无人机在室内环境中自主飞行与避障

Autonomous Flight and Obstacle Avoidance of MAV in Indoor Environment

为使无人机在室内、密林等无GPS或弱GPS信号环境下实现自主飞行并避开飞行过程中的障碍物,基于无迹卡尔曼滤波的方法实现多种机载传感器数据融合,获得无人机飞行状态及周围环境信息;通过信息融合与单目摄像机目标检测等算法,使无人机避开飞行过程中的障碍物并按照顺序穿越特定障碍圈,实现无人机自主飞行与避障功能。在Airsim仿真平台随机生成的场景与现实环境中进行实验。结果表明,在仿真环境中无人机可在560s左右完成自主起飞,准确避开障碍物,依次成功穿越10个障碍圈并自主降落;在现实世界中,当环境发生变化时需调节参数,无人机才可完成穿越障碍圈的任务。

In order to solve the problem that the UAV realizes autonomous flight and avoids obstacles during flight in indoors,jungles and other environments without GPS or weak GPS signals.We use a method based on unscented Kalman filter to integrate a variety of airborne sensor data to obtain the flight status of the drone and surrounding environment information.Through information fusion and monocular camera target detection algorithms,the drone can complete the task of avoiding obstacles during the flight and traversing the specific obstacle circle in order.We experimented with scenes randomly generated by the Airsim simulation platform and the real world environment.The experimental results show that in the simulation environment,the UAV can complete the task of autonomous takeoff,accurately avoid obstacles,successfully cross 10 obstacle circles in numerical order and make autonomous landing in 560s;in the real world,when the environment changes,the parametersneed to be adjusted tocomplete the task of crossing the obstacle circle.The method of monocular camera target detection and sensor data fusion can realize the autonomous flight and obstacle avoidance function of the drone.