Abstract: | In the past few years advanced driver assistance systems (ADAS) have been in focus of many research projects and development efforts of car manufacturers. Aiming on the capability of autonomous driving, positioning plays one major role. While precision and accuracy of uncoupled or loosely-coupled single frequency GPS navigation satisfies the requirements of route guidance, future ADAS will be more demanding towards availability, accuracy and integrity. The Institute of Flight Guidance (IFF) is currently involved in research projects evaluating the performance of unaided traditional GNSS receivers coupled with vehicle sensor measurements such as odometers in a tightly-coupled architecture as shown at the ION GNSS 2011([1]). Besides these involvements the IFF has developed a general purpose software-based GNSS receiver allowing full access to signal processing routines which has been already used to evaluate deep-coupled approaches as shown for example at the ION ITM 2013([2]). The vehicle state estimation of the tightly-coupled navigation is often derogated by GNSS signal shadowing and reflections due to buildings or plants, especially in urban scenarios. Moving a receiver under these conditions, the positioning solution can suffer from the low number of received signals, from frequent changes of the received constellation and from measurements affected by multipath. Even measurement datas without the minimum number of four signals can occur, which is required for a valid position. In off-the-shelf receivers the frequency normal is implemented as a quartz oscillator. The stability of these oscillators lies several magnitudes below the stability of atomic oscillators that are used in GNSS satellites. For a short period of time however, it is often assumed that the receiver clock drift stays on a constant value and the clock error can thereby be predicted during signal outages. In that way three satellite signals would be sufficient for a positioning solution. Using a recursive state estimator, like the Kalman filter, even two or one signals could support state estimation. Nevertheless even under constant temperatures there is a considerable variation of the clock drift which leads to fast drifting range measurements. Mainly this effect is due to accelerations impacting on the quartz. Since the accelerations of the vehicle are well known from vehicle dead reckoning sensors, short term effects can be modelled to adjust the measurements. To identify the character of the acceleration impact on a receiver clock error, the authors have passed several tests in static and dynamic scenarios. As different receivers pursue diverging strategies to control the clock drift and thereby the clock error, measurements from three different receivers were analysed. This three receivers representing three different price segments as well as different clock control strategies. In the NovAtel OEM6, which is located in the high-end price segment, the quartz can be controlled by voltage and is thereby settled within a control loop, minimizing the clock error based on the estimation from the positioning solution. At the low-price segment the ublox 4 utilizes a control strategy which is not well documented, but it allows the clock to drift until a certain clock bias is reached. Then the clock jumps several milliseconds. Furthermore the behaviour of a USB-GNSS-Front-End along with the IFF software GNSS receiver was analysed, because an open control loop is ensured in this case. |
Published in: |
Proceedings of the 27th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2014) September 8 - 12, 2014 Tampa Convention Center Tampa, Florida |
Pages: | 917 - 924 |
Cite this article: | Escher, Martin, Haak, Ulrich, Hecker, Peter, "Evaluation of GNSS Receiver Clock Drift Predictability in Urban Scenarios Based on Vehicle Dead Reckoning Sensor Data," Proceedings of the 27th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2014), Tampa, Florida, September 2014, pp. 917-924. |
Full Paper: |
ION Members/Non-Members: 1 Download Credit
Sign In |