分体式飞行汽车立体环境感知系统设计及试验研究
Design and Experiment Research of 3D Environmental Perception System for Split-type Flying Vehicle
李颖 1王荣煊 2万成麟 2王伟达 1杨超 1徐彬1
作者信息
- 1. 北京理工大学机械与车辆学院 北京 100081;北京理工大学重庆创新中心 重庆 401120
- 2. 北京理工大学机械与车辆学院 北京 100081
- 折叠
摘要
飞行汽车能够根据道路与天气情况在地面行驶和空中飞行两种模式之间进行切换,对提升交通效率、适应复杂地形及降低交通设施成本具有重要意义.分体式飞行汽车由电动垂直起降飞行器、智能操控座舱和自动驾驶底盘三个独立模块构成,具备地面无人行驶与空中自主飞行两种模式.首先,基于分体式飞行汽车自动驾驶、多模块自主对接引导等功能任务需求,提出立体环境感知系统传感器配置与部署方案并标定了传感器.其次,设计多传感器融合环境感知算法,实现目标检测、可行驶区域分割及多模块自主对接定位.最后,在分体式飞行汽车高性能边缘计算设备进行实车验证.试验表明该感知系统能取得高效的环境感知效果,为分体式飞行汽车的自动驾驶及自主对接引导提供新思路.
Abstract
Flying vehicles can switch between ground-driving and air-flying modes according to road and weather conditions,which is of great significance for improving traffic efficiency,adapting to complex terrain,and reducing the cost of transportation facilities.The split-type flying vehicle is composed of three independent modules:an electric vertical take-off and landing aircraft,an intelligent control cockpit,and an autopilot chassis.It has two functions,unmanned driving on the ground and autonomous flight in the air.First of all,based on task requirements,such as automatic driving and multi-module autonomous docking guidance,the sensor configuration and deployment scheme of the 3D environmental perception system is proposed and the sensors are calibrated.Secondly,a multi-sensor fusion environment perception algorithm is designed to realize object detection,drivable area segmentation,and multi-module autonomous docking and positioning.Finally,the real-world verification is carried out on the high-performance edge computing equipment of the split-type flying vehicle.The experimental results show that the perception system can achieve an effective environmental perception effect,which provides a new solution for the automatic driving and autonomous docking guidance of split-type flying vehicles.
关键词
分体式飞行汽车/环境感知/传感器标定与算法测试Key words
split-type flying vehicle/environmental perception/sensor calibration and algorithm testing引用本文复制引用
基金项目
国家自然科学基金(52102449)
国家自然科学基金(52275047)
国家自然科学基金(51975048)
重庆市自然科学基金(cstc2021jcyjmsxmX0879)
出版年
2024
NETL
NSTL
万方数据