Performance Evaluation of Ionospheric Spatial Gradient Monitoring using Radio-over-Fiber Devices and a Single Clock Receiver for Extended Baselines
Takayuki Yoshihara, Susumu Saito, Shinji Saitoh, Electronic Navigation Research Institute (ENRI), National Institute of Maritime, Port and Aviation Technology (MPAT), Japan
Alternate Number 2
GBAS (Ground-Based Augmentation System) is a system based on local differential correction technique to support aircraft precision approach. In its safety design, a large ionospheric spatial gradient is one of the most important integrity risks. Therefore, ionospheric threat model, which represents the upper limit that the system should assume in GBAS operation, is defined. Regarding the threat model in the low magnetic latitude region, a common model was established through activities of ICAO (International Civil Aviation Organization) APAC (Asia-Pacific) ISTF (Ionospheric Studies Task Force) . As our further works of the threat model, there are two issues, which are its improvement using a long-term data set and its customization in consideration of local ionospheric conditions in Japan. Another importance related to ionospheric spatial gradient is real-time monitoring. ICAO SARPs (Standards And Recommended Practices) for ground subsystem of category III (CAT-III) GBAS with single frequency signal of GPS L1-C/A (GAST-D; GBAS Service Type D), which supports aircraft precision approach and landing including rollout, is scheduled to be effective in 2018. GAST-D requires an ionospheric gradient monitor for its ground subsystem to detect ionospheric anomaly. For development of ionospheric threat model and the integrity monitor for GAST-D ground subsystem, a method to estimate ionospheric spatial gradient accurately is essential.
To accomplish our purpose with the above two viewpoints, we employ an estimation method of SF-CBCA (Single-frequency Carrier-based and Code-aided technique), which determines a ISD (Ionospheric delay in common-satellite Single Difference) together with a receiver clock difference between a pair of receivers [2,3]. It should be noted that ionospheric spatial gradient is calculated by dividing a ISD by a baseline length between the two receivers. Because a ISD increases in proportional to baseline length with a constant gradient, it is expected that a ISD is more accurately estimated with larger baseline if random noise is the same. Therefore, we consider to use a monitoring system of ionospheric spatial gradient with a long baseline not only to develop ionospheric threat model but also to improve performance of the ionospheric spatial gradient monitor in the low magnetic latitude region. Although a typical baseline length of GBAS reference stations is several hundred meters, we proposed a new setup to simplify error models in the SF-CBCA by reduction of receiver clock contribution to single difference measurements with a baseline length up to several km . In the previous study, we used a multi-port GNSS receiver with a single clock to remove almost part of receiver clock differences from ISD estimation and Radio-over-Fiber (RoF) devices to input two GNSS antenna signals to the receiver. To analyze preliminary performance of the ISDs, we used each data set with and without a common clock. The both data sets were collected by the same-type antennas with a baseline length of 2.3 km in New Ishigaki airport.
The results of the previous study showed that there were not so large differences between the both data sets even under disturbed conditions. Moreover, reliability of estimated ISDs using an evaluation parameter of fix rate in ambiguity resolution process was worse in the case with a common clock against our initial expectation, when plasma bubble events were observed. The results also showed that it was needed to check data quality and preform data screening especially for the data set with a common clock to compare with another data set. In this study, we conduct data quality control at first. Next, we perform the same analysis for the data set with a common clock to examine whether fix rate is improved or not including reliability of estimated clock error components, comparing with the data set without a common clock. We are also improving a clock error model in the SF-CBCA for process of the data set with a common clock. Namely, we think that clock error component should be estimated with a more tight constraint or not estimated. We anticipate that the improved method makes better performance of ISD estimation under disturbed conditions. Although the previous study examined the data set of several days under quiet and disturbed conditions, we plan to expand analysis data period to about a year in this study.
 S. Saito, et al., “Ionospheric delay gradient model for GBAS in the Asia-Pacific region”, GPS Solut (2017) 21: 1937. https://doi.org/10.1007/s10291-017-0662-1
 S. Fujita et al., “Determination of ionospheric gradients in short baselines by using single frequency measurements,” J. Aero. Astro. Avi., A-42, pp.269-275, 2010.
 S. Saito, et al., “Absolute gradient monitoring for GAST-D with a single-frequency carrier-phase based and code-aided technique,” Proc. ION GNSS 2012, 2184-2190, 2012.
 T. Yoshihara, et al., “Long Baseline Precise Ionospheric Gradient Measurements and its Application to GBAS”, Proc. of the ION 2017 Pacific PNT, pp. 885-896, Honolulu, Hawaii, May 2017.