Bourns College of Engineering

Differential Global Navigation Systems (DGNSS) has been demonstrated to provide reliable, high-quality range correction information enabling real-time navigation with centimeter to sub-meter accuracy, which is required for applications such as connected and autonomous vehicles. However, DGNSS requires a local reference station near each user. For a continental or global scale implementation, this information dissemination approach would require a dense network of reference stations whose construction and maintenance would be prohibitively expensive. Precise Point Positioning (PPP) affords more flexibility as a public service for GNSS receivers, but its State Space Representation (SSR) format is not supported by most receivers in the field or on the market.

This article proposes a novel Virtual Network (VN)DGNSS approach and an optimization algorithm that is key to its implementation. The approach capitalizes on the existing PPP infrastructure without the need for new physical reference stations. Specifically, no reference station is needed in the local vicinity of any user. By connecting to public GNSS SSR data services, a VNDGNSS server maintains current information about satellite code bias, satellite orbit and clock, and atmospheric models. Construction of the RTCM Observation Space Representation (OSR) messages from this SSR information requires both the signal time-of-transmission and the satellite position at that time which are consistent with the time-of-reception for each client. This article presents an algorithm to determine these quantities. Then the VNDGNSS computes and transmits RTCM OSR messages to user receivers. This approach achieves global dissemination coverage for real-time navigation without the need for additional local base stations near each user.

The results of real-time stationary and moving platform evaluations are included, using u-blox M8P and ZED-F9P receivers. The performance surpasses the Society of Automotive Engineering (SAE) specification (68% of horizontal error < 1.5 m and vertical error < 3 m) and shows significantly better horizontal performance than GNSS OS. The moving tests also show better horizontal performance than the ZED-F9P receiver with SBAS enabled and achieve the lane-level accuracy (95% of horizontal errors less than 1 meter).

This article is a tutorial describing the process and issues related to developing a state-space model for the stochastic errors affecting an Inertial Measurement Unit (IMU). The starting point is the instrument error characterization data sheet provided by the manufacturer, which is typically either an Allan Variance graph or the parameters extracted from that graph. Along with this tutorial, supplementary software is available for open source distribution that calculates and plots the Allan Variance for a given set of data; extracts optimal parameters (e.g., $Q_N$, $Q_B$, $Q_K$, and $T_B$) from Allan Variance data for a given model structure; and, computes and simulates of the discrete-time equivalent model.