This paper discusses the use of perception technology in autonomous driving systems, which involves the use of sensors such as LiDAR, cameras, and millimeter wave radar to interact with the external environment. However, single sensors have limitations, and multi-sensor fusion has become a trend in autonomous driving systems. The successful application of sensor fusion requires overcoming several challenges, such as the need for highly sophisticated algorithms and computational capabilities and the ability to navigate efficiently in complex and changing environments. The paper proposes a method to improve the perception effect of low-speed unmanned vehicles by fusing laser and vision sensor data. Firstly, the paper proposes a positioning method based on NDT point cloud alignment algorithm and high precision map to achieve low-cost and high-precision positioning of low-speed unmanned vehicles. Secondly, the paper proposes a target detection method of LiDAR and camera fusion to achieve efficient and fast detection of obstacles during unmanned vehicle travel. Finally, a multi-sensor fusion-based low-speed unmanned vehicle platform is designed and implemented to verify the effectiveness and superiority of the proposed method. The innovative research accomplished in this paper helps to improve the sensing accuracy and robustness of low-speed unmanned vehicles, which has positive significance and reference value for its practical application in real environments.
Research on Unmanned System Environment Perception System Methodology
Lect. Notes in Networks, Syst.
International Workshop on Advances in Civil Aviation Systems Development ; 2023 ; Kyiv, Ukraine May 29, 2023 - May 31, 2023
Proceedings of the International Workshop on Advances in Civil Aviation Systems Development ; Chapter : 17 ; 219-233
2023-07-21
15 pages
Article/Chapter (Book)
Electronic Resource
English
Unmanned environment perception and navigation system based on digital twin technology
European Patent Office | 2023
|