Real-time Simulation of a 2.5D Radar

Kurzfassung

DLR's project ALLFlight (Assisted Low Level Flight and Landing on Unprepared Landing Sites) aims towards assisting a helicopter pilot while landing on adverse, unprepared sites. This includes landing during brown-out or white-out conditions or in unexplored surroundings with no or few data about landing obstacles present. To reach this aim a selection of sensors has been chosen, including a 2.5D radar used for dust and fog penetrating terrain scans. In order to develop and test sensory data fusion and display methods it is necessary to have a ground based simulation available for all sensors that can be used together with a flight simulator. Over the recent years, DLR has developed a suite of computer graphics based, real-time simulators for most of the sensors. We present an extension of the simulator suite to produce simulated 2.5D terrain scanning radar data. Other than more "image-like" sensors, like ladar, a 2.5D radar has a very distinct scanning pattern generated by the complex mechanical movement of the antenna sweep. Therefore, the generated data is not delivered in image frames but rather as a sequence of packages representing measurements along a deformed figure-eight path. First, we generate a cube of 3D measurements inside computer memory making use of graphics hardware acceleration. In a second process the in-memory data is rescanned along a path that resembles the movement of the radar antenna. Using state-of-the-art PC hardware the simulation approach can guarantee data rates comparable to real existing radar devices, thus allowing real-time simulation. This allows to evaluate and test algorithms developed to post-process and fuse the generated data. The implementation is based on the graphics framework OpenGL and makes use of current graphics technology like shaders. We explain implementation details and we discuss how to integrate the simulation module into the existing simulation infrastructure. Finally, we present image sequences and data streams generated by the approach.