ION GNSS 2012
Session A3: Geodesy, Surveying & RTK for Civil Applications

Title: GNSS Antenna Orientation and Detection of Multipath Signals from Direction of Arrival
Author(s): D.E. Grimm, and R. Mautz, Institute of Geodesy and Photogrammetry, Switzerland
Room: 103/104 (NCC)

Objectives This work presents a concept to determine the directions of arrival (DOA) of the satellites´ signals using a customary GNSS antenna. Knowing the DOA with regard to the antenna has two main advantages. First, the absolute orientation of the antenna can be derived without the need for an auxiliary orientation sensor such as a compass or gyroscope. The orientation of the antenna is essential for many navigation applications, in particular in situations where a baseline of two or more antennas is impractical. In addition, the knowledge of the antenna orientation can be used to implement correction models of the antenna phase centre variation (PCV) for high precision GNSS applications. Secondly, the DOA of signals from individual satellites can be used to detect indirect signals resulting from multipath. Such signals can be excluded for the position calculation to improve the position accuracy and remove systematic biases. Methodology Because the DOA is not a directly measurable quantity when using a single, ordinary GNSS antenna, an indirect technique based on influencing the carrier to noise density (C/N0) of the received signals is proposed. To modify the C/N0, a sector of the GNSS antenna is covered by a partial electromagnetic-wave attenuation shield, consisting of microwave absorbing material. The attenuation shield has the shape of a circle segment with the size of a quarter of the antenna. The shield does not suppress the signals completely; it only dampens the received energy of those satellites which are affected by the shield by a factor of a few decibels. The damping factor of the attenuating material is chosen to be weak enough to maintain satellite tracking and hence to keep the positioning function, but should also be sufficiently strong to significantly influence the C/N0. The shield causes a distinct, repeatable inhomogeneity in the antenna characteristic, comparable to a directional antenna. The characteristic gain pattern of the received C/N0 establishes a geometric relation between the DOA with the broadcasted positions of the GNSS satellites and the position of the shield. When the position of the shield is known in relation to the antenna, the antenna orientation can be determined. However, from a single position of the shield the antenna orientation can only be derived very roughly and unreliably. Therefore, the attenuation shield has been designed to rotate horizontally above the antenna, thereby attenuating the received signals of all satellites in view. An encoder provides the angular position of the shield in relation to the antenna. The speed of the rotation is chosen such that a data set of 720 signal-strength values can be obtained for each satellite during one rotation. To calculate the orientation of the antenna, the data series are compared with a previously captured reference data set. For the reference data set, the attenuation shield is mounted on an antenna with known orientation. The reference data is collected over one astronomic day, while the attenuation shield is rotating continuously. Finally, the reference data is averaged and cleaned from outliers. The calculation of the current antenna´s orientation is carried out by cross-correlation between the reference and the current data series. The angular value responsible for the maximal value for the correlation coefficient is used to determine the horizontal orientation of the GNSS antenna. Because the amount of the received energy depends also on the elevation of the satellites, the correlation is carried out for each elevation bin separately between 0 and 75 degrees elevation. Given that the DOA of each satellite´s signal can be measured, the presented concept can be used to detect indirect signals reaching the antenna after reflection. Once the antenna orientation is known, the DOA of each signal can be compared with the values calculated from the ephemerides. Evidently, if the same signal is received from multiple directions, multipath must be present. Moreover, even if only a signal from a single direction is received, the proposed method can indicate whether the signal is arriving directly or is coming from a ´wrong´ direction indicating a reflected signal. This last situation is particularly interesting when only the reflected signal reaches the antenna and the direct signal is not visible. Actual and Anticipated Results Current results show that a measurement uncertainty of the antenna orientation below 1 deg can be reached under optimal conditions and assuming an observation time of at least 2 hours. However, when the requirement on the measurement uncertainty is below 4 deg, a measuring time of 2 minutes is sufficient. If the orientation measurement is carried out at a different location from where the reference data set had been collected, the measurement uncertainty of the orientation increases up to 4 deg due to location-specific propagation effects. The influence on the position uncertainty originating from the attenuation shield is less than 2 mm when averaging over a number of full revolutions of the attenuation shield, and can thus be neglected in most cases. In terms of multipath detection, a first investigation revealed that it is possible to distinguish between unaffected signals from satellites arriving at the antenna directly and signals affected by multipath. Furthermore it is possible to determine different directions of arrival of multiple signals. An investigation on how much the exclusion of reflected signals can improve the GNSS position uncertainty is under way. The presented approach is to be compared with other multipath mitigation strategies. Conclusions Knowing from which direction the different signals arrive at the antenna has the potential to determine the absolute orientation of an GNSS antenna but also to improve the general understanding of GNSS signal propagation and reflection in real environments. This knowledge can be used to identify and exclude disturbing signals. An improvement is expected for the accuracy and reliability of the GNSS position solution in challenging environments. In comparison to approaches using antenna arrays with beam forming capability, an ordinary geodetic GNSS antenna can be used for DOA determination. The proposed method affects the position determination only marginally.

Previous Abstract Return to Session A3