Christian Siebert, Institute of Communications and Navigation, German Aerospace Center (DLR), & Chair of Navigation, RWTH Aachen University, Germany; Andriy Konovaltsev, Institute of Communications and Navigation, DLR, Germany; Michael Meurer, Institute of Communications and Navigation, DLR, & Chair of Navigation, RWTH Aachen University, Germany

View Abstract Sign in for premium content


Multipath propagation is to this day a major source of error in Global Navigation Satellite Systems (GNSSs), in particular in urban environments. Here, additional delayed signal replicas reach the receiver via one or multiple reflections from nearby obstacles. This causes biased range estimates in conventional receivers leading to Position, Velocity, and Time (PVT) errors. New safety-relevant applications as unmanned aerial vehicles or autonomous cars raise demand for high accuracy and fail-safe positioning systems. Deviations of several meters or outages in these areas are unacceptable. A novel proposal is made for an effective multipath rejection. The solution is based on a multicorrelator receiver using an Extended Kalman Filter (EKF) as an iterative solver replacing the conventional Delay Locked Loop (DLL) code tracking. The underlying signal model of the EKF incorporates the radio propagation channel between satellite and receiver inherently considering the reception of reflected multipath signals. Therewith, not only additional resilience against multipath propagation phenomena can be achieved while performing the code pseudorange measurement, but also an estimate for the Channel Impulse Response (CIR) of the radio channel is obtained. The functionality of the proposed approach has been demonstrated with simulations. A hardware GNSS constellation simulator has been used to test the algorithm under realistic conditions. The performance has been compared against the conventional DLL based GNSS receiver. It turned out that the effect of multipath in the pseudoranges as well as the PVT solution has been widely mitigated. In addition to that, it has been demonstrated that the estimation of the CIR of a radio channel affected by strong multipath delivers promising results.