Experimental Tests of the Obstacles Detection Technique in the Hemispherical Area for an Underground Explorer UAV

This paper presents details of experimental tests of an approach to measure the distance to obstacles in a hemispherical area around a vertical takeoff and landing unmanned aerial vehicle (UAV) that has been designed to explore the hard-to-reach areas such as caves, hollows, grottoes, or mines. The idea of the explorer UAV emerged from the need to draw a sketch of the rooms before a man entered, no matter the interior is a newly discovered cave or building. Thus, the system for point cloud data acquisition has been designed for feeding simultaneous localization and mapping algorithm. The developed system is based on an infrared depth sensor that returns data as a range image. The sensor is rotated at a constant speed so that the measurements are performed at predefined angles, which results with the data about distances to objects in a hemisphere around the position of the sensor. The system was tested in static and dynamic laboratory conditions. Static tests include measurements of the distance to the fixed obstacles made around a sensor placed on the fixed platform. In dynamic tests, the platform with the sensor and the obstacle is also linearly moved relative to each other. The laboratory stand was designed, where the platform on the trolley was moved along the linear trail with controlled speed. Results obtained from those tests reassure its suitability as a basis for the future development of the collision avoidance system and three-dimensional modeling sensor.