8/2/33droneobstacle avoidancecomputer visiondepth estimationtransformerknowledge distillation |
| spellingShingle | Ning Zhang Francesco Nex George Vosselman Norman Kerle End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration Drones drone obstacle avoidance computer vision depth estimation transformer knowledge distillation |
| title | End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration |
| title_full | End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration |
| title_fullStr | End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration |
| title_full_unstemmed | End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration |
| title_short | End-to-End Nano-Drone Obstacle Avoidance for Indoor Exploration |
| title_sort | end to end nano drone obstacle avoidance for indoor exploration |
| topic | drone obstacle avoidance computer vision depth estimation transformer knowledge distillation |
| url | https://www.mdpi.com/2504-446X/8/2/33 |
| work_keys_str_mv | AT ningzhang endtoendnanodroneobstacleavoidanceforindoorexploration AT francesconex endtoendnanodroneobstacleavoidanceforindoorexploration AT georgevosselman endtoendnanodroneobstacleavoidanceforindoorexploration AT normankerle endtoendnanodroneobstacleavoidanceforindoorexploration |