Multi-Band Multi-GNSS RTK for Mass-Market Applications
C�cile Mongr�dien, Alex Parkins, Chris Hide, Marten Str�m, D. Ammann
- 发表年份
- 2018
- 引用次数
- 4
摘要
Until recently, high costs and complexity have confined RTK and multi-band GNSS technologies to low-volume niche markets, such as surveying or construction. However, emerging high volume markets, such as precision agriculture, robotic guidance systems, or autonomous driving require high-precision performance that is both affordable and optimized for low power consumption. In the beginning of 2016, several industry players, including u-blox, introduced L1-based GNSS receivers with integrated RTK technology. u-blox’ products, the NEO-M8P rover and reference modules, offer an end-to-end GPS/GLONASS (or GPS/BeiDou) L1 RTK solution that delivers high-precision performance while remaining both low-cost and energy-efficient. Such products are therefore well suited for mass-market applications. However, single-band RTK intrinsically suffers from well-known limitations such as slow convergence, difficulty to detect and correct cycle slips, or necessity to operate either in close vicinity of a reference station or in a dense reference station network. A new generation of receivers using multi-band technology is now introduced to address these limitations. This paper focuses on the ZED-F9P, a multi-band RTK module that can concurrently support all major GNSS constellations. It discusses the trade-offs associated with designing a very precise multi-GNSS multi-band RTK module solution while keeping cost, size, and power consumption as low as possible. It presents the results of several test campaigns. These are used both to highlight the benefits of dual-band and increased satellite availability in challenging environments, and to benchmark the performance of the ZED-F9P modules against existing high-end products. The integration of a second band alongside the legacy L1 has the potential to double the number of available measurements and significantly improve the availability, reliability, and accuracy of the navigation solution. However, despite the modernization and expansion efforts undertaken by all major GNSS programs, civilian signal availability remains limited outside the L1 band. While new and modernized civil signals in the L5 frequency offer attractive characteristics, especially in terms of multipath mitigation, there are currently more civil signals available in the L2 band. The ZED-F9P module first aims to maximize the availability of signals from the second and therefore supports signals from the L2 band in preference to the L5 band, that is, GPS L2C, GLONASS L2OF, Galileo E5b, and BeiDou B2I. With the exception of Galileo E5b, these signals are narrowband signals that are computationally less expensive to process than their broadband L5 counterparts. Their use therefore makes the solution more energy-efficient. The paper will go into more details about why such a combination of signals was found to be superior to the alternative of using signals from the L1 and L5 bands. To better distribute the computational load and increase processing efficiency, the u-blox core RTK technology relies on a dual-filter implementation (patents pending) that decouples the ambiguity estimation from the main navigation filter. In this way, the computationally expensive ambiguity estimation algorithms can run at a lower rate than the main navigation process. Considering that the computational cost of the ambiguity estimation algorithms grow rapidly with the number of ambiguity and that, in a multi-GNSS dual-band receiver like the ZED-F9P, the number of ambiguities can easily reach 40 or more, this dual-filter implementation can lead to significant computational savings, especially when operating at high navigation rate. As will be detailed in the paper, the reliability of this dual-filter implementation depends on carefully managed inter-filter communication and cycle-slip monitoring. To spread the computational load further and to simplify the receiver frequency plan, the ZED-F9P module combines two separate chips that are respectively tuned t
关键词
相关论文
Statistical Learning Theory
Yuhai Wu, Vladimir Vapnik
1999
Artificial intelligence: a modern approach
1995
Fractional Differential Equations
Igor Podlubný
2025
Applied Nonlinear Control
Jean-Jacques Slotine, Weiping Li
1991