Page 44 - 卫星导航2021年第1-2合期

Du et al. Satell Navig (2021) 2:3 Page 11 of 22

combination or subgroup of observations) (Blanch et al. guities and contaminated with cycle slips, resulting
2019; Ge et al. 2017; Imparato et al. 2018a). in extra failure modes that need to be monitored.
(2) PPP needs precise products and correction mod-
Integrity of PPP in ITS context els, and hence the nominal error models and threat
Position accuracy and integrity requirements of ITS models used in integrity monitoring should be care-
ITS applications generally require lane-level (sub-metre) fully developed.
accuracy to enable autonomous driving. Some applica- (3) PPP usually requires recursive processing, such
tions may even need dm-level accuracy (Green et al. as the use of Kalman fltering, involving dynamic
2013; Stephenson et al. 2011). For position integrity, models with process noise, whereas GNSS SPS uses
although there are some discussions and preliminary simple “snapshot” integrity monitoring methods.
statement of ITS requirements (European GNSS Agency
2015, 2018; Reid et al. 2019; Salós et al. 2010), no stand- Yet there is limited literature on PPP integrity, and their
ardised or generally-accepted specifcations, nor mature monitoring methods are still under investigation.
methodology for ITS applications are currently available Apart from above problematic aspects, complex urban
(Zhu et al. 2018). environments make PPP integrity monitoring for ITS
Te basic principle for integrity requirements is much more challenging. Te main difculties are in
that they should be defned according to the relevant the following two aspects (Bryant 2019; Imparato et al.
safety standards, e.g. International Organization for 2018a; Navarro et al. 2016; Zhu et al. 2018):
Standardization (ISO) 26262 and ISO/Publicly Avail-
able Specifcation (PAS) 21448 - Safety of the Intended (1) Multipath, NLOS errors, and signal interference
Function (SOTIF) (Kafka 2012; ISO 2018, 2019; Koo- occur frequently and have signifcant efects in
pman et al. 2019). However, as integrity requirements urban environments, for which appropriate sto-
are highly dependent on applications, e.g. Advanced chastic models and threat models are extremely dif-
Driver-Assistance Systems (ADAS), collision avoid- fcult to develop.
ance, and diferent levels of autonomous driving, the (2) Harsh environments also mean low redundancy in
specifcations for various ITS applications are manda- the number of observations and high probability of
tory (Zhu et al. 2018). Furthermore, there are many multiple faults occurring at the same time.
practical factors that need to be considered when defn-
ing the ITS requirements, such as country and region, Integrity information on real-time products or cor-
road geometry, vehicle type/size, driving speed, and rections for PPP can be generated at the network-end
data latency (Reid et al. 2019). by using the measurements from a GNSS ground track-
Te integrity indicators should also be tailored accord- ing network, like the SIS integrity generation by GBAS
ing to specifc ITS requirements. Especially for AL and or SBAS. Te faults in diferent corrections, e.g. orbit
PL in land applications, users are mainly concerned with and clock correction and ionospheric correction, can be
horizontal positions rather than vertical ones. HAL/ monitored separately by forming the measurements into
HPL should be further decomposed into along-track (or diferent monitors which are mainly sensitive to specifc
longitudinal) AL/PL and cross-track (or lateral) AL/PL errors (Weinbach et al. 2018). Nevertheless, the integ-
(Imparato et al. 2018a; Reid et al. 2019). Furthermore, rity of network-generated products/corrections is rarely
the test statistics and associated thresholds should be discussed in the literature. Currently, none of the correc-
adapted for ITS applications (El-Mowafy 2019). tions provided by IGS-RTS include integrity information,
although URA is reserved according to Radio Technical
Commission for Maritime Services-State Space Repre-
Integrity monitoring for PPP sentation (RTCM-SSR) protocol for future integrity capa-
GNSS PPP integrity shares some common aspects with bility (Cheng et al. 2018; IGS 2019). A preliminary study
GNSS SPS integrity in terms of defnition, indicators, and was done by Cheng et al. (2018) to investigate the strat-
basic monitoring procedure. PPP integrity can be moni- egy of URA characterisation based on the analysis of the
tored at both the system-level and user-level. However, real-time orbit and clock corrections from CNES (Cen-
integrity monitoring for PPP must additionally consider tre National D’Etudes Spatiales). A few service providers,
the following aspects (Bryant 2019; Feng et al. 2009; Pas- including SBAS systems such as those from Trimble and
nikowski 2015; Romay and Lainez 2012): the Quasi-Zenith Satellite System (QZSS), provide integ-
rity information on their correction services (Hirokawa
(1) PPP involves more observations, especially carrier- et al. 2016; Weinbach et al. 2018). Trimble Center-
phase measurements, which are biased by the ambi- Point RTX correction service utilises diferent types of