It was found that a newly developed portable collision alert radar receives reflections from the ground while flying. In this paper a method is developed that uses range and Doppler information from these reflections. This information is used to compute height and velocity information relative to the terrain, something which is not possible with existing hardware. The method was tested on a local flight in the Netherlands, with a prototype of the radar. Flight state results were compared with those of a GPS tracker on board. It was found that the velocity can be computed within meters-per-second accuracy. Height differences are due to the measurement method, measuring directly from the ground surface (radar) or relative to a database (GPS). If developments in microwave sensing techniques continue to improve the hardware, flight state estimation by radar can become an option for pilots who do not want to be dependent on the correctness of a terrain model, but who measure the terrain shape independently.@en

A detailed situation awareness of the local environment is essential for safe flight in General Aviation. When operating under Visual Flight Rules, eyesight is crucial for maintaining situation awareness and objects may be overlooked. Technical solutions such as Flarm have been sought, but they only work on a basis of co-operation: obstacles without the proper equipment are invisible. Recent developments in the field of radar technology, partly empowered by the demand for sensors for autonomous cars, have improved the size and power consumption of available hardware. Today, the hardware exists to build a portable primary radar system for situation awareness. In this paper the results are presented of efforts to build the first portable primary radar for general, which has to be lightweight, cheap and have a low power consumption. The focus in this paper is on the software design of such a radar system. The physical principles of radar sensing are described, as well as the scientific steps needed to provide situation awareness. The hardware and software for the radar are both built and tested, and the results of these tests are presented. A flight experiment is performed with a small aircraft flying past a stationary radar on a small hill. It is found that the radar is capable of detecting the aircraft up to a distance of at least 3 kilometers. 3D localization is performed and the location determined by the radar was on average 46 meters away from the aircraft position as measured by satellite navigation, relative to a total distance of about 1000 meters from the radar. A low-pass filter can be applied on the raw results in order to improve the location estimation further. Future research will focus on bringing the portable radar in motion while operating.

@en

Although the main goal of a newly developed Collision Alert Radar is to observe airborne targets, it was found that reflections of the ground are received by the radar. The radar is carried on board of the aircraft, and the ground reflections may be used to detect flight information with respect to the terrain, something which is not possible with existing hardware. In this paper a method is developed which makes use of range and Doppler information from ground reflections, in order to provide the pilot with height and velocity information. The method was tested on a local flight in the Netherlands, with a prototype of the radar on-board. State results were compared to those of a GPS tracker on board. It was found that the horizontal and vertical components of the velocity were found with a standard deviation of about 3m/s, and the height estimates had a standard deviation of 23m. Also, a discrepancy of 36m between the GPS and radar height estimates was found, which was caused by a fault in the GPS earth surface model, which was no problem for the radar. It is concluded that the quality of radar state estimates is approaching that of GPS measurements. The rapid developments in microwave sensing techniques can help the radar to surpass the quality of GPS in the coming years. If that happens, state estimation by radar can become an option for pilots who do not want to be dependent on the correctness of a terrain model, but who measure the terrain shape independently. @en

Many pilots in General Aviation use electronic add-ons aids in flight, which rely on satellite navigation information. This navigation information is often a single point of failure which is undesirable since the pilot relies on the information. This paper presents the results of research whether a novel mobile radar station can be used to validate the navigation results from the GPS. The radar transmits signals to the ground, and compares the locations of the reflections to a digital map such as Google maps. A test flight was performed with a radar system on board. Fifteen different methods for processing the images were investigated, and it was found that Ridge Operators and Entropy Detection are good methods to extract similar features in Google and radar images. These algorithms were always successful in picking the single correct GPS coordinate out of a pool of 300 false ones within 150m of the correct answer, except when the aircraft was making a turn and the radar was pointed to the sky. It is concluded that a ground-scanning radar on board can be used to validate the results of a GPS, provided that the radar can observe recognizable features that can be compared to a digital map. The type of image processing used to extract the data is crucial for the application. @en

Object tracking is performed when surveillance applications have multiple observations of an object over time. An example of such a surveillance application is mounting a wideangle Frequency Modulated Continuous Wave (FMCW) radar system on board of a General Aviation aircraft. This is done in order to observe its environment in detail, including noncooperative objects such as birds and windmills. Data generated by such a system follows different physical laws than the images of standard visual applications. In this paper, a novel tracking algorithm is introduced which is tailor-made for FMCW applications. The algorithm is tested in a simulated crowded general aviation airspace, and the resulting tracks are qualitatively and quantitatively analysed. The proposed algorithm performs better than a traditional algorithm on all aspects, but tracking errors can still be made in rare cases. The proposed algorithm can be used in conjunction with research focusing on observation quality or assignment problems. @en