Autonomous Path Planning For Micro Aerial Vehicles In 3D Environment

With the rapid development of artificial intelligence field,Micro Aerial Vehicle(MAV)has been further popularized in human production and life,but most aircrafts at this stage can only fly in open environments with human remote control.How to let the MAV equipped with different sensors and algorithms have the ability to fly autonomously and complete more complicated and changeable tasks has become a hotspot in the field of mobile robots in recent years,and the core of which is MAV's autonomous path planning.The path planning environment can be divided into known environment and unknown environment.In this paper,the path planning method based on a priori environment map and the 3D path planning method for unknown environment are proposed respectively for these two environments.The path planning method based on a priori environment map needs a priori map reconstructed by using a Velodyne VLP-16 3D laser and 3D-NDT algorithm,along with the perception and path planning of local dynamic environment,an improved A * algorithm is used to realize the fixed height autonomous flight of the MAV on the prior flight environment.The improved A * algorithm mainly includes the optimization of the evaluation function,the introduction of security strategies,the smoothing of the path and the acceleration of the algorithm,and a simulation experiment has been done to show the differences between the traditional A * algorithm and the improved one.In this path planning method,a 2D laser sensor is used to obtain local maps and the strategy of paths set in advance is applied to adapt the local environment,finally the global path planning is done.Aiming at 3D path planning in unknown environment,this paper proposes a 3D path planning algorithm based on binocular stereo vision.In this method,a binocular vision sensor is used to obtain the local environment information in real time,and the distribution of obstacles in the environment can be analyzed with the using of depth images acquired by the binocular vision.By drawing on the idea of dynamic window approach algorithm,the 3D path planning algorithm converts the global path to the superposition of local paths by using a kind of 3D path planning model when finding the optimal path in local environment.By using a binocular vision sensor to obtain the environment information,the 3D path planning algorithm can find the local optimal path quickly,and finally realize the MAV's autonomous flight in unknown 3D environments.